xvEPA
  United States
  Environmental Protection
  Agency
Office of Transportation                    EPA420-R-04-007
and Air Quality                       May 2004
             Final Regulatory Analysis:
             Control of Emissions from
             Nonroad Diesel Engines

-------
                                     EPA420-R-04-007
                                           May 2004
Final Regulatory Impact Analysis:
    Control of Emissions from
     Nonroad Diesel Engines
       Assessment and Standards Division
      Office of Transportation and Air Quality
      U.S. Environmental Protection Agency

-------
CHAPTER 1: Industry Characterization
     1.1 Characterization of Engine Manufacturers
         1.1.1 Engines Rated between 0-19 kW (0 and 25 hp)	
         1.1.2 Engines Rated between 19 and 56 kW (25 and 75 hp)	
         1.1.3 Engines Rated between 56 and 130 kW (75 and 175 hp) . .
         1.1.4 Engines Rated between 130 and  560 kW (175 and 750 hp)
         1.1.5 Engines Rated over 560 kW (750 hp)	
     1.2 Characterization of Equipment Manufacturers 	
         1.2.1 Equipment Using Engines Rated under 19 kW (0 and 25 hp) 	
         1.2.2 Equipment Using Engines Rated between 19 and 56 kW (25 and 75 hp)
         1.2.3 Equipment Using Engines Rated between 56kW and 130 kW (75 and 175 hp)
         1.2.4 Equipment Using Engines Rated between 130 and 560 kW (175 and 750 hp) .
                                                                                           -1
                                                                                           -2
                                                                                           -2
                                                                                           -2
                                                                                           -2
                                                                                           -3
                                                                                           -3
                                                                                           -4
                                                                                           -6
                                                                                           -7
                                                                                           -9
         1.2.5 Equipment Using Engines Rated over 560 kW (750 hp)  	1-11
     1.3 Refinery Operations	1-12
         1.3.1 The Supply-Side	1-12
         1.3.2 The Demand Side	1-19
         1.3.3 Industry Organization	1-26
         1.3.4 Markets and Trends  	1-30
     1.4 Distribution and Storage Operations  	1-35
         1.4.1 The Supply-Side	1-35
         1.4.2 The Demand-Side	1-37
         1.4.3 Industry Organization	1-37
         1.4.4 Markets and Trends  	1-38

CHAPTER 2: Air Quality, Health, and Welfare Effects
     2.1 Particulate Matter	2-3
         2.1.1 Health Effects of Particulate Matter	2-4
         2.1.2 Attainment and Maintenance of the PM10 and PM2 5 NAAQS:  Current and Future Air
             Quality	2-16
         2.1.3 Environmental Effects of Particulate Matter  	2-38
     2.2 Air Toxics  	2-55
         2.2.1 Diesel Exhaust PM  	2-55
         2.2.2 Gaseous Air Toxics	2-75
     2.3 Ozone	2-88
         2.3.1 Health Effects of Ozone 	2-89
         2.3.2 Attainment and Maintenance of the 1-Hour and 8-Hour Ozone NAAQS  	2-92
         2.3.2 Attainment and Maintenance of the 1-Hour and 8-Hour Ozone NAAQS  	2-93
         2.3.3 Welfare Effects Associated with Ozone and its Precursors  	2-118
     2.4 Carbon Monoxide  	2-121

CHAPTER 3: Emission Inventory
     3.1 Nonroad Diesel Baseline Emission Inventory Development	3-2
         3.1.1 Land-Based Nonroad Diesel Engines—PM25, NOX, SO2, VOC, and CO Emissions ... 3-2
         3.1.2 Land-Based Nonroad Diesel Engines—Air Toxics Emissions	3-15
         3.1.3 Commercial Marine Vessels and Locomotives  	3-16
         3.1.4 Recreational Marine Engines  	3-21
         3.1.5 Fuel Consumption for Nonroad Diesel Engines	3-24
     3.2 Contribution of Nonroad Diesel Engines to National Emission Inventories	3-26
         3.2.1 Baseline Emission Inventory Development	3-26
         3.2.2 PM25 Emissions  	3-28
         3.2.3 NOX Emissions 	3-28
         3.2.4 SO9 Emissions	3-29

-------
         3.2.5 VOC Emissions  	3-29
         3.2.6 CO Emissions	3-29
     3.3 Contribution of Nonroad Diesel Engines to Selected Local Emission Inventories	3-37
         3.3.1 PM25 Emissions  	3-37
         3.3.2 NOX Emissions  	3-41
     3.4 Nonroad Diesel Controlled Emission Inventory Development	3-43
         3.4.1 Land-Based Diesel Engines—PM25, NOX, SO2, VOC, and CO Emissions  	3-43
         3.4.2 Land-Based Diesel Engines—Air Toxics Emissions	3-52
         3.4.3 Commercial Marine Vessels and Locomotives  	3-53
         3.4.4 Recreational Marine Engines 	3-55
     3.5 Projected Emission Reductions from the Final Rule	3-58
         3.5.1 PM25 Reductions	3-58
         3.5.2 NOX Reductions  	3-66
         3.5.3 SO2 Reductions	3-68
         3.5.4 VOC and Air Toxics Reductions 	3-75
         3.5.5 CO Reductions  	3-78
         3.5.6 PM25 and SO2 Reductions from the 15 ppm Locomotive and Marine (LM) Fuel Program
          	!	3-79
         3.5.7 SO2 and Sulfate PM Reductions from Other Nonhighway Fuel	3-81
     3.6 Emission Inventories Used for Air Quality Modeling	3-86

CHAPTER 4: Technologies and Test Procedures for Low-Emission Engines
     4.1 Feasibility of Emission Standards 	4-1
         4.1.1 PM Control Technologies	4-2
         4.1.2 NOx Control Technologies	4-19
         4.1.3 Can These Technologies Be Applied to Nonroad Engines and Equipment?  	4-70
         4.1.4 Are the Standards for Engines >25 hp and <75 hp Feasible?	4-82
         4.1.5 Are the Standards for Engines <25 hp Feasible? 	4-94
         4.1.6 Meeting the Crankcase Emission Requirements	4-101
         4.1.7 Why Do We Need 15 ppm Sulfur Diesel Fuel? 	4-101
     4.2 Transient Emission Testing	4-110
         4.2.1 Background and Justification	4-110
         4.2.2 Data Collection and Cycle Generation	4-113
         4.2.3 Composite Cycle Construction	4-126
         4.2.4 Cycle Characterization Statistics 	4-128
         4.2.5 Cycle Normalization/Denormalization Procedure  	4-129
         4.2.6 Cycle Performance Regression Statistics	4-130
         4.2.7 Constant-Speed, Variable-Load Equipment Considerations  	4-130
         4.2.8 Cycle Harmonization   	4-134
         4.2.9 Cold-Start Transient Test Procedure	4-146
         4.2.10 Applicability of Component Cycles to Nonroad Diesel Market	4-148
         4.2.11 Final Certification Cycle Selection Process 	4-151
     4.3 Steady-State Testing  	4-152
         4.3.1 Ramped Modal Cycle	4-153
         4.3.2 Transportation Refrigeration Unit Test Cycle 	4-166
     4.4 Not-to-Exceed Testing	4-169

CHAPTER 5: Fuel Standard Feasibility
     5.1 The Blendstocks and Properties of Non-Highway Diesel Fuel  	5-1
         5.1.1 Blendstocks Comprising Non-highway Diesel Fuel and their Sulfur Levels  	5-1
         5.1.2 Current Levels of Other Fuel Parameters in Non-highway Distillate	5-2
     5.2 Evaluation of Diesel Fuel  Desulfurization Technology	5-4

-------
         5.2.1 Introduction to Diesel Fuel Sulfur Control  	5-4
         5.2.2 Conventional Hydrotreating	5-5
         5.2.3 Process Dynamics Isotherming 	5-18
         5.2.4 Phillips S-Zorb Sulfur Adsorption 	5-22
         5.2.5 Chemical Oxidation and Extraction  	5-23
     5.3 Feasibility of Producing 500 ppm Sulfur NRLM Diesel Fuel in 2007	5-25
         5.3.1 Expected use of Desulfurization Technologies for 2007 	5-25
         5.3.2 Lead-time Evaluation	5-26
     5.4 Feasibility of Producing 15 ppm Sulfur NRLM in 2010 and 2012	5-34
         5.4.1 Expected use of Desulfurization Technologies in 2010 and 2012 	5-34
         5.4.2 Lead-time Evaluation	5-37
     5.5 Distribution Feasibility Issues	5-38
         5.5.1 Ability of Distribution System to Accommodate the Need for Additional Product
             Segregations That Could Result from This Rule	5-38
         5.5.2 Limiting Sulfur Contamination	5-61
         5.5.3 Handling Practices for Distillate Fuels that Become Mixed in the Pipeline Distribution
             System	5-63
     5.6 Feasibility of the Use  of a Marker in Heating Oil	5-66
     5.7 Impacts on the Engineering and Construction Industry  	5-73
         5.7.1 Design and Construction Resources Related to Desulfurization Equipment	5-74
         5.7.2 Number and Timing of Revamped and New Desulfurization Units	5-75
         5.7.3 Timing of Desulfurization Projects Starting up in the Same Year	5-76
         5.7.4 Timing of Design and Construction Resources Within a Project	5-76
         5.7.5 Projected Levels  of Design and Construction Resources	5-78
     5.8 Supply of Nonroad, Locomotive, and Marine Diesel Fuel  (NRLM)  	5-82
     5.9 Desulfurization Effect on Other Non-Highway Diesel Fuel Properties	5-90
         5.9.1 Fuel Lubricity	5-90
         5.9.2 Volumetric Energy Content  	5-93
         5.9.3 Fuel Properties Related to Storage and Handling  	5-95
         5.9.4 Cetane Index and Aromatics	5-95
         5.9.5 Other Fuel Properties 	5-96
     Appendix 5A: EPA's Legal Authority for Adopting Nonroad,  Locomotive, and Marine Diesel Fuel
         Sulfur Controls  	5-99
CHAPTER 6: Estimated Engine and Equipment Costs
     6.1 Methodology for Estimating Engine and Equipment Costs	6-2
     6.2 Engine-Related Costs	6-5
         6.2.1 Engine Fixed Costs	6-5
         6.2.2 Engine Variable Costs  	6-25
         6.2.3 Engine Operating Costs  	6-49
     6.3 Equipment-Related Costs	6-60
         6.3.1 Equipment Fixed Costs	6-61
         6.3.2 Equipment Variable Costs  	6-69
         6.3.3 Potential Impact of the Transition Provisions for Equipment Manufacturers  	6-72
     6.4 Summary of Engine and Equipment Costs  	6-74
         6.4.1 Engine Costs	6-74
         6.4.2 Equipment Costs	6-77
         6.4.3 Engine and Equipment Costs on a Per Unit Basis  	6-78
     6.5 Weighted Average Costs for Example Types of Equipment	6-82
         6.5.1 Summary of Costs for Some Example Types of Equipment 	6-82
         6.5.2 Method of Generating Costs for a Specific Piece of Equipment  	6-86

-------
         6.5.3 Costs for Specific Examples from the Proposal	6-89
    6.6 Residual Value of Platinum Group Metals  	6-90

CHAPTER 7: Estimated Costs of Low-Sulfur Fuels
    7.1 Production and Consumption of NRLM Diesel Fuel  	7-2
         7.1.1 Overview	7-2
         7.1.2 Distillate Fuel Production and Demand in 2001	7-6
         7.1.3 Distillate Fuel Production and Demand in 2014	7-18
         7.1.4 Sensitivity Cases	7-45
         7.1.5 Methodology for Annual Distillate Fuel Demand: 1996 to 2040	7-62
         7.1.6 Annual Distillate Fuel Demand and Sulfur Content	7-67
    7.2 Refining Costs	7-86
         7.2.1 Methodology	7-86
         7.2.2 Refining Costs  	7-157
    7.3 Cost of Lubricity Additives	7-188
    7.4 Cost of Distributing Non-Highway Diesel Fuel  	7-189
         7.4.1 New Production Segregation at Bulk Plants  	7-190
         7.4.2 Reduction in Fuel Volumetric Energy Content 	7-192
         7.4.3 Handling of Distillate Fuel Produced from Pipeline Interface	7-194
         7.4.4 Fuel Marker Costs	7-200
         7.4.5 Distribution and Marker Costs Under Alternative Sulfur Control Options  	7-205
    7.5 Total Cost of Supplying NRLM Fuel Under the Two-Step Program 	7-206
    7.6 Potential Fuel Price Impacts 	7-208

CHAPTER 8: Estimated Aggregate Cost and Cost per Ton of Reduced Emissions
    8.1 Projected Sales and Cost Allocations  	8-2
    8.2 Aggregate Engine Costs	8-3
         8.2.1 Aggregate Engine Fixed Costs	8-3
         8.2.2 Aggregate Engine Variable Costs	8-7
    8.3 Aggregate Equipment Costs 	8-11
         8.3.1 Aggregate Equipment Fixed Costs	8-11
         8.3.2 Aggregate Equipment Variable Costs	8-13
    8.4 Aggregate Fuel Costs and Other Operating Costs	8-15
         8.4.1 Aggregate Fuel Costs 	8-16
         8.4.2 Aggregate Oil-Change Maintenance Savings	8-18
         8.4.3 Aggregate CDPF Maintenance, CDPF Regeneration, and CCV Maintenance Costs . . 8-20
         8.4.4 Summary of Aggregate  Operating Costs  	8-22
         8.4.5 Summary of Aggregate  Operating Costs Associated with a Fuel-only Scenario	8-24
    8.5 Summary of Aggregate Costs of the Final Rule  	8-26
    8.6 Emission Reductions 	8-29
    8.7 Cost per Ton	8-30
         8.7.1 Cost per Ton for the NRT4 Final Rule	8-30
         8.7.2 Cost per Ton for the NRLM Fuel-only Scenario 	8-34
         8.7.3 Costs and Costs per Ton for Other Control Scenarios  	8-37
         8.7.4 Costs per Ton Summary	8-63
    Appendix 8A: Estimated Aggregate Cost and Cost per Ton of Sensitivity Analyses	8-65
    Appendix 8B: Fuel Volumes used throughout Chapter 8  	8-89
CHAPTER 9: Cost-Benefit Analysis	9-1
    9.1 Time Path of Emission Changes for the Final Standards	9-8
    9.2 Development of Benefits Scaling Factors Based on Differences in Emission Impacts
         Between the Final Standards and Modeled Preliminary Control Options  	9-11
    9.3 Summary of Modeled Benefits and Apportionment Method	9-12

-------
         9.3.1 Overview of Analytical Approach	9-16
         9.3.2 Air Quality Modeling	9-17
         9.3.3 Health Impact Functions  	9-19
         9.3.4 Economic Values for Health Outcomes	9-23
         9.3.5 Welfare Effects  	9-24
         9.3.6 Treatment of Uncertainty	9-28
         9.3.7 Model Results  	9-29
         9.3.8 Apportionment of Benefits to NOx, SO2, and Direct PM Emissions Reduction§-39
    9.4 Estimated Benefits of Final Nonroad Diesel Engine Standards in 2020 and 2030 . .  9-42
    9.5 Development of Intertemporal Scaling Factors and Calculation of Benefits Over Tirfie48
    9.6 Comparison of Costs and Benefits	9-53
    Appendix 9A: Benefits Analysis of Modeled Preliminary Control Option	9-77
    Appendix 9B: Supplemental Analyses Addressing Uncertainties in the Concentration-
         Response and Valuation Functions for Paniculate Matter Health Effects	9-205
    Appendix 9C: Sensitivity Analyses of Key Parameters in the Benefits Analysis	9-253
    Appendix 9D:  Visibility Benefits Estimates for Individual Class I Areas	9-271

CHAPTER 10: Economic Impact Analysis
    10.1 Overview and Results  	10-1
         10.1.1 What is an Economic Impact Analysis?	10-1
         10.1.2 What Methodology Did EPA Use in this Economic Impact Assessment?	10-1
         10.1.3 What are the key features of the NDEIM?	10-4
         10.1.4 Summary of Economic Analysis  	10-13
    10.2 Economic Methodology	10-27
         10.2.1 Behavioral Economic Models	10-27
         10.2.2 Conceptual Economic Approach  	10-28
         10.2.3 Key Modeling Elements	10-36
         10.2.4 Estimation of Social Costs	10-46
    10.3 NDEIM Model Inputs and Solution Algorithm 	10-49
         10.3.1 Description of Product Markets  	10-50
         10.3.2 Market Linkages	10-57
         10.3.3 Baseline Economic Data	10-57
         10.3.4 Calibrating the Fuel  Spillover Baseline 	10-66
         10.3.5 Compliance Costs	10-66
         10.3.6 Growth Rates 	10-79
         10.3.7 Market Supply and Demand Elasticities 	10-79
         10.3.8 Model Solution  	10-83
    10.4 Estimating Impacts	10-86
    Appendix 10A: Impacts on the Engine Markets 	10-92
    Appendix 10B: Impacts on Equipment Markets	10-101
    Appendix IOC: Impacts on Application Markets 	10-152
    Appendix 10D: Impacts on the Nonroad Fuel Market	10-158
    Appendix 10E: Time Series of Social Cost	10-163
    Appendix 10F: Model Equations	10-167
    Appendix 10G: Elasticity Parameters for Economic Impact Modeling	10-172
    Appendix 10H: Derivation of Supply Elasticity 	10-188
    Appendix 101: Sensitivity Analysis	10-189

CHAPTER 11: Small-Business Flexibility Analysis
    11.1 Overview of the Regulatory Flexibility Act 	11-1
    11.2 Need for the Rulemaking and Rulemaking Objectives  	11-2

-------
     11.3 Issues Raised by Public Comments	11-2
         11.3.1 Comments Regarding Small Business Engine and Equipment Manufacturers  	11-3
         11.3.2 Comments Regarding Small Fuel Refiners, Distributors, and Marketers	11-3
     11.4 Description of Affected Entities  	11-6
         11.4.1 Description of Nonroad Diesel Engine and Equipment Manufacturers	11-7
         11.4.2 Description of the Nonroad Diesel Fuel Industry	11-9
     11.5 Projected Reporting, Recordkeeping, and Other Compliance Requirements of the Regulation
      	 11-10
     11.6 Steps to Minimize Significant Economic Impact on Small Entities  	11-11
         11.6.1 Transition and Hardship Provisions for Small Engine Manufacturers	11-12
         11.6.2 Transition and Hardship Provisions for Nonroad Diesel Equipment Manufacturers 11-16
         11.6.3 Transition and Hardship Provisions for Nonroad Diesel Fuel Small Refiners	11-20
         11.6.4 Transition and Hardship Provisions for Nonroad Diesel Fuel Small Distributors and
              Marketers	11-26
     11.7 Conclusion	11-27

CHAPTER 12: Regulatory Alternatives
     12.1 Overview 	12-1
     12.2 Description of Options	12-2
         12.2.1 One-Step Options	12-3
         12.2.2 Two-Step Options	12-7

-------

-------
                                         List of Acronyms
ABT
AEO
AGME
APA
AT
BSFC
CAA
CCV
CDPF
CFR
CI
CMV
CO
DF
DI
DOC
EF
EGR
EIA
EIA
FR
FTC
GPA
GDP
HC
HD2007
hp
IDI
IRFA
kW
L&M
MPP
NAICS
NDEIM
NMHC
Averaging, Banking, and Trading
Annual Energy Outlook
Above-ground mining equipment
Administrative Procedures Act
Aftertreatment
Brake Specific Fuel Consumption
Clean Air Act
Closed crankcase ventilation
Catalyzed diesel particulate filter
Code of Federal Regulations
Compression-Ignition
Commercial Marine Vessel
Carbon monoxide
Deterioration Factor
direct injection
Diesel oxidation catalyst
Emission Factor
Exhaust gas recirculation
U. S. Energy Information Administration
Economic Impact Analysis
Federal Register
Federal Trade Commission
Geographic Phase-In Area
Gross domestic product
Hydrocarbons
Heavy-duty 2007 refers to the final rule setting emission standards for 2007 and later engines
Horsepower
Indirect injection
Initial Regulatory Flexibility Analysis
kilowatt
Locomotive and marine
marginal physical product
North American Industry Classification System
Nonroad Diesel Economic Impact Model
Non-methane hydrocarbons

-------
NPV



NR



NRLM



O&M



OMB



PADD



PM



ppm



PSR



PTD



R&D



RFA



RIA



SBA



SBAR



SBREFA



SER



SIC



stds



TAP



TPEM



ULSD



VMP



VOC



ZHL
Net present value



Nonroad



Nonroad, Locomotive, and Marine diesel fuel



operating and maintenance



Office of Management and Budget



Petroleum Administration Districts for Defense



Particulate matter



Parts per million



Power Systems Research



Product Transfer Document



Research and Development



Regulatory Flexibility Act



Regulatory Impact Analysis



Small Business Administration



Small Business Advocacy Review



Small Business Regulatory Enforcement Fairness Act



Small Entity  Representative



Standard Industrial Classification



standards



Transient Adjustment Factor



Transition program for engine manufacturers (see 40 CFR 89.102 and the proposed 40 CFR



Ultra Low Sulfur Diesel



value of marginal product



Volatile organic compounds



Zero-Hour Emission Level

-------
                                                                  Executive Summary
                             Executive Summary
   The Environmental Protection Agency (EPA) is adopting requirements to reduce emissions
of particulate matter (PM), oxides of nitrogen (NOX), and air toxics from nonroad diesel engines.
This rule includes emission standards for new nonroad diesel engines.  The rule also reduces the
level of sulfur for diesel fuels used in nonroad engines, locomotive engines, and marine engines.
The reduction in sulfur for nonroad diesel fuel will enable the use of advanced emission-control
technology that new nonroad diesel engines will use to achieve the emission reductions called
for under the engine standards in this final rule.  In addition, the reduction in sulfur will provide
important public health and welfare benefits by reducing emissions of PM and SO2 from
nonroad, locomotive and marine diesel engines.

   This executive summary describes the relevant air-quality issues, highlights the new Tier 4
emission standards and fuel requirements, and gives an overview of the analyses in the rest of
this document.

Air Quality Background and Estimated Environmental Impact of the Final Rule

   Emissions from nonroad, locomotive, and marine diesel engines contribute greatly to a
number of serious air pollution problems and would continue to do so in the future absent further
reduction measures. Such emissions lead to adverse health and welfare effects associated with
ozone, PM, NOX, SOX, and volatile organic compounds,  including toxic compounds. In addition,
diesel exhaust is of specific concern because it is likely to be carcinogenic to humans by
inhalation, as well as posing a hazard from noncancer respiratory effects.  Ozone, NOX, and PM
also cause significant public welfare harm, such as damage to crops, eutrophication, regional
haze, and soiling of building materials.

   Millions of Americans  continue to live in areas with unhealthy air quality that may endanger
public health and welfare.  There are approximately 159 million people living in areas that either
do not meet the 8-hour ozone National Ambient Air Quality  Standards (NAAQS) or contribute
to violations in other counties  as noted in EPA's recent nonattainment designations for part or all
of 474 counties.  In addition, approximately 65 million people live in counties where air quality
measurements violate the PM2 5 NAAQS. These numbers do not include the tens of millions of
people living in areas where there is a significant future  risk of failing to maintain or achieve the
ozone or PM2 5 NAAQS. Federal, state, and local governments are working to bring ozone and
PM levels into compliance with the NAAQS attainment  and maintenance plans.  The reductions
included in this final rule will play a critical part in these actions. Reducing regional emissions
of SOX is critical to this strategy for attaining the PM NAAQS and meeting regional haze goals  in
our treasured national parks.  SOX levels can themselves  also pose a respiratory hazard.

   In 1996,  emissions from land-based nonroad diesel engines,  locomotive engines, and marine


                                         ES-1

-------
Final Regulatory Support Document
diesel engines were estimated to be about 40 percent of the total mobile-source inventory of
PM2 5 (particulate matter less than 2.5 microns in diameter) and 25 percent of the NOX inventory.
Absent this final rule, these contributions would be expected to grow to 44 percent and 47
percent by 2030 for PM2 5 and NOX, respectively.  By themselves, land-based nonroad diesel
engines are a very large part of the mobile-source PM25 inventory for diesel engines,
contributing about 47 percent in 1996, and growing to 70 percent by 2020 without this final rule.

    The requirements in this rule will result in substantial benefits to public health and welfare
and the environment through significant reductions in NOX and PM, as well as nonmethane
hydrocarbons (NMHC), carbon monoxide (CO), SOX and air toxics. By 2030, this program will
reduce annual emissions of NOX and PM by 738,000 and 129,000 tons, respectively.  We
estimate these annual emission reductions will prevent 12,000 premature deaths, over 8,900
hospitalizations, 15,000 nonfatal heart attacks, and approximately 1 million days that people
miss work because of respiratory symptoms.  The overall quantifiable benefits will total over $83
billion annually by 2030, with a 30-year net present value of $805 billion.

    A comparison of the rule's quantified costs and quantified benefits indicates that estimated
benefits (approximately $80 billion per year)  are much larger than estimated costs (roughly $2
billion per year). This favorable result was found to be robust in a variety of sensitivity and
uncertainty analyses.  The favorable net benefits are particularly impressive since there are a
substantial number of health and environmental advantages of the rule that could not be
quantified.  In the final Regulatory Impact Analysis, the Agency has done extensive analysis to
identify, describe and quantify the degree of uncertainty in the benefit estimates (see Chapter 9).
This analysis suggests that the high end of the uncertainty range for this rule's estimated benefits
could exceed the low end of the range by a factor  of 20.  In addition, illustrative calculations
indicate that the uncertainty range could span two orders of magnitude using the preliminary
results of an EPA-OMB  collaborative study on expert judgment for the relative risk of mortality
from PM exposure. Despite the uncertainty inherent in the benefit-cost analysis for this rule,  the
results strongly support a conclusion that the benefits will substantially exceed costs.

Engine Emission Standards

    Tables 1 through 4 show the Tier 4 emission standards and when they apply. For most
engines, these standards  are similar in stringency to the final standards included in the 2007
highway diesel program  and are expected to require the use of high-efficiency aftertreatment
systems.  As shown in the Table 2, we are phasing in many of the standards over time to address
considerations of lead time, workload, and overall feasibility.  In addition, the final rule includes
other provisions designed to address the transition to meeting the long-term Tier 4 standards.
                                          ES-2

-------
                                                                           Executive Summary
                  Table 1—Tier 4 PM Standards (g/bhp-hr) and Schedule
Engine Power
hp<25 (kW<19)
25750 (kW>560)
Model Year
2008
0.30 a
0.22"


2009




2010




2011



0.01
2012


0.01

2013

0.02


see Table 3
Notes:
a For air-cooled, hand-startable, direct injection engines under 11 hp, a manufacturer may instead delay
implementation until 2010 and demonstrate compliance with a less stringent PM standard of 0.45 g/bhp-hr,
subject also to additional provisions discussed in section II. A3.a of the preamble.
b  A manufacturer has the option of skipping the 0.22 g/bhp-hr PM standard for all 50-75 hp engines. The 0.02
g/bhp-hr PM standard would then take effect one year earlier for all 50-75 hp engines, in 2012.
                 Table 2—Tier 4 NOx and NMHC Standards and Schedule
Engine Power
25 750 (kW>560)
Standard (g/bhp-hr)
NOx
NMHC
3.5NMHC+NOxb
0.30
0.30
0.14
0.14
Phase-in Schedule2 (model year)
2011


50%
2012

50% c
50%
2013
100%
50% c
50%
2014

100%c
100%
see Table 3
Notes:
a  Percentages indicate production required to comply with the Tier 4 standards in the indicated model year.
b  This is the existing Tier 3 combined NMHC+NOx standard level for the 50-75 hp engines in this category. In
2013 it applies to the 25-50 hp engines as well.
0  Manufacturers may use banked Tier 2 NMHC+NOx credits to demonstrate compliance with the 75-175 hp
engine NOx standard in this model year. Alternatively, manufacturers may forego this special banked credit
option and instead meet an alternative phase-in requirement of 25/25/25% in 2012, 2013, and 2014 through
December 30, with 100% compliance required beginning December 31, 2014.  See sections III.A and II.A.2.b of
the preamble.
                                               ES-2

-------
Final Regulatory Support Document
             Table 3 - Tier 4 Alternative NOx Phase-in Standards (g/bhp-hr)
Engine Power
75 < hp < 175 (56 < kW<130)
175 < hp < 750 (130 < kW < 560)
NOx Standard
(g/bhp-hr)
1.7 a
1.5
 Notes:
 "Under the option identified in footnote b of Table 2, by which manufacturers may meet an alternative phase-in
 requirement of 25/25/25% in 2012, 2013, and 2014 through December 30, the corresponding alternative NOx
 standard is 2.5 g/bhp-hr.
              Table 4—Tier 4 Standards for Engines Over 750 hp (g/bhp-hr)

engines used in:
generator sets <1200 hp
generator sets >1200 hp
all other equipment
2011
PM
0.075
0.075
0.075
NOx
2.6
0.50
2.6
NMHC
0.30
0.30
0.30
2015
PM
0.02
0.02
0.03
NOx
0.50
no new
standard
no new
standard
NMHC
0.14
0.14
0.14
   EPA has also taken steps to ensure that engines built to these standards achieve effective real-
world emission control including the transient duty cycle (both cold-start and hot-start testing),
steady-state duty cycles, and Not-to-Exceed standards and test procedures.  The Not-to-Exceed
provisions are modeled after the highway program, with which much of the industry has gained
some level  of experience.

Feasibility of Meeting Tier 4 Emission Standards

       For the past 30 or more years, emission-control development for gasoline vehicles and
engines has concentrated most aggressively on aftertreatment technologies (i.e., in-exhaust
catalyst technologies). These devices currently provide as much as or more than 95 percent of
the emission control on a gasoline vehicle.  In contrast, the emission-control development work
for highway and nonroad diesel engines has concentrated on improvements to the engine itself to
limit the emissions formed in the engine (engine-out control technologies).

   During  the past 15 years, however, more development effort has been put into catalytic
exhaust emission-control devices for diesel engines, particularly in the area of particulate matter
(PM) control. Those developments, and recent developments in diesel NOx exhaust emission-
control devices, make the widespread commercial use of highly efficient diesel exhaust emission
controls feasible.  EPA has recently set new emission standards for diesel engines installed in
                                          ES-4

-------
                                                                   Executive Summary
highway vehicles based on the emission-reduction potential of these devices. These devices will
also make possible a level of emission control for nonroad diesel engines that is similar to that
attained by gasoline catalyst systems. However, without the same ultra-low-sulfur diesel fuel
that will be used by highway engines, these technologies cannot be implemented.

    The primary focus of the Tier 4 program is the transfer of catalyst based emission control
technologies developed for on-highway diesel engines to nonroad engines. This RIA
summarizes extensive analyses evaluating the effectiveness  of these new emission control
technologies and the specific challenges to further develop these technologies for nonroad
applications.  The RIA concludes that for a very significant fraction of nonroad diesel engines
and equipment, the application of advanced catalyst based emission control technology is
feasible in the Tier 4 timeframe given the availability of 15 ppm sulfur diesel fuel.

    Although the primary focus of the Tier 4 emissions program and the majority of the analyses
contained in this RIA are directed at the application of catalytic emission control technologies
enabled by 15 ppm sulfur diesel fuel, there are also important elements of the program based
upon continuing improvements in engine-out emission controls.  Like the advanced catalytic
based technologies, these engine-out emission solutions for nonroad diesel engines rely upon
technologies already applied to on-highway diesel engines.  Additionally, these technologies
form the basis for the Tier 3 emission standards for some nonroad diesel engines in other size
categories.

Controls on the Sulfur Content of Diesel Fuel

    We are finalizing the a two-step sulfur standard for nonroad, locomotive and marine
(NRLM) diesel fuel that will achieve significant, cost-effective sulfate PM and SO2 emission
reductions. These emission reductions will, by themselves, provide dramatic environmental and
public health benefits which far outweigh the cost of meeting the standards necessary to achieve
them. In addition, the final sulfur standards for nonroad diesel fuel will enable advanced high
efficiency emission control technology to be applied to nonroad engines. As a result, these
nonroad fuel sulfur standards, coupled with our program for more stringent emission standards
for new nonroad engines and equipment, will also achieve dramatic NOx and PM emission
reductions. Sulfur significantly inhibits or impairs the function of the diesel exhaust emission
control devices which will generally be necessary for nonroad diesel engines to meet the
emission standards in this final rule.  With the 15 ppm sulfur standard for nonroad diesel fuel, we
have concluded that this emission control technology will be available for model year 2011 and
later nonroad diesel engines to achieve the NOx and PM emission standards adopted in this final
rule. The benefits of this final rule also include the sulfate PM and SO2 reductions achieved by
establishing the same standard for the sulfur content of locomotive and marine diesel fuel.

    The fuel sulfur requirements established under this final rule are similar to the sulfur limits
established for highway diesel fuel in prior rulemakings - 500 ppm in 1993 ( 55 FR 34120,
August 21, 1990) and  15 ppm in 2006 (66 FR 5002, January 18, 2001). Beginning June 1, 2007,
refiners will be required to produce NRLM diesel fuel with a maximum sulfur content of 500

                                         ES-5

-------
Final Regulatory Support Document
ppm. Then, beginning June 1, 2010, the sulfur content will be reduced for nonroad diesel fuel to
a maximum of 15 ppm. The sulfur content of locomotive and marine diesel fuel will be reduced
to 15 ppm beginning June 1, 2012. The program contains certain provisions to ease refiners'
transition to the lower sulfur standards and to enable the efficient distribution of all diesel fuels.

   The final program also contains provisions to smooth the refining industry's transition to the
low sulfur fuel requirements, encourage earlier introduction of cleaner burning fuel, maintain the
fuel distribution system's flexibility to fungibly distribute similar products, and provide an outlet
for off-specification distillate product.  These provisions, which will maintain, and even enhance,
the health and environmental benefits of this rule, include the 2012 date for locomotive and
marine diesel fuel, early credits for refiners and importers and special provisions for small
refiners, transmix processors, and entities in the fuel distribution system.

Feasibility of Meeting Diesel Fuel Sulfur Standards

       We conclude that it is feasible for refiners to meet the 500 ppm and 15 ppm sulfur cap
standards for nonroad, locomotive and marine diesel fuel (NRLM).  We project that refiners will
use conventional desulfurization technology for complying with the 500 ppm sulfur standard in
2007, which is the same technology used to produce 500 ppm sulfur highway diesel fuel today.
Refiners complying with the 500 ppm sulfur NRLM diesel fuel standard will have about the
same amount of lead time refiners had in complying with the highway diesel fuel  standard, when
it took affect in 1993, and they can draw on their experience gained from complying with the
1993 highway sulfur standard. Thus we conclude that refiners producing 500 ppm NRLM diesel
fuel will have sufficient leadtime. For complying with the 15 ppm  sulfur cap standards
applicable to nonroad diesel fuel in 2010 and to locomotive and marine diesel fuel in 2012,
refiners will be able to use the experience gained from complying with the 15 ppm highway
diesel fuel standard which begins to take effect in 2006. Furthermore, refiners will have ample
lead time of at least six years before they will have to begin to produce 15 ppm sulfur nonroad
diesel fuel. For complying with both the 15 ppm sulfur standard for nonroad diesel fuel in 2010
and the locomotive and marine diesel fuel  in 2012, we expect many refiners to utilize lower cost
advanced desulfurization technologies which have recently been commercialized.  Others will
rely on extensions of conventional hydrotreating technology which most refiners are planning on
using to comply with the 15 ppm cap for highway diesel fuel in 2006.  These technologies will
enable refiners to achieve the 15 ppm NRLM sulfur standards.

       We do not expect any new significant issues regarding the feasibility of distributing
NRLM fuels that meet the sulfur standards in this rule. The highway diesel program
acknowledged that limiting sulfur contamination during the distribution of 15 ppm diesel fuel
would be a significant challenge to industry. Industry is already taking the necessary steps to
rise to this challenge to distribute highway diesel fuel meeting a 15 ppm sulfur standard by the
2006 implementation date for this standard. Thus, we believe that any issues regarding limiting
sulfur contamination during the distribution of 15 ppm sulfur nonroad, and locomotive/marine
diesel fuel will have been resolved a  number of years before the implementation of the  15 ppm
sulfur standard for these fuels (in 2010 and 2012 respectively).

                                         ES-6

-------
                                                                   Executive Summary
       The fuel program in this rule is structured in such a way to maximize fuel fungibility and
minimize the need for additional segregation of products in the fuel distribution system. Thus,
this rule will only result in the need for a limited number of additional storage tanks at terminals
and bulk plants in the interim, and in the long run will result in a simplified overall product slate
that needs to be distributed.
Estimated Costs and Cost-Effectiveness

       There are approximately 600 nonroad equipment manufacturers using diesel engines in
several thousand different equipment models.  There are more than 50 engine manufacturers
producing diesel engines for these applications.  Fixed costs consider engine research and
development, engine tooling, engine certification, and equipment redesign.  Variable costs
include estimates for new emission-control hardware. Near-term and long-term costs for some
example pieces of equipment are shown in Table 5. Also shown in Table 5  are typical prices for
each piece of equipment for reference. See Chapter 6 for detailed information related to our
engine and equipment cost analysis.

      Table 5— Long-Term Costs for Several Example Pieces of Equipment ($2002)a

Horsepower
Displacement (L)
Incremental Engine &
Equipment Cost
Long Term
Near Term
Estimated Equipment
Priceb
GenSet
9hp
0.4
$120
$180
$4,000
Skid/Steer
Loader
33 hp
1.5
$790
$1,160
$20,000
Backhoe
76 hp
3.9
$1,200
$1,700
$49,000
Dozer
175 hp
10.5
$2,560
$3,770
$238,000
Agricultural
Tractor
250 hp
7.6
$1,970
$3,020
$135,000
Dozer
503 hp
18
$4,140
$6,320
$618,000
Off-
Highway
Truck
1000 hp
28
$4,670
$8,610
$840,000
1 Near-term costs include both variable costs and fixed costs; long-term costs include only variable costs and represent
       those costs that remain following recovery of all fixed costs.
       Our estimated costs related to changing to ultra-low-sulfur fuel take into account all of
the necessary changes in both refining and distribution practices. We have estimated the cost of
producing 500 ppm sulfur NRLM fuel to be, on average, 2.1 to 3.5 cents per gallon.  Average
costs for 15 ppm sulfur NR fuel during the years 2010 through 2012 are estimated to be an
additional 2.5 cents per gallon for a combined cost of 5.8 cents per gallon.  Average costs for 15
ppm sulfur NRLM fuel are estimated to be an additional 1.2 cents per gallon for a combined cost
of 7.0 cents per gallon for the years 2014 and beyond.  All of these fuel costs are summarized in
Table 6.  These ranges consider variations in regional issues in addition to factors that are
specific to individual refiners.  In addition, engines running on low-sulfur fuel will have reduced
maintenance expenses that we estimate will be equivalent to reducing the cost of the fuel by 2.9
                                          ES-7

-------
Final Regulatory Support Document
to 3.2 cents per gallon.
                    Table 6—Increased Cost of Providing Nonroad,
          Locomotive and Marine Diesel Fuel (cents per gallon of affected fuel
Specification
500 ppm NRLM
500 ppm NRLM
500 ppm NRLM
1 5 ppm Nonroad
15 ppm NRLM
15 ppm NRLM
Year
2007-10
2010-12
2012-14
2010-12
2012-14
2014+
Refining Costs
(c/gal)
1.9
2.7
2.9
5.0
5.6
5.8
Distribution &
Additive Costs (c/gal)
0.2
0.6
0.6
0.8
0.8
1.2
Total Costs
(c/gal)
2.1
3.3
3.5
5.8
6.4
7.0
       Chapter 8 describes the analysis of aggregating the incremental fuel costs, operating
costs, and the costs for producing compliant engines and equipment, operating costs. Table 7
compares these aggregate costs with the corresponding estimated emission reductions to present
cost-per-ton figures for the various pollutants.

       Table 7—Aggregate Cost per Ton for the Proposed Two-Step Fuel Program
    and Engine Program—2004-2036 Net Present Values at 3% Discount Rate ($2002)
Pollutant
NOx+NMHC
PM
SOX
Aggregate Discounted Lifetime
Cost per ton
$1,010
$11,200
$690
Economic Impact Analysis

       As described in Chapter 10, we prepared an Economic Impact Analysis (EIA) to estimate
the economic impacts of this rule on producers and consumers of nonroad engines and
equipment and fuels, and related industries. The EIA has two parts: a market analysis and a
welfare analysis.  The market analysis explores the impacts of the proposed program on prices
and quantities of affected products. The welfare analysis focuses on changes in social welfare
and explores which entities will bear the burden of the proposed program.  The EIA relies on the
Nonroad Diesel Economic Impact Model (NDEIM).  The NDEEVI uses a multi-market analysis
framework that considers interactions between 62 regulated markets and other markets to
estimate how compliance costs can be expected to ripple through these markets.

       As shown in Table 8, the market impacts of this rule suggest that the overall economic

                                        ES-8

-------
                                                                   Executive Summary
impact on society is expected to be small, on average. According to this analysis, price increases
of goods and services produced using equipment and fuel affected by this rule (the application
marktets) are expected to average about 0.1 percent per year.  Output decrease in the application
markets are expected to average less than 0.02 percent for all years. The price increases for
engines, equipment, and fuel are expected to be about 20 percent, 3 percent, and 7 percent,
respectively (total impact averaged over the relevant years). The number of engines and
equipment produced annually is expected to decrease by less than 250 units, and the amount of
fuel produced annually is expected to decrease by less than 4 million gallons.

             Table 8—Summary of Expected Market Impacts, 2013 and 2020
Market
Engines
Equipment
Application
markets*
Nonroad Fuel
Markets
Loco/Marine
Transportation
2013
Average
engineering
cost per unit
$1,052
$1,198
—
$0.06
—
Price change
21.4%
2.9%
0.10%
6.0%
0.01%
Quantity
change
-0.014%
-0.017%
-0.015%
-0.019%
-0.007
2036
Average
engineering
cost per unit
$931
$962
—
$0.07
—
Price change
18.2%
2.5%
0.10%
7.0%
0.01%
Quantity
change
-0.016%
-0.018%
-0.016%
-0.022%
-0.008
Commodities in the application markets are normalized; only percentage changes are presented
       The welfare analysis predicts that consumers and producers in the application markets are
expected to bear the burden of this proposed program. In 2013, the total social costs of the rule
are expected to be about $1.5 billion.  About 83 percent of the total social costs is expected to be
borne by producers and consumers in the application markets, indicating that the majority of the
costs associated with the rule are expected to be passed on in the form of higher prices. When
these estimated impacts are broken down, 58.5 percent are expected to be borne by consumers in
the application markets and 41.5 percent are expected to be borne by producers in the application
markets. Equipment manufacturers are expected to bear about 9.5 percent of the total social
costs. These are primarily the costs associated with equipment redesign. Engine manufacturers
are expected to bear about 2.8 percent; this is primarily the fixed costs for R&D.  Nonroad fuel
refiners are expected to bear about 0.5 percent of the total social costs.  The remaining 4.2
percent is accounted for by locomotive and marine transportation services.

       Total social costs continue to increase over time and are projected to be about $2.0 billion
by 2030 and $2.2 billion in 2036 ($2002). The increase is due to the projected annual growth in
the engine and equipment populations. Producers and consumers in the application markets are
                                          ES-9

-------
Final Regulatory Support Document
expected to bear an even larger portion of the costs, approximately 96 percent.  This is consistent
with economic theory, which states that, in the long run, all costs are passed on to the consumers
of goods and services.

Impact on Small Businesses

       Chapter 11 discusses our Final Regulatory Flexibility Analysis, which evaluates the
potential impacts of new engine standards and fuel controls on small entities. Before issuing our
proposal, we analyzed the potential impacts of this rule on small entities.  As a part of this
analysis, we interacted with several small entities representing the various affected sectors and
convened a Small Business Advocacy Review Panel to gain feedback and advice from these
representatives. This feedback was used to develop regulatory alternatives to address the
impacts of the rule on small businesses. Small entities raised general concerns related to
potential difficulties and costs of meeting the upcoming standards.

       The Panel consisted  of members from EPA, the Office of Management and Budget, and
the  Small Business Administration's Office of Advocacy. We either proposed or requested
comment on the Panel's recommendations.  Chapter 11 discusses the options recommended in
the  Panel Report, the regulatory alternatives we considered in the proposal, and the provisions
we  are adopting in the final  rule. We have adopted several provisions that give small  engine and
equipment manufacturers and small refiners several compliance options aimed specifically at
educing the burden on these small entities.  In general the options are similar to small entity
provisions adopted in prior rulemakings where EPA  set standards for nonroad diesel engines and
controlled the level of sulfur in highway gasoline and diesel fuel. These provisions will reduce
the  burden on small entities that must meet this rule's requirements.

Alternative Program Options

       In the course of developing our final program, we investigated several alternative
approaches to both the engine and fuel programs. These alternative program options included
variations in:
             The applicability of aftertreatment-based standards for different horsepower
             categories
             The phase-in schedule for engine standards
             The start date for the diesel fuel sulfur standard
       •     The use of a single-step instead of a two-step approach to fuel sulfur standards
       •     The applicability of the very-low fuel sulfur standards to fuel used by locomotives
             and marine engines

       Chapter 12 includes  a complete description of twelve alternative program options.  The
draft RIA contained an assessment of technical feasibility, cost, cost-effectiveness, inventory
impact, and health and welfare benefits for each alternative.  We refer the reader to the detailed
evaluations of the options presented in the Draft Regulatory Impact Analysis.
                                         ES-10

-------
CHAPTER 1: Industry Characterization
   1.1 Characterization of Engine Manufacturers  	1-1
       1.1.1 Engines Rated between 0-19 kW (0 and 25 hp)	1-2
       1.1.2 Engines Rated between 19 and 56 kW (25 and 75 hp)	1-2
       1.1.3 Engines Rated between 56 and 130 kW (75 and 175 hp)	1-2
       1.1.4 Engines Rated between 130 and 560 kW (175 and 750 hp)	1-2
       1.1.5 Engines Rated over 560 kW (750 hp)	1-3
   1.2 Characterization of Equipment Manufacturers 	1-3
       1.2.1 Equipment Using Engines Rated under 19 kW (0 and 25 hp)  	1-4
       1.2.2 Equipment Using Engines Rated between 19 and 56 kW (25 and 75 hp) 	1-6
       1.2.3 Equipment Using Engines Rated between 56kW and 130 kW (75 and 175 hp)  . 1-7
       1.2.4 Equipment Using Engines Rated between 130 and 560 kW (175 and 750 hp)  . . 1-9
       1.2.5 Equipment Using Engines Rated over 560 kW (750 hp)	1-11
   1.3 Refinery Operations	1-12
       1.3.1 The Supply-Side	1-12
       1.3.2 The Demand Side	1-19
       1.3.3 Industry Organization	1-26
       1.3.4 Markets and Trends  	1-30
   1.4 Distribution and Storage Operations	1-35
       1.4.1 The Supply-Side	1-35
       1.4.2 The Demand-Side	1-37
       1.4.3 Industry Organization	1-37
       1.4.4 Markets and Trends  	1-38

-------
                                                            Industry Characterization
              CHAPTER 1: Industry Characterization

   In understanding the impact of emission standards on regulated industries, it is important to
assess the nature of the regulated and otherwise affected industries.  The industries affected are
the nonroad diesel engine and equipment manufacturing, oil-refining, and fuel-distribution
industries. This chapter provides market share information for the above industries.  This
information is provided for background purposes. The information presented in this chapter will
be most helpful for those unfamiliar with the engine/equipment industry and/or the oil refining
and fuel-distribution industries.

   Nonroad engines are generally distinguished from highway engines in one of four ways:  (1)
the engine is used in a piece of motive equipment that propels itself in addition to performing an
auxiliary function (such as a bulldozer grading a construction site); (2) the engine is used in a
piece of equipment that is intended to be propelled as it performs its function (such as a
lawnmower); (3) the engine is used in a piece of equipment that is stationary when in operation
but portable ( such as a generator or compressor) or (4) the engine is used in a piece of motive
equipment that propels itself, but is primarily used for off-road functions (such as an off-highway
truck).

   The nonroad  category is also different from other mobile source categories because: (1) it
applies to a wider range of engine sizes and power ratings; (2) the pieces of equipment in which
the engines are used are extremely diverse; and (3) the same engine can be used in widely
varying equipment applications (for example, the same engine used in a backhoe can also be
used in a drill rig or in an air compressor).

   A major consideration in regulating nonroad engines is the lack of vertical integration in this
field. Although some nonroad engine manufacturers also produce equipment that rely on their
own engines, most engines are sold to various equipment manufacturers over which the original
engine manufacturer has minimal control. A characterization of the industry affected by this
rulemaking must therefore include equipment manufacturers as well as engine manufacturers.

   Sections  1.1 and 1.2 characterize the nonroad engine and equipment industries based on
different manufacturers and their products and the diversity of the manufacturer pool for the
various types of equipment. They describe the nonroad diesel engine market and related
equipment markets by power category.  Additional information related to engine/equipment
profiles, including employment figures, production costs, information  on engine component
materials and firm characteristics, are available in the docket.1

1.1 Characterization of Engine Manufacturers

   For purposes of discussion, the characterization of nonroad engine manufacturers is arranged
by the power categories used to define the new emission standards.  The information detailed in
this section was derived from the Power Systems Research database and trade journals.2 We

                                          1-1

-------
Final Regulatory Support Document
recognize that the PSR database is not comprehensive, but have not identified a better source to
provide consistent data for identifying additional companies. The sales figures presented in this
chapter pertain to both mobile and stationary nonroad equipment. The former forms the bases
for cost and other analyses such as included in Chapters 6 and 10.

1.1.1 Engines Rated between 0-19 kW (0 and 25 hp)

   In year 2000, sales of engines in this category comprised 18 percent (approximately 135,828
units) of the nonroad market. The largest manufacturers of engines in this category are Kubota
(36,601 units) and Yanmar (32,126 units). Seventy three percent of Yanmar's engines are four-
cycle, water-cooled, indirect-injection models. A majority of Kubota's engines are also four-
cycle, water-cooled indirect-injection models. Another major manufacturer in this category is
Kukje with 21,216 units.

1.1.2 Engines Rated between 19 and 56 kW (25 and 75 hp)

   This is the largest category, comprised of 38 percent of engines with approximately 281,157
units sold in year 2000.  Direct-injection (DI) engines account for 59 percent of this category
with 165,427 units. Yanmar has  approximately 19 percent of the DI market share, followed by
Deutz (16%), Kubota (13%), Hatz (12%), Isuzu(10%) ,Caterpillar/Perkins(10% ) and Deere
(8%). Kubota dominates the Indirect-injection (IDI) market with 51 percent of sales , followed
by Daewoo Heavy Industries (12%), Ihi-Shibaura (12%), Isuzu(8%) and Caterpillar/Perkins
(5%). Ag tractors, generator sets, skid-steer loaders and refrigeration and air conditioning units
are the largest selling engines in this power range.

1.1.3 Engines Rated between 56 and 130  kW (75 and 175 hp)

   In year 2000, manufacturers sold approximately  206,028 engines in this power range.  This
represents the second-largest category of nonroad engines with 28 percent of the total  market.
Almost all  of these engines are DI. The top three manufacturers are John Deere (28%),
Caterpillar/Perkins (20%) and Cummins (17%).  Other manufacturers include Case/ New
Holland, Deutz, Hyundai Motor, Isuzu, Toyota and Komatsu. The engines in this power range
are used  mostly in agricultural equipment such as ag tractors. The second-largest use for these
engines is in construction equipment such as tractor/loader/backhoes and skid-steer loaders.

1.1.4 Engines Rated between 130 and 560 kW (175 and 750 hp)

   Engines in this power range rank fourth (15% of the total market) in nonroad diesel engines
sales with approximately 108,172 units sold in year 2000. Almost all  of these are DI engines.
Deere has approximately 32 percent of the DI market, followed by Caterpillar/Perkins (22%),
Cummins (21%), Case/New Holland (8%),Volvo (4%), and then by Komatsu and Detroit Diesel
(each 3%). The  largest selling engines in this category are used in agricultural equipment (ag
tractors),  followed by construction equipment (wheel loaders, bulldozers, and excavators).
                                          1-2

-------
                                                           Industry Characterization
1.1.5 Engines Rated over 560 kW (750 hp)

   This is the smallest nonroad category with approximately 5,633 engines comprising 1
percent of the total nonroad market and consist of all DI engines.  Caterpillar is the largest
manufacturer (44%), followed by Cummins (19%), Komatsu (18%), and Detroit Diesel (11%).
Power generation is the principal application in this range, followed by large off-highway trucks
and other types of construction equipment such as crawlers , wheel loaders and bulldozers.

1.2 Characterization of Equipment Manufacturers

   Nonroad equipment can be grouped into several categories.  This section considers the
following seven segments: agriculture, construction, general industrial, lawn and garden,
material handling, pumps and compressors, and welders and generator sets. Engines used in
locomotives, marine applications, aircraft, recreational vehicles, underground mining equipment,
and all spark-ignition engines within the above categories are not included in this rulemaking.
Table 1.2-1 has examples of the types of nonroad equipment that will be impacted by this
rulemaking, arranged by category.

                                      Table 1.2-1
                      Sampling of Nonroad Equipment Applications
Segment
Agriculture
Construction
General Industrial
Lawn and Garden
Pumps and Compressors
Material Handling
Welders and Generators
Applications
Ag Tractor
Baler
Combine
Bore/drill Rig
Crawler
Excavator
Grader
Off-highway Tractor
Concrete/Ind. Saw
Crushing Equipment
Lawn and Garden
Tractor
Air Compressor
Hydro Power Unit
Pressure Washer
Aerial Lift
Crane
Generator Set, Welder
Sprayer
Windrower
Other Ag Equipment
Off -highway Truck
Paver
Plate Compactor
Roller
Wheel Loader/Dozer
Oil Field Equipment
Refrigeration/ AC
Commercial Mower
Pump
Gas Compressor
Forkhft
Terminal Tractor
Lt Plant/Signal Board

T amper/Rammer
Scraper
Skid- Steer Loader
Trencher
Scrubber/sweeper
Rail Maintenance
Trimmer/edger/cutter
Irrigation Set
Rough-Terrain Forklift

   Based on power rating rating of the engines, a fraction of applications such as air
compressors, generator sets, hydropower units, irrigation sets, pumps and welders is considered
                                          1-2

-------
Final Regulatory Support Document
to be stationary and therefore not subject to EPA emission standards for nonroad engines.
However, the tables in Sections 1.2.1 to 1.2.5 account for all equipment manufactured, whether
stationary or mobile within an engine power category.

   For purposes of discussion, nonroad equipment is grouped into five power ranges similar to
those used for characterizing nonroad engines. This section explores the characteristics of
nonroad equipment applications and the companies involved in manufacturing these equipment.
This analysis includes several numerical summaries of different categories.

1.2.1 Equipment Using Engines Rated under 19 kW (0 and 25 hp)

   The applications with the most sales are ag tractors followed by generator sets. There are
about 29 total applications with engines rated under 19 kW. The six leading manufacturers
produce 46 percent of the equipment in this category. Their collective sales volume over five
years (1996 to 2000) was approximately 251,000 pieces of equipment in a market that has a
five-year total sales volume of 551,000. These manufacturers and the major equipment types
manufactured by them are shown in Table 1.2-2.

                                       Table 1.2-2
                         Characterization of the  Top 6 Equipment
                      Manufacturers for Engines Rated below 19 kW
Original Equipment
Mnniifnp.tiirpr
Ingersoll-Rand
Deere & Company
Korean Gen-sets
China Gen-sets
SDMO
Kubota Corp.
Major Equipment Manufactured
Refrigeration/AC, Skid-steer loaders,
and Excavators
Agricultural tractors, Commercial
mowers, Lawn & garden tractors
Generator Sets
Generator Sets
Generator Sets
Ag tractors,Lawn & garden tractors
Commercial mowers
Average
Annual Sinlps
13,394
11,042
9,970
5,559
5,191
5,117
Percentage
of Mnrlcpt
12%
10%
9%
5%
5%
5%
Engine
P.hflrnp.tpriyntinn*
W,NA, I
W,NA, I
W,NA, I
W,NA,D/ 1
W/A,NA, D/I
W,NA,I
*W=water-cooled, A=air-cooled,O=oil cooled;NA=naturally aspirated,T=turbocharged;I=indirect
injection,D=direct injection.
    Sales for these top six OEMs are typified  by generator sets, skid-steer loaders, ag tractors,
commercial mowers, and refrigeration/air conditioning units.  The sales of the equipment are
listed in Table 1.2-3. The top six manufacturers have equipment that are typical of the market.
Fifty-six OEMs produce 92 percent of the equipment in this power range.
                                           1-4

-------
                                          Industry Characterization
                      Table 1.2-3
Equipment Sales Distribution for Engines Rated below 19 kW
Application Description

Generator sets
Agricultural tractors
Commercial mowers
Refrigeration/AC
Welders
Light plants/Signal boards
Skid-steer loaders
Lawn & garden tractors
Pumps
Rollers
Pressure washers
Plate compactors
Utility vehicles
Aerial lifts
Excavators
Mixers
Scrubbers/sweepers
Commercial turf equipment
Finishing equipment
Other general industrial equipment
Tampers/rammers
Tractor/loader/backhoes
Dumpers/tenders
Air compressors
Hydraulic power units
Trenchers
Concrete/industrial saws
Irrigation sets
Wheel loaders/bulldozers
Other agricultural equipment
Surfacing equipment
Bore/drill rigs
Listed Total
Grand Total
Five-year sales Volume
n 996.7000s!
171,435
59,863
59,713
57,668
32,284
28,239
23,685
17,879
16,262
12,063
11,959
11,535
8,502
7,058
6,118
4,639
2,829
2,627
2,351
2,334
2,156
1,794
1,689
1,516
797
776
733
614
502
426
362
275


Average Annual
Sains
34,287
11,973
11,943
11,534
6,457
5,648
4,737
3,576
3,252
2,413
2,392
2,307
1,700
1,412
1,224
928
566
525
470
467
431
359
338
303
159
155
147
123
100
85
72
55
110,137
110,289
Percentage of Total
Sains
31.1
9.5
9.5
9.2
5.1
4.5
3.8
2.8
2.6
1.9
1.9
1.8
1.4
1.1
1.0
0.7
0.4
0.4
0.4
0.4
0.3
0.3
0.3
0.2
0.1
0.1
0.1
0.1
0.1
0.1
0.1
0.0
91.4
100.0
                         1-5

-------
Final Regulatory Support Document
1.2.2 Equipment Using Engines Rated between 19 and 56 kW (25 and 75 hp)

   All market segments are represented within the 19 to 56 kW range. They are made up of 55
applications and about 17 percent of total sales are by Ingersoll- Rand. For the 19 to 56 kW range,
the equipment uses either direct-injection or indirect-injection engines that are water-cooled or oil-
cooled and are either naturally aspirated or turbocharged. The six leading manufacturers produce
53 percent of the equipment in this category.  These manufacturers are listed in Table 1.2-4. They
manufacture equipment typical of the market, such as agricultural tractors, generator sets, skid-steer
loaders and refrigeration/AC.  These top selling applications represent about 70 percent of the
market as seen in Table 1.2-5.  The top 90 percent of the market is supplied  by 60 different
companies.

                                       Table 1.2-4
                          Characterization of the Top 6 Equipment
                  Manufacturers for Engines Rated between 19 and 56 kW
Original Equipment Manufacturer
Ingersoll-Rand
Case New Holland
Thermadyne Holdings
Deere & Company
Kubota Corp.
United Technologies Co.
Major Equipment Manufactured
Refrigeration A/C, Skid-steer
loaders, Air compressors
Agricultural tractors, Skid-steer
loaders
Generator sets
Agricultural tractors, Skid-steer
loaders, Commercial mowers
Agricultural tractors, Excavators,
Wheel Loaders, Bulldozers
Refrigeration/AC
Average
Annual Sinlps
40,199
23,194
19,090
17,752
14,391
12,484
Percentage of
"Mark-fit
17%
10%
8%
7%
6%
5%
Engine
Phflrnp.tpriyntinn*
W/O,NA/T,D/I
W/O,NA/T,D/I
A,NA,D
W,NA/T,D
W,NA/T,D/I
W,NA,D/I
    *W=water-cooled, A=air-cooled,O=oil cooled;NA=naturally aspirated, T=turbocharged, I=indirect injection,
    D=direct injection.
                                           1-6

-------
                                                           Industry Characterization
                                      Table 1.2-5
          Equipment Sales Distribution across Applications between 19 and 56 kW
Application Description
Agricultural tractors
Generator sets
Skid-steer loaders
Refrigeration/AC
Welders
Commercial mowers
Air compressors
Trenchers
Aerial lifts
Forklifts
Rollers
Excavators
Rough terrain forklifts
Scrubbers/sweepers
Light plants/signal boards
Pumps
Bore/drill rigs
Utility vehicles
Wheel Loaders/bulldozers
Pressure washers
Pavers
Commercial turf
Tractor/loader/backhoes
Irrigation sets
Concrete/industrial saws
Other general industrial
Chippers/grinders
Crushing/processing equipment
Hydraulic power units
Terminal tractors
Surfacing equipment
Dumpers/tenders
Listed Total
Grand Total
Five-year sales
Volume
n 996.7000s!
286,295
223,960
177,925
142,865
60,035
47,735
33,840
26,465
25,810
23,480
18,010
16,485
13,530
11,770
11,720
9,290
9,000
8,460
6,985
6,700
6,395
5,760
5,115
4,300
3,400
3,400
2,625
2,305
1,950
1,765
1,490
1,055


Average Annual
Sales
57,259
44,792
35,585
28,573
12,007
9,547
6,768
5,293
5,162
4,696
3,602
3,297
2,706
2,354
2,344
1,858
1,800
1,692
1,397
1,340
1,279
1,152
1,023
860
680
680
525
461
390
353
298
211
239,984
241,710
Percentage of
Total Sales
24%
19%
15%
12%
5.0%
3.9%
2.8%
2.2%
2.1%
1.9%
1.5%
1.4%
1.1%
1.0%
1.00%
0.77%
0.74%
0.70%
0.58%
0.55%
0.53%
0.48%
0.42%
0.36%
0.28%
0.28%
0.22%
0.19%
0.16%
0.15%
0.12%
0.09%
99.3%
100.0%
1.2.3 Equipment Using Engines Rated between 56kW and 130 kW (75 and 175 hp)

   Engines rated between 56 and 130 kW are all direct-injection engines that are either water-
cooled (94% ), oil-cooled (4%) or air-cooled (2%).  The six leading manufacturers produce 49
percent of the equipment in this category. Their collective sales volume over five years (1996 to
2000) was approximately 440,000 pieces of equipment in a market that has a five-year total sales
volume of 905,000. These manufacturers are shown in Table 1.2-6.
                                         1-7

-------
Final Regulatory Support Document
                                        Table 1.2-6
                          Characterization of the Top 6 Equipment
         Manufacturers for Engines Rated between 56kW and 130 kW (75 and 175 hp)
Original Equipment
Manufacturer
Case New Holland
Deere & Company
Caterpillar
Ingersoll-Rand
Agco
Landini Holding
Major Equipment Manufactured
Ag Tractors, Combines, Crawlers, Skid-steer
loaders, Tractors/loaders/backhoes
Ag Tractors, Combines, Wheel
Loaders/Dozers
Generator Sets, Scrapers, Crawlers,
Excavators, Wheel loaders, bulldozers,
Graders, Rough terrain fork-lifts
Air compressors, Rollers, Bore/drill rigs
Agricultural tractors, Combines, Sprayers
Agricultural tractors
Average
Annual Sains
26,717
25,648
13,670
10,169
6,182
5,467
Percentage of
Market
15%
14%
8%
6%
3%
3%
Engine
Charap.tnriyation*
W,T,D
W,T,D
W,T/N,D
W,T,D
W/A,T,D
W,T/N,D
    *W=water-cooled, A=air-cooled,O=oil cooled;NA=naturally aspirated, T=turbocharged, I=indirect injection,
    D=direct injection.
    Sales of these top six OEMs are typified by agricultural tractors, tractors/loaders/backhoes,
generator sets, skid-steer loaders, rough terrain fork-lifts, excavators, air compressors and
crawlers. The sales of these equipment are listed in Table 1.2-7.  The top six manufacturers have
engines that are typical of the market.  Seventy-two OEMs produce 90 percent of the equipment
in this power range.

-------
                                                            Industry Characterization
                                      Table 1.2-7
          Equipment Sales Distribution across Applications between 56 and 130 kW
Application Description
Agricultural tractors
Tractor/loader/backhoes
Generator sets
Skid-steer loaders
Rough terrain forklfts
Excavators
Air compressors
Crawlers
Forklifts
Wheel Loaders/bulldozers
Rollers
Commercial turf equipment
Other general industrial
Scrubbers/sweepers
Irrigation sets
Windrowers
Pumps
Sprayers
Listed Total
Grand Total
Five-yr sales Volume
n 996.7000s!
185,315
106,780
103,490
74,040
56,770
50,140
32,080
30,260
29,705
27,520
23,195
17,425
16,580
16,005
15,745
11,385
10,265
8,830


Average
Annual Sains
37,063
21,356
20,698
14,808
11,354
10,028
6,416
6,052
5,941
5,504
4,639
3,485
3,316
3,201
3,149
2,277
2,053
1,766
163,108
181,094
Percentage of
Total Sains
20%
12%
11%
8.2%
6.3%
5.5%
3.5%
3.3%
3.3%
3.0%
2.6%
1.9%
1.8%
1.8%
1.7%
1.3%
1.1%
1.0%
90.1%
100.0%
1.2.4 Equipment Using Engines Rated between 130 and 560 kW (175 and 750 hp)

   For the 130  to 560 kW range (where 560 kW is included in the range), most of the
equipment uses  direct-injection engines that are water-cooled and turbocharged. A few are
naturally aspirated. The six leading manufacturers produce 56 percent of the equipment in this
category.  These manufacturers are listed in Table 1.2-8. Their products have the following
applications  : ag tractors, combines, generator sets, wheel loaders/bull dozers, which is typical of
the market.

   The 130  to 560 kW range is characterized by applications as shown in Table 1.2-9. They
represent about 94 percent of the market.  The top 90 percent of this market is supplied by 60
OEMs.
                                          1-9

-------
Final Regulatory Support Document
                                       Table 1.2-8
                   Characterization of the Top 6 Equipment Manufacturers
                        for Engines Rated between 130 and 560 kW
Original Equipment
Manufacturer
Deere & Company
Case New Holland
Caterpillar
Komatsu
Ingersoll-Rand
Agco
Major Equipment Manufactured
Ag Tractors, Combines, Wheel
Loaders/bulldozers
Ag Tractors, Combines, Crawlers, Generator
Sets, Scrapers, Crawlers,
Excavators,wheel loaders/dozers, graders
Crawlers, Excavators,Graders, Wheel
Loaders/Dozers
Air Compressors, Rollers, Bore/Drill Rigs
Ag Tractors, Combines, Sprayers
Average
Annual Sains
27,990
14,778
13,151
4,941
3,683
3,194
Percentage
of Market
27%
14%
13%
5%
4%
3%
Engine
C,h a ra p.tnri ya ti on *
W,T,D
W,T,D
W,T/N,D
W,T,D
W,T,D
W/A,T,D
   *W=water-cooled, A=air-cooled,O=oil cooled;NA=naturally aspirated, T=turbocharged, I=indirect injection,
   D=direct injection.

                                       Table 1.2-9
          Equipment Sales Distribution across Applications between 130 and 560 kW
Application Description
Agricultural tractors
Generator sets
Wheel loaders/bulldozers
Combines
Excavators
Crawlers
Air compressors
Graders
Sprayers
Terminal ractors
Forest equipment
Pumps
Off -highway trucks
Cranes
Scrapers
Bore/drill rigs
Irrigation sets
Rollers
Other agricultural equipment
Chippers/grinders
Other construction equipment
Listed Total
Grand Total
Five-yr sales Volume
n QQfi-?or>m
149,589
57,400
43,475
35,743
35,166
28,478
20,884
14,814
12,193
12,141
12,101
9,901
9,377
9,356
7,097
7,047
6,835
6,055
5,935
4,669
4,142


Average Annual
Sa1p«
29,918
11,480
8,695
7,149
7,033
5,696
4,177
2,963
2,439
2,428
2,420
1,980
1,875
1,871
1,419
1,409
1,367
1,211
1,187
934
828
98,480
492,398
Percentage of
Total Sa1p«
29.0%
11.0%
8.3%
6.8%
6.7%
5.4%
4.0%
2.8%
2.3%
2.3%
2.3%
1.9%
1.8%
1.8%
1.4%
1.3%
1.3%
1.2%
1.1%
0.9%
0.8%
94.0%
100.0%
                                           1-10

-------
                                                             Industry Characterization
1.2.5 Equipment Using Engines Rated over 560 kW (750 hp)

    The largest engines, those rated over 560 kW, are produced only for the nonroad market
segments of construction equipment and welders and generators. As much as 35 percent of the
equipment in this power range is manufactured by Caterpillar. Most equipment manufacturers
must buy engines from another company. For most power categories, the Power Systems
Research database estimates that between 5 and 25 percent of equipment sales are from
equipment manufacturers that also produce engines.  Since vertically integrated manufacturers
are typically very large companies, such as  John Deere and Caterpillar, the companies that make
up this fraction of the market are in a distinct minority.

   As in the previous category, the equipment rated over 560 kW uses mostly turbocharged,
direct-injection engines that are water-cooled.  The leading six manufacturers produce 81 percent
of the equipment in this power range.  These manufacturers are shown in Table  1.2-10.
Although generator sets make up the majority of equipment sold in this range, a fraction of them
are considered stationary and are therefore not impacted by this rulemaking. Off-highway trucks
, wheel loaders/dozers and crawlers also have significant sales (see Table 1.2-11).

                                      Table 1.2-10
    Characterization of the Top 6 Equipment Manufacturers for Engines Rated over 560 kW
Original Equipment
Manufacturer
Caterpillar
Komatsu
Multiquip
Kohler
Cummins
Onis Visa
Major Equipment Manufactured
Generator Sets, Off -highway trucks,
crawler tractors
Crawlers, Wheel Loaders/Dozers, Off-
Highway Trucks
Generator Sets
Generator Sets
Generator Sets
Generator Sets
Average
Annual Sales
1,857
1,376
336
335
325
107
Percentage of
Market
35%
26%
6%
6%
6%
2%
Engine
Characterization*
W,T,D
W,T,D
W,T,D
W,T,D
W,T,D
W,T,D
   *W=water-cooled, A=air-cooled,O=oil cooled;NA=naturally aspirated, T=turbocharged, I=indirect injection,
   D=direct injection.
                                          1-11

-------
Final Regulatory Support Document
                                       Table 1.2-11
               Equipment Sales Distribution across Applications over 560 kW
Application Description
Generator sets
Off -highway trucks
Crawlers
Wheel loaders/bulldozers
Off -highway tractors
Excavators
Oil field equipment
Chippers/grinders
Listed Total
Grand Total
Five-yr sales Volume
n 996.7000s!
14,237
4,048
3,857
2,567
542
371
225
132


Average Annual
Sains
2,847
810
771
513
108
74
45
26
5,196
5,241
Percentage of Total
Sains
54%
15%
15%
9.8%
2.1%
1.4%
0.9%
0.5%
99.1%
100.0%
    Section 1.3 characterizes the U.S. petroleum refinery industry, market structure and trends
as it pertains to distillate fuels, including nonroad diesel fuel.  In addition, it covers refinery
operations that are directly impacted by this final rule.  Section 1.4 discusses distribution of
refined petroleum products through pipelines from refineries,  as well as storage operations for
these products.  Sections 1.3 and 1.4 are both are based on a report prepared by RTI under EPA
contract, which is available in the docket.3

1.3 Refinery Operations

1.3.1 The Supply-Side

    This section describes the supply side of the petroleum refining industry, including the
current refinery production processes and raw materials used.  It also discusses the need for
potential changes in refinery production created by this final rule.  Finally, it describes the three
primary categories of petroleum products affected by the rule  and the ultimate costs of
production currently faced by the refineries.

    Refinery Production Processes/Technology.  Petroleum refining is the thermal and
physical separation of crude oil into its major distillation fractions, followed by further
processing (through a series of separation and chemical  conversion steps) into highly valued
finished petroleum products. Although refineries are extraordinarily complex and each site has a
unique configuration, we will describe a generic set of unit operations that are found in most
medium and large facilities.  A detailed discussion of these processes can be found in EPA's
sector notebook of the petroleum refining industry (EPA, 1995); simplified descriptions are
available on the web sites of several major petroleum producers (Flint Hills Resources, 2002;
Chevron, 2002).
                                           1-12

-------
                                                              Industry Characterization
   Figure 1.3-1 shows the unit operations and major product flows in a typical refinery. After
going through an initial desalting process to remove corrosive salts, crude oil is fed to an
atmospheric distillation column that separates the feed into several fractions.  The lightest
boiling range fractions are processed through reforming and isomerization units into gasoline or
diverted to lower-value uses such as LPG and petrochemical feedstocks. The middle-boiling
fractions make up the bulk of the aviation and distillate fuels produced from the crude. In most
refineries, the undistilled liquid (called bottoms) is sent to a vacuum still to further fractionate
this heavier material.  Bottoms from the vacuum distillation can be further processed into
low-value products such as residual fuel oil, asphalt, and petroleum coke.

   A portion of the bottoms from the atmospheric distillation, along with distillate from the
vacuum still, are processed further in a catalytic cracking unit or in a hydrocracker.  These
operations break large hydrocarbon molecules into smaller ones that can be converted to high-
value gasoline and middle distillate products. Bottoms from the vacuum still are increasingly
processed in a coker to produce saleable coke and gasoline and diesel fuel blendstocks. The
cracked molecules are processed further in combining  operations (alkylation, for example),
which combine small  molecules into larger, more useful entities, or in reforming, in which
petroleum molecules are reshaped into higher quality species.  It is in the reforming operation
that the octane rating of gasoline is increased to the desired level for final  sale. A purification
process called hydrotreating helps remove chemically bound sulfur from petroleum products and
is critically important for refineries to process their refinery streams into valuable products and
to achieve the low sulfur levels required under the regulation.

   For each of the major products, several product streams from the refinery will be blended
into a finished mixture. For example, diesel fuel typically has a straight-run fraction from crude
distillation, distillate from the hydrocracker, light-cycle oil from the catalytic cracker, and
hydrotreated gas oil from the coker. Several auxiliary  unit operations are  also needed in the
refinery complex, including hydrogen generation, catalyst handing and regeneration, sulfur
recovery, wastewater treatment, and blending and storage tanks. Table  1.3-1  shows average
yields of major products from U.S. refineries.
                                           1-13

-------
     Final Regulatory Support Document
                                      Figure 1,3-1
                                 The Modern Refinery
     Gas and
Light Gasoline
 It-
Ends
Mont
                                     LPG
                                              light Straight Run Gasoline


                                                Isonw nation  bometate
                                                   Pfont

                                                             Refontwle
Crude
   Oil
                     Kerosene
          vacuum   VGO
         DHtiiUUHin
                                                 Alkylation
                                                   Plant
    Residuum

                               Light Cyde Oil
                         i
               Coker
                                                   u
          AsphaH
                 t
                Coke
                                                                                           Jet
                                                                                           Diesel
                                                                                         .  Fi.cl
                                                                                           ow
     Source: Chevron. 2002. Diesel Fuel Refining and Chemistry. As accessed on August 19, 2002.
        www.chevron.com/prodserv/fuels/bulletin/diesel/L2_4_2rf.htm.
                                                1-14

-------
                                                           Industry Characterization
                                        Table 1.3-1
                Yields of Major Petroleum Products from Refinery Operations
Product
Crude Feed
Gasoline
Highway diesel fuel
Jet Fuel
Petroleum Coke
Residual Fuel Oil
LPGas
Home heating oil
Asphalt
Nonroad diesel fuel
Other Products
Total
Gallons per Barrel of Crude
42.0
19.4
6.3
4.3
2.0
1.9
1.9
1.6
1.4
0.8
4.0
43.6
Percentage of Total Feed*
100.0%
46.0%
15.0%
10.0%
5.0%
4.5%
4.5%
4.0%
3.0%
2.0%
9.5%
104.0%
    *Note: Total exceeds 100 percent due to volume gain during refining.
    Source: Calculated from EIA data in Petroleum Supply Annual 2001. U. S. Department of Energy, Energy
       Information Administration (EIA).  2002a.  Petroleum Supply Annual 2001, Tables 16, 17, and 20.
       Washington, DC.
   Potential Changes in Refining Technology Due to the Final Rule. Regulations requiring
much lower levels of sulfur for both gasoline and highway diesel fuel will come into effect over
the next few years.  To meet these challenges, refineries are planning to add hydrotreater units to
their facilities, route more intermediate product fractions through existing hydrotreaters, and
operate these units under more severe conditions to reduce levels of chemically bound sulfur in
finished products.  As has been documented in economic impact analyses for the gasoline and
highway diesel rules, these changes will require capital investments for equipment, new piping,
and in-process storage; increased use of catalyst and hydrogen; and modifications to current
operating strategies.

   The addition of lower sulfur limits for nonroad diesel fuel will result in additional refinery
changes similar in nature to those required for highway diesel fuel. Product streams formerly
sent directly to blending tanks will need to be routed through the hydrotreating operation to
reduce their sulfur level.  In addition, because an increasing fraction of the total volumetric
output of the facility must meet ultra-low sulfur requirements, flexibility will be somewhat
reduced. For example, it will become more difficult to sell off spec products if errors or
equipment failures occur during operation.

   Types of Products. The major products made at petroleum refineries are unbranded
commodities, which must meet established specifications for fuel value, density, vapor pressure,
                                           1-15

-------
Draft Regulatory Support Document
sulfur content, and several other important characteristics.  As Section 1.3.2 describes, they are
transported through a distribution network to wholesalers and retailers, who may attempt to
differentiate their fuel from competitors based on the inclusion of special additives or purely
through adroit marketing.  Gasoline and highway diesel are taxed before final sale, whereas
nonroad fuel is not. To prevent accidental or deliberate misuse, nonroad diesel fuel must be
dyed before final sale.

   A total of $158 billion of petroleum products were sold in the 1997 census year, accounting
for a nontrivial 0.4 percent of GDP.  Table 1.3-2 lists the primary finished products produced; as
one might expect, the percentages are quite close to the generic refinery output shown in Table
1.3-1. Motor gasoline is the dominant product, both in terms of volume and value, with almost
three billion barrels produced in 1997. Distillate fuels accounted for less than half as much as
gasoline, with 1.3 billion barrels produced in the United States in the same year.  Data from the
Energy Information Administration (EIA) suggest that 60 percent of that total is low-sulfur
highway diesel, with the remainder split between nonroad diesel and heating oil. Jet fuel, a
fraction slightly heavier than gasoline, is the third most important product, with a production
volume  of almost 600 million barrels.

                                       Table 1.3-2
                  Types of Petroleum Products Produced by U.S. Refineries
Products
Liquified Refinery Gases
Finished Motor Gasoline
Finished Aviation
Jet Fuel
Kerosene
Distillate Fuel Oil
Residual Fuel Oil
Naphtha for Feedstock
Other Oils for Feedstock
Special Naphthas
Lubricants
Waxes
Petroleum Coke
Asphalt and Road Oil
Still Gas
Miscellaneous
Total
Total Produced
(thousand barrels)
243,322
2,928,050
6,522
558,319
26,679
1,348,525
263,017
60,729
61,677
18,334
63,961
6,523
280,077
177,189
244,432
21,644
6,309,000
Percentage of Total
3.9%
46.4%
0.1%
8.8%
0.4%
21.4%
4.2%
1.0%
1.0%
0.3%
1.0%
0.1%
4.4%
2.8%
3.9%
0.3%
100.0%
                                          1-16

-------
                                                          Industry Characterization
   Primary Inputs. Crude oil is the dominant input in the manufacture of refined petroleum
products, accounting for 74 percent of material cost, or about $95 billion in 1997, according to
the latest Economic Census (U.S. Census Bureau, 1999).  The census reported almost equal
proportions of imported and domestic crude in that year, with 2.5 billion barrels imported and
2.8 billion barrels originating from within the United States.  More recent data published by the
EIA show a higher import dependence in the most recent year, with 3.4 billion barrels, or 61.7
percent, imported out of a total of 5.5 billion barrels used by refineries during 2001 (EIA,
2002a).

   Crude oil extracted in different regions of the world have quite different characteristics,
including the mixture of chemical species present, density and vapor pressure, and sulfur
content.  The cost of production and the refined product output mix vary considerably depending
on the type of crude processed. A light, sweet crude oil, such as that found in Nigeria, will
process very differently from a heavy, sulfur-laden Alaska or Arabian crude.  The ease of
processing any particular material is reflected in its purchase price, with sweet crudes selling at a
premium. The result of these variations is that refineries are frequently optimized to run only
certain types of crude; they  may be unable or unwilling to switch to significantly different feed
materials.

   In addition to crude oil,  refineries may also feed to their refineries hydrocarbon by-products
purchased from chemical companies and other refineries and/or semiprocessed fuel oils  imported
from overseas.  In 1997, the Census reported that these facilities purchased $11 billion of
hydrocarbons and imported $2.4 billion of unfinished oils. Other significant raw materials
purchased include $600 million for precious metal catalysts and more than $800 million in
additives.

   Costs of Production. According to the latest Economic Census, there were 244 petroleum
refining establishments in the United States in 1997, owned by 123 companies and employing
64,789 workers. Data from EIA using a more stringent definition show 164 operable refineries
in 1997, a number that fell to 153 by January 1, 2002. As seen in Table 1.3-3, value of
shipments in 2000 was $216 billion, up from $158 billion in the 1997 census year. The  costs of
refining are divided into the main input categories of labor, materials, and capital expenditures.
Of these categories, the cost of materials represents about 80 percent of the total value of
shipments, as defined by the Census, varying from year to year as crude petroleum prices change
(see Table 1.3-4). Labor and capital expenditures tend to be more stable, each accounting for 2
to 4 percent of the value of  shipments.

                                       Table 1.3-3
                                          1-17

-------
Draft Regulatory Support Document
                  Description of Petroleum Refineries—Census Bureau Data
NAICS324110—
Petroleum Refineries
2000
1999
1998
1997
1992 (reported as SIC 2911)
Establishments
(NA)
(NA)
(NA)
244
232
Companies
(NA)
(NA)
(NA)
123
132
Employment
62229
63619
64920
64789
74800
Value of Shipments
($106)
$215,592
$144,292
$118,156
$157,935
$136,239
Sources:
1992 data from U.S. Census Bureau.  1992 Census of Manufactures, Industry Series MC920I-29A. Table 1A.
1997 data from U.S. Census Bureau, 1997 Economic Census - Manufacturing, Industry Series EC97M-3241A, Table 1.
1998-2000 data from U.S. Census Bureau, Annual Survey of Manufactures-2000, 2000, Statistics for Industry Groups
    and Industries MOO(AS)-1, Table 2.
                                         Table 1.3-4
                     Petroleum Refinery Costs of Production, 1997-2000
Petroleum Refinery
Costs of Production
Cost of Materials (106)
as percent of shipment
value
Cost of Labor (106)
as % of shipment value
Capital Expenditures (106)
as % of shipment value
1997
$127,555
80.4%
$3,885
2.4%
$4,244
2.7%
1998
$92,212
78.0%
$3,965
3.4%
$4,169
3.5%
1999
$114,131
79.1%
$3,983
2.8%
$3,943
2.7%
2000
$178,631
82.9%
$3,995
1.9%
$4,453
2.1%
       Source: U.S. Census Bureau, Annual Survey of Manufactures. 2000. 2000 Statistics for Industry Groups
           and Industries MOO(AS)-1, Tables 2 and 5.
    Refinery Production Practices.  Refining, like most continuous chemical processes, has
high fixed costs from the complex and expensive capital equipment installed.  In addition,
shutdowns are very expensive, because they create large amounts of off-specification product
that must be recycled and reprocessed before sale.  As a result, refineries attempt to operate 24
hours per day, 7 days per week, with only 2 to 3 weeks of downtime per year. Intense focus on
cost-cutting has led to large increases in capacity utilization over the past several years.  A
Federal Trade Commission (FTC) investigation into the gasoline price spikes  in the Midwest
during the summer of 2000 disclosed an average utilization rate of 94 percent during that year,
and EIA data from 2001 show that a 92.6 percent utilization rate was maintained in 2001 (FTC,
2001;EIA, 2002a).
                                            1-18

-------
                                                          Industry Characterization
   Because of long lead times in procuring and transporting crude petroleum and the need to
schedule pipeline shipments and downstream storage, refinery operating strategies are normally
set several weeks or months in advance.  Once a strategy is established for the next continuous
run, it is difficult or impossible to change it. Exact proportions of final products can be altered
slightly, but at a cost of moving away from the optimal cost profile established initially. The
economic and logistical drivers combine to generate an extremely low supply elasticity. One
recent study estimated the supply elasticity for refinery products at 0.24 (Considine, 2002). The
FTC study discussed above concluded that refiners had little or no ability to respond to the
shortage of oxygenated gasoline in the Midwest in the summer of 2000, even with some advance
warning that this would occur.

1.3.2 The Demand Side

   This section describes the demand side of the market for refined  petroleum products, with a
focus on the distillate fuel oil industry. It discusses the primary consumer markets identified and
their distribution by end use and PADD.  This section also considers substitution possibilities
available in each of these markets and the feasibility and costs of these substitutions. Figure
1.3-2 is a map of the five  PADD regions.

   Uses and Consumers.  Gasoline, jet fuel, and distillate fuel oils  account for almost 80
percent of the value of refinery product shipments, with gasoline making up about 51 percent
(U.S. Census Bureau, 1999). Actual and relative net production volumes of these three major
products, along with residual fuel oils, are shown in Table 1.3-5, broken out by PADD and for
the country as a whole. PADD III, comprising the states of Texas, Louisiana, Arkansas,
Alabama, Mississippi, and New Mexico, is a net exporter of refined  products, shipping them
through pipelines to consumers on the East Coast and also to the Midwest. Compared with
gasoline production patterns, distillate production is slightly lower in PADD V (the West Coast)
and higher in PADD II (the Midwest).

   The primary end-use markets for distillate and residual fuel oils are divided by EIA as
follows:
       •   residential—primarily fuel oil for home (space) heating;
       •   commercial—high-sulfur diesel fuel, low-sulfur diesel fuel, and fuel oil for space
          heating;
       •   industrial—low-sulfur diesel fuel for highway use, high-sulfur diesel fuel for nonroad
          use, and residual fuel oil for operating steam boilers and turbines (power generation);
       •   oil companies—mostly fuel oil and some residual fuel for internal use;
       •   farm—almost exclusively high-sulfur diesel fuel;
       •   electric utility—residual fuel and distillate fuel oil for power generation;
       •   railroad—high-sulfur diesel fuel and low-sulfur diesel fuel  used for locomotives;
       •   vessel bunking—combination of fuel oil and residual fuel for marine engines;
       •   on-highway diesel—low-sulfur diesel fuel  for highway trucks and automobiles;
          military—high-sulfur diesel fuel sales to the Armed Forces; and
       •   off-highway diesel—high-sulfur diesel fuel and low-sulfur  diesel fuel used in
          construction and other industries.

                                          1-19

-------
Draft Regulatory Support Document
                                       Figure 1.3-2
                          PADD Districts of the United States
                        Petroleum Administration Defense Districts (PADDs)
HAWAII
   As Table 1.3-6 indicates, the highway diesel fuel usage of 33.1 billion gallons represents the
bulk of distillate fuel usage (58 percent) in 2000. Residential distillate fuel usage, which in the
majority is fuel oil, accounts for 11 percent of total usage in 2000.  Nonroad diesel fuel is
primarily centered on industrial, farm, and off-highway diesel (construction) usage. In 2000,
these markets consumed about 13 percent of total U.S. distillate fuels.

   To determine the regional consumption of distillate fuel usage, 2000 sales are categorized by
PADDs.  As shown in Table 1.3-7, PADD I (the East Coast) consumes the greatest amount of
distillate fuel at 20.9 billion gallons. However, residential, locomotive, and vessel bunking
consumers account for 6.4 billion gallons of the distillate fuel consumed, which means that at
least one-third of the total consumed in PADD I is due to fuel oil and not to diesel fuel
consumption.
                                         1-20

-------
                                                               Industry Characterization
                                           Table 1.3-5
              Refinery Net Production of Gasoline and Fuel Oil Products by PADD
Motor Gasoline
PADD
I
II
III
IV
V
Total
Quantity
(l,000bbl)
369,750
641,720
1,306,448
97,869
512,263
2,928,050
Percent
(%)
12.6%
21.9%
44.6%
3.3%
17.5%
100.0%
Distillate Fuel Oil
Quantity
(l,000bbl)
170,109
316,023
629,328
54,698
178,367
1,348,525
Percent
(%)
12.6%
23.4%
46.7%
4.1%
13.2%
100.0%
Jet Fuel
Quantity
(l,000bbl)
30,831
80,182
288,749
9,787
148,770
558,319
Percent
(%)
5.5%
14.4%
51.7%
1.8%
26.6%
100.0%
Residual Fuel Oil
Quantity
(l,000bbl)
38,473
24,242
132,028
4,151
64,123
263,017
Percent
(%)
14.6%
9.2%
50.2%
1.6%
24.4%
100.0%
Source: U.S. Department of Energy, Energy Information Administration (ElA). 2002a. Petroleum Supply Annual 2001,
    Tables 16, 17, and 20. Washington, DC. Table 17.
                                           Table 1.3-6
                              Distillate Fuel Oil by End Use (2000)
End Use
Residential
Commercial
Industrial
Oil Company
Farm
Electric Utility
Railroad
Vessel Bunking
Highway Diesel
Military
Off-Highway Diesel
Total
2000 Usage (thousand gallons)
6,204,449
3,372,596
2,149,386
684,620
3,168,409
793,162
3,070,766
2,080,599
33,129,664
233,210
2,330,370
57,217,231
Percentage Share (%)
10.8%
5.9%
3.8%
1.2%
5.5%
1.4%
5.4%
3.6%
57.9%
0.4%
4.1%
100.0%
    Source: U.S. Department of Energy, Energy Information Administration (El A). 2001b. Fuel Oil and Kerosene
       Sales, 2000, Tables 7-12. Washington, DC.
                                              1-21

-------
Draft Regulatory Support Document
                                       Table 1.3-7
                         Distillate Fuel Oil by End Use and PADD
PADD (Thousand Gallons)
End Use
Residential
Commercial
Industrial
Oil Company
Farm
Electric Utility
Railroad
Vessel Bunking
Highway Diesel
Military
Off -highway Diesel
Total
I
5,399,194
2,141,784
649,726
19,101
432,535
304,717
499,787
490,150
10,228,244
70,801
669,923
20,905,962
II
628,414
568,089
600,800
41,727
1,611,956
133,971
1,232,993
301,356
11,140,616
36,100
608,307
16,904,329
III
1,117
346,578
420,400
560,905
552,104
194,786
686,342
1,033,333
5,643,703
9,250
516,989
9,965,507
IV
38,761
102,905
241,146
29,245
220,437
8,492
344,586
173
1,474,611
4,163
180,094
2,644,613
V
136,962
213,240
237,313
33,643
351,377
151,196
307,059
255,586
4,642,490
112,895
355,056
6,796,817
    Table 1.3-8 presents a closer look at on-highway consumption of distillate fuel, which is
entirely low-sulfur diesel fuel.  PADD I (the East Coast) and PADD II (the Midwest) consume
almost 65 percent of all U.S. distillate fuel sold for on-highway use.

    Table 1.3-9 shows that residential consumption of distillate fuel (primarily fuel oil) is
centered in PADD I (the East Coast).  Fuel-oil-fired furnaces and water heaters in New York and
New England consume most of this heating oil; in most of the rest of the country, residential
central heating is almost universally provided by natural gas furnaces or electric heat pumps. A
comparison of Tables  1.3-5 and 1.3-9 reveals that PADD I produces far less distillate fuel oil
than it consumes.  The balance is made up by shipments from PADD III and imports from
abroad.

    Tables  1.3-10,  1.3-11, and 1.3-12 focus on diesel sales for industrial, agricultural, and
construction use.  Industrial use of diesel fuel is fairly evenly spread across PADDs. PADD II
(the Midwest) has the  highest percentage of diesel usage at 28 percent, while PADD V (the West
Coast) has the lowest percentage  at 11 percent. In contrast,  agricultural purchases of diesel are
in the great majority (51 percent) centered in PADD II (the Midwest). For construction only,
distillate fuel sales are available, but these sales are assumed to be principally diesel fuel.
Construction usage of diesel fuel, as with industrial usage, is fairly evenly spread across PADDs,
with the exception of PADD IV.  PADD IV represents only 8 percent of total construction usage.
                                          1-22

-------
                                       Industry Characterization
                     Table 1.3-8
 Sales for Highway Use of Distillate Fuel by PADD (2000)
PADD
I
II
III
IV
V
Total
Distillate Usage
(Thousand Gallons)
10,228,244
11,140,616
5,643,703
1,474,611
4,642,490
33,129,664
Share of
Distillate Fuel Used
30.9%
33.6%
17.0%
4.5%
14.0%
100.0%
                     Table 1.3-9
Sales for Residential Use of Distillate Fuel by PADD (2000)
PADD
I
II
III
IV
V
Total
Distillate Usage
(Thousand Gallons)
5,399,194
628,414
1,117
38,761
136,962
6,204,448
Share of
Distillate Fuel Used
87.0%
10.1%
0.0%
0.6%
2.2%
100.0%
                     Table 1.3-10
    Industrial Use of Distillate Fuel by PADD (2000)
PADD
I
II
III
IV
V
Total
Distillate Usage
(Thousand Gallons)
649,726
600,800
420,400
241,146
237,313
2,149,385
Share of
Distillate Fuel Used
30.2%
28.0%
19.6%
11.2%
11.0%
100.0%
                        1-23

-------
Draft Regulatory Support Document
                                       Table 1.3-11
               Adjusted Sales for Farm Use of Distillate Fuel by PADD (2000)
PADD
I
II
III
IV
V
Total
Distillate Usage
(Thousand Gallons)
432,535
1,611,956
552,104
220,437
351,377
3,168,409
Share of
Distillate Fuel Used
13.6%
50.9%
17.4%
7.0%
11.1%
100.0%
                                       Table 1.3-12
          Sales for Construction Use of Off-Highway Distillate Fuel by PADD (2000)
PADD
I
II
III
IV
V
Total
Distillate Usage
(Thousand Gallons)
510,876
549,299
394,367
150,060
295,235
1,899,837
Share of
Distillate Fuel Used
26.9%
28.9%
20.8%
7.9%
15.5%
100.0%
    Substitution Possibilities in Consumption For engines and other combustion devices
designed to operate on gasoline, there are no practical substitutes, except among different grades
of the same fuel. Because EPA regulations apply equally to all gasoline octane grades, price
increases will not lead to substitution or misfueling.  In the case of distillate fuels, it is currently
possible to substitute between low-sulfur diesel fuel, high-sulfur diesel fuel, and distillate fuel
oil, although higher sulfur levels are associated with increased maintenance and poorer
performance.

    With the consideration of more stringent nonroad fuel and emission regulations, substitution
will become less likely.  Switching from nonroad ultralow-sulfur diesel to highway ultralow-
sulfur diesel is not financially attractive,  because of the taxes levied on the highway product.
Misfueling with high-sulfur fuel oil will  rapidly degrade the performance of the exhaust system
of the affected engine, with negative consequences for maintenance and repair costs.

1.3.3 Industry Organization

    To determine the ultimate effects of the rule, it is important to have a good understanding of
the overall refinery industry structure.  The degree of industry concentration, regional patterns of
                                           1-24

-------
                                                         Industry Characterization
production and shipment, and the nature of the corporations involved are all important aspects of
this discussion. In this section, we look at market measures for the United States as a whole and
by PADD region.

   Market Structure—Concentration.  There is a great deal of concern among the public
about the nature and effectiveness of competition in the refining industry. Large price spikes
following supply disruptions and the tendency for prices to slowly fall back to more reasonable
levels have created suspicion of coordinated action or other market imperfections in certain
regions. The importance of distance in total delivered cost to various end-use markets also
means that refiners incur a wide range of costs in serving some markets; because the price is set
by the highest cost producer serving the market as long  as supply and demand are in balance,
profits are made by the low-cost producers in those markets.

   Market concentration is  measured in a variety of ways by antitrust regulators in the
Department of Justice (DOJ) and Federal Trade Commission (FTC), including four-firm
concentration ratios (CR4) and the Herfindahl-Hirschman Index (HHI).  The CR4 is simply the
combined market share of the four largest sellers in a given market, a very intuitive
concentration measure. The HHI, which is currently used by the DOJ's Antitrust Division and
the FTC, is constructed by summing up the squared market shares, in percentage terms, of all
competitors in the market. According to these agencies' 1997 Horizontal Merger Guidelines, a
market with an HHI under 1,000 is considered "unconcentrated," one with an HHI between
1,000 and 1,800 is "moderately concentrated," and one with a measure over 1,800 is "highly
concentrated" (DOJ, 1997).

   The merger guidelines assume that high concentration offers the potential for firms to
influence prices through coordinated action on prices. Still it is possible for highly concentrated
markets to behave competitively if firms are unwilling or unable to coordinate their actions or if
potential entry can serve to limit price increases. The RTI report presents the calculated HHI
values for diesel engine markets.
   There is, however, no convincing evidence in the literature that markets should be modeled
as imperfectly competitive.  The FTC study cited earlier concluded that the extremely low
supply and demand elasticities made large price movements likely and inevitable given
inadequate supply or unexpected increases in demand. Nevertheless, their economic analysis
found no evidence of collusion or other anticompetitive behavior in the summer of 2000.
Furthermore, the industry is not highly concentrated on a nationwide level or within regions.
The 1997 Economic Census presented the following national concentration information:
four-firm concentration ratio (CR) of 28.5 percent, eight-firm concentration ratio of 48.6 percent,
and an HHI of 422.  Merger guidelines followed by the FTC and Department of Justice consider
little potential for pricing power in an industry with an HHI below 1,000.

   Two additional considerations were important in making a determination as to whether we
can safely assume that refineries act as price-takers in their markets.  First, with greater
concentration in regional or local markets than at the national level, as well as with significant

                                          1-25

-------
Draft Regulatory Support Document
transport costs, competition from across the country will not be effective in restraining prices.
Secondly, several large mergers have occurred since the 1997 Economic Census was conducted,
all of which have prompted action by the FTC to ensure that effective competition was retained.

    To investigate these issues, RTI estimated concentration measures that are not based on
refinery-specific production figures (which are not available), but rather on crude distillation
capacity, which is the industry's standard measure of refinery size.  We aggregated the total
capacity controlled by each corporate parent, both at the PADD level and nationwide, and then
calculated CR-4, CR-8, and HHI figures. The results are presented in Table 1.3-13.

                                       Table 1.3-13
            2001 Concentration Measures for Refineries Based on Crude Capacity
PADD
I
II
III
IV (current)
IV (future)
V
National
Quantity
1,879,400
3,767,449
8,238,044
606,650
606,650
3,323,853
17,815,396
CR-4
71.6%
54.6%
48.8%
59.6%
45.4%
61.3%
41.89%
CR-8
91.3%
78.2%
68.0%
90.1%
80.5%
90.9%
65.50%
HHI
1,715
1,003
822
1,310
918
1,199
644
   Note:  Quantity is crude distillation capacity in thousands of barrels per stream day.
   Source:U.S. Department of Energy, Energy Information Administration (EIA).  2002b. Refinery Capacity Data
       Annual. As accessed on September 23, 2002. http://www.eia.doe.gov/
       oil_gas/petroleum/data_publications/ refinery_capacity_data/refcap02.dbf.  Washington, DC.  See text
       discussion.
    The data in this table provide several interesting conclusions:
       •  The current and future state of PADD IV shows the impact of FTC oversight to
          maintain competition.  As part of approving the Phillips-Conoco merger, the FTC
          ordered the merged company to divest two refineries in PADD IV—Commerce City,
          Colorado, and Woods Cross, Utah.  Once those divestitures take place, the
          concentration levels will drop below 1,000, a level that is not generally of concern.
       •  The only region that is highly concentrated is PADD I, which is generally dominated
          by two large refineries. In this  case, however, imports of finished petroleum
          products, along with shipments from PADD III, should prevent price-setting behavior
          from emerging in this market.  Table 1.3-14 shows imports of refined products for
          PADD I and the entire country. About 90 percent of total U.S. imports of gasoline
          and distillate fuels come into PADD I, aided by inexpensive ocean transport. It is
          reasonable to assume that any attempts to set  prices by the dominant refineries would
          be defeated with increased imports.
                                           1-26

-------
                                                          Industry Characterization
                                       Table 1.3-14
                            PADD I and Total U.S. Imports of
               Gasoline and Fuel Oil Products by Top Five Countries of Origin
Finished Motor Gasoline
Top Five Countries of
Origin
Venezuela
Brazil
Canada
Russia
Virgin Islands (U.S.)
Sum of Top Five
Total
Percentage of Total
U.S. Imports
PADD I
Import
21,017
8,286
41,711
869
38,135
110,018
153,633
92.6%
Total U.S.
Import
21,257
8,286
43,778
968
38,882
113,171
165,878

Distillate Fuel Oil
PADD I
Import
16,530
1,472
30,350
10,345
30,810
89,507
112,318
89.4%
Total U.S.
Import
16,530
1,832
35,165
10,345
31,540
95,412
125,586

Residual Fuel
PADD I
Import
17,667
8,361
9,483
174
13,412
49,097
91,520
85.0%
Total U.S.
Import
18,341
9,105
11,723
1,051
13,502
53,722
107,688

Source: U.S. Department of Energy, Energy Information Administration (ElA). 2002a. Petroleum Supply Annual 2001.
   Tables 16, 17, and 20. Washington, DC. Table 20.
       •   Markets in PADDs II and III, which are not overly concentrated or geographically
          isolated, should behave competitively, with little potential for price-setting among its
          refineries.
       •   The four large mergers (Exxon-Mobil, BP-Amoco, Chevron-Texaco, and
          Phillips-Conoco) have not increased nationwide concentration to a level of concern
          for competitive reasons.

   Market Structure—Firms  and Facilities. PADD III has the greatest number of refineries
affected by the final rule and will account for the largest volume of low-sulfur nonroad diesel
fuel. Tables 1.3-15 and 1.3-16 present the number of operating refineries and the number of
crude distillation units in each PADD; output volumes were presented in Table 1.3-5. PADD III
also accounts for 45 to 50 percent of U.S. refinery net production of finished motor gasoline,
distillate fuel oil, and residual fuel oil. Similarly, PADD IV has the fewest number of affected
facilities and accounts for the smallest share of distillate production.  Still, because compliance
costs per unit of output are likely to depend on refinery scale, the small size and geographic
isolation of the PADD IV refineries  suggest that the financial impact may be greatest on these
operations.
                                          1-27

-------
Draft Regulatory Support Document
                                      Table 1.3-15
                        Number of Petroleum Refineries by PADD
PADD
I
II
III
IV
V
Total
Number of Facilities
16
28
54
14
32
144
Percentage of Total
11.1%
19.4%
37.5%
9.7%
22.2%
100.0%
                                      Table 1.3-16
                     Number of Crude Distillation Facilities by PADD
PADD
I
II
III
IV
V
Total
Number of Facilities
12
26
50
16
35
139
Percentage of Total
8.6%
18.7%
36.0%
11.5%
25.2%
100.0%
   According to the EIA Petroleum Supply Annual 2001, the top three owners of crude
distillation facilities are ExxonMobil Corp. (11 percent of U.S. total), Phillips Petroleum Corp.
(10 percent), and BP PLC (9 percent).  Tablel.3-17 gives an overview of the top refineries in
each PADD, in descending order of total crude distillation capacity. As operating refineries
attempt to run at full utilization rates, this measure should correlate directly to total output.
Information is not available on actual production of highway diesel, nonroad diesel, and other
distillate fuels for each refinery. Note that PADD III has more than 50 percent of the total crude
distillation capacity, as well as the three largest single facilities.

   Firm Characteristics. Many of the large integrated refineries are owned by major
petroleum producers, which are among the largest corporations in the United States. According
to Fortune Magazine's Fortune 500 list, ExxonMobil is the second largest corporation in the
world, as well as in the United States. Chevron Texaco ranks as the eighth largest U.S.
corporation, placing it fourteenth in the world. The newly merged Phillips and Conoco entity
will rank in the top 20 in the United States, and six more U.S. petroleum firms make the top 500.
BP Amoco (fourth worldwide) and Royal Dutch Shell (eighth worldwide) are foreign-owned, as
is Citgo (owned by Petroleos de Venezuela).
                                          1-28

-------
                                                          Industry Characterization
   Many of the smallest refineries are small businesses. A total of 21 facilities owned by 13
different parent companies qualify or have applied for small business status (EPA, 2002).  These
small refineries are concentrated in the Rocky Mountain and Great Plains region of PADD IV,
and their conversion to low-sulfur diesel fuel calls for significant flexibility.

1.3.4 Markets and Trends

   There is considerable diversity in how different markets for distillate fuels have been
growing over the past several years.  Table 1.3-18 shows that residential and commercial use of
fuel oil has been dropping steadily since 1984, while highway diesel use has nearly doubled over
the same period. Farm use of distillate has been flat over the 15-year period, while off-highway
use, mainly for construction, has increased by 40 percent.
                                          1-29

-------
                          Table
Top Refineries in Each PADD by
1.3-17
Total Crude Distillation Capacity
Name
of Company
Sunoco Inc. (R&M)
PADD I phllllPs 66 Co-
Phillips 66 Co.
Motiva Enterprises LLC
Sunoco Inc.
TOTAL
BP Products North America, Inc.
PADD II Philip 66 Co.
Flint Hills Resources LP
ExxonMobil Refg & Supply Co.
Marathon Ashland Petro LLC
Conoco Inc.
Marathon Ashland Petro LLC
Williams Refining LLC
TOTAL
Location Crude Distillation
of Facilities Capacity (barrels/day)
Philadelphia
Linden
Trainer
Delaware City
Marcus Hook

Whiting
Wood River
Saint Paul
Joliet
Catlettsburg
Ponca City
Robinson
Memphis

PA
NJ
PA
DE
PA

IN
IL
MN
IL
KY
OK
IL
TN

330,000
250,000
180,000
175,000
175,000
1,576,600
410,000
288,300
265,000
235,500
222,000
194,000
192,000
180,000
3,428,053
Percentage of Total
PADD Crude
Distillate Capacity
20.9%
15.9%
11.4%
11.1%
11.1%
100.0%
12.0%
8.4%
7.7%
6.9%
6.5%
5.7%
5.6%
5.3%
100.0%
Percentage of Total U.S.
Crude Distillate Capacity
2.0%
1.5%
1.1%
1.1%
1.1%
9.7%
2.5%
1.8%
1.6%
1.4%
1.4%
1.2%
1.2%
1.1%
21.1%
                                                                                      (continued)

-------
                   Table 1.3-17 (continued)
Top Refineries in Each PADD by Total Crude Distillation Capacity
Name
of Company
ExxonMobil Refg & Supply Co.
ExxonMobil Refg & Supply Co.
BP Products North America, Inc.
PADD III ExxonMobil Refg & Supply Co.
Deer Park Refg Ltd Ptnrshp
Citgo Petroleum Corp.
Chevron U.S. A. Inc.
Flint Hills Resources LP
Lyondell Citgo Refining Co. Ltd.
Premcor Refg Group Inc
Conoco Inc.
Phillips 66 Co.
Motiva Enterprises LLC
Marathon Ashland Petro LLC
Motiva Enterprises LLC
Motiva Enterprises LLC
Phillips 66 Co.
Valero Refining Co. Texas
Chalmette Refining LLC
Atofina Petrochemicals Inc.
Total
Location Crude Distillation
of Facilities Capacity (barrels/day)
Bay town
Baton Rouge
Texas City
Beaumont
Deer Park
Lake Charles
Pascagoula
Corpus Christi
Houston
Port Arthur
Westlake
Belle Chasse
Port Arthur
Garyville
Norco
Convent
Sweeny
Texas City
Chalmette
Port Arthur

TX
LA
TX
TX
TX
LA
MS
TX
TX
TX
LA
LA
TX
LA
LA
LA
TX
TX
LA
TX

516,500
488,500
437,000
348,500
333,700
326,000
295,000
279,300
274,500
255,000
252,000
250,000
245,000
232,000
228,000
225,000
213,000
204,000
182,500
178,500
7583080
Percentage of Total
PADD Crude
Distillate Capacity
6.8%
6.4%
5.8%
4.6%
4.4%
4.3%
3.9%
3.7%
3.6%
3.4%
3.3%
3.3%
3.2%
3.1%
3.0%
3.0%
2.8%
2.7%
2.4%
2.4%
100.0%
Percentage of Total U.S.
Crude Distillate Capacity
3.2%
3.0%
2.7%
2.1%
2.1%
2.0%
1.8%
1.7%
1.7%
1.6%
1.6%
1.5%
1.5%
1.4%
1.4%
1.4%
1.3%
1.3%
1.1%
1.1%
46.7%
                                                                                  (continued)

-------
                                                          Table 1.3-17 (continued)
                                   Top Refineries in Each PADD by Total Crude Distillation Capacity

PADD IV


PADDV



Total U.S.
Name
of Company
Conoco Inc.
Sinclair Oil Corp.
Conoco Inc.
TOTAL
BP West Coast Products LLC
Chevron U.S. A. Inc.
BP West Coast Products LLC
Chevron U.S. A. Inc.
Williams Alaska Petro Inc.
TOTAL
(excluding Virgin Islands)
Location
of Facilities
Commerce City
Sinclair
Billings

Los Angeles
El Segundo
Cherry Point
Richmond
North Pole


CO
WY
MO

CA
CA
WA
CA
AK


Crude Distillation
Capacity (barrels/day)
62,000
62,000
60,000
567,370
260,000
260,000
225,000
225,000
197,928
3,091,198
16,246,301
Percentage of Total
PADD Crude Percentage of Total U.S.
Distillate Capacity Crude Distillate Capacity
2.0%
2.0%
1.9%
18.4%
8.4%
8.4%
7.3%
7.3%
6.4%
100.0%

0.4%
0.4%
0.4%
3.5%
1.6%
1.6%
1.4%
1.4%
1.2%
19.0%
100.0%
Source:U.S. Department of Energy, Energy Information Administration (EIA). 2002b. Refinery Capacity Data Annual.  As accessed on September 23, 2002.
    http://www.eia.doe.gov/oil_gas/petroleum/data_publications/refinery_capacity_data/refcap02.dbf.  Washington, DC.

-------
                                                             Table
                          Sales of Distillate Fuel Oils to End Users
1.3-18
1984-1999 (thousands of barrels per day)
Resi-
Year dential
1984 450
1985 471
1986 476
1987 484
1988 498
1989 489
1990 393
1991 391
1992 406
1993 429
1994 413
1995 416
1996 436
1997 423
1998 367
1999 381
Com-
mercial
319
294
280
279
269
252
228
226
218
218
218
216
223
210
199
196
Off-
Indust- Oil Electric Rail- Vessel Highway Highway All
rial Co. Farm Utility road Bunkering Diesel Military Diesel Other Total
153 59 193 45 225 110 1,093 45 109 44 2,845
169 57 216 34 209 124 1,127 50 105 12 2,868
175 49 220 40 202 133 1,169 50 111 9 2,914
190 58 211 42 205 145 1,185 58 113 5 2,976
170 57 223 52 212 150 1,304 64 119 4 3,122
167 55 209 70 213 154 1,378 61 107 2 3,157
160 63 215 48 209 143 1,393 51 116 (s) 3,021
152 59 214 39 197 141 1,336 54 110 (s) 2,921
144 51 228 30 209 146 1,391 42 113 (s) 2,979
128 50 211 38 190 133 1,485 31 127 (s) 3,041
136 46 209 49 200 132 1,594 34 130 (s) 3,162
132 36 211 39 208 129 1,668 24 126 — 3,207
137 41 217 45 213 142 1,754 24 134 — 3,365
141 41 216 42 200 137 1,867 22 136 — 3,435
147 37 198 63 185 139 1,967 18 142 — 3,461
142 38 189 60 182 135 2,091 19 140 — 3,572
Source: U.S. Department of Energy, Energy Information Administration (EIA). 2001a.  Annual Energy Review, 2000, Table 5-13.  Washington, DC.

-------
Draft Regulatory Support Document
1.4 Distribution and Storage Operations

    Refined petroleum products, including gasoline, distillates, and jet fuel, are transported by
barge and truck and through pipelines from refineries to the wholesale and retail networks in the
major markets of the United States.  The most important of these routes is the 86,500-mile
pipeline network, operated by nearly 200 separate companies (AOPL, 2000; FERC, 2002).
Terminals and other storage facilities are located near refineries, along pipelines at breakout
stations, and at bulk plants near major consumer markets.  There are currently more than 1,300
terminals for refined products in the United States (API, 2002).

1.4.1 The Supply-Side

   Pipelines are constructed of large-diameter welded steel pipe and typically buried
underground. Pumps at the source provide motive force for the 3 to 8 miles per hour flow in the
piping network (API, 1998; AOPL, 2000).  Periodically, the line pressure is boosted at
strategically placed pumping stations, which are often located at breakout points for intermediate
distribution of various components.  The product is moved rapidly enough to ensure turbulent
flow, which prevents back-mixing of components. Figure 1.4-1 shows a typical configuration of
several refined components on the Colonial Pipeline, a major artery connecting East Texas
producing sites to Atlanta, Charlotte, Richmond, and New Jersey.

   The pipelines do not change the physical form of the petroleum products that they carry and
add value only by moving the products closer to markets.  Operating  costs of transporting
products in a pipeline are quite small, so most of the cost charged to customers  represents
amortization of capital costs for construction. According to the 1997 Economic Census,
revenues for pipeline transportation, NIACS code 48691, were $2.5 billion, of which only $288
million represented wags and salaries (U.S. Census Bureau, 2000). Almost all pipeline
companies act as a common carrier (they do not take ownership of the products they transport),
so their revenues and economic value added are equivalent. Census data for storage operations
are not broken down in enough detail to permit estimation of revenues or value added.
                                         1-34

-------
                                                            Industry Characterization
                                      Figure 1.4-1
    Typical Sequence in which Products are Batched While in Transit on Colonial System
                                               Reformulated Regular Gasoline

                                             • Low Sulfur Diesel Fuel
                                             O Kerosine/Jet Fuel
                                             • High Sulfur Diesel Fuel
                                             © Conventional Regular Gasoline
                                             O All Premium Grades
                          • Compatible Interfaces      • Reformulated Regular Gasoline
                          D Transmix
                           (Interface material which
                            must be reprocessed)
    The most important impact of additional EPA regulation on the distribution network has been
to increase the number of different products handled by each pipeline.  Although some concern
has been expressed by these firms in relation to the gasoline and highway diesel regulations, the
incremental effect of reducing sulfur content for nonroad diesel should be minor. The Colonial
Pipeline mentioned previously currently handles 38 grades of motor gasoline, 16 grades of
distillate products, 7 grades of kerosene-type fuels (including jet fuel), and an intermediate
refinery product, light cycle oil (Colonial, 2002).

    As Figurel.4-1  shows, these pipelines are shipping low-sulfur gasoline, low-sulfur diesel
fuel, and high-sulfur nonroad fuel in the same pipeline.  In most cases, the interface (mixing
zone) between products is degraded to the poorer quality material. When they begin handling
ultralow-sulfur diesel fuel and gasoline, they may be forced to downgrade more interface
material to nonroad or fuel oil and will need to carefully prevent contamination in storage tanks
and pumping stations.

    Importantly, changeover to ultralow-sulfur diesel fuel for nonroad applications will not add
additional complexity to their operations.  We expect there to be no physical difference between
15 ppm diesel fuel destined for the highway market and 15 ppm diesel fuel destined for the off-
highway market prior to the terminal level when dye must be added to off-highway diesel fuel to
denote its untaxed status.  This will allow pipeline operators to ship such fuels in fungible
batches.  Consequently, the introduction of 15 ppm off-highway diesel should not result in
increased difficulty in limiting sulfur contamination during the transportation of ultra-low sulfur
products.  Pipeline operators will continue to have a market for the downgraded mixing zone

                                           1-35

-------
Draft Regulatory Support Document
material generated during the shipment of 15 ppm diesel fuel by pipeline.  After the 15 ppm
standard for highway diesel fuel and the comparable fuel standards in this final rule take effect,
the pipelines that transport the majority of the nation's diesel fuel are projected to continue to
carry high-sulfur diesel fuel and/or 500 ppm diesel fuel.  These pipelines will blend their
downgraded 15 ppm diesel into the 500 ppm and/or high-sulfur diesel fuel that they ship.  A
fraction of the pipelines are projected to carry only a single grade of diesel fuel (15 ppm fuel)
after the HD2007 rule takes effect. These pipelines currently carry only 500 ppm highway diesel
fuel. In the HD2007 rule, we projected that these pipelines will install an additional storage tank
to contain the relatively low volumes of downgraded 15  ppm diesel fuel generated during
pipeline transportation of the product. We projected that this downgraded material will be sold
into the off-highway diesel market. The new regulation  of nonroad diesel fuel will not change
this practice.  We expect these pipeline operators to continue finding a market for the
downgraded 15 ppm fuel, either as 500 ppm off-highway diesel fuel or for use in stationary
diesel engines.

1.4.2 The Demand-Side

       Demand for  distribution through pipelines (versus barge or truck movement) is driven by
cost differentials with these alternate means of transportation. The National Petroleum Council
estimated in a comprehensive 1989 report that water transport of a gallon of petroleum products
was about three times as expensive per mile as transport via pipeline, and truck transportation
was up to 25 times as expensive per mile (National Petroleum Council, 1989).  A recent pipeline
industry publication shows that pipelines handle around  60 percent of refined petroleum product
movements, with 31 percent transported by water, 5.5 percent by truck, and 3.5 percent by rail
(AOPL, 2001).

   Pipeline transport charges make up only a small portion of the delivered cost of fuels.
Industry publications cite costs of about $1 per barrel, equal to 2.5 cents per gallon, for a 1600
mile transfer from Houston to New Jersey, and about 2 cents per gallon for a shipment of 1100
miles from Houston to Chicago (AOPL, 2002; Allegro, 2001). Although average hauls are
shorter and somewhat more expensive per mile, average transport rates are on the order of 0.06
to 0.18 cents per barrel  per mile.

1.4.3 Industry Organization

   Just as it has with other transportation modes defined by site-specific assets and high fixed
costs, the federal government has traditionally regulated pipelines as common carriers.  Unlike
railroad and long-haul trucking,  however, pipeline transport was not deregulated during the
1980s, and the Federal Energy Regulatory Commission (FERC) still sets allowable tariffs for
pipeline movements. A majority of carriers, therefore, compete as regulated monopolies.

   Most pipelines are permitted small annual increases in rates without regulatory  approval,
typically limited to 1 percent less than the increase in the producer price index (PPI).  If
regulatory changes caused significant cost increases, for instance from the addition of tankage to
handle two grades of nonroad diesel fuel, pipeline operators would have to engage in a rate case

                                          1-36

-------
                                                           Industry Characterization
with FERC to pass their increased costs along to consumers. If they chose not to request rate
relief, the pipelines would absorb any costs above the allowable annual increases.

1.4.4 Markets and Trends

   Pipeline firms have seen slowly rising demand for their services over the past several years.
The latest available data, from the 1996 to 1999 period, are displayed in Table 1.4-1. Pipelines
have not only captured most of the overall increase in total product movements, they have also
taken some share away from water transport during the period. Railroad shipments have grown
as well, but from a very small base.

                                        Table 1.4-1
                   Trends in Transportation of Refined Petroleum Products

Pipelines
Water Carriers
Motor Carriers
Railroads
Totals
1996
280.9
154.1
28.0
16.0
479.0
1997
279.1
148.3
26.0
16.2
469.6
1998
285.7
147.1
26.7
16.2
475.7
1999
296.6
147.5
27.6
18.2
489.9
Percentage Change
1996-1999
5.6%
-4.3%
-1.4%
13.8%
2.2%
       Note: All figures, except percentages, in billions of ton-miles.
       Source:  Association of Oil Pipe Lines (AOPL). 2001.  Shifts in Petroleum Transportation. As accessed on
          November 20, 2002. www.aopl.org/pubs/facts.html.
                                           1-37

-------
Draft Regulatory Support Document
References to Chapter 1

1. RTI. (2003). Industry Profile for Nonroad Diesel Tier 4 Rule - Final Report.  Prepared for the
U.S. Environmental Protection Agency. EPA Contract Number 68-D-99-024, April 2003.  A
copy can be found in Docket A-2001-28, Document No. II-A-182.

2. Power Systems Research(PSR). 2002. OELink Sales Database.

3. RTI. (2003). Economic Impact Analysis for Nonroad Diesel Tier 4 Rule.  Prepared for the
U.S. Environmental Protection Agency.  EPA Contract No. 68-D-99-024, April 2003. A copy
can be found in Docket A-2001-28, Document II-A-115.
Allegro Energy Group. 2001.  How Pipelines Make the Oil Market Work—Their Networks,
    Operations, and Regulations. New York: Allegro. A copy can also be found in Docket A-
    2001-28, Document No. II-A-137.

American Petroleum Institute (API).  1998.  "All About Petroleum." As accessed on November
    20, 2002. api-ec.api.org/filelibrary/AllAboutPetroleum.pdf A copy can also be found in
    Docket A-2001-28, Document No.  II-A-172.

American Petroleum Institute (API).  2001.  "Pipelines Need Operational Flexibility to Meet
    America's Energy Needs." As accessed on November 20, 2002.  api-
    ep.api.org/industry/index.cfm. A copy can also be found in Docket A-2001-28, Document
    No. II-A-173.

American Petroleum Institute (API).  2002.  "Marketing Basic Facts."  As accessed on
    September 25, 2002. www.api.org/industry/marketing/markbasic.htm.

Association of Oil Pipe Lines (AOPL).  2000. "Fact Sheet: U.S. Oil Pipe Line Industry." As
    accessed on November 20, 2002. www.aopl.org/pubs/pdf/fs2000.pdf

Association of Oil Pipe Lines (AOPL).  2001. "Shifts in Petroleum Transportation." As
    accessed on November 20, 2002. www.aopl.org/pubs/facts.html.

Association of Oil Pipe Lines (AOPL).  2002. "Why Pipelines?" As accessed on November 20,
    2002. www.aopl.org/about/pipelines.html.

Business & Company  Resource Center. http://www.gale.com/servlet/Item
    DetailServlet?region=9&imprint=000&titleCode=GAL49&type=l&id=l 15085. A copy of
    information describing this database can be found in Docket A-2001-28, Document No.11-
    B-42.
                                        1-38

-------
                                                        Industry Characterization
Chevron. 2002. "Diesel Fuel Refining and Chemistry." As accessed on August 19, 2002.
    www.chevron.com/prodserv/fuels/bulletin/diesel/L2_4_2rf.htm.

Colonial. 2002. "Frequently Asked Questions." As accessed on September 24, 2002.
    www.colpipe.com/ab_faq.asp.

Considine, Timothy J.  2002. "Inventories and Market Power in the World Crude Oil Market."
    As accessed on November 1, 2002. http://www.personal.psu.edu/faculty/ c/p/
    cpw/resume/fmvmarketpower.html.

Dun & Bradstreet.  Million Dollar Directory,  http://www.dnb.com/dbproducts/description/
    0,2867,2-223-1012-0-223-142-177-l,OO.html.

Federal Energy Regulatory Commission (FERC).  2002. FERC Form No. 6, Annual Report of
    Oil Pipelines. www.ferc.fed.us/oil/oil_list.htm. A  copy of this webpage can be found in
    Docket A-2001-28, Document No. II-B-43.

Flint Hills Resources. 2002. "Refining Overview." As accessed on September 10, 2002.
    www.fhr.com/Refmingl01/default.asp.

Federal Trade Commission (FTC). Midwest Gasoline Price Investigation, March 29, 2001, p.7.
    As accessed September 25, 2002.  www.ftc.gov/os/2001/03/mwgasrpt.htm. A copy can also
    be found in Docket A-2001-28, Document No. II-A-23.

Freedonia Group.  2001. "Diesel Engines and Parts in the United States to 2005—Industry
    Structure."  http://www.freedoniagroup.com/scripts/cgiip.exe/WService=freedonia
    /abstract.html?ARTNUM=l 153.

Hoover's Online, http://www.hoovers.com/.

National Petroleum Council.  1989.  "Petroleum Storage and Transportation." System
    Dynamics. Volume II. Washington, DC: National Petroleum Council.

U.S. Department of Agriculture, National Agricultural Statistics Service (USDA-NASS).  2002.
    Agricultural Statistics 2002.  See especially Tables 15-1, 15-2, 9-39, 9-40. Washington, DC:
    U.S. Department of Agriculture. This document can be accessed at
    http://www.usda.gov/nass/pubs/agr02/acr02.htm. A copy of selected tables can be found in
    Docket A-2001-28, Document No. II-A-  174.

U.S. Department of Energy, Energy Information Administration (EIA). 200 la.  Annual Energy
    Review, 2000.  See especially Table 5.13. Washington, DC: Department of Energy.
    DOE/EIA-0384(2000), August 2001. This document can be found at
    http://tonto.eia.doe.gov/FTPROOT/multifuel/038400.pdf A copy of this document can also
    be found in Docket A-2001-28, Document No. II-A-175.
                                         1-39

-------
Draft Regulatory Support Document
U.S. Department of Energy, Energy Information Administration (EIA). 2001b. Fuel Oil and
    Kerosene Sales, 2000, Tables 7-12.  Washington, DC: Department of Energy. DOE-EIA-
    0535(00), Distribution Category UC-950, September 2001. A copy of this document can
    also be found at
    http://www.eia.doe.gov/pub/oil_gas/petroleum/data_publications/fuel_oil_and_kerosene
    _sales/historical/2000/foks_2000.html. A copy can also be found in Docket A-2001-28,
    Document No. II-A-176.

U.S. Department of Energy, Energy Information Administration (EIA). 2002a. Petroleum
    Supply Annual 2001. Volumes 1 and 2. See especially Volume I, Table 16(page 48).
    Washington, DC: Department of Energy. DOE/EIA-0340(01)/1 and DOE/EIA-0340(01)/2.
    June 2002. A copy of these documents can also be found at
    http ://www. eia. doe. gov/oil_gas/petroleum/data_publications/petroleum_supply_annual/psa_
    volumel/psa_volumel.html  and
    http://www.eia.doe.gov/oil_gas/petroleum/data_publications/petroleum_supply_annual/psa_
    volumel/psa_volume2.html. A copy of these documents can also be found in Docket A-
    2001-28, Documents No. II-A-165 and II-A-177.

U.S. Department of Energy, Energy Information Administration (EIA). 2002b. Refinery
    Capacity Data Annual.  As accessed on September 23, 2002.
    http://www.eia.doe. gov/oil_gas/petroleum/data_publications/refinery_capacity_data/refcapa
    city.html. Washington, DC: Department of Energy. A copy of this document is also
    available in Docket A-2001-28, Document No. II-A-178.

U.S. Environmental Protection Agency.  1995a. EPA Office of Compliance Sector Notebook
    Project: Profile of the Motor Vehicle Assembly Industry. EPA310-R-95-009. Washington,
    DC: U.S. Environmental Protection Agency.

U.S. Environmental Protection Agency (EPA).  1995b.  Profile of the Petroleum Refining
    Industry. EPA Industry Sector Notebook Series. U.S. Environmental Protection Agency.

U.S. Environmental Protection Agency (EPA).  2000. Heavy-Duty Standards/Diesel Fuel RIA.
    EPA420-R-00-026. Washington, DC: U.S. Environmental Protection Agency.

U.S. Environmental Protection Agency (EPA).  2002. Highway Diesel Progress Review.
    EPA420-R-02-016. Washington, DC: EPA Office of Air and Radiation.
U.S. Census Bureau. 1992 Census of Manufactures, Industry Series, Petroleum and Coal
    Products. MC92-I-29A. Table 1 A. A copy of this document is available at
    http://www.census.gOv/prod/l/manmin/92mmi/mci29af.pdf. A copy can also be found in
    Docket A-2001-28, Document No. II-A-179.

U.S. Census Bureau. 1997 Economic Census, Manufacturing, Industry Series, Petroleum
    Refineries, EC97M-3241A, Table 1.  A copy of this document can be found at

                                         1-40

-------
                                                       Industry Characterization
    http://www.census.gov/prod/ec97/97m3241a.pdf. A copy can also be found in Docket A-
    2001-28, Document No. II-A-180.

U.S. Census Bureau, 2002 Annual Survey of Manufactures, Statistics for Industry Groups and
    Industries MOO(AS)-1, MOO(AS)-1, Table 2. A copy of this document can be found at
    http://www.census.gov/prod/2002pubs/mOOas-l.pdf. A copy  can also be found in Docket
    A-2001-28, Document No. II-A-181.
                                        1-41

-------
CHAPTER 2: Air Quality, Health, and Welfare Effects
   2.1 Particulate Matter	2-3
       2.1.1 Health Effects of Particulate Matter	2-4
       2.1.2 Attainment and Maintenance of the PM10 and PM2 5 NAAQS:  Current and Future
          Air Quality	2-16
          2.1.2.1 Current PM Air Quality	2-16
          2.1.2.2 Risk of Future Violations	2-26
       2.1.3 Environmental Effects of Particulate Matter  	2-38
          2.1.3.1  Visibility Degradation	2-39
          2.1.3.2 Other Effects 	2-51
   2.2 Air Toxics  	2-55
       2.2.1 Diesel Exhaust PM	2-55
          2.2.1.1 Potential Cancer Effects of Diesel Exhaust	2-55
          2.2.1.2 Other Health Effects of Diesel Exhaust	2-59
          2.2.1.3 Diesel Exhaust PM Ambient Levels 	2-61
          2.2.1.4 Diesel Exhaust PM Exposures  	2-71
       2.2.2 Gaseous Air Toxics	2-75
          2.2.2.1 Benzene 	2-79
          2.2.2.2  1,3-Butadiene 	2-82
          2.2.2.4 Acetaldehyde  	2-85
          2.2.2.6 Polycyclic Organic Matter	2-87
          2.2.2.7 Dioxins	2-88
   2.3 Ozone	2-88
       2.3.1 Health Effects of Ozone  	2-89
       2.3.2 Attainment and Maintenance of the 1-Hour and 8-Hour Ozone NAAQS 	2-92
       2.3.2 Attainment and Maintenance of the 1-Hour and 8-Hour Ozone NAAQS 	2-93
          2.3.2.1 1-Hour Ozone Nonattainment and Maintenance Areas and Concentration 2-95
          2.3.2.2 8-Hour Ozone Levels: Current Nonattainment and Future Concentrations 2-97
          2.3.2.3 Potentially Counterproductive Impacts on Ozone Concentrations from NOx
              Emission Reductions	2-113
       2.3.3 Welfare Effects Associated with Ozone and its Precursors  	2-118
   2.4 Carbon Monoxide  	2-121
          2.4.1  General Background       	2-121
          2.4.2  Health Effects of CO 	2-122
          2.4.3  CO Nonattainment	2-122

-------
                                              Air Quality, Health, and Welfare Effects
    CHAPTER 2: Air Quality, Health,  and Welfare Effects
   With this rulemaking, we are acting to extend highway types of emission controls to another
major source of diesel engine emissions: nonroad land-based diesel engines. This final rule sets
out emission standards for nonroad land-based diesel engines - engines used mainly in
construction, agricultural, industrial and mining operations - that will achieve reductions in
particulate matter (PM) and NOx standards in excess of 95 percent and 90 percent, respectively.
This action also regulates nonroad diesel fuel for the first time by reducing sulfur levels in this
fuel more than 99 percent to 15 part per million (ppm). The diesel fuel sulfur requirements will
decrease PM and sulfur dioxide (SO2) emissions for land-based diesel engines, as well as for
three other nonroad source  categories: commercial marine diesel vessels, locomotives, and
recreational marine diesel engines.

   These sources are significant contributors to atmospheric pollution of (among other
pollutants) PM, ozone and a variety of toxic air pollutants. In 1996, emissions from these four
source categories were estimated to be 40 percent of the mobile source inventory for PM25 and
25 percent for NOx. Without further control beyond those we have already adopted, by the year
2030, these sources will emit 44 percent of PM25 from mobile sources, and 47 percent of NOx
emissions from mobile sources. Thus, reducing emissions from nonroad sources is critically
important to achieving the nation's air quality goals.

   In 2030, we estimate that this program will reduce over 129,000 tons PM2 5 and 738,000 tons
of NOx. It will also virtually eliminate nonroad diesel SO2 emissions, which amounted to
approximately 236,000 tons in 1996, and would otherwise grow to approximately 379,000 tons
by 2030.

   These dramatic reductions in nonroad emissions are a critical part of the effort by Federal,
State, local and Tribal governments to reduce the health related impacts of air pollution and to
reach attainment of the National Ambient Air Quality Standard (NAAQS) for PM and ozone, as
well  as to improve environmental effects such as visibility.  These emission reductions will be
directly helpful to the 474 partial and full counties nationwide that have been recently designated
as nonattainment areas for the 8-hour ozone standard and the PM2 5 areas that will be designated
later this year.  Based on the most recent monitoring data available for this rule, such problems
are widespread in the United States. There are almost 65 million people living in 120 counties
with PM25 levels exceeding the PM25 NAAQS (based on 2000-2002), and about 159 million
people living in 474 partial and full counties that are in nonattainment for either failing to meet
the 8-hour ozone NAAQS or for contributing to poor air quality in a nearby area. Figure 2.-1
illustrates the widespread nature of these problems. Shown in this figure are counties exceeding
either or both of the PM25 NAAQS or designated 8-hour ozone nonattainment areas plus
mandatory Federal Class I areas, which have particular needs for reductions in haze.
                                          2-1

-------
Final Regulatory Impact Analysis
    As described in Chapter 9, the air quality improvements expected from this rulemaking will
produce major benefits to human health and welfare, with a combined value in excess of three
quarters of a trillion dollars between 2007 and 2036. By the year 2030, we expect that this rule
will annually prevent approximately 12,000 premature deaths and 15,000 nonfatal heart attacks.
By 2030, it will  also prevent 13,000 annual acute bronchitis attacks in children, 280,000 upper
and lower respiratory symptoms in children, nearly 1 million lost work days among adults
because of their own symptoms, and 5.9 million days where adults have to restrict their activities
due to symptoms in 2030.

                    Figure 1-1. Air Quality Problems are Widespread.
        | 8 Houi Ozone Noiigency guidance
    In this chapter and chapter 3, we describe in more detail the air pollution problems associated
with emissions from nonroad diesel engines and air quality information that we are relying upon
in this rulemaking. To meet these emission standards, engine manufacturers directly control
emissions of NOx, PM, non-methane hydrocarbons (NMHC), and to a lesser extent, carbon
monoxide (CO).  Gaseous air toxics from nonroad diesel engines will also decrease as a
consequence of the new emission standards. In addition, there will be a substantial reduction in
                                            2-2

-------
                                               Air Quality, Health, and Welfare Effects
SO2 emissions resulting from the decreasing sulfur level in diesel fuel. SO2 is transformed in
the atmosphere to form PM (sulfate) and can also pose a public health hazard in the gas phase.

    From a public health perspective, we are primarily concerned with nonroad engine
contributions to atmospheric levels of particulate matter in general (diesel PM in particular),
various gaseous air toxics emitted by diesel  engines, and ozone.A We will first review important
public health effects caused by these pollutants, briefly describing the human health effects, and
we will then review the current and expected future ambient levels of directly or indirectly
caused pollution.  Our presentation will show that substantial further reductions  of these
pollutants,  and the underlying emissions from nonroad diesel engines, will be needed to protect
public health.

    Following discussion of health effects, we will discuss a number of welfare effects associated
with emissions from diesel engines.  These effects include atmospheric visibility impairment,
ecological and property damage caused by acid deposition, eutrophication and nitrification of
surface waters, environmental threats posed by polycyclic organic matter (POM) deposition, and
plant and crop damage from ozone. Once again, the information available to us  indicates a
continuing need for further nonroad emission reductions to bring about improvements in air
quality.

2.1 Particulate Matter

    Particulate matter (PM) represents a broad class of chemically and physically diverse
substances. It can be principally characterized as discrete particles that exist in the condensed
(liquid or solid) phase spanning several orders of magnitude in size. PM10 refers to particles with
an aerodynamic diameter less than or equal to a nominal 10 micrometers. Fine particles refer to
those particles with an aerodynamic diameter less than or equal to a nominal 2.5 micrometers
(also known as PM2 5), and coarse fraction particles are those particles with  an aerodynamic
diameter greater than 2.5 microns, but less than or equal to a nominal 10 micrometers. Ultrafine
PM refers to particles with diameters of less than 100 nanometers (0.1 micrometers).  The health
and environmental effects of PM are in some cases related to the size of the particles.
Specifically, larger particles (greater than 10 micrometers) tend to be deposited nasally and in
the larger conducting airways, and they are removed by the respiratory clearance mechanisms
whereas smaller particles (PM10) are deposited deeper in the lungs.  Also, fine particles scatter
light obstructing visibility.
   In addition to directly emitted particles, nonroad diesel engines currently emit high levels of
NOx, which reacts in the atmosphere to form secondary PM2 5 (namely ammonium nitrate).
   AAmbient PM from nonroad diesel engine is associated with the direct emission of diesel PM
and sulfate PM, and with PM formed indirectly in the atmosphere by NOx and SO2 emissions
(and to a lesser extent NMHC emissions). Both NOx and NMHC can participate in the
atmospheric chemical reactions that produce ozone.

                                             20
                                             -j

-------
Final Regulatory Impact Analysis
Nonroad diesel engines also emit SO2 and HC, which react in the atmosphere to form secondary
PM25 (namely sulfates and organic carbonaceous PM25).  Both types of directly and indirectly
formed particles from nonroad engines are found principally in the fine fraction.  Thus, this
discussion will focus on fine particles (PM2 5). Ambient fine particles are a complex mixture
generally composed of sulfate, nitrate, chloride, ammonium compounds, organic carbon,
elemental carbon, and metals. Fine particles can remain in the atmosphere for days to weeks and
travel through the atmosphere hundreds to thousands of kilometers, while coarse particles
generally tend to deposit to the earth within minutes to hours and within tens of kilometers from
the emission source.

2.1.1 Health Effects of Particulate Matter

   Scientific studies show ambient PM concentrations  (which are attributable to a number of
sources including diesel) contribute to a series of adverse health effects. These health effects are
discussed in detail in the EPA Air Quality Criteria Document for PM (PM Criteria Document) as
well as the draft updates of this document released in the past year.1 EPA's Health Assessment
Document for Diesel Engine Exhaust (Diesel HAD) also reviewed health effects information
related to diesel exhaust as a whole including diesel PM, which is one component of ambient
PM.2  We are relying on the data and conclusions in these documents regarding the effects of
particulate matter. We also present additional recent studies. Taken together this information
supports  the conclusion that PM-related emissions from nonroad diesel engines have been
associated with adverse health effects.

   We received a number of public comments on specific health studies, and we are relying on
the discussions and conclusions  presented in the PM Criteria Document and Diesel HAD in
which EPA prepared detailed evaluations of the body of scientific information and subjected
those evaluations  to extensive public and expert peer review.  Additional information is
available in the Summary and Analysis of Public Comments that accompanies this final rule.

   2.1.1.1 Short-Term Exposure-Mortality and Morbidity Studies

   As detailed in the PM Criteria Document, health effects associated with short-term variation
in ambient PM have been indicated by numerous epidemiologic studies showing associations
between  exposure and increased hospital admissions for ischemic heart disease,3 heart failure,4
respiratory disease,5'6'7'8 including chronic obstructive pulmonary disease (COPD) and
pneumonia.9'10'n  Short-term elevations in ambient PM have also been associated with increased
cough, lower respiratory symptoms, and decrements in lung function.12'13'14  Short-term
variations in ambient PM have also been associated with increases in total and cardiorespiratory
daily mortality in  individual cities15'16'17'18 and in multi-city studies.19'20'21

   Several studies specifically address the contribution of PM from mobile sources  in these
time-series studies.  Analyses incorporating source apportionment by factor analysis with daily
time-series studies of daily death also established a specific influence of mobile source-related
PM2 5 on  daily mortality22 and a  concentration-response function for mobile source-associated
PM25 and daily mortality.23 Another recent study in 14 U.S. cities examined the  effect of PM10

                                           2-4

-------
                                              Air Quality, Health, and Welfare Effects
exposures on daily hospital admissions for cardiovascular disease (CVD). They found that the
effect of PM10 was significantly greater in areas with a larger proportion of PM10 coming from
motor vehicles, indicating that PM10 from these sources may have a greater effect on the toxicity
of ambient PM10 when compared with other sources.24

   In 2002, questions were raised about the default convergence criteria and standard error
calculations made using generalized additive models (GAM), which has been the statistical
model of choice in many of the time-series epidemiologic studies.  A number of time-series
studies were reanalyzed using alternative methods, typically GAM with more stringent
convergence criteria and an alternative model such as generalized linear models (GLM) with
natural smoothing splines. Since then, the Health Effects Institute convened an expert panel to
review the results of  and the results of the reanalyses have been compiled and reviewed in a
recent HEI publication.25  In most, but not all, of the reanalyzed studies, it was found that risk
estimates were reduced and confidence intervals increased with the use of GAM with more
stringent convergence criteria or GLM analyses; however, the reanalyses generally did not
substantially change the findings of the original studies, and the changes in risk estimates with
alternative analysis methods were much smaller than the variation in effects across studies.  The
HEI review committee concluded the following:

   a.  While the number of studies showing an association of PM with mortality was slightly
       smaller, the PM association persisted in the majority of studies.
   b.  In some of the large number of studies in which the PM association persisted, the
       estimates of PM effect were substantially smaller.
   c.  In the few studies in which investigators performed further sensitivity analyses, some
       showed marked sensitivity of the PM effect estimate to the degree of smoothing and/or
       the specification of weather.

   As discussed in Chapter 9, examination of the original studies used in our economic benefits
analysis found that the health endpoints that are potentially  affected by the GAM issues include:
reduced hospital admissions, reduced lower respiratory symptoms, and reduced premature
mortality due to short-term PM exposures. It is important to note that the benefits estimates
derived from the long-term exposure studies, which account for a major share of the economic
benefits described in  Chapter 9, are not affected.  Similarly, the time-series studies and case-
crossover studies employing generalized linear models or other parametric methods are not
affected.

   2.1.1.2 Long-Term Exposure Mortality and Morbidity Studies

   Short-term studies provide one way of examining the effect of short-term variations in air
quality on morbidity  and mortality. However,  they do not allow for an evaluation of the effect of
long-term exposure to air pollution on human mortality and morbidly.26 Longitudinal cohort
studies allow for analysis of such effects.

   As discussed in the  PM Criteria Document, the newer morbidity studies that combine the
features of cross-sectional and cohort studies provide the best evidence for chronic exposure

                                           2-5

-------
Final Regulatory Impact Analysis
effects.  The Gauderman et al. studies both found significant decreases in lung function growth
among southern California school children to be related to PM2 5 and/or PM10 levels.27,28
However, Peters et al. reported no relationship between respiratory symptoms and annual
average PM10 levels in 12 southern California communities.29 Long-term (months to years)
exposure to PM was linked with decreased lung function and increased incidence of respiratory
disease such as bronchitis (PM Criteria Document 1996, p. V-26, Abbey et al. 1995).  The
results of studies using long-term and short-term PM exposure data were reported to be
consistent with one another. In addition, toxicology studies using surrogate particles or PM
components, generally at high concentrations, and autopsy studies of humans and animals
reported evidence of pulmonary effects, including morphological damage (e.g., changes in
cellular structure of the airways) and changes in resistance to infection.

   Additional data are available regarding long-term PM exposures and mortality. To date, four
major cohorts in the U.S. have examined mortality and long-term exposure to PM25. These
studies are described in detail in the PM Criteria Document and we are relying on the analyses
and conclusions in that document for these studies. Many of the issues raised in public comment
are addressed by the Criteria Document (as detailed in the Summary and Analysis of public
comments document.) In addition to the U.S. studies, there are additional data from Europe and
Canada.  A cohort in the Netherlands evaluated exposure to mobile source-related pollutants.30
Another study examines exposure-mortality relationships with income in southern Ontario,
Canada.31

   Two major U.S. cohort studies, the Harvard Six Cities and the American Cancer Society
studies,  suggest an association between exposure to ambient PM2 5 measured in the city of
residence and premature mortality from cardiorespiratory causes.32'33 As discussed in the PM
Criteria Document, these two prospective cohort studies tracked health outcomes in discrete
groups of people over time. Subsequent reanalysis of these studies have confirmed the findings
of these articles, and a recent extension of the ACS cohort study found statistically significant
increases in lung cancer mortality risk associated with ambient PM2 5.34 This most recent finding
is of special interest in this rulemaking, because of the association of diesel exhaust and lung
cancer in occupational studies of varying design.

   More recently, the Adventist Health Study on Smog (AHSMOG) in California indicated that
long-term exposure to PM10 resulted in a significant risk of premature mortality in men, although
risks were not elevated among women.35 In another AHSMOG analysis, ambient PM2 5 estimates
made from visibility data at an  airport were used to compare the  effects of PM10 and PM25 for the
cohort.36 No statistically significant increase in risk was observed with any component of PM.
Among  men, the PM2 5 coefficient on mortality from all natural causes was consistently larger
than the coarse fraction of PM10.  Among women, no elevation in mortality risk was found for
any PM index.

   Another study evaluated in the PM Criteria Document examining long-term exposure to
ambient PM and mortality is the Electric Power Research Institute (EPRI)-Washington
University mortality study in American Veterans.37 The Veterans Study was originally designed
as a means of assessing the efficacy of anti-hypertensive drugs in reducing morbidity and

                                          2-6

-------
                                              Air Quality, Health, and Welfare Effects
mortality in a population with pre-existing high blood pressure (in this case, male veterans)
(Lipfert et al., 2000). Unlike previous long-term analyses, this study found some associations
between premature mortality and ozone but found inconsistent results for PM indicators. A
variety of issues associated with the study design, including sample representativeness and loss
to follow up, make this cohort a poor choice for extrapolating to the general public.
Furthermore, the selective nature of the population in the veteran's cohort and methodological
weaknesses may have resulted in estimates of relative risk that are biased relative to a relative
risk for the general population.

   The Hoek et al. (2002) study examines a cohort of residents of the Netherlands who were
recruited as part of the Netherlands Cohort study on Diet and Cancer (NLCS).38 Five thousand
study participants were selected at random from the larger cohort, which consisted of persons
aged 55 to 69 in 1986, with follow up until 1994.  In 1986, all participants filled out
questionnaires on diet and other risk factors.  All participants with full questionnaire data were
included in the study. Each participants' home address was mapped by  street address.
Individual exposures to ambient pollutants were assigned by matching residential address to an
exposure metric via geographic information  system (GIS). "Black smoke" - widely used in
Europe as a surrogate of particulate elemental carbon - and NO2 had been previously assessed as
a function of regional background, urban background, and contribution from local traffic based
on proximity to busy roads.39  Results of the  survival analysis indicated that residential black
smoke predicted from regional, urban, and intra-urban variation was associated with a relative
risk (RR) of cardiopulmonary mortality per 10 ug/m3 of 1.71 (with a 95  percent confidence
interval (CI) of [1.10, 2.67]) and an RR for all-cause mortality of 1.31 [0.95, 1.80].  In a model
including background black smoke and proximity to a major roadway, the cardiopulmonary
mortality RR associated with living near a busy road was 1.95 [1.09, 3.51]. This study is of
particular interest in this rule, because of the strong focus on mobile source pollutants in the
exposure assessment portion of the study. This study also highlights the "near-roadway" health
concerns, discussed later.

   The Six Cities, ACS, AHSMOG, Veterans, and NLCS Studies are discussed in detail in the
draft PM Criteria Document and revised Chapter 8. We are relying on the evaluations and
conclusions presented in those documents. The long-term exposure health effects of PM are
summarized in Table 2.1.1-1, which is taken directly from Table 9-11 of the draft Air Quality
Criteria Document referenced earlier that was released in 2003. This document is continuing to
undergo expert and public review. One study discussed below does not appear in the PM
Criteria Document because it was published  after the date required for inclusion in the Criteria
Document.40

   Finklestein et al. (2003) examined a cohort of 5,228 residents of the Hamilton-Burnling area
of southern Ontario, Canada who had been referred for lung function testing between 1985 and
1999.41 The study was not a random sample of the population in the Hamilton-Burlington area.
Total non-accidental and cardiopulmonary mortalities between 1992 and 1999 were determined
based on the Ontario Mortality Registry.  The subjects'  age, sex, postal code, body mass index,
and pulmonary function test results were matched with disease diagnosis via the Ontario Health
Insurance Plan. Canada's health insurance system allowed the investigators to determine disease

                                           2-7

-------
Final Regulatory Impact Analysis
diagnoses during the follow-up period. Postal codes were used to assign "ecological" variables
of census-derived mean household income, 24-hour average total suspended particulate (TSP)
measured every 6 days, and SO2 measured continuously during the mid-1990's  Air monitoring
data came from 9 TSP and 23 SO2 monitors, which were subject to spatial interpolation
techniques. Postal code-specific pollutant concentrations were assigned using GIS.  Analysis of
the  air quality data indicated that TSP and SO2 tended to be higher in low-income areas. The
study group was divided into higher and lower income and pollution strata, based on the median
income, and TSP and SO2 levels at the postal code level. Compared to the high-income, low-
pollution group, all other groups had significantly elevated mortality relative risks with income,
and each pollutant (in one-pollutant models) was associated with increased risk. Age appeared
as an effect modifier, with attenuated effects at elevated age.

    The 1996 PM AQCD indicated that past epidemiologic studies of chronic PM exposures
collectively indicate increases in mortality to be associated with long-term exposure to airborne
particles of ambient origins. The PM effect size estimates for total mortality from these studies
also indicated that a substantial portion of these deaths reflected cumulative PM impacts above
and beyond those exerted by acute exposure events.

    Several advances have been made in terms of further analyses and/or reanalyses of several
studies of long-term PM exposure effects on total, cardiopulmonary, or lung cancer mortality.
The Harvard Six Cities analyses (as confirmed by the HEI reanalyses) and the recent extension
of the ACS study by Pope et al. (2002) probably provide the most credible and precise estimates
of excess mortality risk associated with long-term PM2.5 exposures in the United States.

    2.1.1.3 Long-Term Exposures and Physiological Response in Individuals

    Several studies examined in the PM Criteria Document have examined the effect of long-
term exposure to air pollution on individual physiological and  organ structure.  These studies
provide insight into the biological pathways by which air pollution may act to produce adverse
health effects.  The studies below provide examples of the types of studies examined in the PM
Criteria Document.

    Studies in Vancouver, BC, and Mexico City, Mexico, have demonstrated increased retention
of PM25 in the lungs of residents of the more highly polluted Mexico City.42 More recently,
comparisons of non-smoking women in Mexico City and Vancouver have shown that particle
retention in the lungs of Mexico City women was associated with small airways remodeling.43
In another study, dogs autopsied in the Mexico City and other  less-polluted areas showed that
dogs in  more polluted areas showed greater respiratory and cardiac pathology indicative of long-
term inflammatory stress.44'45

    One recent study (not addressed in the PM Criteria Document) was conducted in Leicester,
UK studying lung cells (alveolar macrophages (AM)) obtained from children undergoing
elective surgery.46 The cells were examined by electron microscope, and the study reported that
in all children, some of the AMs contained particles, ranging from 1 to 16 percent of total AM
collected. Of particular note, the authors found that a significantly higher fraction of the AM

                                          2-8

-------
                                               Air Quality, Health, and Welfare Effects
collected from children living on main roads contained particles as compared to children living
on quiet residential roads, and that these particles were composed of single and chain aggregates
of ultrafine carbon particles that appeared to be combustion-related.  This study is of particular
relevance to this rule, given the evidence that exposure to mobile source PM results in greater
concentrations of PM in the lung.  Given the elevated exposures to carbonaceous  PM in
occupations that work with nonroad diesel engines (discussed below), this study provides a link
between nonroad PM exposure an potential  lung and systemic health effects.

   2.1.1.4 Studies of Short-Term Exposures and Physiological Response in Individuals

   A number of studies have investigated biological processes and physiological effects that
may underlie the epidemiologic findings of earlier studies. This research has found associations
between short-term changes in PM exposure with changes in heart beat, force, and rhythm,
including reduced heart rate variability (HRV), a measure of the autonomic nervous system's
control of heart function.47'48'49'50'51'52  The findings indicate associations between measures of
heart function and PM measured over the prior 3 to 24 hours or longer.  Decreased HRV has
been shown to be associated with coronary heart disease and cardiovascular mortality in both
healthy and  compromised populations.53'54'55'56

   Other studies have investigated the association between PM and such  systemic factors such
as inflammation, blood coagulability and viscosity. It is hypothesized that PM-induced
inflammation in the lung may activate a "non-adaptive" response by the immune  system,
resulting in increased markers of inflammation in the blood and tissues, heightened blood
coagulalability, and leukocyte count in the blood. A number of studies have found associations
between controlled exposure to either concentrated or ambient PM or diesel exhaust exposure
and pulmonary inflammation.57'58> 59'60  A number of studies have also shown evidence of
increased blood markers of inflammation, such as C-reactive protein, fibrinogen,  and white
blood cell count associated with inter-day variability in ambient PM.61'62'63'64 These blood
indices have been associated with coronary heart disease and cardiac events such  as heart
attack.65'66  Studies have also shown that repeated or chronic exposures to urban PM were
associated with increased severity of atherosclerosis, microthrombus formation, and other
indicators of cardiac risk.67'68

   The recent studies examining inflammation, heart rate and rhythm in relation to PM provide
some  evidence into the mechanisms by which  ambient PM may cause injury to the heart. New
epidemiologic data have indicated that short-term changes in ambient PM mass is associated
with adverse cardiac outcomes like myocardial infarction (MI) or ventricular arrythmia.69'70
These studies provide additional evidence that ambient PM2 5 can cause both acute and chronic
cardiovascular injury, which can result in death or non-fatal effects.
                                           2-9

-------
Final Regulatory Impact Analysis
                                            Table 2.1.1-1
                   Effect Estimates per Increments'1 in Long-term Mean Levels of
               Fine and Inhalable Particle Indicators From U.S. and Canadian Studies
          Type of Health
        Effect and Location
      Indicator
Change in Health Indicator per
     Increment in PM*
Range of City
PM Levels**
Means (ug/m3)
 Increased Total Mortality in Adults
                          Relative Risk (95% CI)
     Six City1
PM15/10 (20 ug/m3)
PM25 (10 ug/m3)
SO; (15 ug/m3)
      1.18(1.06-1.32)
      1.13(1.04-1.23)
      1.46(1.16-2.16)
    18-47
    11-30
    5-13
ACS Studyc
(151 U.S. SMSA)

Six City ReanalysisD

ACS Study ReanalysisD

ACS Study Extended
Analyses12
Southern CaliforniaE



Vetrans CohortR
Increased Bronchitis in Children
Six CityF
Six City0
24 City11
24 City11
24 City11
24 City11
Southern California1
12 Southern California
communities'1
(all children)
12 Southern California
communitiesK
(children with asthma)
PM25 (10 ug/m3)
SO; (15 ug/m3)
PM15/10 (20 ug/m3)
PM25 (10 ug/m3)
PM15/10 (20 ug/m3)
(dichot)
PM25 (10 ug/m3)
PM25 (10 ug/m3)
PM10 (20 ug/m3)
PM10 (cutoff =
30 days/year
>100 ug/m3)
PM10 (20 ug/m3)
PM10 (cutoff =
30 days/year
>100 ug/m3)
PM25 (10 ug/m3)

PM15/10 (50 ug/m3)
TSP (100 ug/m3)
H+ (100 nmol/m3)
SO; (15 ug/m3)
PM2, (25 ug/m3)
PM10 (50 ug/m3)
SO; (15 ug/m3)
PM10 (25 ug/m3)
Acid vapor (1.7 ppb)
PM10(19ug/m3)
PM25 (15 ug/m3)
Acid vapor (1.8 ppb)
1.07(1.04-1.10)
1.10(1.06-1.16)
1.19(1.06-1.34)
1.13(1.04-1.23)
1.04(1.01-1.07)
1.07(1.04-1.10)
1.04(1.01-1.08)
1.091 (0.985-1. 212) (males)
1.082 (1.008-1. 162) (males)
0.950 (0.873-1.033) (females)
0.958 (0.899-1.021) (females)
0.90 (0.85, 0.954; males)
Odds Ratio (95% CI)
3.26(1.13, 10.28)
2.80(1.17,7.03)
2.65(1.22,5.74)
3.02(1.28,7.03)
1.97(0.85,4.51)
3.29(0.81, 13.62)
1.39(0.99, 1.92)
0.94(0.74, 1.19)
1.16(0.79, 1.68)
1.4(1.1, 1.8)
1.4(0.9,2.3)
1.1 (0.7, 1.6)
9-34
4-24
18.2-46.5
11.0-29.6
58.7(34-101)
9.0-33.4
21.1 (SD=4.6)
51 (±17)

51 (±17)

5.6-42.3

20-59
39-114
6.2-41.0
18.1-67.3
9.1-17.3
22.0-28.6
—
28.0-84.9
0.9-3.2 ppb
13.0-70.7
6.7-31.5
1.0-5.0 ppb
                                              2-10

-------
Type of Health
Effect and Location
Increased Cough in Children
12 Southern California
communities'
(all children)
12 Southern California
communitiesK
(children with asthma)
10 Canadian
Communities5
Indicator

PM10 (20 ug/m3)
Acid vapor (1.7 ppb)
PM10 (20 ug/m3)
PM25 (10 ug/m3)
Acid vapor (1.8 ppb)
PM10 (20 ug/m3)
Change in Health Indicator per
Increment in PM*
Odds Ratio (95% CI)
1.05 (0.94, 1.16)
1.13 (0.92, 1.38)
1.1 (0.7, 1.8)
1.2 (0.8, 1.8)
1.4(0.9,2.1)
1.19(1.04,1.35)
Range of City
PM Levels**
Means (ug/m3)

28.0-84.9
0.9-3. 2 ppb
13.0-70.7
6.7-31.5
1.0-5.0 ppb
13-23
Increased Wheeze in Children
10 Canadian
Communities5
Increased Airway Obstruction
Southern CaliforniaL
PM10 (20 ug/m3)
in Adults
PM10 (20ug/m3)
1.35(1.10,1.64)

1.09 (0.92, 1.30)
13-23

NR
Decreased Lung Function in Children
Six CityF
Six City0
24 CityM
24 CityM
24 CityM
24 CityM
12 Southern California
communitiesN
(all children)
12 Southern California
communitiesN
(all children)
12 Southern California
communities0
(4th grade cohort)
12 Southern California
communities0
(4th grade cohort)
PM15/10 (50 ug/m3)
TSP (100 ug/m3)
H+ (52 nmoles/m3)
PM,, (15 ug/m3)
SO; (7 ug/m3)
PM10 (17 ug/m3)
PM10 (25 ug/m3)
Acid vapor (1.7 ppb)
PM10 (25 ug/m3)
Acid vapor (1.7 ppb)
PM10 (51.5 ug/m3)
PM25 (25.9 ug/m3)
PM10.25 (25.6 ug/m3)
Acid vapor (4.3 ppb)
PM10 (51.5 ug/m3)
PM25 (25.9 ug/m3)
PM10.25 (25.6 ug/m3)
Acid vapor (4.3 ppb)
NS Changes
NS Changes
- 3.45% (-4.87, -2.01) FVC
-3.21% (-4.98, -1.41) FVC
-3.06% (-4.50, -1.60) FVC
-2.42% (-4.30, -.0.51) FVC
-24.9 (-47.2, -2.6) FVC
-24.9 (-65.08, 15.28) FVC
- 32.0 (-58.9, -5.1) MMEF
-7.9 (-60.43, 44.63) MMEF
-0.58 (-1.14, -0.02) FVC growth
-0.47 (-0.94, 0.01) FVC growth
-0.57 (-1.20, 0.06) FVC growth
-0.57 (-1.06, -0.07) FVC growth
- 1.32 (-2.43, -0.20) MMEF growth
- 1.03 (-1.95, -0.09) MMEF growth
- 1.37 (-2.57, -0.15) MMEF growth
- 1.03 (-2.09, 0.05) MMEF growth
20-59
39-114
6.2-41.0
18.1-67.3
9.1-17.3
22.0-28.6
28.0-84.9
0.9-3. 2 ppb
28.0-84.9
0.9-3. 2 ppb
NR
NR

-------
           Type of Health
         Effect and Location
      Indicator
                                   Range of City
Change in Health Indicator per       PM Levels * *
      Increment in PM*             Means (ug/m3)
 Lung Function Changes in Adults
      Southern California1"
      (% predicted FEVb
      females)
PM10 (cutoff of
54.2 days/year
>100 ug/m3)
      Southern California1"        PM10 (cutoff of
      (% predicted FEVb males)   54.2 days/year
                                 >100 ug/m3)
      Southern California1"        PM10 (cutoff of
      (% predicted FEVb males   54.2 days/year
      whose parents had asthma,   >100 ug/m3)
      bronchitis, emphysema)
      Southern California1"        SOJ (1.6 ug/m3)
      (% predicted FEVi,
      females)
      Southern California1"        SOJ (1.6 ug/m3)
      (% predicted FEV,, males)	
   +0.9 % (-0.8, 2.5)
                             +0.3 % (-2.2, 2.8) FEVj
  -7.2 % (-11.5, -2.7)
                                  Not reported
                             -1.5% (-2.9, -0.1) FEVJ
52.7 (21.3, 80.6)
                                  54.1(20.0,80.6)
                                                            54.1 (20.0, 80.6)
                                   7.4(2.7, 10.1)
                                   7.3 (2.0, 10.1)
      *Results calculated using PM increment between the high and low levels in cities, or other PM increments
      given in parentheses; NS Changes = No significant changes.
      **Range of mean PM levels given unless, as indicated, studies reported overall study mean (min, max), or
      mean (±SD); NR=not reported.
      *** Results only for smoking category subgroups.

a Schwartz, J.; Dockery, D. W.; Neas, L. M. (1996) Is daily mortality associated specifically with fine particles?  J. Air
    Waste Manage. Assoc.  46: 927-939.
b Ostro, B. D.; Broadwin,  R.; Lipsett, M. J. (2000) Coarse and fine particles and daily mortality in the Coachella  Valley,
    California: a follow-up study. J. Exposure Anal. Environ. Epidemiol. 10: 412-419.
0 Lippmann, M.; Ito, K.; Nadas, A.; Burnett, R. T. (2000) Association of particulate matter components with daily
    mortality and morbidity in urban populations. Cambridge, MA: Health Effects Institute; research report no. 95.
d Lipfert, F. W.; Morris, S. C.; Wyzga, R. E. (2000) Daily mortality in the Philadelphia metropolitan area and
    size-classified particulate matter. J. Air Waste Manage. Assoc.: 1501-1513.
e Mar, T. F.; Norris, G. A.; Koenig, J. Q.; Larson, T. V. (2000) Associations between air pollution and mortality in
    Phoenix, 1995-1997.  Environ. Health Perspect. 108: 347-353.
f Smith, R. L.; Spitzner, D.; Kim,  Y.; Fuentes, M. (2000) Threshold dependence of mortality effects for fine and coarse
    particles in Phoenix, Arizona. J. Air Waste Manage. Assoc. 50: 1367-1379.
8 Fairley, D. (1999) Daily  mortality and air pollution in Santa Clara County, California: 1989-1996. Environ. Health
    Perspect. 107:637-641.
h Burnett, R. T.; Brook, I; Dann,  T.; Delocla,  C.; Philips, O.; Cakmak, S.; Vincent, R.; Goldberg, M. S.; Krewski, D.
    (2000) Association between particulate- and gas-phase components of urban air pollution and daily mortality in eight
    Canadian cities. In: Grant, L. D., ed. PM2000: particulate matter and health. Inhalation Toxicol. 12(suppl. 4): 15-39.
1 Burnett, R. T.; Cakmak,  S.; Brook, J. R.; Krewski, D. (1997) The role of particulate size and chemistry in the association
    between summertime ambient air pollution and hospitalization for cardiorespiratory diseases. Environ. Health
    Perspect. 105: 614-620.
' Burnett, R. T.; Smith-Doiron, M.; Stieb, D.; Cakmak, S.; Brook,  J. R. (1999) Effects of particulate and gaseous air
    pollution on cardiorespiratory hospitalizations. Arch. Environ. Health 54: 130-139.
k Tolbert, P. E.; Klein, M.; Metzger, K. B.; Peel, J.; Flanders, W. D.; Todd, K.; Mulholland, J. A.; Ryan, P. B.; Frumkin,
    H. (2000) Interim results of the study of particulates and health in Atlanta (SOPHIA). J. Exposure Anal. Environ.
    Epidemiol. 10: 446-460.
1 Sheppard, L.; Levy, D.; Norris, G.; Larson, T. V.; Koenig, J. Q. (1999) Effects of ambient air pollution on nonelderly
    asthma hospital admissions in Seattle, Washington, 1987-1994. Epidemiology 10: 23-30.
m Schwartz, J.; Neas, L. M. (2000) Fine particles are more strongly associated than coarse particles with acute respiratory
    health effects in schoolchildren. Epidemiology. 11: 6-10.

-------
                                                Air Quality, Health, and Welfare Effects
"Naeher, L. P.; Holford, T. R.; Beckett, W. S.; Belanger, K.; Triche, E. W.; Bracken, M. B.; Leaderer, B. P. (1999)
    Healthy women's PEF variations with ambient summer concentrations of PM10, PN25, SO42_, H+, and O3. Am. J.
    Respir. Grit. Care Med. 160: 117-125.
0 Zhang, H.; Triche, E.; Leaderer, B. (2000) Model for the analysis of binary time series of respiratory symptoms. Am. J.
    Epidemiol.  151:  1206-1215.
p Neas, L. M.; Schwartz, J.; Dockery, D. (1999) A case-crossover analysis of air pollution and mortality in Philadelphia.
    Environ. Health Perspect. 107: 629-631.
q Moolgavkar, S. H. (2000) Air pollution and hospital admissions for chronic obstructive pulmonary disease in three
    metropolitan areas in the United States. In: Grant, L. D., ed. PM2000: particulate matter and health.  Inhalation
    Toxicol. 12(suppl. 4): 75-90.
RLipfert et al. 2000b
sHoweletal. 2001
    2.1.1.6 Roadway-Related Exposure and Health Studies

    A recent body of studies has suggested a link between residential proximity to heavily-
trafficked roadways (where diesel engines are operated) and adverse health effects.  While many
of these studies did not measure PM specifically, they include potential exhaust exposures which
include mobile source PM because they employ exposure indices such as roadway proximity or
traffic volumes.

    Based on extensive emission characterization studies and as reviewed in the EPA Diesel
HAD (Health Assessment Document for Diesel Exhaust), diesel PM is found principally in the
fine fraction (both primary and secondarily formed PM).71'72 In addition, in the Diesel HAD, we
note that the particulate characteristics in the zone around nonroad diesel engines is likely to be
substantially the same as published air quality measurements made along busy roadways.  This
conclusion supports the relevance of health effects associated with on-road diesel engine-
generated PM to nonroad applications. Thus, near roadway studies are relevant to understanding
potential health impacts of emissions from nonroad diesel engines.

    Specifically, in a recent body of studies, scientists have examined  health effects associated
with living near major roads. As discussed above,  a Dutch cohort study recently developed
estimates of the relative risk of cardiopulmonary and all-cause mortality associated with living
near a busy roadway.73 The study found a statistically significant excess risk of cardiopulmonary
mortality of 95 percent (i.e., a relative risk of 1.95, 95% CI: 1.09-3.52) associated with living
near a busy road. A recent British ecological study examined mortality attributable to stroke in
England and Wales.74  After adjusting for potential confounders, the study found a significantly
greater rate of mortality in men and women living within 200 meters of a busy road of 7 percent
[95% CI on RR: 1.04 to 1.09] and 4 percent [95% CI on RR: 1.02-1.06], respectively. Risks
decreased with increased distance from roadways.  However, being an ecological study design, it
is impossible to rule out confounding variables.

    Other studies relate the incidence or prevalence of respiratory health outcomes to roadway
proximity.  Several studies have found positive associations between respiratory symptoms and
residential roadway proximity or traffic volume.  Most recently, a study  in U.S. veterans living
                                            2-13

-------
Final Regulatory Impact Analysis
in southeastern Massachusetts found significant increases in self-reported respiratory symptoms
among subjects living within 50 meters of a major road.75

   A Dutch cohort study following infants from birth found that traffic-related pollutant
concentrations found positive associations with respiratory symptoms, several illnesses, and
physician-diagnosed asthma, the last of which was significant for diagnoses prior to 1 year of
age.76

   In a case-control study of children under 14 years old in San Diego, CA, with asthma
diagnosis was confirmed by Medicaid claims, no associations between odds of physician
diagnosis of asthma and traffic was found.77 However, a case-based analysis of the data
associated traffic flows with an increased number of medical  visits among children with asthma.

   A case-control study of children aged 4 to 48 months diagnosed with wheezing bronchitis
included exposures predicted from traffic data, dispersion models of NO2 as a marker of mobile
source emissions, and included separate exposures for home and day care.78 Analyses found that
cases had significantly elevated NO2 exposures compared with controls, but only among girls.  A
significant trend with NO2 was reported.

   Two cross-sectional studies of self-reported wheezing and allergic rhinitis symptoms in
German aged 12 to 15 years found increased prevalence of wheezing and allergic rhinitis based
on subject-reported frequency of truck traffic.79,80

   A cross-sectional study in the Netherlands examined self-reported respiratory diagnoses,
allergies, and respiratory symptoms in association with annual truck and automobile density,
living within 100 meters of a freeway, and indoor measures of air pollution (black smoke,
NO2).81 The study found associations for truck traffic density with wheeze and asthma attacks
in girls but not boys.  Associations among girls but not boys were also found for homes within
100 m of a freeway and chronic  cough, wheeze, and rhinitis.  Physician-diagnosed asthma was
not associated with traffic-related exposures.  Physician-diagnosed allergy was inversely
associated with NO2 and black smoke.

   A cross-sectional study in Surrey, England, compared city wards  transected by freeways and
those not transected by  freeways.82 Respiratory symptoms in the past year and self-reported
diagnosis of asthma by  a physician was not associated with any respiratory metric.

   A recent review of epidemiologic studies examining associations between asthma and
roadway proximity concluded that some coherence was evident in the literature, indicating that
asthma, lung function decrement, respiratory symptoms,  and  atopic illness appear to be higher
among people living near busy roads.83 Other studies have shown children living near roads with
high truck traffic density have decreased lung function and greater prevalence of lower
respiratory symptoms compared with children living on other roads.84

   Another recently published study from Los Angeles,  CA,  found that maternal residence near
heavy traffic during pregnancy is associated with adverse birth outcomes, such as preterm birth

                                          2-14

-------
                                              Air Quality, Health, and Welfare Effects
and low birth weight.85 However, these studies are not specifically related to PM, but to fresh
emissions from mobile sources, which includes other components as well.

   Other studies have shown that living near major roads results in substantially higher
exposures to ultrafine particles. A British study found that in the lungs of children living near
major roads in Leicester, UK, a significantly higher proportion of the alveolar macrophages
contained PM compared with children living on quiet streets.86 All particles observed in the
lungs of children were carbon particles under 0.1 um, which are known to be emitted from diesel
engines and other mobile sources. This study is consistent with recent studies of ultrafine
particle concentrations around major roads in Los Angeles, CA and Minnesota, which found that
concentrations of the smallest particles were substantially elevated near roadways with  diesel
traffic.87'88'89

   The particulate characteristics in the zone around nonroad diesel engines is not likely to
differ substantially from published air quality measurements made along  busy roadways; thus,
these studies are relevant to the diesel exhaust emissions from nonroad diesel engines. While
these studies do not specifically examine nonroad diesel engines, several  observations may be
drawn.  First, nonroad diesel engine emissions are similar in their emission characteristics to on-
road motor vehicles. Secondly, exposures from nonroad engines may actually  negatively bias
these studies, because exposures from nonroad sources are not accounted for, and therefore
reduce the study's statistical power. Third, certain populations that are exposed directly to fresh
nonroad diesel exhaust are exposed at greater concentrations than those found in studies among
the general population. These groups include workers in the construction, timber, mining,  and
agriculture industries, and members of the general population that spend  a large amount of time
near areas where diesel engine emissions are most densely clustered, such as residents in
buildings near large construction  sites.

2.1.2 Attainment and Maintenance of the PM10 and PM25 NAAQS: Current and Future
Air Quality

   2.1.2.1 Current PM Air Quality

   There are NAAQS for both PM10 and PM25.  Violations of the annual PM2 5 standard are
much more widespread than are violations of the PM10 standards. Emission reductions  needed to
attain the PM2 5 standards will also assist in attaining and maintaining compliance with the PM10
standards. Thus, since most PM emitted by nonroad diesel engines is in the fine fraction of PM,
the emission controls resulting from this final rule will contribute to attainment and maintenance
of the existing PM NAAQS. More broadly, the new standards will benefit public health and
welfare through reductions in direct diesel PM and reductions of NOx, SOx, and HCs that
contribute to secondary formation of PM.  As described above, diesel particles from nonroad
diesel engines are a component of both coarse and fine PM, but fall mainly in the fine (and even
ultrafine) size range.
                                          2-15

-------
Final Regulatory Impact Analysis
    The emission reductions from this final rule will assist States as they work with EPA through
implementation of local controls including the development and adoption of additional controls
as needed to help their areas attain and maintain the standards.

    2.1.2.1.1PM10Levels

    The current NAAQS for PM10 were established in 1987. The primary (health-based) and
secondary (public welfare based) standards for PM10 include both short- and long-term NAAQS.
The short-term (24-hour) standard of 150 |ig/m3 is not to be exceeded more than once per year
on average over three years.  The long-term standard specifies an expected annual arithmetic
mean not to exceed  50 |ig/m3 averaged over three years.

    Currently, 29.3 million people live in PM10 nonattainment areas, including moderate and
serious areas.  There are presently 56 moderate PM10 nonattainment areas with a total population
of 6.6 million.90  The attainment date for the initial moderate PM10 nonattainment areas,
designated by law on November 15, 1990, was December 31, 1994.  Several additional PM10
nonattainment areas were designated on January 21, 1994, and the attainment date for these areas
was December 31, 2000.

    There are 8 serious PM10 nonattainment areas with a total affected population of 22.7 million.
According to the Act, serious PM10 nonattainment areas must attain the standards no later than
10 years after designation.  The initial serious PM10 nonattainment areas were designated January
18, 1994 and had an attainment date set by the Act of December 31, 2001. The Act provides that
EPA may grant extensions of the serious area attainment dates of up to 5 years, provided that the
area requesting the extension meets the requirements of Section 188(e) of the Act.  Five serious
PM10 nonattainment areas (Phoenix, Arizona; Clark County (Las Vegas), NV; Coachella Valley,
South Coast (Los Angeles), and Owens Valley, California) have received extensions of the
December 31, 2001  attainment date and thus have new attainment dates of December 31, 2006.

    Many PM10 nonattainment areas continue to experience exceedances. Of the 29.3 million
people living in designated PM10 nonattainment areas, approximately 24.5 million people are
living in nonattainment  areas with measured air quality violating the PM10 NAAQS in 2000-
2002. Among these are 8 serious areas listed in Table 1.2-1 and 6 moderate areas: Nogales, AZ,
Imperial Valley,  CA, Mono Basin, CA, Coso Junction, CA,B Ft. Hall, ID, and El Paso, TX.
   BOn August 6, 2002, EPA finalized certain actions affecting the Searles Valley, California, PM10 nonattainment
area, which is located in the rural high desert and includes portions of Inyo, Kern, and San Bernardino Counties. The
action splits the Searles Valley nonattainment area into three separate areas: Coso Junction, Indian Wells Valley and
Trona. EPA's action also determines that the Trona area attained the PM-10 standards by December 31, 1994. On
May 7, 2003, EPA finalized approval of the Indian Wells Moderate Area and Maintenance Plan and redesignated the
area from nonattainment to attainment for paniculate matter (PM-10).

Source:  http://www.epa.gov/region9/air/searlespm/index.html
                                           2-16

-------
                                              Air Quality, Health, and Welfare Effects
Serious
                                       Table 1.2-1
                                        Nonattainment Areas
Area
Owens Valley, CA
Phoenix, AZ
Clark County, NV (Las Vegas)
Coachella Valley, CA
Los Angeles South Coast Air Basin, CA
San Joaquin Valley, CA
Walla Walla, WA
Washoe County, NV (Reno)
Total Population
Attainment
Date
December 3 1,2006
December 3 1,2006
December 3 1,2006
December 3 1,2006
December 3 1,2006
2001
2001
2001
2000
Population
7,000
3,111,876
1,375,765
225,000
14,550,521
3,080,064
10,000
339,486
2000-2002 Measured
Violation
Yes
Yes
Yes
Yes
Yes
Yes
No
No
22.7 million
   In addition to these designated nonattainment areas, there are 16 unclassified areas, where
6.2 million live, for which States have reported PM10 monitoring data for 2000-2002 period
indicating a PM10 NAAQS violation. An official designation of PM10 nonattainment indicates
the existence of a confirmed PM10 problem that is more than a result of a one-time monitoring
upset or a result of PM10 exceedances attributable to natural events.  We  have not yet excluded
the possibility that  one or the other of these is responsible for the monitored violations in 2000-
2002 in these 16 unclassified areas. We adopted a policy in 1996 that allows areas whose PM10
exceedances are attributable to natural events to remain unclassified if the State is taking all
reasonable measures to safeguard public health regardless of the sources of PM10 emissions.
Areas that remain unclassified areas are not required to submit attainment plans, but we work
with each of these areas to understand the nature of the PM10 problem and to determine what best
can be done to reduce it.

   2.1.2.1.2 PM25  Levels

   The need for reductions in the levels of PM2 5 is widespread. Figure 2.1.1-4 below shows
PM25 monitoring data highlighting locations measuring concentrations above the level of the
NAAQS.  As can be seen from that figure, high ambient levels are widespread throughout the
country.  In addition, there may be counties without monitors that exceed the level of the
standard.  A listing of available measurements by county can be found in the air quality technical
support document (AQ TSD) for the rule.

   The NAAQS for PM25 were established in 1997 (62 FR 38651, July  18, 1997).  The short
term (24-hour) standard is set at a level of 65 |ig/m3 based on the 98th percentile concentration
averaged over three years. (The air quality statistic compared with the standard is referred to as
                                          2-17

-------
Final Regulatory Impact Analysis
the "design value.") The long-term standard specifies an expected annual arithmetic mean not to
exceed 15 |ig/m3 averaged over three years.

   Current PM2 5 monitored values for 2000-2002 indicate that 120 counties in which almost 65
million people live have annual design values that violate the PM2 5 NAAQS. In total, this
represents 23 percent of the counties and 37 percent of the population with levels above the
NAAQS in the areas with monitors that met completeness criteria. An additional 32 million
people live in 91 counties that have air quality measurements within 10 percent of the level of
the standard.  These areas, though not currently violating the standard, will also benefit from the
additional reductions from this rule in order to ensure long-term maintenance.  There are another
204 counties where 21 million people live that had incomplete data.

   Figure 2.1.2-1 is a map of currently available PM25 monitoring data, highlighting monitor
locations near or above the annual PM25 NAAQS.  As can be seen from this  figure, high ambient
levels are widespread throughout the East  and California.
                                          2-18

-------
                                 Air Quality, Health, and Welfare Effects
                      Figure 2.1.2-1
  PMzs County Design  Values, 2000-2002
                     Data from AQS 7/9/03
Counties with at least 1 complete site w/ DV > 15.0 (violate the NAAQS) [120]
Counties with at least 1 complete site w/ DV > 13.5 and < 15.0 (within 10% of the NAAQS) [91 ]
Counties with at least 1 complete site w/ DV < 13.5 [313]
Counties without a complete site [204]
                             2-19

-------
Final Regulatory Impact Analysis
   Further insights into the need for reductions from this rule can be gained by evaluating
counties at various levels above the level of the NAAQS. As shown in Table 2.1.1-3 of the 64.9
million people currently living in counties with measurements above the NAAQS, 18.8 million
live in counties above 20 |ig/m3.  In Section 2.1.2.2, we discuss that absent additional controls,
our modeling predicts there will continue to be large numbers of people living in counties with
PM levels above the standard.

                                       Table 2.1.1-3
       2000-2002 Monitored Populationa Living in Counties with Annual Average13 PM2 5
                                  Concentrations Shown
Measured 2000-2002
Annual Average PM2 5
Concentration
(|ig/m3N)
>25
>20 <=25
>15 <=20
<=15
Number of Counties
Within The
Concentration
Range
2
6
112
404
2000 Population Living in
Monitored Counties
Within The Concentration
Range (Millions, 2000
Census Data)
3.3
15.5
46.1
110.9
a Monitored population estimates represent populations living in counties with monitors producing data that meet the
   NAAQS data completeness requirements for 2000 - 2002. This analysis excludes the 204 counties whose
   monitoring data do not meet the completeness criteria.
b Annual average represents the monitor reading with the highest average in each monitored county.
c The monitored population is 175.7 million (or 62 percent of the U.S. Census total county-based 2000 population for the
   U.S. of 281.4 million).
    Chemical composition of ambient PM25 also underscores the contribution of emissions from
the engines subject to this rule and points to the need for reductions. Data on PM25 composition
are available from the EPA Speciation Trends Network and the IMPROVE Network for
September 2001 to August 2002 covering both urban and rural areas in numerous regions of the
United States. The relative contribution of various chemical components to PM25 varies by
region of the country. Figure 2.1.2-2 shows the levels and composition of ambient PM2 5 in some
urban areas. Figure 2.1.2-3  shows the levels and composition of PM25 in rural areas where the
total PM2 5 levels are generally lower. These data show that carbonaceous PM2 5 makes up the
major component for PM25 in both urban and rural areas in the Western United States.
Carbonaceous PM2 5  includes both elemental and organic carbon. Nonroad engines, especially
nonroad diesel engines, contribute significantly to ambient PM2 5 levels, largely through
emissions of carbonaceous PM25. For the Eastern and middle United States, these data show that
carbonaceous PM2 5 is a major contributor to ambient PM2 5 both urban and rural areas. In some
eastern areas, carbonaceous PM2 5 is responsible for up to half of ambient PM2 5 concentrations.
                                           2-20

-------
                           Figure 2.1.2-2
Annual Average PM2 5 Species and Concentrations in Selected Urban Areas
                   (September 2001- August 2002)

-------
                                   Figure 2.1.2-3
          Annual Average PM2.5 Concentration and Species in Rural Areas
                            (September 2001 - August 2002)
    Sulfate
    Ammonium
    Nitrate
    TCM
    Crustal
         /  \
         \    /
1.71 7.91  14.11

-------
                                              Air Quality, Health, and Welfare Effects
   Another important component of PM in the West is nitrates, which are formed from NOx.
Nitrates are especially prominent in the California area where it is responsible for about a quarter
of the ambient PM25 concentrations. Nonroad diesel engines also emit high levels of NOx,
which reacts in the atmosphere to form secondary PM2 5 (namely ammonium nitrate).  Sulfate
plays a lesser role in these western regions by mass, but it remains important to visibility
impairment discussed below. Nonroad diesel engines also emit SO2 and HC, which react in the
atmosphere to form secondary PM2 5 (namely sulfates and organic carbonaceous PM2 5). Sulfate
is also a major contributor to ambient PM25 in the Eastern United States and in some areas make
greater contributions than carbonaceous PM2 5.

   From Figures 2.1.2-2 and 2.1.2-3, one can compare the levels and composition of PM25 in
various urban areas and a corresponding rural area.  This comparison, in Figure 2.1.2-4, shows
that much of the excess PM2 5 in urban areas (annual average concentration at urban monitor
minus annual average concentration at corresponding rural monitor) is indeed from
carbonaceous PM.91'92 See the AQ TSD for details.

   The ambient PM monitoring networks account for both directly emitted PM as well as
secondarily formed PM. Emission inventories, which account for directly emitted PM and PM
precursors separately, also show that mobile source PM emissions, including that from nonroad
diesel engines, is a major contributor to total PM emissions. Nationally, this final rule will
significantly reduce emissions of carbonaceous PM. NOx emissions, a prerequisite for
formation of secondary nitrate aerosols, will also be reduced. Nonroad diesel engines are major
contributors to both of these pollutants.  The new requirements in this rule will also reduce SOx
and HC.  Nonroad diesel engines emissions also contribute to national SOx and HC emission
inventories, but to a lesser degree than for PM and NOx.  The emission inventories are
discussed in detail in Chapter 3.

   As discussed in Sections 2.2.2.6 and 2.1, diesel PM also contains small quantities of
numerous mutagenic and carcinogenic compounds associated with the particles (and also organic
gases). In addition, while toxic trace metals emitted by nonroad diesel engines represent a very
small portion of the national emissions of metals (less than one percent) and a small portion of
diesel PM (generally less than one percent of diesel PM), we note that several trace  metals of
potential toxicological significance and persistence in the environment are emitted by diesel
engines.  These trace metals include chromium, manganese, mercury and nickel.  In addition,
small amounts of dioxins have been measured in highway engine diesel exhaust, some of which
may partition into the particulate phase;  dioxins are a major health concern but diesel engines are
a minor contributor to overall dioxin emissions.   Diesel engines also emit polycyclic organic
matter (POM), including polycyclic aromatic hydrocarbons (PAH), which can be present in both
gas and particle phases of diesel exhaust. Many PAH compounds are classified by EPA as
probable human carcinogens.
                                          2-23

-------
                                               Figure 2.1.2-4
                 Composition of Urban Excess PM2.5 at Selected Sites (September 2001 - August 2002)
                             (Source: U.S. EPA (2004) AQ TSD; Rao and Frank (2003))
              Salt Lakd City
F res TO

-------
                                                Air Quality, Health, and Welfare Effects
       2.1.2.2 Risk of Future Violations

    2.1.2.2.1 PMAir Quality Modeling and Methods

    In conjunction with this rulemaking, we performed a series of PM air quality modeling
simulations for the continental U.S. The model simulations were performed for five emission
scenarios: a 1996 baseline projection, a 2020 baseline projection and a 2020 projection with
nonroad controls, a 2030 baseline projection and a 2030 projection with nonroad controls.
Further discussion of this modeling, including evaluations of model performance relative to
predicted future air quality, is provided in the AQ Modeling TSD.

    The model outputs from the 1996, 2020 and 2030 baselines, combined with current air
quality data, were used to identify areas expected to exceed the PM2 5 NAAQS in 2020 and 2030.
These areas became candidates for being determined to be residual exceedance areas that will
require additional emission reductions to attain and maintain the PM25 NAAQS.  The impacts of
the nonroad controls were determined by comparing the model results in the future year control
runs against the baseline simulations of the same year.  We note that there are significant SO2
benefits from sulfur reductions in home heating oil fuel that  are not accounted for in our
modeling. This modeling supports the conclusion that there is a broad set of areas with predicted
PM25 concentrations at or above 15 |ig/m3 between 1996 and 2030 in the baseline scenarios
without additional emission reductions.

    The air quality modeling performed for this rule was based upon an improved version of the
modeling system  used in the HD Engine/Diesel Fuel rule (to address peer-review comments)
with the addition  of updated inventory estimates for 1996, 2020 and 2030.

    A national-scale version of the REgional Model System for Aerosols and Deposition
(REMSAD) was utilized to estimate base and future-year PM concentrations over the contiguous
United States for  the various emission scenarios.  Version 7 of REMSAD was used for this
rulemaking. REMSAD was designed to calculate the concentrations of both inert and
chemically reactive pollutants in the atmosphere that affect annual  particulate concentrations and
deposition over large spatial scales.0 Because it accounts for spatial and temporal variations as
well as differences in the reactivity of emissions, REMSAD  is useful for evaluating the impacts
of the final rule on PM concentrations in the United States. The following sections provide an
overview of the PM modeling completed as part of this rulemaking.  More detailed information
is included in the  AQ Modeling TSD, which is located in the docket for this rule.
    c  Given the potential impact of the final rule on secondarily formed particles it is important to employ a
Eulerian model such as REMSAD. The impact of secondarily formed pollutants typically involves primary
precursor emissions from a multitude of widely dispersed sources, and chemical and physical processes of pollutants
that are best addressed using an air quality model that employs an Eulerian grid model design. Thus, comments from
industry that EPA's methodology form computing benefits over time is based on unsupportable assumptions such as
that there will be no interactions between precursors and directly emitted PM in the formation of secondary PM and
that EPA excludes consideration of non-linearities in its air quality modeling are incorrect. This air quality modeling
for 2020 and 2030 does incorporate the nonlinear interactions between NOx, SO2, and direct PM.

                                           2-25

-------
Final Regulatory Impact Analysis
   The PM air quality analyses employed the modeling domain used previously in support of
Clear Skies air quality assessment. The domain encompasses the lower 48 States and extends
from 126 degrees to 66 degrees west longitude and from 24 degrees to 52 degrees north latitude.
The model contains horizontal grid-cells across the model domain of roughly 36 km by 36 km.
There are 12 vertical layers of atmospheric conditions with the top of the modeling domain at
16,200 meters.

   The simulation periods modeled by REMSAD included separate full-year application for
each of the five emission scenarios (1996 base year, 2020 base, 2020 control, 2030 baseline,
2030 control) using the 1996 meteorological inputs described below.

   The meteorological data required for input into REMSAD (wind, temperature, surface
pressure, etc.) were obtained from a previously developed 1996 annual run of the Fifth-
Generation National Center for Atmospheric Research (NCAR) / Penn State Mesoscale Model
(MM5). A postprocessor called MM5- REMSAD was developed to convert the MM5 data into
the appropriate REMSAD grid coordinate systems and file formats.  This postprocessor was used
to develop the hourly average meteorological input files from the MM5 output.  Documentation
of the MM5REMSAD code  and further details on the development of the input files is contained
in Mansell (2000).93 A more detailed description of the development of the meteorological input
data is provided in the AQ Modeling TSD, which is located in the docket for this rule.

   The modeling specified initial species concentrations and lateral boundary conditions to
approximate background concentrations of the species; for the lateral boundaries the
concentrations varied (decreased parabolically) with height.  These initial conditions reflect
relatively clean background  concentration values. Terrain elevations and land use information
was obtained from the U.S. Geological Survey database at 10 km resolution  and aggregated to
the roughly 36 km horizontal resolution used for this REMSAD application.  The development
of model inputs is discussed in greater detail in the AQ Modeling TSD, which is available in the
docket for this rule.

   2.1.2.2.2 Model Performance Evaluation

   The purpose of the base year PM air quality modeling was to reproduce the atmospheric
processes resulting in formation and dispersion of fine particulate matter across the United
States. An operational model performance evaluation for PM2 5 and its related speciated
components (e.g., sulfate, nitrate,  elemental carbon etc.) for 1996 was performed in order to
estimate the ability of the modeling system to replicate base year concentrations.

   This evaluation is comprised principally of statistical assessments of model versus observed
pairs. The robustness of any evaluation is directly proportional to the amount and quality of the
ambient data available for comparison. Unfortunately, for 1996 there were few PM25 monitoring
networks with available data for evaluation of the Nonroad PM modeling. Critical limitations of
the existing databases are a lack of urban monitoring sites with speciated measurements and poor
geographic representation of ambient concentration in the Eastern United States.
                                         2-26

-------
                                              Air Quality, Health, and Welfare Effects
    The largest available ambient database for 1996 comes from the IMPROVE network.
IMPROVE is a cooperative visibility monitoring effort between EPA, federal land management
agencies, and state air agencies. Data are collected at Class I areas across the United States
mostly at national parks, national wilderness areas, and other protected pristine areas.94 There
were approximately 60 IMPROVE sites that had complete annual PM2 5 mass and/or PM2 5
species data for 1996. Using the 100th meridian to divide the Eastern and Western United States,
42 sites were located in the West and  18 sites were in the East.

    The observed IMPROVE data used for the performance evaluation consisted of PM2 5 total
mass, sulfate ion, nitrate ion, elemental carbon, organic aerosols, and crustal material (soils).
The REMSAD model output species were postprocessed in order to achieve compatibility with
the observation species.

    The principal evaluation statistic used to evaluate REMSAD performance is the "ratio of the
means." It is defined as the ratio of the average predicted values over the average observed
values.  The annual average ratio of the means was calculated for five individual PM25 species as
well as for total PM2 5 mass. The metrics were calculated for all IMPROVE sites across the
country as well as for the East and West individually.  Table 2.1.2-1 shows the ratio of the
annual means. Numbers greater than  1 indicate overpredictions compared with ambient
observations (e.g. 1.23 is a 23 percent overprediction). Numbers less than 1 indicate
underpredictions.

                                      Table 2.1.2-1
    Model Performance Statistics for REMSAD PM2S Species Predictions: 1996 Base Case
IMPROVE PM Species
PM2 5, total mass
Sulfate ion
Nitrate ion
Elemental carbon
Organic aerosols
Soil/Other
Ratio of the Means (annual average concentrations)
Nationwide
0.68
0.81
1.05
1.01
0.55
1.38
Eastern U.S.
0.85
0.9
1.82
1.23
0.58
2.25
Western U.S.
0.51
0.61
0.45
0.8
0.53
0.88
Note: The dividing line between the West and East was defined as the 100th meridian.
   When considering annual average statistics (e.g., predicted versus observed), which are
computed and aggregated over all sites and all days, REMSAD underpredicts fine particulate
mass (PM25) by roughly 30 percent.  PM25 in the Eastern United States is slightly
underpredicted, while PM2 5 in the West is underpredicted by about 50 percent.  Eastern sulfate is
slightly underpredicted, elemental carbon is slightly overpredicted, while nitrate and crustal are
                                          2-27

-------
Final Regulatory Impact Analysis
largely overpredicted. This is balanced by an underprediction in organic aerosols. Overall the
PM2 5 performance in the East is relatively unbiased due to the dominance of sulfate in the
observations. Western predictions of sulfate, nitrate, elemental carbon, and organic aerosols are
all underpredicted.

   REMSAD performance is relatively good in the East. The model is overpredicting nitrate,
but less so than in previous model applications. The overpredictions in soil/other concentrations
in the East can largely be attributed to overestimates of fugitive dust emissions.  The model is
performing well for sulfate, which is the dominant PM2 5 species in most of the East.  Organic
aerosols are underpredicted in both the East and West.  There is a large uncertainty in the current
primary organic inventory  as well as the modeled production of secondary organic aerosols.

   REMSAD is underpredicting all species in the West. The dominant species in the West is
organic aerosols.  Secondary formation of sulfate, nitrate, and organics appears to be
underestimated in the West.  Additionally, the current modeling inventory does not contain
wildfires, which may be a  significant source of primary  organic carbon in the West.

   It should be noted that  PM2 5 modeling is an evolving science. There have been few regional
or national scale model applications for primary and secondary PM. Unlike ozone modeling,
there is essentially no database of past performance statistics against which to measure the
performance of this modeling.  Given the  state of the science relative to PM modeling, it is
inappropriate to judge PM model performance using criteria derived for other pollutants, like
ozone.  Still, the performance of this air quality modeling is encouraging, especially considering
that the results are limited  by our current knowledge of PM science and chemistry, and by the
emission inventories for primary PM and  secondary PM precursor pollutants. EPA and others
are only beginning to understand the limitations and uncertainties in the current inventories and
modeling tools. Improvements to the tools are being made on a continuing basis.

   2.1.2.2.3 Results with Areas at Risk of Future PM2 5  Violations

   Our air quality modeling performed for this rulemaking also indicates that the present
widespread number of counties with annual averages above 15 |ig/m3 are likely to persist in the
future in the absence of additional controls.  For example, in 2020 based on emission controls
currently adopted or expected to be in place, we project that 66 million people will live in 79
counties with average PM25 levels at and above 15 |ig/m3. In 2030, the number of people
projected to live in areas exceeding the PM25 standard is expected to increase to 85 million in
107 counties. An additional 24 million people are projected to live in counties with annual
averages within 10 percent of the standard in 2020, and  17 million people are projected to live in
counties with annual averages within 10 percent of the standard in 2030. The AQ Modeling
TSD lists the specifics.

   Our modeling also indicates that the reductions from this final rule will make a substantial
                                          2-28

-------
                                                Air Quality, Health, and Welfare Effects
contribution to reducing these potential exposures.0 In 2020, we estimate that the number of
people living in counties with PM2 5 levels above the NAAQS will be reduced from 66 million to
60 million living in 67 counties.  That is a reduction of 9 percent in potentially exposed
population and 15 percent of the number of counties. In 2030, there will be an estimated
reduction from 85 million people to 71 million living in 84 counties.  This represents an even
greater improvement than projected for 2020 because of the fleet turnover and corresponds to a
16 percent reduction in potentially exposed population  and a 21 percent of the number of
counties.  Furthermore, our modeling also shows that the emission reductions will assist areas
with future maintenance of the standards.

    Table 2.1.2-2 lists the counties with 2020 and 2030 projected annual PM25 design values that
violate the annual standard. Counties  are marked with  an "V"  in the table if their projected
design values are greater than or equal to 15.05 |ig/m3.  The current 3-year average design values
of these counties are also  listed.  Recall that we project future design values only for counties
that have  current design values, so this list is limited to those counties with 1999-2001 ambient
monitoring data sufficient to calculate current 3-year design values.
    DThe results illustrate the type of PM changes for the preliminary control option, as discussed in Section 3.6.
The analysis differs from the modeled control case based on public comment and updated information; however, we
believe that the net results would approximate future emissions, though we anticipate the PM reductions might be
smaller.  We also note that our modeling does not account for substantial reductions in SO2 associated with sulfur
reductions in home heating oil.

                                            2-29

-------
Final Regulatory Impact Analysis
                                   Table 2.1.2-2
                  Counties with 2020 and 2030 Projected Annual PM2.5
               Design Values in Violation of the Annual PM2.5 Standard.a'b
State
AL
AL
AL
AL
AL
AL
AL
AL
AL
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CT
DE
DC
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
County
DeKalb
Houston
Jefferson
Mobile
Montgomery
Morgan
Russell
Shelby
Talladega
Fresno
Imperial
Kern
Los Angeles
Merced
Orange
Riverside
San Bernardino
San Diego
San Joaquin
Stanislaus
Tulare
New Haven
New Castle
Washington
Bibb
Chatham
Clarke
Clayton
Cobb
DeKalb
Dougherty
Floyd
Fulton
Hall
Muscogee
Paulding
Richmond
1999-2001
Design Value
(ug/m3)b
16.8
16.3
21.6
15.3
16.8
19.1
18.4
17.2
17.8
24
15.7
23.7
25.9
18.9
22.4
29.8
25.8
17.1
16.4
19.7
24.7
16.8
16.6
16.6
17.6
16.5
18.6
19.2
18.6
19.6
16.6
18.5
21.2
17.2
18
16.8
17.4
2020
Base

V
V

V
V
V
V
V
V

V
V
V
V
V
V
V

V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
Control"


V

V
V
V
V
V
V

V
V
V
V
V
V
V

V
V
V
V
V
V
V
V
V
V
V
V
V
V

V
V
V
2030
Base
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
Control"
V
V
V
V
V
V
V
V
V
V

V
V
V
V
V
V
V

V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
Population
in 2000
64,452
88,787
662,047
399,843
223,510
111,064
49,756
143,293
80,321
799,407
142,361
661,645
9,519,338
210,554
2,846,289
1,545,387
1,709,434
2,813,833
563,598
446,997
368,021
824,008
500,265
572,059
153,887
232,048
101,489
236,517
607,751
665,865
96,065
90,565
816,006
139,277
186,291
81,678
199,775
                                       2-30

-------
State
GA
GA
IL
IL
IL
IL
IL
IN
IN
IN
IN
KY
KY
LA
LA
MD
MD
MD
MA
MI
MS
MO
MT
NJ
NJ
NY
NY
NC
NC
NC
NC
NC
NC
NC
NC
NC
OH
OH
OH
OH
OH
OH
OH
County
Washington
Wilkinson
Cook
Du Page
Madison
St Clair
Will
Clark
Lake
Marion
Vanderburgh
Jefferson
Kenton
East Baton Rouge
West Baton Rouge
Baltimore
Prince Georges
Baltimore City
Suffolk
Wayne
Jones
St Louis City
Lincoln
Hudson
Union
Bronx
New York
Catawba
Davidson
Durham
Forsyth
Gaston
Guilford
McDowell
Mecklenburg
Wake
Butler
Cuyahoga
Franklin
Hamilton
Jefferson
Lawrence
Lucas
1999-2001
Design Value
(ug/m3)b
16.5
18.1
18.8
15.4
17.3
17.4
15.9
17.3
16.3
17
16.9
17.1
15.9
14.6
14.1
16
17.3
17.8
16.1
18.9
16.6
16.3
16.4
17.5
16.3
16.4
17.8
17.1
17.3
15.3
16.2
15.3
16.3
16.2
16.8
15.3
17.4
20.3
18.1
19.3
18.9
17.4
16.7
2020
Base
V
V
V

V
V
V
V
V
V

V




V
V
V
V
V
V
V
V

V
V
V
V



V

V

V
V
V
V
V
V
V
Control3
V
V
V

V
V

V
V


V




V
V

V


V
V


V

V





V


V
V
V
V
V
V
2030
Base
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
Control"
V
V
V

V
V
V
V
V
V

V

V


V
V

V
V
V
V
V
V
V
V
V
V

V

V

V

V
V
V
V
V
V
V
Population
in 2000
21,176
10,220
5,376,741
904,161
258,941
256,082
502,266
96,472
484,564
860,454
171,922
693,604
151,464
412,852
21,601
754,292
801,515
651,154
689,807
2,061,162
64,958
348,189
18,837
608,975
522,541
1,332,650
1,537,195
141,685
147,246
223,314
306,067
190,365
421,048
42,151
695,454
627,846
332,807
1,393,978
1,068,978
845,303
73,894
62,319
455,054

-------
State
OH
OH
OH
OH
OH
OH
PA
PA
PA
PA
SC
SC
TN
TN
TN
TN
TN
TX
TX
UT
VA
WV
WV
WV
WV
WV
WI
County
Mahoning
Montgomery
Scioto
Stark
Summit
Trumbull
Allegheny
Delaware
Philadelphia
York
Greenville
Lexington
Davidson
Hamilton
Knox
Shelby
Sullivan
Dallas
Harris
Salt Lake
Richmond City
Brooke
Cabell
Hancock
Kanawha
Wood
Milwaukee
1999-2001
Design Value
(ug/m3)b
16.4
17.6
20
18.3
17.3
16.2
21
15
16.6
16.3
17
15.6
17
18.9
20.4
15.6
17
14.4
15.1
13.6
14.9
17.4
17.8
17.4
18.4
17.6
14.5
Number of Violating Counties b
Population of Violating Counties0
2020
Base

V
V
V
V

V

V

V


V
V



V


V
V
V
V
V

79
65,821,000
Control"

V
V
V
V

V

V

V


V
V



V


V
V
V
V


67
60,453,500
2030
Base
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
107
85,525,600
Control"

V
V
V
V

V

V

V

V
V
V



V


V
V
V
V
V

84
71,375,600
Population
in 2000
257,555
559,062
79,195
378,098
542,899
225,116
1,281,666
550,864
1,517,550
381,751
379,616
216,014
569,891
307,896
382,032
897,472
153,048
2,218,899
3,400,578
898,387
197,790
25,447
96,784
32,667
200,073
87,986
940.164


a As described in Chapter 3, the final control case differs from the modeled control case based on public comment and
    updated information; however, we believe that the net results would approximate future emissions, although we
    anticipate the design value improvements would be smaller. In our modeling, we do not account for SO2 reductions
    related to sulfur reductions in home heating oil.
b Projections are made only for counties with monitored design values for 1999-2001. These were the most current data
    at the time the analyses were performed. Counties with insufficient data or lacking monitors  are excluded.
c Populations are based on 2020 and 2030 estimates rounded to nearest hundred.  See the AQ Modeling TSD for details.

-------
                                               Air Quality, Health, and Welfare Effects
    Table 2.1.2-3 lists the counties with 2020 or 2030 projected annual PM25 design values that
do not violate the annual standard, but are within 10 percent of it. Counties are marked with an
"X" in the table if their projected design values are greater than or equal to 13.5 5 i-ig/m3, but less
than 15.05 |ig/m3.  Counties are marked with an "V" in the table if their projected design values
are greater than or equal to 15.05 |ig/m3. The 1999-2001 design values of these counties are also
listed. These are counties that are not projected to violate the standard, but to be close to it, so
the final rule will help ensure that these counties continue to meet the standard in either the base
or control case for at least one of the years analyzed.
                                           2-33

-------
Final Regulatory Impact Analysis
                                    Table 2.1.2-3
           Counties with 2020 and 2030 Projected Annual
                   within Ten Percent of the Annual PM2
PM2.5 Design Values
.5 Standard.3'b
State
AL
AL
AL
AL
AL
AR
AR
CA
CA
CA
CA
CA
CT
DE
GA
IL
IL
IL
IN
IN
IN
IN
IN
IN
IN
KY
KY
KY
KY
KY
KY
KY
KY
LA
LA
LA
LA
County
Alabama
DeKalb
Houston
Madison
Mobile
Crittenden
Pulaski
Butte
Imperial
Kings
San Joaquin
Ventura
Fairfield
Sussex
Hall
Du Page
Macon
Will
Elkhart
Floyd
Howard
Marion
Porter
Tippecanoe
Vanderburgh
Bell
Boyd
Bullitt
Campbell
Daviess
Fayette
Kenton
Pike
Caddo
Calcasieu
East Baton Rouge
Iberville
1999-2001
Design Value
(ue/nrY
15.5
16.8
16.3
15.5
15.3
15.3
15.9
15.4
15.7
16.6
16.4
14.5
13.6
14.5
17.2
15.4
15.4
15.9
15.1
15.6
15.4
17
13.9
15.4
16.9
16.8
15.5
16
15.5
15.8
16.8
15.9
16.1
13.7
12.7
14.6
13.9
2020
Base
X
X
V

X
X
X

X
X
X
X


V
X
X
V
X
X
X
V

X
X
X
X

X
X
X
X
X


X
X
Control"
X
X
X

X
X
X

X

X
X


X
X
X
X

X

X


X
X
X



X
X
X


X

2030
Base
X
V
V
X
V
X
X
X
V
X
V
X
X
X
V
V
X
V
X
X
X
V
X
X
V
X
X
X
X
X
X
V
X
X
X
V
X
Control"
X
V
V

V
X
X
X
X
X
X
X


V
X
X
V
X
X
X
V

X
X
X
X

X
X
X
X
X
X

V
X
Population
in 2000
14,254
64,452
88,787
276,700
399,843
50,866
361,474
203,171
142,361
129,461
563,598
753,197
882,567
156,638
139,277
904,161
114,706
502,266
182,791
70,823
84,964
860,454
146,798
148,955
171,922
30,060
49,752
61,236
88,616
91,545
260,512
151,464
68,736
252,161
183,577
412,852
33,320
                                       2-34

-------
State
LA
LA
LA
MD
MA
MA
MI
MS
MS
MS
MS
MS
MO
MO
MO
MO
MO
NJ
NJ
NY
NC
NC
NC
NC
NC
NC
NC
NC
NC
NC
NC
NC
NC
NC
OH
OH
OH
OH
OH
PA
PA
PA
PA
County
Jefferson
Orleans
West Baton Rouge
Baltimore
Hampden
Suffolk
Kalamazoo
Forrest
Hinds
Jackson
Jones
Lauderdale
Jackson
Jefferson
St Charles
St Louis
St Louis City
Mercer
Union
Bronx
Alamance
Cabarrus
Catawba
Cumberland
Durham
Forsyth
Gaston
Guilford
Haywood
McDowell
Mitchell
Orange
Wake
Wayne
Butler
Lorain
Mahoning
Portage
Trumbull
Berks
Cambria
Dauphin
Delaware
1999-2001
Design Value
Cue/nrY
13.6
14.1
14.1
16
14.1
16.1
15
15.2
15.1
13.8
16.6
15.3
13.9
15
14.6
14.1
16.3
14.3
16.3
16.4
15.3
15.7
17.1
15.4
15.3
16.2
15.3
16.3
15.4
16.2
15.5
14.3
15.3
15.3
17.4
15.1
16.4
15.3
16.2
15.6
15.3
15.5
15
2020
Base

X
X
X

V
X
X
X

V
X

X
X

V
X
X
V
X
X
V
X
X
X
X
V
X
X
X

X

V
X
X
X
X
X

X
X
Control3


X
X

X

X


X
X

X


X

X
X
X
X
X

X
X
X
X

X


X

X

X
X
X
X


X
2030
Base
X
X
V
V
X
V
X
X
X
X
V
X
X
X
X
X
V
X
V
V
X
X
V
X
V
V
V
V
X
V
X
X
V
X
V
X
V
X
V
X
X
X
V
Control3
X
X
X
X

X
X
X
X
X
V
X

X
X

V
X
V
V
X
X
V
X
X
V
X
V
X
X
X

X

V
X
X
X
X
X

X
X
Population
in 2000
455,466
484,674
21,601
754,292
456,228
689,807
238,603
72,604
250,800
131,420
64,958
78,161
654,880
198,099
283,883
1,016,315
348,189
350,761
522,541
1,332,650
130,800
131,063
141,685
302,963
223,314
306,067
190,365
421,048
54,033
42,151
15,687
118,227
627,846
113,329
332,807
284,664
257,555
152,061
225,116
373,638
152,598
251,798
550,864

-------
State
PA
PA
PA
SC
SC
SC
SC
TN
TN
TN
TN
TN
TX
UT
VA
VA
VA
VA
WV
WV
WV
WV
WI
WI
County
Lancaster
Washington
York
Georgetown
Lexington
Richland
Spartanburg
Davidson
Roane
Shelby
Sullivan
Sumner
Dallas
Salt Lake
Bristol City
Richmond City
Roanoke City
Virginia Beach Cit
Berkeley
Marshall
Ohio
Wood
Milwaukee
Waukesha
1999-2001
Design Value
Cue/nrY
16.9
15.5
16.3
13.9
15.6
15.4
15.4
17
17
15.6
17
15.7
14.4
13.6
16
14.9
15.2
13.2
16
16.5
15.7
17.6
14.5
14.1
Number of Counties within 10%b
Population of Counties within 10%°
2020
Base
X

X

X
X
X
X
X
X
X
X
X
X

X


X
X
X
V
X

70
23,836,400
Control3
X

X

X
X
X
X
X
X
X

X


X


X
X

X
X

62
24,151,800
2030
Base
X
X
V
X
V
X
X
V
X
V
V
X
V
V
X
V
X
X
X
X
X
V
V
X
64
16,870,300
Control3
X

X

X
X
X
V
X
X
X
X
X
X
X
X


X
X
X
V
X

70
24,839,600
Population
in 2000
470,658
202,897
381,751
55,797
216,014
320,677
253,791
569,891
51,910
897,472
153,048
130,449
2,218,899
898,387
17,367
197,790
94,911
425,257
75,905
35,519
47,427
87,986
940,164
360.767


a As described in Chapter 3, the final control case differs from the modeled control case based on public comment and
    updated information; however, we believe that the net results would approximate future emissions, although we
    anticipate the design value improvements would be smaller.  In our modeling, we do not account for SO2
    reductions related to sulfur reductions in home heating oil.
b Projections are made only for counties with monitored design values for 1999-2001. These were the most current data
    at the time the analyses were performed. Counties with insufficient data or lacking monitors are excluded.
c Populations are based on 2020 and 2030 estimates rounded to nearest hundred. See the AQ Modeling TSD for details.
    We estimate that the reduction of this final rule will produce nationwide air quality
improvements in PM levels.  On a population-weighted basis, the average change in future-year
annual averages is projected to decrease by 0.42 |ig/m3 in 2020, and 0.59 |ig/m3 in 2030.

    While the final implementation process for bringing the nation's air into attainment with the
PM2 5 NAAQS is still being completed in a separate rulemaking action, the basic framework is
well defined by the statute. EPA has requested that States and Tribes submit their
recommendations by February 15, 2004.  EPA's current plans call for designating PM2 5
attainement and nonattainment areas in December 2004. Following designation, Section 172(b)
of the Clean Air Act allows states up to 3 years to submit a revision to their state implementation

-------
                                              Air Quality, Health, and Welfare Effects
plan (SIP) that provides for the attainment of the PM2 5 standard. Based on this provision, states
could submit these SIPs in late-2007.  Section 172(a)(2) of the Clean Air Act requires that these
SIP revisions demonstrate that the nonattainment areas will attain the PM2 5 standard as
expeditiously as practicable but no later than 5 years from the date that the area was designated
nonattainment.  However, based on the severity of the air quality problem and the availability
and feasibility of control measures, the Administrator may  extend the attainment date "for a
period of no greater than  10 years from the date of designation as nonattainment." Based on
section 172(a) provisions in the Act, we expect that areas will need to attain the PM25 NAAQS
in the 2010 (based on 2007 - 2009 air quality data) to 2015 (based  on 2012 to 2014 air quality
data) time frame, and then be required to maintain the NAAQS thereafter.

    Since the emission reductions from this final rule will begin in this  same time frame, the
projected reductions in nonroad emissions will be used by states in meeting the PM2 5 NAAQS.
States and state organizations have told EPA that they need nonroad diesel engine reductions in
order to be able to meet and maintain  the PM2 5 NAAQS as well  as visibility regulations,
especially in light of the otherwise increasing emissions from nonroad sources without more
stringent standards.95'96'97 The following are sample comments from states and state
associations on the proposed rule, which corroborate that this rule is a critical element in States'
NAAQS attainment efforts. Fuller information can be found in the Summary and Analysis of
Comments.

    - "Unless emissions from nonroad diesels are sharply reduced,  it is very likely that many
    areas of the country will be unable to attain and maintain health-based NAAQS for ozone
    and PM." (STAPPA/ALAPCO)
    - "Adoption of the proposed regulation ... is necessary for the protection of public health in
    California and to comply with air  quality standards." (California Air Resources Board)
    - "The EPA's  proposed regulation is necessary if the West is to make reasonable progress
    towards improving  visibility in our nation's Class I areas." (Western Regional Air
    Partnership (WRAP))
    - "Attainment of the NAAQS for ozone and PM25 is of immediate concern to the states in the
    northeast region....Thus, programs ... such as the proposed rule for nonroad diesel engines are
    essential."  (NESCAUM)

    Furthermore, this rule ensures that nonroad diesel emissions  will continue to decrease as the
fleet turns over in the years beyond 2014; these reductions will be important for maintenance of
the NAAQS following  attainment.  The future reductions are also important to achieve visibility
goals, as discussed below.

2.1.3 Environmental Effects of Particulate Matter

    In this section, we discuss public welfare effects of PM and its precursors including visibility
impairment, acid deposition, eutrophication and nitrification, POM deposition, materials
damage, and soiling.
                                          2-37

-------
Final Regulatory Impact Analysis
    2.1.3.1 Visibility Degradation

    Visibility can be defined as the degree to which the atmosphere is transparent to visible
light.98  Visibility impairment has been considered the "best understood and most easily
measured effect of air pollution."99 Fine particles are the major cause of reduced visibility in
parts of the United States.  Haze obscures the clarity, color, texture, and form of what we see.
Visibility is an important effect because it has direct significance to people's enjoyment of daily
activities in all parts of the country. Visibility is also highly valued in significant natural  areas
such as national parks and wilderness areas, because of the special emphasis given to protecting
these lands now and for future generations.

    Scattering and absorption by both gases and particles decrease light transmittance. Size and
chemical composition of particles strongly affects their ability to scatter or absorb light.  The
same particles (sulfates, nitrates, organic  carbon, smoke, and soil dust) comprising PM2 5, which
are linked to serious health effects and environmental effects (e.g., ecosystem damage), can also
significantly degrade visual air quality. (For data on chemical composition of particles in sleeted
urban and rural areas, see Figures 2.1.2-2 and 2.1.2-3 above.) Sulfates contribute to visibility
impairment especially on the haziest days, accounting in the rural Eastern United States for more
than 60 percent of annual average light extinction on the best days and up to 86 percent of
average light extinction on the haziest days. Nitrates and elemental carbon each typically
contribute 1 to 6 percent of average light extinction on haziest days in rural locations in the
Eastern United States.100

     To quantify changes in visibility, the analysis presented in this chapter computes a light-
extinction coefficient, based on the work of Sisler, which shows the total fraction of light that is
decreased per unit distance.101  This coefficient  accounts for the scattering and absorption of light
by both particles and gases, and accounts for the higher extinction efficiency of fine particles
compared with coarse particles. Visibility can be described in terms of visual range, light
extinction or deciview.E  Visibility impairment  also has  a temporal dimension in that impairment
might relate to a short-term excursion or to longer periods (e.g., worst 20 percent of days  or
annual average levels).  More detailed discussions of visibility effects are contained in the EPA
Criteria Document for PM.102

    Visibility effects are manifest in two principal ways: (1) as local impairment (e.g., localized
hazes and plumes) and (2) as regional haze. The emissions from engines covered by this  rule
contribute to both types of visibility impairment.
    EVisual range can be defined as the maximum distance at which one can identify a black object against the
horizon sky. It is typically described in miles or kilometers. Light extinction is the sum of light scattering and
absorption by particles and gases in the atmosphere. It is typically expressed in terms of inverse megameters (Mm"1),
with larger values representing worse visibility. The deciview metric describes perceived visual changes in a linear
fashion over its entire range, analogous to the decibel scale for sound. A deciview of 0 represents pristine
conditions. The higher the deciview value, the worse the visibility, and an improvement in visibility is a decrease in
deciview value.

                                            2-38

-------
                                               Air Quality, Health, and Welfare Effects
   Local-scale visibility degradation is commonly in the form of either a plume resulting from
the emissions of a specific source or small group of sources, or it is in the form of a localized
haze such as an urban "brown cloud." Plumes are comprised of smoke, dust, or colored gas that
obscure the sky or horizon relatively near sources. Impairment caused by a specific source or
small group of sources has been generally termed as "reasonably attributable."

   The second type of impairment, regional haze, results from pollutant emissions from a
multitude of sources located across a broad geographic region.  It impairs visibility in every
direction over a large area, in some cases over multi-state regions. Regional haze masks objects
on the horizon and reduces the color and contrast of nearby objects.103

   On an annual average basis, the concentrations of non-anthropogenic fine PM are generally
small when compared with concentrations of fine particles from anthropogenic sources.104
Anthropogenic contributions account for about one-third of the average extinction coefficient in
the rural West and more than 80 percent in the rural East.105 In the Eastern United States,
reduced visibility is mainly attributable to secondarily formed particles, particularly those less
than a few micrometers in diameter (e.g.,  sulfates).  While secondarily formed particles still
account for a significant amount in the West, primary emissions contribute a larger percentage of
the total paniculate load than in the East.  Because of significant differences related to visibility
conditions in the Eastern and Western United States, we present information about visibility by
region.  Furthermore, it is important to note that even in those areas with relatively low
concentrations of anthropogenic fine particles, such as the Colorado plateau, small increases in
anthropogenic fine particle concentrations can lead to significant decreases in visual range.  This
is one of the reasons mandatory Federal Class I areas have been given special consideration
under the  Clean Air Act.  The 156 mandatory Federal Class I areas are displayed on the map in
Figure 2-1 above.

   EPA determined that emissions from nonroad engines significantly  contribute to air pollution
that may be reasonably anticipated to endanger public health and welfare for visibility effects in
particular (67 FR 68242, November 8, 2002).  The primary and PM-precursor emissions from
nonroad diesel engines subject to this rule contribute to these effects. To demonstrate this, in
addition to the inventory information in Chapter 3, we present information about both general
visibility impairment related to ambient PM levels across the country, and we also analyze
visibility conditions in mandatory Federal Class I areas. Accordingly, in this section, for both
the nation and for mandatory Federal Class I areas, we discuss the types of effects, current and
future visibility conditions absent the projected emission reductions, and the changes we
anticipate from the projected emission reductions. We conclude that the projected emission
reductions will improve visibility conditions across the country and in particular in mandatory
Federal  Class I areas.

   2.1.3.1.1  Visibility Impairment Where People Live, Work and Recreate

   Good visibility is valued by people throughout the country - in the places they live, work,
and enjoy recreational activities. However, unacceptable visibility impairment occurs in many
areas throughout the country. In this section, in order to estimate the magnitude of the visibility

                                          2-39

-------
Final Regulatory Impact Analysis
problem, we use monitored PM25 data and modeled air quality accounting for projected
emissions from nonroad diesel engines absent additional controls. The air quality modeling is
discussed in Section 2.1.2 above and in the AQ Modeling TSD.106 The engines covered by this
rule contribute to PM2 5 levels in areas across the country with significant visibility impairment.

   The secondary PM NAAQS is designed to protect against adverse welfare effects such as
visibility impairment. In 1997, the secondary PM NAAQS was set as equal to the primary
(health-based) PM NAAQS (62 Federal Register No. 138, July 18, 1997). EPA concluded that
PM can and does produce adverse effects on visibility in various locations, depending on PM
concentrations and factors such as chemical composition and average relative humidity.  In
1997, EPA demonstrated that visibility impairment is an important effect on public welfare and
that visibility impairment is experienced throughout the United States, in multi-state regions,
urban areas, and remote Federal Class I areas.

   The updated monitored data and air quality modeling presented below confirm that the
visibility situation identified during the NAAQS review in 1997 is still likely to exist.
Specifically, there will still likely be  a broad number of areas that are above the annual PM2 5
NAAQS in the Northeast, Midwest, Southeast and California , such that the determination in the
NAAQS rulemaking about broad visibility impairment and related benefits from NAAQS
compliance are still relevant. Thus, levels above the fine PM NAAQS cause  adverse welfare
impacts, such as visibility impairment (both regional and localized impairment). EPA recently
confirmed this in our determination about nonroad engines significant contribution to
unacceptable visibility impairment (67 FR 68251, November 8, 2002).

   In addition, in setting the PM NAAQS, EPA acknowledged that levels of fine particles below
the NAAQS may also contribute to unacceptable visibility impairment and regional haze
problems in some areas, and Clean Air Act Section 169 provides additional authorities to remedy
existing impairment and prevent future impairment in the 156 national parks, forests and
wilderness  areas labeled as mandatory Federal Class I areas (62 FR at 38680-81, July  18, 1997).

   In making determinations about the level of protection afforded by the secondary PM
NAAQS, EPA considered how the Section 169 regional haze program and the secondary
NAAQS would function together.107  Regional strategies, such as this rule, are expected to
improve visibility in many urban and non-Class I areas as well. Visibility impairment in
mandatory Federal Class I areas is discussed in Section 2.1.4.

       2.1.3.1.1.1 Current Areas Affected by Visibility Impairment: Monitored Data

   The need for reductions  in the levels of PM2 5 is widespread, as discussed above and  shown
in Figure 2-1.  Currently, high ambient PM2 5 levels are measured throughout the country. Fine
particles may remain suspended for days or weeks and travel hundreds to thousands of
kilometers, and thus fine particles emitted  or created in one county may contribute to ambient
concentrations in a neighboring region.108

   Without the effects of pollution, a natural visual range is approximately 120 to 180 miles

                                         2-40

-------
                                               Air Quality, Health, and Welfare Effects
(200 to 300 kilometers) in the West and 45 to 90 miles (75 to 150 kilometers) in the East.109
However, over the years, in many parts of the United States, fine particles have significantly
reduced the range that people can see.  In the West, the visibility range is 33 to 90 miles (53 to
144 kilometers), and in the East, the current range is only 14 to 24 miles (22 to 38 kilometers).110

   Current PM2 5 monitored values for 2000-2002 indicate that almost 65 million people in  120
counties live in areas where design values of PM25 annual levels are at or above the PM25
NAAQS.  This represents 23 percent of the counties and 37 percent of the population in the areas
with monitoring data that met completeness requirements and had levels above the NAAQS.
Thus, at least these populations (plus others who travel to these areas) would likely be
experiencing visibility impairment that is unacceptable.  Emissions of PM and its precursors
from nonroad diesel engines contribute to this unacceptable impairment.

   An additional 32 million people live in 91 counties that have air quality measurements for
2000-2002 within 10 percent of the level of the PM standard. These areas, though not currently
violating the standard, will also benefit from the additional reductions from this final rule to
ensure long-term maintenance of the standard and to prevent deterioration in visibility
conditions.

   Although we present the annual average to represent national visibility conditions, visibility
impairment can also occur on certain days or other shorter periods. As discussed below, the
Regional Haze program targets the worst 20 percent of days in a year. The projected emission
reductions from this rule are also needed to improve visibility on the worst days.

       2.1.3.1.1.2 Areas Affected by Future Visibility Impairment

   Because the chemical composition of PM and other atmospheric conditions affect visibility
impairment, we used the REMSAD air quality model to project visibility conditions in 2020  and
2030 to estimate visibility impairment directly as changes in deciview.  One of the inputs to the
PM modeling described above is a projection of future emissions from nonroad diesel engines
absent additional controls. Thus, we are able to demonstrate that the nonroad diesel emissions
contribute to the projected visibility impairment and that there continues to be a need for
reductions from those engines.

   As described above, based on this modeling and absent additional controls, we predicted that
in 2020, there will be 79 counties with a population of 66 million where annual PM2 5 levels are
above 15 |ig/m3.in In 2030, this number will rise to 107 counties with a population  of 85 million
in the absence of additional controls.  Section 2.1.2 and the AQ Modeling TSD provides
additional details.

   Based upon the light-extinction coefficient, we also calculated a unitless visibility index or
deciview. As shown in Table 2.1.3-1,  in 2030 we estimate visibility in the East to be about
20.54 deciviews (or visual range of 50 kilometers) on  average, with poorer visibility in urban
areas, compared with the visibility conditions without man-made pollution of 9.5 deciviews (or
visual range of 150 kilometers). Likewise, we estimate visibility in the West to be about 8.83

                                          2-41

-------
Final Regulatory Impact Analysis
deciviews (or visual range of 162 kilometers) in 2030, compared with the visibility conditions
without anthropogenic pollution of 5.3 deciviews (or visual range of 230 kilometers). Thus, in
the future, a substantial percent of the population may experience unacceptable visibility
impairment in areas where they live, work and recreate.

                                       Table 2.1.3-1
                  Summary of Future National (48 state) Baseline Visibility
                     Conditions Absent Additional Controls (Deciviews)
Regions*
Eastern U.S.
Urban
Rural
Western U.S.
Urban
Rural
Predicted 2020
Visibility
(annual average)
20.27
21.61
19.73
8.69
9.55
8.5
Predicted 2030
Visibility
(annual average)
20.54
21.94
19.98
8.83
9.78
8.61
Natural Background
Visibility
9.5
5.3
       ' Eastern and Western Regions are separated by 100 degrees north longitude. Background visibility conditions
          differ by region.
    The emissions from nonroad diesel engines contribute to this visibility impairment as
discussed in Chapter 3. Nonroad diesel engines emissions contribute a large portion of the total
PM emissions from mobile sources and anthropogenic sources, in general. These emissions
occur in and around areas with PM levels above the annual PM2 5 NAAQS.  The nonroad
engines subject to this rule contribute to these effects as well as localized visibility impairment.
Thus, the emissions from these sources contribute to the unacceptable current and anticipated
visibility impairment.

       2.1.3.1.1.3 Future Improvements in Visibility from the Projected Emission Reductions

    For this rule, we also modeled a preliminary control scenario that illustrates the likely
emission reductions. As public comment and additional data regarding technical feasibility and
other factors became available, our judgment about the controls that are feasible has evolved.
Thus, the preliminary control option differs from what we are proposing,  as summarized in
Section 3.6.  It is important to note that these changes would not affect our estimates of the
baseline conditions without additional controls described above. In our air quality modeling, we
did not account for SO2 reductions from reductions in sulfur levels in home heating oil. We
anticipate that the nonroad diesel emission reductions from this final rule together with other
strategies would improve the projected visibility impairment, and we conclude that there
continues to be a need for reductions from those engines.
                                           2-42

-------
                                              Air Quality, Health, and Welfare Effects
   Based on our modeling, we predict that in 2020, there will be 12 counties with a population
of 6 million that come into attainment with the annual PM25 because of the improvements in air
quality from the emission reductions resulting from this final rule. In 2030, an estimated total of
24 counties (12 additional counties) with a population of 14 million (8 million additional people)
will come into attainment with the annual PM25 because of the improvements in air quality from
this final rule.  There will also be emission reductions in counties with levels close to the air
quality standards that will improve visibility conditions and help them maintain the standards.
All of these areas and their populations will experience improvements in visibility as well as
health effects, as described earlier.

   We estimate that the emission reductions resulting from this final rule will produce
nationwide air quality improvements in PM levels. On a population-weighted basis, the average
change in future-year annual averages will be a decrease of 0.33 |ig/m3 in 2020, and 0.46 |ig/m3
in 2030.  These reductions are discussed in more detail in Section 2.1.2 above.

   We can also calculate these improvement in visibility as decreases in deciview value. As
shown in Table 2.1.3-2, in 2030 we estimate visibility in the East to be about 20.54 deciviews (or
visual range of 50 kilometers) on average, with poorer visibility in urban areas.  Emission
reductions from this final rule in 2030 will improve visibility by an estimated 0.33 deciviews.
Likewise, we estimate visibility in the West to be about 8.83 deciviews (or visual  range of 162
kilometers) in 2030, and we estimate that emission reductions from this final rule  in 2030 will
improve visibility by 0.25 deciviews.  These improvements are needed in conjunction with other
sulfur reduction strategies in the East and a combination of strategies in the West to make
reasonable progress toward visibility goals.112  Thus, this final rule is an important part of
strategies to improve visibility in areas where they live, work and recreate.
                                          2-43

-------
Final Regulatory Impact Analysis
                                       Table 2.1.3-2
                    Summary of Future National Visibility Improvements
           from Nonroad Diesel Emission Reductions (Annual Average Deciviews)
Regions8
Eastern U.S.
Urban
Rural
Western U.S.
Urban
Rural
2020
Predicted Baseline
2020 Visibility
20.27
21.61
19.73
8.69
9.55
8.5
Predicted 2020
Control Visibility15
20.03
21.37
19.49
8.51
9.3
8.33
2030
Predicted Baseline
2030 Visibility
20.54
21.94
19.98
8.83
9.78
8.61
Predicted 2030
Control Visibility15
20.21
21.61
19.65
8.58
9.43
8.38
a Eastern and Western Regions are separated by 100 degrees north longitude. Background visibility conditions differ by
    region.
b The results illustrate the type of visibility improvements for the preliminary control option, as discussed in Section 3.6.
    The analysis in Chapter 3 differs based on updated information; however, we believe that the net results would
    approximate future PM emissions, although we anticipate the annual average visibility improvements would be
    smaller.
    2.1.3.1.2 Visibility Impairment in Mandatory Federal Class I Areas

    Achieving the annual PM2 5 NAAQS will help improve visibility across the country, but it
will not be sufficient to meet the statutory goal of no manmade impairment in the mandatory
Federal Class I areas (64 FR 35722, July 1, 1999 and 62 FR 38680, July 18, 1997).  In setting the
NAAQS, EPA discussed how the NAAQS in combination with the regional haze program,  is
deemed to improve visibility consistent with the goals of the Act.113 In the East,  there are and
will continue to be sizable areas above 15 |ig/m3 and where light extinction is significantly  above
natural background. Thus, large areas of the Eastern United States have air pollution that is
causing and will continue to cause unacceptable visibility problems. In the West, scenic vistas
are especially important to public welfare. Although the annual PM2 5 NAAQS is met in most
areas outside of California, virtually the entire West is in close proximity to  a scenic mandatory
Federal Class I area protected by  169 A  and 169B of the Act.

    The 156 Mandatory Federal Class I  areas are displayed on the map in Figure 2-1 above.
These areas include many of our best known and most treasured natural areas, such as the Grand
Canyon, Yosemite, Yellowstone,  Mount Rainier,  Shenandoah, the  Great Smokies, Acadia,  and
the Everglades. More than 280 million visitors come to enjoy the scenic vistas and unique
natural features including the night sky  in these and other park and wilderness areas each year.
                                           2-44

-------
                                                Air Quality, Health, and Welfare Effects
    In the 1990 Clean Air Act amendments, Congress provided additional emphasis on regional
haze issues (see section 169B). In 1999 EPA finalized a rule that calls for States to establish
goals and emission reduction strategies for improving visibility in all 156 mandatory Class I
national parks and wilderness areas.  In this rule, EPA established a "natural visibility" goal.114
In that rule, EPA also encouraged the States to work together in developing and implementing
their air quality plans.  The regional haze program is focused on long-term emissions decreases
from the entire regional emission inventory comprised of major and minor stationary sources,
area sources and mobile sources.  The regional haze program is designed to improve visibility
and air quality in our most treasured natural areas so that these areas may be preserved and
enjoyed by current and future generations. At the same time, control strategies designed to
improve visibility in the national parks and wilderness areas will improve visibility over broad
geographic areas, including other recreational sites, our cities and residences. In the PM
NAAQS rulemaking, EPA also anticipated the need in addition to the NAAQS and Section  169
regional haze program to continue to address localized impairment that may relate to unique
circumstances in some Western areas. For mobile sources, there may also be a need for a
Federal role in reduction of those emissions, in particular, because mobile source engines are
regulated primarily at the Federal level.

    The regional haze program calls for states to establish goals for improving visibility in
national parks and wilderness areas to improve visibility on the haziest 20 percent of days and to
ensure that no degradation occurs on the clearest 20 percent of days (64 FR 35722.  July 1,
1999). The rule requires states to develop long-term strategies including enforceable measures
designed to meet reasonable progress goals toward natural visibility conditions.  Under the
regional haze program, States can take credit for improvements in air quality achieved as a result
of other Clean Air Act programs, including national mobile-source programs.F

       2.1.3.1.2.1 Current Mandatory Federal Class I Areas Affected by Visibility Impairment:
       Monitored Data

    Detailed information about current and historical visibility conditions in mandatory Federal
Class I areas is summarized in the EPA Report to Congress and the recent EPA Trends Report.115
The conclusions draw upon the Interagency Monitoring of Protected Visual Environments
(IMPROVE) network data.116 The National Park Service report also describes the state of
national park visibility conditions and discusses the need for improvement.117

    As described in the EPA Trends Report 1999, most of the IMPROVE sites in the
intermountain West and Colorado Plateau have annual average impairment of 12 deciviews or
    F Although a recent court case, American Corn Growers Association v. EPA, 291F.3d 1(D.C .Cir 2002), vacated
the Best Available Retrofit Technology (BART) provisions of the Regional Haze rule, the court denied industry's
challenge to EPA's requirement that state's SIPS provide for reasonable progress towards achieving natural visibility
conditions in national parks and wilderness areas and the "no degradation" requirement. Industry did not challenge
requirements to improve visibility on the haziest 20 percent of days. The court recognized that mobile source
emission reductions would need to be a part of a long-term emission strategy for reducing regional haze.  A copy of
this decision can be found in Docket A-2000-01, Document IV- A-l 13.

                                           2-45

-------
Final Regulatory Impact Analysis
less, with the worst days ranging up to 17 deciviews (compared with 5.3 deciviews of natural
background visibility).118 Several other western IMPROVE sites in the Northwest and California
experience levels on the order of 16 to 23 deciviews on the haziest 20 percent of days. Many
rural locations in the East have annual average values exceeding 21 deciviews, with average
visibility levels on the haziest days up to 32 deciviews.

    Although there have been general trends toward improved visibility, progress is still needed
on the haziest days.  Specifically, as discussed in the EPA Trends Report, in the 10 Class I areas
in the Eastern United States, visibility on the haziest 20 percent of days remains significantly
impaired with a mean visual range of 23 kilometers for 1999, as compared with 84  kilometers for
the clearest days in 1999. In the 26 Class I reported areas in the Western United States, the
conditions for the haziest 20 percent of days degraded between 1997 and 1999 by 17 percent.
However, visibility on the haziest 20 percent of days in the West remains relatively unchanged
over the 1990s with the mean visual range for 1990 (80 kilometers) nearly the same as the 1990
level (86 kilometers).

       2.1.3.1.2.2 Mandatory Federal Class I Areas Affected by Future Visibility Impairment

    As part of the PM air quality modeling described above, we modeled future visibility
conditions in the mandatory Federal Class I areas absent additional controls. The results by
region are summarized in Table 2.1.3-3.  In Figure 2.1.3-1, we define the regions used in this
analysis.119 These air quality results show that visibility is impaired in most mandatory Federal
Class I areas and additional reductions from engines subject to this rule are needed to achieve the
goals of the Clean Air Act of preserving natural conditions in mandatory Federal Class I areas.
                                          2-46

-------
                                                   Air Quality, Health, and Welfare Effects
                                          Table 2.1.3-3
        Summary of Future Baseline Visibility Conditions in Mandatory Federal Class I
           Areas Absent Additional Emission Reductions (Annual Average Deciview)
Class I Regions a
Eastern
Southeast
Northeast/Midwest
Western
Southwest
California
Rocky Mountain
Northwest
National Class I Area
Average
Predicted 2020 Visibility
19.72
21.31
18.30
8.80
6.87
9.33
8.46
12.05
11.61
Predicted 2030 Visibility
20.01
21.62
18.56
8.96
7.03
9.56
8.55
12.18
11.80
Natural Background
Visibility
9.5
5.3

1 Regions are depicted in Figure 1-5.1. Background visibility conditions differ by region based on differences in relative
   humidity and other factors: Eastern natural background is 9.5 deciviews (or visual range of 150 kilometers) and in
   the West natural background is 5.3 deciviews (or visual range of 230 kilometers).
                                              2-47

-------
                                                           Figure 2.1.3-1
                                       Visibility Regions for the Continental United States
                            Study Region
                            Transfer Region
Note: Study regions were represented in the Chestnut and Rowe (1990a, 1990b) studies used in evaluating the benefits of visibility improvements.

-------
                                                 Air Quality, Health, and Welfare Effects
       2.1.3.1.2.3 Future Improvements in Mandatory Federal Class I Visibility from the
       Projected Emission Reductions

    The overall goal of the regional haze program is to prevent future and remedy existing
visibility impairment in mandatory Federal Class I areas.  As shown by the future deciview
estimates in Table 2.1.3-4, additional emission reductions will be needed from the broad set of
sources that contribute, including the emissions from engines subject to this rule.  The table also
presents the results from our modeling of  a preliminary control scenario that illustrates the likely
reductions from the final rule.  Emission reductions from nonroad diesel engines are needed to
achieve the goals of the Act of preserving natural conditions in mandatory Federal Class I areas.
These reductions are a part of the overall strategy to achieve the visibility goals of the Act and
the regional haze program.

                                        Table 2.1.3-4
        Summary of Future Visibility Improvements15 in Mandatory Federal Class I Areas
            from Nonroad Diesel Emission Reductions (Annual Average Deciviews)
Mandatory Federal
Class I Regions3
Eastern
Southeast
Northeast/Midwest
Western
Southwest
California
Rocky Mountain
Northwest
National Class I Area
Average
2020
Predicted Baseline
2020 Average
Visibility
19.72
21.31
18.30
8.80
6.87
9.33
8.46
12.05
11.61
Predicted 2020
Control Average
Visibility15
19.54
21.13
18.12
8.62
6.71
9.12
8.31
11.87
11.43
2030
Predicted Baseline
2030 Average
Visibility
20.01
21.62
18.56
8.96
7.03
9.56
8.55
12.18
11.80
Predicted 2030
Control Average
Visibility15
19.77
21.38
18.32
8.72
6.82
9.26
8.34
11.94
11.56
a Regions are presented in Figure 2.1.3-1 based on Chestnut and Rowe (1990a, 1990b) study regions.
b The results illustrate the type of visibility improvements for the preliminary control option, as discussed in Section 3.6.
    The final control scenario described in Chapter 3 differs from the modeled scenario based on public comment and
    updated information; however, we believe that the net results would approximate future PM emissions, although we
    anticipate the annual average visibility improvements would be smaller.
                                            2-49

-------
Final Regulatory Impact Analysis
2.1.3.2 Other Effects

   2.1.3.2.1  Acid Deposition

   Acid deposition, or acid rain as it is commonly known, occurs when SO2 and NOx react in
the atmosphere with water, oxygen, and oxidants to form various acidic compounds that later fall
to earth in the form of precipitation or dry deposition of acidic particles.120  It contributes to
damage of trees at high elevations and  in extreme cases may cause lakes and streams to become
so acidic that they cannot support aquatic life. In addition, acid deposition accelerates the decay
of building materials and paints, including irreplaceable buildings, statues,  and sculptures that
are part of our nation's cultural heritage.  To reduce damage to automotive paint caused by acid
rain and acidic dry deposition,  some manufacturers use acid-resistant paints, at an average cost
of $5 per vehicle—a total of near $80 million per year when applied to all new cars and trucks
sold in the United States each year.

   Acid deposition primarily affects bodies of water that rest atop soil with a limited ability to
neutralize acidic compounds.  The National Surface Water Survey (NSWS) investigated the
effects of acidic deposition in over 1,000 lakes larger than 10 acres and in thousands of miles of
streams. It found that acid deposition was the primary cause of acidity in 75 percent of the
acidic lakes and about 50 percent of the acidic streams, and that the areas most sensitive to acid
rain were the Adirondacks, the mid-Appalachian highlands, the upper Midwest and the high
elevation West. The NSWS found that approximately 580 streams in the Mid-Atlantic Coastal
Plain are acidic primarily due to acidic deposition. Hundreds of the lakes in the Adirondacks
surveyed in the NSWS have acidity levels incompatible with the survival of sensitive fish
species. Many of the over 1,350 acidic streams in the Mid-Atlantic Highlands (mid-Appalachia)
region have already experienced trout losses due to increased stream acidity.  Emissions from
U.S. sources contribute to acidic deposition in Eastern Canada, where the Canadian government
has estimated that 14,000 lakes are acidic. Acid deposition also has been implicated in
contributing to degradation of high-elevation spruce forests that populate the ridges of the
Appalachian Mountains from Maine to Georgia. This area includes national parks such as the
Shenandoah and Great Smoky Mountain National Parks.

   A study of emission trends and acidity of water bodies in the Eastern United States by the
General Accounting Office (GAO) found that from 1992 to 1999 sulfates declined in 92 percent
of a representative sample of lakes, and nitrate levels increased in 48 percent of the lakes
sampled.121 The decrease in sulfates is consistent with emission trends, but the increase in
nitrates is  inconsistent with the stable levels of nitrogen emissions and deposition.  The study
suggests that the vegetation and land surrounding these lakes have lost some of their previous
capacity to use nitrogen, thus allowing more of the nitrogen to flow into the lakes and increase
their acidity.  Recovery of acidified lakes is expected to take a number of years, even where soil
and vegetation have not been "nitrogen saturated," as EPA called the phenomenon in a 1995
study.122 This situation places a premium on reductions of SOx and especially NOx from all
sources, including nonroad diesel engines, in order to reduce the extent and severity of nitrogen
saturation and acidification of lakes in  the Adirondacks and throughout the United States.
                                          2-50

-------
                                               Air Quality, Health, and Welfare Effects
   The SOx and NOx reductions from this rule will help reduce acid rain and acid deposition,
thereby helping to reduce acidity levels in lakes and streams throughout the country and help
accelerate the recovery of acidified lakes and streams and the revival of ecosystems adversely
affected by acid deposition. Reduced acid deposition levels will also help reduce stress on
forests, thereby accelerating reforestation efforts and improving timber production.
Deterioration of our historic buildings and monuments, and of buildings, vehicles, and other
structures exposed to acid rain and dry acid deposition also will be reduced, and the costs borne
to prevent acid-related damage may also decline.  While the reduction in sulfur and nitrogen acid
deposition will be roughly proportional to the reduction in SOx and NOx emissions,
respectively, the precise impact of this rule will differ across different areas.

   2.1.3.2.2 Eutrophication and Nitrification

   Eutrophication is the accelerated production of organic matter, particularly algae, in a water
body.  This increased growth can cause numerous adverse ecological effects and economic
impacts, including nuisance algal blooms, dieback of underwater plants due to reduced light
penetration, and toxic plankton blooms. Algal and plankton blooms can also reduce the level  of
dissolved oxygen, which can also adversely affect fish and shellfish populations.

   In 1999, the National Oceanic and Atmospheric Administration (NOAA) published the
results of a five year national assessment of the severity and extent of estuarine eutrophication.
An estuary is defined as the inland arm of the sea that meets the mouth of a river.  The 138
estuaries characterized in the study  represent more than 90 percent of total estuarine water
surface area and the total number of U.S. estuaries. The study found that estuaries with moderate
to high eutrophication conditions represented 65 percent of the estuarine surface area.
Eutrophication is of particular concern in coastal areas with poor or stratified circulation
patterns, such as the Chesapeake Bay, Long Island Sound, or the Gulf of Mexico. In such areas,
the "overproduced" algae tends to sink to the bottom and decay, using all or most of the available
oxygen and thereby reducing or eliminating populations of bottom-feeder fish and shellfish,
distorting the normal population balance between different aquatic organisms, and in extreme
cases causing dramatic fish kills.

   Severe and persistent eutrophication often directly impacts  human activities. For example,
losses in the nation's fishery resources may be directly  caused by fish kills associated with low
dissolved oxygen and toxic blooms. Declines in tourism  occur when low dissolved oxygen
causes noxious smells and floating mats of algal blooms create unfavorable aesthetic conditions.
Risks to human health increase when the toxins from algal blooms accumulate in  edible fish and
shellfish, and when toxins become airborne, causing respiratory problems due to inhalation.
According to the NOAA report, more than half of the nation's estuaries have moderate to high
expressions of at least one of these symptoms - an indication that eutrophication is well
developed in more than half of U.S. estuaries.

   In recent decades, human activities have greatly accelerated nutrient inputs, such as nitrogen
and phosphorous, causing excessive growth of algae and  leading to degraded water quality  and
associated impairments of freshwater and estuarine resources for human uses.123 Since 1970,

                                          2-51

-------
Final Regulatory Impact Analysis
eutrophic conditions worsened in 48 estuaries and improved in 14. In 26 systems, there was no
trend in overall eutrophication conditions since 1970.124 On the New England coast, for
example, the number of red and brown tides and shellfish problems from nuisance and toxic
plankton blooms have increased over the past two decades, a development thought to be linked
to increased nitrogen loadings in coastal waters. Long-term monitoring in the United States,
Europe, and other developed regions of the world  shows a substantial rise of nitrogen levels in
surface waters, which are highly correlated with human-generated inputs of nitrogen to their
watersheds.

   Between 1992 and 1997, experts surveyed by National Oceanic and Atmospheric
Administration (NOAA) most frequently recommended that control strategies be developed for
agriculture, wastewater treatment, urban runoff, and atmospheric deposition.125 In its Third
Report to Congress on the Great Waters, EPA reported that atmospheric deposition contributes
from 2 to 38 percent of the nitrogen load to certain coastal waters.126  A review of peer reviewed
literature in 1995 on the subject of air deposition suggests a typical contribution of 20 percent or
higher.127 Human-caused nitrogen loading to the Long Island Sound from the atmosphere was
estimated at 14 percent by a collaboration of federal and state air and water agencies in 1997.128
The National Exposure Research Laboratory,  U.S. EPA, estimated based on prior studies that 20
to 35 percent of the nitrogen loading to the Chesapeake Bay is attributable to atmospheric
deposition.129 The mobile source portion of atmospheric NOx contribution to the Chesapeake
Bay was modeled at about 30 percent  of total  air deposition.130

   Deposition of nitrogen from nonroad diesel engines contributes to elevated nitrogen levels in
waterbodies. The new emission standards for nonroad diesel engines will reduce total NOx
emissions by 738,000 tons in 2030. The NOx reductions will reduce the airborne nitrogen
deposition that contributes to eutrophication of watersheds, particularly in aquatic systems where
atmospheric deposition of nitrogen represents a significant portion of total nitrogen loadings.

   2.1.3.2.3 Polycyclic Organic Matter (POM) Deposition

   EPA's Great Waters Program has identified 15 pollutants whose deposition to water bodies
has contributed to the overall contamination loadings to the these  Great Waters.131 One of these
15 compounds, a group known as poly cyclic organic matter (POM), are compounds that are
mainly adhered to the particles emitted by mobile  sources and later fall to earth in the form of
precipitation or dry deposition of particles.  The mobile source contribution of the seven most
toxic POM is at least 62 tons/year and represents only those POM that are adhered to mobile
source paniculate emissions.132 The majority  of these emissions are produced by diesel engines.

   POM is generally defined as a large class of chemicals consisting of organic compounds
having multiple benzene rings and a boiling point  greater than 100° C. Polycyclic aromatic
hydrocarbons are a chemical class that is a subset  of POM.  POM are naturally occurring
substances that are byproducts of the incomplete combustion of fossil fuels and plant and animal
biomass (e.g., forest fires).  Also, they occur as byproducts from steel and coke productions and
waste incineration.
                                          2-52

-------
                                               Air Quality, Health, and Welfare Effects
   Evidence for potential human health effects associated with POM comes from studies in
animals (fish, amphibians, rats) and in human cells culture assays.  Reproductive, developmental,
immunological, and endocrine (hormone) effects have been documented in these systems. Many
of the compounds included in the class of compounds known as POM are classified by EPA as
probable human carcinogens based on animal data.

   The new emission standards will reduce not only the PM emissions from land-based nonroad
diesel engines, but also the deposition of the POM adhering to the particles, thereby reducing
health effects of POM in lakes and streams, accelerating the recovery of affected lakes and
streams, and reviving adversely affected ecosystems.

   2.1.3.2.4 Materials Damage and Soiling

   The deposition of airborne particles can also reduce the aesthetic appeal of buildings and
culturally important articles through soiling, and can contribute directly (or in conjunction with
other pollutants) to structural damage by means of corrosion or erosion. Particles affect materials
principally by promoting and accelerating the corrosion of metals, by degrading paints, and by
deteriorating building materials such as concrete and limestone. Particles contribute to these
effects because of their electrolytic, hygroscopic, and acidic properties, and their ability to sorb
corrosive gases (principally sulfur dioxide). The rate of metal  corrosion depends on a number of
factors,  including the deposition rate and nature of the pollutant; the influence of the metal
protective corrosion film; the amount of moisture present; variability in the electrochemical
reactions; the presence and concentration of other surface electrolytes; and the orientation of the
metal surface.

   Paints undergo natural weathering processes from exposure to environmental factors such as
sunlight, moisture, fungi, and varying temperatures.  In addition to the natural environmental
factors,  studies show paniculate matter exposure may give painted surfaces a dirty appearance.
Several  studies also suggest that particles serve as carriers of other more corrosive pollutants,
allowing the pollutants to reach the underlying surface or serve as concentration sites for other
pollutants.  A number of studies have shown some correlation between particulate matter and
damage to automobile finishes. A number of studies also support the conclusion that gaseous
pollutants contribute to the erosion rates of exterior paints.

   Damage to calcareous stones (i.e., limestone, marble and carbonated cemented stone) has
been attributed to deposition of acidic particles.  Moisture and  salts are considered the most
important factors in building material damage. However, many other factors (such as normal
weathering and microorganism damage) also seem to play a part in the  deterioration of inorganic
building materials. The relative importance of biological, chemical, and physical mechanisms
has not been studied to date.  Thus, the relative contribution of ambient pollutants to the damage
observed in various building stone is not well quantified.  Under high wind conditions,
particulates result in slow erosion of the surfaces, similar to sandblasting.

   Soiling is the accumulation of particles on the surface of an exposed material resulting in the
degradation of its appearance.  When such accumulation produces  sufficient changes in

                                          2-53

-------
Final Regulatory Impact Analysis
reflection from opaque surfaces and reduces light transmission through transparent materials, the
surface will become perceptibly dirty to the human observer. Soiling can be remedied by
cleaning or washing, and depending on the soiled material, repainting.

2.2 Air Toxics

2.2.1 Diesel Exhaust  PM

   A number of health studies have been conducted regarding diesel exhaust including
epidemiologic studies  of lung cancer in groups of workers, and animal studies focusing on non-
cancer effects specific to diesel exhaust. Diesel exhaust PM (including the associated organic
compounds that are generally high molecular-weight hydrocarbon types, but not the more
volatile  gaseous hydrocarbon compounds) is generally used as a surrogate measure for diesel
exhaust.

   2.2.1.1 Potential Cancer Effects of Diesel Exhaust

   In addition to its contribution to ambient PM inventories, diesel exhaust is of specific
concern because it has been judged to pose a lung cancer hazard for humans as well as a hazard
from noncancer respiratory effects such as pulmonary inflammation.

   In 2001, EPA completed  a rulemaking on mobile source air toxics with a determination that
diesel particulate matter and diesel exhaust organic gases be identified as a Mobile Source Air
Toxic (MSAT).133 This determination was based on a draft of the Diesel HAD on which the
Clean Air Scientific Advisory Committee (CAS AC) of the Science Advisory Board had reached
closure. Including both diesel PM and diesel exhaust organic gases in the determination was
made in order to be precise about the components of diesel exhaust expected to contribute to the
observed cancer and non-cancer health effects. Currently available science, while suggesting an
important role for the particulate phase component of diesel exhaust, does not attribute the likely
cancer and noncancer health  effects independently to diesel particulate matter as distinct from
the gas phase components (EPA, 2001). The purpose of the MSAT list is to provide a screening
tool that identifies compounds emitted from motor vehicles or their fuels for which further
evaluation of emission controls is appropriate.

   EPA released its final "Health Assessment Document for Diesel Engine Exhaust"  (the EPA
Diesel HAD), referenced earlier. There, diesel exhaust was classified as likely to be
carcinogenic to humans by inhalation at environmental exposures, in accordance with the revised
draft 1996/1999 EPA cancer  guidelines.134 In accordance with earlier EPA guidelines, diesel
exhaust would be similarly classified as a probable human carcinogen (Group  Bl).135'136 A
number of other agencies (National Institute for Occupational Safety and Health, the
International Agency for Research on Cancer, the World  Health Organization, California EPA,
and the U.S. Department of Health and Human Services) have made similar
classifications.137'138'139440441 The Health Effects Institute  has also made numerous studies and
report on the potential carcinogenicity of diesel exhaust.142'143'144 Numerous animal and
                                          2-54

-------
                                              Air Quality, Health, and Welfare Effects
bioassay/genotoxic tests have been done on diesel exhaust.145'146 Also, case-control and cohort
studies have been conducted on railroad engine exposures147'148'149 in addition to studies on truck
workers.150'151'152 Also, there are numerous other epidemiologic studies including some studying
mine workers and fire fighters.153'154

   It should be noted that the conclusions in the EPA Diesel HAD were based on diesel engines
currently in use, including nonroad diesel engines such as those found in bulldozers, graders,
excavators, farm tractor drivers and heavy construction equipment.  As new diesel engines with
significantly less PM exhaust emissions replace existing engines, the conclusions of the EPA
Diesel HAD will need to be reevaluated.

   More specifically, the EPA Diesel HAD states that the conclusions of the document apply to
diesel exhaust in use today including both highway and nonroad engines. The EPA Diesel HAD
acknowledges that the studies were done on engines with older technologies generally for
highway applications and that "there have been changes  in the physical and chemical
composition of some DE [diesel exhaust] emissions (highway vehicle emissions) overtime,
though there is no definitive information to show that the emission changes portend significant
toxicological changes."  The EPA Diesel HAD further concludes that "taken together, these
considerations have led to a judgment that the hazards identified from older-technology-based
exposures are applicable to current-day exposures." The diesel technology used for nonroad
diesel engines typically lags that used for highway engines, which have been subject to PM
standards since 1988.  Thus, the conclusions from the EPA Diesel HAD continue to be relevant
to current nonroad diesel engine emissions.

   Some of the epidemiologic  studies discussed in the EPA Diesel HAD were conducted
specifically on nonroad diesel engine emissions.  In particular, one recent study examined
bulldozer operators, graders, excavators, and full-time farm tractor drivers finding increased
odds of lung cancer.155  Another cohort study of operators of heavy construction equipment also
showed increased lung cancer incidence for these workers.156

   For the EPA Diesel HAD, EPA reviewed 22 epidemiologic studies in detail, finding
increased lung cancer risk in 8  out of 10 cohort studies and 10 out of 12 case-control studies.
Relative risk for lung cancer associated with exposure range from 1.2 to 2.6. In addition, two
meta-analyses of occupational studies of diesel exhaust and lung cancer have estimated the
smoking-adjusted relative risk of 1.35 and 1.47, examining 23 and 30 studies, respectively.157'158
That is, these two studies show an overall increase in lung cancer for the exposed groups of 35
percent and 47 percent compared with the groups not exposed to diesel exhaust.  In the EPA
Diesel HAD, EPA selected 1.4  as a reasonable estimate of occupational relative risk for further
analysis.

   EPA generally derives cancer unit risk estimates to calculate population risk more precisely
from exposure to carcinogens. In the simplest terms, the  cancer unit risk is the increased risk
associated with average lifetime exposure of 1 i-ig/m3. EPA concluded in the Diesel HAD that it
is not possible currently to calculate a cancer unit risk for diesel exhaust due to a variety of
factors that limit the current studies, such as a lack of standard exposure metric for diesel exhaust

                                          2-55

-------
Final Regulatory Impact Analysis
and the absence of quantitative exposure characterization in retrospective studies.

   However, in the absence of a cancer unit risk, the EPA Diesel HAD sought to provide
additional insight into the possible ranges of risk that might be present in the population. Such
insights, while not confident or definitive, nevertheless contribute to an understanding of the
possible public health significance of the lung cancer hazard. The possible risk range analysis
was developed by comparing a typical environmental exposure level to a selected range of
occupational exposure levels and then proportionally scaling the occupationally observed risks
according to the exposure ratio's to obtain an estimate of the possible  environmental risk. If the
occupational and environmental exposures are similar, the environmental risk would approach
the risk seen in the occupational studies whereas a much higher occupational exposure indicates
that the environmental risk is lower than the occupational risk.  A comparison of environmental
and occupational exposures showed that for certain occupations the exposures are similar to
environmental exposures while, for others, they differ by a factor of about 200 or more.

   The first step in this process is to note that the occupational relative risk of 1.4, or a 40
percent from increased risk compared with the typical 5 percent lung cancer risk in the U.S.
population, translates to  an increased risk of 2 percent (or 10"2) for these diesel exhaust exposed
workers.  The Diesel HAD derived  a typical nationwide average environmental exposure level of
0.8 |ag./m3 for diesel PM from highway sources  for 1996. This estimate was based on national
exposure modeling;  the derivation of this exposure is discussed in detail in the EPA Diesel HAD.
Diesel PM is a surrogate for diesel exhaust and, as mentioned above, has been classified as a
carcinogen by some agencies.

   The possible environmental risk range was estimated by taking the relative risks in the
occupational setting, EPA selected  1.4 and converting this to absolute risk of 2% and then
ratioing this risk by differences in the occupational vs environmental exposures of interest. A
number of calculations are needed to accomplish this, and these can be seen in the EPA Diesel
HAD.  The outcome was that environmental risks from diesel exhaust using higher estimates  of
occupational exposure could range from a low of 10"4 to  10"5 or be as  high as 10"3 if lower
estimates of occupational exposure  were used.  Note that the environmental exposure of interest
(0.8 |ag/m3) remains constant in this analysis, while the occupational exposure is a variable. The
range of possible environmental risk is a reflection of the range of occupational exposures that
could be associated with the relative and related absolute risk levels observed in the occupational
studies.

   While these risk estimates are exploratory and not intended to provide  a definitive
characterization of cancer risk, they are useful in gauging the possible range of risk based on
reasonable judgment. It is important to note that the possible risks could also be higher or lower
and a zero risk cannot be ruled out.  Some individuals in the population may have a high
tolerance to exposure from diesel exhaust and low cancer susceptibility.  Also, one cannot rule
out the possibility of a threshold of exposure below which there is no cancer risk, although
evidence has not been seen or substantiated on this point.

   Also, as discussed in the Diesel HAD, there is a relatively small difference between some

                                          2-56

-------
                                               Air Quality, Health, and Welfare Effects
occupational settings where increased lung cancer risk is reported and ambient environmental
exposures. The potential for small exposure differences underscores the concerns about the
potential public hazard, since small differences suggest that environmental risk levels may be
close to those observed in the occupational setting.

   EPA also assessed air toxic emissions and their associated risk (the National-Scale Air
Toxics Assessment or NAT A for 1996), and we concluded that diesel exhaust ranks with other
substances that the national-scale assessment suggests pose the greatest relative risk.159 This
national assessment estimates average population inhalation exposures to diesel PM in  1996  for
nonroad as well as highway sources.  These are the sum of ambient levels in various locations
weighted by the amount of time people spend in each of the locations.  This analysis shows a
somewhat higher diesel exposure level than the 0.8 |ig/m3 used to develop the risk perspective in
the Diesel HAD. The average nationwide NATA mobile exposure levels are 1.44 |ig/m3 total
with highway source contribution of 0.46 |ig/m3 and a nonroad source contribution of 0.98
l-ig/m3. The  average urban exposure was 1.64 |ig/m3 and the average rural exposure was 0.55
l-ig/m3. In five percent  of urban census tracts across the United States, average exposures were
above 4.33 |ig/m3.  The EPA Diesel HAD states that use of the NATA exposure estimates
instead of the 0.8 |ig/m3 estimate results in a similar risk perspective.0

   In summary, even though EPA does not have a specific carcinogenic potency with which to
accurately estimate the carcinogenic impact of diesel exhaust, the likely hazard to humans
together with the potential for significant environmental risks leads us to conclude that diesel
exhaust emissions need to be reduced from nonroad engines in order to protect public health.
The following factors lead to our determination.

   1. EPA has officially designated  diesel exhaust as a likely human carcinogen due to
       inhalation at environmental exposure. Other organizations have made similar
       determinations.
   2. The entire U.S. population is exposed to various levels of diesel exhaust. The higher
       exposures at environmental levels is comparable to some occupational exposure levels,
       so that environmental risk could be the same as, or approach, the risk magnitudes
       observed in the occupational epidemiologic studies.
   3. The possible range of risk for the general U.S. population due to exposure to diesel
       exhaust is 10"3 to  10"5 although the risk could be lower and a zero risk cannot be ruled
       out.

   Thus, the concern for a carcinogenicity hazard resulting from diesel exhaust exposures is
longstanding based on  studies done over many years. This hazard may be widespread due to the
   GIt should be note that, as with any modeling assessment, there are a number of significant limitations and
uncertainties in NATA. These uncertainties and limitations include use of default values to model local conditions,
limitations in emissions data, uncertainties in locating emissions spatially and temporally, and accounting for
atmospheric processes.  NATA limitations and uncertainties are discussed at the following website:
http ://www. epa. gov/ttn/atw/nata/natsalim2. html
                                           2-57

-------
Final Regulatory Impact Analysis
ubiquitous nature of exposure to diesel exhaust.

   2.2.1.2 Other Health Effects of Diesel Exhaust

   The acute and chronic exposure-related effects of diesel exhaust emissions are also of
concern to the Agency. The Diesel HAD established an inhalation Reference Concentration
(RfC) specifically based on animal studies of diesel  exhaust. An RfC is defined by EPA as "an
estimate of a continuous inhalation exposure to the human population, including sensitive
subgroups, with uncertainty spanning perhaps an order of magnitude, that is likely to be without
appreciable risks of deleterious noncancer effects during a lifetime." EPA derived the RfC from
consideration of four well-conducted chronic rat inhalation studies showing adverse pulmonary
effects.160'161-162'163 The diesel RfC is based on a "no observable adverse effect" level of 144
l-ig/m3 that is further reduced by applying uncertainty factors of 3 for interspecies extrapolation
and 10 for human variations in sensitivity.  The resulting RfC derived in the Diesel HAD is  5
l-ig/m3 for diesel exhaust as measured by diesel PM.  This RfC does not consider allergenic
effects such  as those associated with asthma or immunologic effects.  There is growing evidence
that diesel exhaust can exacerbate these effects, but the exposure-response data are presently
lacking to derive an RfC.

   While there have been relatively few human controlled exposure studies associated
specifically with the noncancer impact of diesel PM alone, diesel PM is frequently  part of the
ambient particles studied in numerous epidemiologic studies. Conclusions  that health effects
associated with ambient PM in general are relevant to diesel PM are  supported by studies that
specifically associate observable human noncancer health effects with exposure to diesel PM.
As described in the Diesel HAD, these studies include some of the same health effects reported
for ambient PM, such as respiratory symptoms (cough, labored breathing, chest tightness,
wheezing), and chronic respiratory disease (cough, phlegm, chronic bronchitis and  suggestive
evidence for decreases in pulmonary function).  Symptoms of immunological effects such as
wheezing and increased allergenicity are also seen. Studies in rodents, especially rats, show the
potential for human inflammatory effects in the lung and consequential lung tissue  damage from
chronic diesel exhaust inhalation exposure. The Diesel HAD notes that acute or short-term
exposure to diesel exhaust can cause acute irritation (e.g., eye, throat, bronchial),
neurophysiological symptoms (e.g., lightheadedness, nausea), and respiratory symptoms (cough,
phlegm). There is also evidence for an immunologic effect such as the exacerbation of allergenic
responses to known allergens and asthma-like symptoms.164'165'166'167  The Diesel HAD lists
numerous other studies as well. Also, as discussed in more detail previously, in addition to  its
contribution to ambient PM inventories, diesel PM is of special concern because it has been
associated with an increased risk of lung cancer.

   The Diesel HAD also briefly summarizes health effects associated with ambient PM  and the
EPA's annual NAAQS of 15 |ig/m3. There is a much more extensive body  of human data
showing a wide spectrum of adverse health effects associated with exposure to ambient PM, of
which diesel exhaust is an important component. The RfC is not meant to say that 5 jig/m3
provides adequate public health protection for ambient PM2 5.  In fact, there may be benefits to
                                          2-58

-------
                                                  Air Quality, Health, and Welfare Effects
reducing diesel PM below 5 |ig/m3 since diesel PM is a major contributor to ambient PM25 .H

    Also, as mentioned earlier in the health effects discussion for PM2 5, there are a number of
other health effects associated with PM in general—and motor vehicle exhaust, including that
from diesel engines in particular—that provide additional evidence for the need for significant
emission reductions from nonroad diesel sources.

    As indicated earlier, a number of recent studies have associated living near roadways with
adverse health effects.  Two of the studies cited earlier will be mentioned again here  as examples
of the type of work that has been done.  A Dutch study (discussed earlier by G. Hoek et al.,
2002) of a population of people  55-69 years old found that there was an elevated risk of heart
and lung related mortality among populations living near high traffic roads. A review discussed
earlier of studies (by R. Delfino et al., 2002) of the respiratory health of people living near
roadways included a publication indicating that the risk of asthma and related respiratory
disease appeared elevated in people living near heavy traffic.168  These studies offer evidence
that people exposed most directly to emissions from mobile sources, including those  from diesel
engines, face an elevated risk of illness or death.

    All of these health effects plus the designation of diesel exhaust as a likely human carcinogen
provide ample health justification  for control.

    Public comments from the Building and Construction Trades Department, AFL-CIO, and
International Union of Operating Engineers supported the need to adopt the nonroad  rule noting
that exposure to diesel emissions from nonroad diesel engine poses a great risk to workers in the
construction industry and other occupations, but are highest among construction workers because
they work in close proximity to the exposure source, and are exposed daily to the hazards of
nonroad diesel pollution. In their comments, BCTD noted that construction workers  may be
exposed to hazards generated from work performed by other trades employed by other
contractors because sources of diesel exposure are scattered throughout the site.  They noted
further that in an exposure study, railway workers, heavy equipment operators and miners had
    HIt should again be noted that recent epidemiologic studies of ambient PM2 s do not indicate a threshold of
effects at low concentrations.  For example, the authors of the Pope reanalysis note that, for the range of exposures
considered in their reanalysis, the slope of the concentration-response function appears to be monotonic and nearly
linear, although they cannot exclude the potential for a leveling off or steepening at higher exposure levels. The
EPA Science Advisory Board's Advisory Council for Clean Air Compliance, which provides advice and review of
EPA's methods for assessing the benefits and costs of the Clean Air Act under Section 812 of the Act, has advised
that there is currently no scientific basis for assuming any specific threshold for the PM-related health effects
considered in typical benefits analyses (EPA-SAB-Council-ADV-99-012, 1999). Also, the National Research
Council, in its own review of EPA's approach to benefits analyses, has agreed with this advice. This advice is
supported by the recent literature on health effects of PM exposure (Daniels et al., 2000; Pope, 2000; Pope et al.,
2002, Rossi et al., 1999; Schwartz, 2000, Schwartz, Laden, and Zanobetti 2002  [Schwarz, I; Laden, F.; and
Zanobetti, A. (2002) The Concentration-Response Relation between PM2.5 and Daily Deaths. Environ Health
Perspect 110(10):  1025-1029]) which generally finds no evidence of a non-linear concentration-response
relationship and, in particular, no evidence of a distinct threshold for health effects. The most recent draft of the
EPA Air Quality Criteria for Paniculate Matter (U.S. EPA, 2002)  reports only one study, analyzing data from
Phoenix, AZ, that reported even limited evidence suggestive of a possible threshold for PM2 5 (Smith et al., 2000).

                                             2-59

-------
Final Regulatory Impact Analysis
higher mortality rates from lung cancer and all causes than workers without diesel exposure.
Heavy equipment operators and miners had comparable relative risk for lung cancer, both of
which were over 2.5 times that of non-exposed workers (Boffetta, 1988).

   2.2.1.3 Diesel Exhaust PM Ambient Levels

   Because diesel PM is part of overall ambient PM and cannot be easily distinguished from
overall PM, we do not have direct measurements of diesel PM in the ambient air.  Diesel PM
concentrations are estimated instead using one of three approaches: 1) ambient air quality
modeling based on diesel PM emission inventories; 2) using elemental carbon concentrations in
monitored data as surrogates; or 3) using the chemical mass balance (CMB) model in
conjunction with ambient PM measurements. (Also, in addition to CMB, UNMIX/PMF have
also been used).  Estimates using these three approaches are  described below.  In addition,
estimates developed using the first two approaches above are subjected to a statistical
comparison to evaluate overall reasonableness of estimated concentrations from ambient air
quality modeling. It is important to note that, while there are inconsistencies in some of these
studies on the relative importance of gasoline and diesel PM, the studies discussed in the Diesel
HAD all show that diesel PM is a significant contributor to overall ambient PM.  Some of the
studies differentiate nonroad from highway diesel PM.

   2.2.1.3.1 Toxics Modeling and Methods

   In addition to the general ambient PM modeling conducted for this rulemaking, diesel PM
concentrations for 1996 were estimated as part of the National-Scale Air Toxics Assessment
(NATA; EPA, 2002). In this assessment, the PM inventory developed for the recent regulation
promulgating 2007 heavy duty vehicle standards was used (EPA, 2000).  Note that the nonroad
inventory used in this modeling was based on an older version of the draft NONROAD Model
that showed higher diesel PM than the current version, so the ambient concentrations may be
biased high. Ambient impacts of mobile source emissions were predicted using the Assessment
System for Population Exposure Nationwide (ASPEN) dispersion model.

   From the NATA 1996 modeling, overall mean annual national ambient diesel PM levels of
2.06 |ag/m3 were calculated with a mean of 2.41  in urban counties and 0.74 in rural counties.
Table 2.2.1-1 below summarizes the distribution of average ambient concentrations to  diesel
PM at the national scale. Over half of the diesel PM can be attributed to nonroad diesel engines.
A map of county median concentrations (median of census tract concentrations) from highway
and nonroad sources is provided in Figure 2.2.1-1. We have not generated a map depicting the
estimated geographic distribution of nonroad diesel PM alone. While the high median
concentrations are clustered in the Northeast, Great Lake States and California, areas of high
median concentrations are distributed throughout the United States.
                                          2-60

-------
                              Air Quality, Health, and Welfare Effects
                      Table 2.2.1-1
     Distribution of Average Ambient Concentrations of
Diesel PM at the National Scale in the 1996 NAT A Assessment.

5th Percentile
25th Percentile
Average
75th Percentile
95th Percentile
Onroad Contribution
to Average
Nonroad Contribution
to Average
Nationwide (i-ig/m3)
0.33
0.85
2.06
2.45
5.37
0.63
1.43
Urban (i-ig/m3)
0.51
1.17
2.41
2.7
6.06
0.72
1.69
Rural (i-ig/m3)
0.15
0.42
0.74
0.97
1.56
0.27
0.47
                          2-61

-------
                                            Figure 2.2.1-1
                       Estimated County Median Concentrations of Diesel Particulate Matter
             1996  Estimated  County Median Ambient Concentrations
                Diesel  Particulate Matter  —  United  States Counties
Distribution of U.S. Ambient Concentrations
     HlghsetlnU.fi. ^^^_  15
             95 |     |  1.90
             90
 Percentile   75
             50
             £5
      Lowset In U.S.
                                    County Median Ambient Pollutant Concentration
                                    ( micrograms / cubic meter )
Q.Q14-
                                                                          Source: U.S. EPA / QijQPS
                                                             NATA Nalfonal— Scale Afr Toxfcs Assessment
Source: EPA National-Scale Air Toxics Assessment for 1996. Results should not be used to draw conclusions about local
concentrations. Results are most meaningful at the Regional or National level.

-------
    Air Quality, Health, and Welfare Effects
2-63

-------
Final Regulatory Impact Analysis
   Diesel PM concentrations were also recently modeled across a representative urban area,
Houston, Texas, for 1996,  using the Industrial Source Complex Short Term (ISCST3) model.169
The methodology used to model diesel PM concentrations is the same as the methodology used
for benzene and other hazardous air pollutants, as described in a recent EPA technical report.170
For Harris County, which has the highest traffic density in Houston area, link-based diesel PM
emissions were estimated for highway mobile sources, using diesel PM emission rates developed
for the recent EPA 2007 heavy duty engine and highway diesel fuel sulfur control rule.171  This
link-based modeling approach is designed to specifically account for local traffic patterns within
the urban center, including diesel truck traffic along specific roadways. For other counties in the
Houston metropolitan area, county level emission estimates from  highway vehicles were
allocated to one kilometer grid cells based on total roadway miles. Nonroad diesel  emissions for
Houston area counties were obtained from the inventory done for the 2007 heavy duty rule, and
allocated to one kilometer grid cells using activity surrogates. The modeling in Houston suggests
strong spatial gradients (on the order of a factor of 2-3 across a modeling domain) for diesel PM
and indicates that "hotspot" concentrations can be very high.  Values as high as 8 |ig/m3 at were
estimated at a receptor  versus a 3 |ig/m3 average in Houston. Such "hot spot" concentrations
suggest both a high localized exposure plus higher estimated average annual exposure levels for
urban centers than what has been estimated in assessments such as NATA 1996, which are
designed to focus on regional and national scale averages.  Figure 2.2.1-2 depicts the spatial
distribution of diesel PM concentrations in Houston.
                                          2-64

-------
                                              Air Quality, Health, and Welfare Effects
                                Figure 2.2.1-2
  Annual Average Ambient Concentrations of Diesel PM in Houston, 1996, based on
 Dispersion Modeling Using Industrial Source Complex Short Term (ISCST3) model.

                  Annual Average Concentrations Jigrtn3]. ISC Roads. Houston, TX. 1996
                             DieeelPm. All Sources- no Background
    3260 -
            220
                        240          260         280
                             UTM Zone 15 West-East Distance (km)
                                                                       320
   2.2.1.3.2 Elemental Carbon Measurements
   As shown in Figures 2.1.1-1 to 3, the carbonaceous component is significant in ambient PM.
The carbonaceous component consists of organic carbon and elemental carbon. Monitoring data
on elemental carbon concentrations can be used as a surrogate to determine ambient diesel PM
concentrations. Elemental carbon is a major component of diesel exhaust, contributing to

                                          2-65

-------
Final Regulatory Impact Analysis
approximately 60-80 percent of diesel particulate mass, depending on engine technology, fuel
type, duty cycle, lube oil consumption, and state of engine maintenance.  In most areas, diesel
engine emissions are major contributors to elemental carbon, with other potential sources
including gasoline exhaust, combustion of coal, oil, or wood, charbroiling, cigarette smoke, and
road dust.  Because of the large portion of elemental carbon in diesel paniculate matter, and the
fact that diesel exhaust is one of the major contributors to elemental carbon in most areas,
ambient diesel PM concentrations can be bounded using elemental carbon measurements.

   The measured mass of elemental carbon at a given site varies depending on the measurement
technique used. Moreover, to estimate diesel PM concentration based on elemental carbon level,
one must first estimate the percentage of PM attributable to diesel engines and the percentage of
elemental carbon in diesel PM.  Thus, there are significant uncertainties in estimating diesel PM
concentrations using an elemental carbon surrogate. Also, there are issues with the measurement
methods used for elemental carbon. Many studies used thermal optimal transmission (TOT), the
NIOSH method developed at Sunset laboratories. Other  studies used thermal optical reflectance
(TOR), a method developed by Desert Research Institute. EPA has developed multiplicative
conversion factors to estimate diesel PM concentrations  based on elemental carbon levels.172
Results from several source apportionment studies were used to develop these factors.173'174'175'
176,177, ns, 179 Average conversion factors were compiled together with lower and upper bound
values. Conversion factors (CFs) were calculated by dividing the diesel PM2 5  concentration
reported in these studies by the total organic carbon or elemental carbon concentrations also
reported in the studies. Table 2.2.1-2 presents the minimum, maximum, and  average EC
conversion factors as a function of:

   •   Measurement technique
   •   Eastern or Western United States
       Season
   •   Urban or rural

The reported minimum, maximum, and average values in Table 2.2.1-2 are the minima, maxima,
and arithmetic means of the EC conversion factors across all sites (and seasons, where
applicable) in the given site subset. For the TOT data collected in the East, the minimum,
maximum, and average conversion factors are all equal.  This is because these values were based
only on one study where the data were averaged over sites, by season.180  Depending on the
measurement technique used, and assumptions made in converting elemental carbon
concentration to diesel PM concentration, average nationwide concentrations for current years of
diesel PM estimated from elemental carbon data range from about 1.2 to  2.2 |ig/m3. EPA has
compared these estimates based on elemental carbon measurements with modeled concentrations
in the NATA for 1996.  Results of comparisons of mean percentage differences are presented in
Table 2.2.1-3.  These results show that the two sets of data agree reasonably well, with estimates
for the majority of sites within a factor of 2, regardless of the measurement technique or
methodology for converting elemental carbon to diesel PM concentration. Agreement was better
when modeled concentrations were adjusted to reflect recent changes in the nonroad inventory.
The best model performance based on the fraction of modeled values within 100 % of the
monitored value is for the DPM-maximum value, which reflects changes to the nonroad

                                         2-66

-------
                                              Air Quality, Health, and Welfare Effects
inventory model. The corresponding fractions of modeled values within 100 % of the monitored
value are 73 % for TOR sites, 80 % for TOT sites, and 92 % for TORX sites. All in all, this
performance compares favorably with the model to monitor results for other pollutants assessed
in NAT A, with the exception of benzene, for which the performance of the NATA modeling was
better.

   2.2.1.3.3 Chemical Mass Balance Receptor Modeling and Source Apportionment

   The third approach for estimating ambient diesel PM concentrations uses the chemical mass
balance (CMB) model for source apportionment in conjunction with ambient PM measurements
and chemical source "fingerprints" to estimate ambient diesel PM concentrations. The CMB
model uses a statistical fitting technique to determine how much mass from each source would
be required to reproduce the chemical  fingerprint of each speciated ambient monitor. Inputs to
the CMB model applied to ambient PM2 5 include measurements made at an air monitoring site
and measurements made of each of the source types suspected to affect the site. The CMB
model uses a statistical fitting technique ("effective variance weighted least squares") to
determine how much mass from each source would be required to reproduce the chemical
fingerprint of each speciated ambient monitor. This calculation is based on optimizing the sum
of sources, so that the difference between the ambient monitor and the sum of sources is
minimized. The optimization technique employs "fitting species" that are related to the sources.
The model assumes that source profiles are constant over time, that the sources do not interact or
react in the atmosphere, that uncertainties in the source fingerprints are well-represented, and
that all sources are represented in the model.

   This source apportionment technique presently does not distinguish between highway and
nonroad but, instead, gives diesel PM as a whole.  One can allocate the diesel PM numbers based
on the inventory split between highway and nonroad diesel, although this allocation was not
done in the studies published to date. This source apportionment technique can though
distinguish between  diesel and gasoline PM. Caution in interpreting CMB results is warranted,
as the use of fitting species that are not specific to the sources modeled can lead to misestimation
of source contributions. Ambient concentrations using this approach are generally about 1 i-ig/m3
annual  average. UNMIX/PMF  models show similar results.
                                          2-67

-------
Final  Regulatory Impact Analysis
                                         Table 2.2.1-2
              Summary of Calculated Elemental Carbon (EC) Conversion Factors
              (Conversion factors to convert total EC to diesel PM2 5 concentration)
Ambient
Measurement
Technique: TOT
or TOR
TOT
East or
West
East
East
East
East
West
Season
Fall (Q4)
Spring (Q2)
Summer
(Q3)
Winter (Ql)
Unknown
Location
Type
General
Mixed
Mixed
Mixed
Mixed
Urban
TOT Total
TOR


Winter
Winter
Rural
Urban
Winter Total
TOR Total
Grand Total
MTN3
2.3
2.4
2.1
2.2
1.2
1.2
0.6
0.5
0.5
0.5
0.5
MAY3
2.3
2.4
2.1
2.2
2.4
2.4
1.0
1.0
1.0
1.0
2.4
AVERAGE3
2.3
2.4
2.1
2.2
1.6
2.0
0.8
0.7
0.8
0.8
1.3
Recommended
Conversion Factors
EAST
X
X
X
X


X
X



WEST




X

X
X



Source: ICF Consulting for EPA, 2002, Office of Transportation and Air Quality. Report No. EPA420-D-02-004.
3 Minimum, maximum, or average value across all sites of the estimated conversion factors.

TOT = thermal optimal transmission, the NIOSH method developed at Sunset laboratories.
TOR = thermal optical reflectance, a method developed by Desert Research Institute.
                                             2-68

-------
                                                 Air Quality, Health, and Welfare Effects
                                         Table 2.2.1-3
             Summary of Differences Between the Nearest Modeled Concentration
      of Diesel Pm from the National Scale Air Toxics Assessment and Monitored Values
   Based on Elemental Carbon Measurements (Diesel PM model-to-measurement comparison)
Modeled
Variable3
concnear
concnear2
concnear
concnear2
concnear
concnear2
concnear
concnear2
concnear
concnear2
concnear
concnear2
concnear
concnear2
concnear
concnear2
concnear
concnear2
Monitored
Variable*
TOR
TOR
TORH
TORH
TORL
TORL
TOT
TOT
TOTH
TOTH
TOTL
TOTL
TORX
TORX
TORXH
TORXH
TORXL
TORXL
N
15
15
15
15
15
15
95
95
95
95
95
95
88
88
88
88
88
88
Mean
Modeled
Value
1.56
1.20
1.56
1.20
1.56
1.20
2.61
2.05
2.61
2.05
2.61
2.05
2.31
1.81
2.31
1.81
2.31
1.81
Mean
Monitored
Value
0.94
0.94
1.16
1.16
0.64
0.64
1.73
1.73
2.10
2.10
1.52
1.52
1.70
1.70
2.23
2.23
1.19
1.19
Mean
Difference
0.63
0.26
0.40
0.04
0.92
0.55
0.88
0.32
0.52
-0.05
1.09
0.52
0.61
0.11
0.08
-0.42
1.12
0.62
Mean
%
Diffprpnr.p
100
56
62
26
190
126
80
42
61
27
101
58
47
15
13
-12
110
65
Fraction ol Modeled Values
Within
10%
0.07
0.07
0.00
0.00
0.13
0.07
0.12
0.11
0.11
0.11
0.09
0.09
0.10
0.17
0.11
0.08
0.10
0.14
75%
0.13
0.13
0.07
0.07
0.40
0.33
0.21
0.37
0.22
0.35
0.17
0.32
0.30
0.30
0.26
0.22
0.26
0.31
50%
0.53
0.47
0.40
0.33
0.47
0.47
0.45
0.53
0.46
0.53
0.43
0.52
0.59
0.59
0.60
0.52
0.41
0.52
100%
0.53
0.60
0.60
0.73
0.53
0.53
0.68
0.77
0.74
0.80
0.63
0.72
0.78
0.85
0.84
0.92
0.65
0.74
Source: ICF Consulting for EPA, 2002, Office of Transportation and Air Quality. Report No. EPA420-D-02-004.

a Modeled variable:
    concnear   Nearest modeled DPM concentration from the 1996 NAT A
    concnear2   Nearest modeled DPM concentration with NATA concentrations adjusted to be consistent with
              changes to the nonroad inventory model
b Monitored variable:
    TOR   EC value multiplied by TOR average correction factor
    TORH  EC value multiplied by TOR maximum correction factor
    TORL   EC value multiplied by TOR minimum correction factor
    TOT    EC value multiplied by TOT average correction factor
    TOTH  EC value multiplied by TOT maximum correction factor
    TOTL   EC value multiplied by TOR minimum correction factor
    TORX  TOR values plus the TOR equivalent values multiplied by TOR average correction factor
    TORXH TOR values plus the TOR equivalent values multiplied by TOR maximum correction factor
    TORXL TOR values plus the TOR equivalent values multiplied by TOR minimum correction factor
    Because of the correlation of diesel and gasoline exhaust PM emissions in time and space,
chemical molecular species that provide markers for separation of these sources have been
sought. Recent advances in chemical analytical techniques have facilitated the development of
sophisticated molecular source profiles, including detailed speciation of organic compounds,
which allow the apportionment of particulate matter to gasoline  and diesel sources with
increased certainty. As mentioned previously, however, caution in  interpreting CMB results is
warranted.  Markers that have been used in CMB receptor modeling have included elemental
carbon, polycyclic aromatic hydrocarbons (PAHs), organic acids, hopanes, and steranes.
                                             2-69

-------
Final Regulatory Impact Analysis
   It should be noted that since receptor modeling is based on the application of source profiles
to ambient measurements, this estimate of diesel PM concentrations includes the contribution
from on-highway and nonroad sources of diesel PM, although no study to date has included
source profiles from nonroad engines. Engine operations, fuel properties, regulations, and other
factors may distinguish nonroad diesel engines from their highway counterparts.

   In addition, this model accounts for primary emissions of diesel PM only; the contribution of
secondary aerosols is not included.  The role of secondarily formed organic PM in urban PM2 5
concentrations is not known, particularly from diesel engines.

   The first major application of organic tracer species in applying the CMB model evaluated
ambient PM20 in Los Angeles, CA sampled in 1982.181 This study was the first to distinguish
gasoline and diesel exhaust. CMB model application at four sites in the Los Angeles area
estimated ambient diesel PM20  concentrations to be 1.02-2.72 |ig/m3. Note that diesel PM
estimates are derived from source profiles measured on in-use diesel trucks.

   Another major study examining diesel exhaust separately from  gasoline exhaust and other
sources is the Northern Front Range Air Quality Study (NFRAQS).182 This study was conducted
in the metropolitan Denver,  CO area during 1996-1997. The NFRAQS study employed a
different set of chemical species, including PAHs and other organics to produce  source profiles
for a diverse range of mobile sources, including "normal emitting"  gasoline vehicles, cold start
gasoline vehicles, high emitting gasoline vehicles, and diesel  vehicles.  Average source
contributions from diesel engines in NFRAQS were  estimated to be 1.7 |ig/m3 in an urban area,
and 1.2 |ig/m3 in a rural area. Source profiles in this study were based on highway vehicles.

   The CMB model was applied in California's San Joaquin Valley during winter 1995-1996.183
The study employed similar source tracers as the earlier study of Los Angeles PM2.0, in addition
to other more specific markers.  Diesel PM source contribution estimates in Bakersfield, CA
were 3.92 and 5.32 during different measurement periods. Corresponding  estimates in Fresno,
CA were 9.68 and 5.15 |ig/m3.  In the Kern Wildlife  Refuge, diesel PM source contribution
estimates were 1.32  and 1.75 |ig/m3 during the two periods.

   The CMB model was applied in the Southeastern United States on data collected during the
Southeastern Aerosol Research and Characterization (SEARCH) study (Zheng et al., 2002).
Modeling was conducted on data collected during April, July, and October 1999 and January
2000. Examining ambient monitors in urban, suburban, and rural areas, the modeled annual
average contribution of primary diesel emissions to ambient PM25 was 3.20-7.30 |ig/m3 in
N. Birmingham, AL, 1.02-2.43  |ig/m3 in Gulfport, MS, 3.29-5.56 |ig/m3 in Atlanta, GA, and
1.91-3.07 |ag/m3 in Pensacola, FL, which together represented the urban sites in the study.
Suburban sites in the study were located outside Pensacola, FL (1.08-1.73  |ig/m3).  Rural sites
were located in Centreville,  AL (0.79-1.67 |ig/m3), Oak Grove, MS (1.05-1.59 |ig/m3), and
Yorkville, GA (1.07-2.02 ng/m3).

   The CMB model was applied to ambient PM2 5 data collected during a severe photochemical
smog event during 1993 in Los Angeles using organic tracers.184 Modeled concentrations of

                                          2-70

-------
                                              Air Quality, Health, and Welfare Effects
diesel contributions to PM2 5 during this episode were conducted for Long Beach (8.33 |ig/m3),
downtown Los Angeles (17.9 i-ig/m3), Azusa (14.9 i-ig/m3), and Claremont, CA (7.63 |ig/m3).

   While these studies provide an indication that diesel exhaust is a substantial contributor to
ambient PM2 5 mass, they should still be viewed with caution.  CMB modeling depends on
ensuring the use of highly specific tracer species. If sources, such as nonroad diesel engines, are
chemically different from other sources, including highway  diesel trucks, the CMB model can
misestimate  source contributions. Nevertheless, these studies provide information that is
complementary to source-oriented air quality modeling (discussed above). From these studies, it
is apparent that diesel exhaust is a substantial contributor to ambient PM2 5, even in remote and
rural areas.

   2.2.1.4 Diesel Exhaust PM Exposures

   Exposure of people to diesel exhaust depends on their various activities, the time spent in
those activities, the locations where these activities occur, and the levels of diesel exhaust
pollutants (such as PM) in those locations. While ambient levels are specific for a particular
location, exposure levels account for such factors as a person moving from location to location,
proximity to the emission source, and whether the exposure occurs in an enclosed environment.

   2.2.1.4.1 Occupational Exposures

   Diesel particulate exposures have been measured for a number of occupational groups over
various years but generally for more recent years (1980s and later) rather than earlier years.
Occupational exposures had a wide range varying from 2 to 1,280 |ig/m3 for a variety of
occupational groups including miners, railroad workers, firefighters,  air port crew, public transit
workers, truck mechanics, utility linemen, utility winch truck operators, fork lift operators,
construction workers, truck dock workers, short-haul truck drivers, and long-haul truck drivers.
These individual studies are discussed in the Diesel HAD.

   The highest exposure to diesel PM is for workers in coal mines and noncoal mines, which are
as high a 1,280 i-ig/m3, as discussed in the Diesel HAD. The National Institute of Occupational
Safety and Health (NIOSH) has estimated a total of 1,400,000 workers are occupationally
exposed to diesel exhaust from on-road and nonroad equipment.

   Many measured or estimated occupational exposures are for on-road diesel engines and some
are for school buses.185'186> 187'188  Also,  some (especially the higher ones) are for occupational
groups (fork lift operator, construction workers, or mine workers) who would be exposed to
nonroad diesel exhaust.  Sometimes, as is the case for the nonroad engines, there are only
estimates of exposure based on the length of employment or similar factors rather than a i-ig/m3
level. Estimates for exposures to diesel PM for diesel fork lift operators have been made that
range from 7 to 403 |ig/m3 as reported in the Diesel HAD. In addition, the Northeast States for
Coordinated Air Use Management (NESCAUM) measured occupational exposures to  particulate
and elemental carbon near the operation of various diesel non-road equipment. Exposure groups
include agricultural farm operators, grounds  maintenance personnel (lawn and garden

                                          2-71

-------
Final Regulatory Impact Analysis
equipment), heavy equipment operators conducting multiple job tasks at a construction site, and
a saw mill crew at a lumber yard.  Samples will be obtained in the breathing zone of workers. In
a recently released interim report on occupational health risks from diesel engine exposure,
pollution inside the cabs of heavy diesel equipment were shown to be up to 16 times higher than
federal health recommendations. The diesel PM was estimated to exist at levels that pose risk of
chronic inflammation and lung damage in exposed individuals (NESCAUM, 2003).

   In public comments from the Building and Trade Department, AFL-CIO, they note their
research center, the Center to Protect Workers' Rights,  has sponsored research conducted by
the Construction Occupational Health Program (COHP) at University of Massachusetts at
Lowell which documents diesel emissions exposure among a number of trades employed on  a
major highway project underway in Boston, MA. Over 260 personal samples of diesel exposure
were collected among laborers (116); operating engineers (113) and other trades including
ironworkers (15), carpenters (9), piledrivers (5), boilermakers (1), plumbers (1) and surveyors
(1). Exposures associated with specific work processes were also documented.  Using the
American Conference of Governmental Industrial Hygienists Threshold Limit Value (TLV) for
diesel exhaust as elemental carbon of 20 ug/m3 as proposed in 2002, the percentage of samples
exceeding the TLV overall was 14 percent (Woskie, 2002; ACGIH, 2002). It should be noted
that much of this project involves  construction of underground tunnels. However, work in
enclosed and/or poorly ventilated work areas is common in construction.

   One recent study found that construction workers in Ontario are exposed to elevated
concentrations of elemental carbon (EC) measured by thermal-optical transmission (TOT),
which the authors used as a surrogate for diesel exhaust.189 Task-based exposure measurements
were made  corresponding to engine use.  Demolition laborers were exposed to between 4.9
to 146 ug/m3 of EC-TOT while operating compressors, performing excavation and cleanup, and
in tearing down structures.  Construction equipment operating engineers were exposed to 4.3 to
7.8 ug/m3 EC-TOT while operating their machinery. Painters in new commercial construction
were exposed to between 3.6 to 9.0 ug/m3 EC-TOT, as a result of operating mixers. While these
concentrations are substantially higher than those seen in typical urban air, it is difficult to assign
these EC-TOT measurements to diesel engines, and the study authors did not indicate  the fuel
source of the equipment used.  However, it is likely that many of the engines in this study were
diesel engines.

   2.2.1.4.2 Ambient Exposures in the General Population

   Currently, personal exposure monitors for PM cannot differentiate diesel from other PM.
Thus, we use modeling to estimate exposures.  Specifically, exposures for the general  population
are estimated by first conducting dispersion modeling of both highway and nonroad diesel
emissions, described above, and then by conducting exposure modeling. The most
comprehensive modeling for cumulative on-road and non-road exposures to diesel PM is the
NATA. This assessment calculates exposures of the national population as a whole to a variety
of air toxics, including diesel PM. As discussed previously, the ambient levels are calculated
using the ASPEN dispersion model. As discussed above, the preponderance of modeled diesel
PM concentrations are within a factor of 2 of diesel PM concentrations estimated from elemental

                                         2-72

-------
                                              Air Quality, Health, and Welfare Effects
carbon measurements.190 This comparison adds credence to the modeled ASPEN results and
associated exposure assessment.

   The modeled concentrations for calendar year 1996 are used as inputs into an exposure
model called the Hazardous Air Pollution Exposure Model (HAPEM4) to calculate exposure
levels.  Average exposures calculated nationwide are 1.44 |ig/m3 with levels of 1.64 |ig/m3 for
urban counties and 0.55 |ig/m3 for rural counties.  Again, nonroad diesel emissions account for
over half of the this exposure.  Table 2.2.1-4 summarizes the distribution of average exposure
concentrations to diesel PM at the national scale in the 1996 NATA assessment. Figure 2.2.1-3
presents a map of the distribution of median exposure concentrations for U.S. counties.

                                      Table 2.2.1-4
                    Distribution of Average Exposure Concentrations to
               Diesel PM at the National Scale in the 1996 NATA Assessment.

5th Percentile
25th Percentile
Average
75th Percentile
95th Percentile
Onroad Contribution to Average
Nonroad Contribution to Average
Nationwide (|_ig/m3)
0.16
0.58
1.44
1.73
3.68
0.46
0.98
Urban (|_ig/m3)
0.29
0.81
1.64
1.91
4.33
0.52
1.12
Rural (|J,g/m3)
0.07
0.29
0.55
0.67
1.08
0.21
0.34
                                         2-73

-------
                                                     Figure 2.2.1-3
                          Estimated County Median Exposure Concentrations of Diesel Particulate Matter
                        1996 Estimated  County Median Exposure Concentration
                           Diesel Particulate Matter —   United  States Counties
                  Saeram*
                      Distribution of U.S. hi halation Exposure Concentration
                                   Highest In U.S. _ - . 10.2
                                          95 I - 1 1J35
o~   „*•,!„
Percent! le
                                                        ^  County Median Exposure Concentration
                                                        5                / etjbie meter)

                                                                                   Source: U, S, EPA / QAQPS
                                                                           Ha If on al— Scale Afr Toxfcs Assessment
Source: EPA National-Scale Air Toxics Assessment for 1996. Results should not be used to draw conclusions about local exposure
concentrations.  Results are most meaningful at the Regional or National level.

-------
                                              Air Quality, Health, and Welfare Effects
As explained earlier, the fact that these levels are below the 5 |ig/m3 RfC (which is based on
limited animal studies on diesel PM) does not necessarily mean that there are no adverse health
implications from overall PM2 5 exposure The health studies for the PM2 5 NAAQS are far more
encompassing than the limited animal studies used to develop the RfC for diesel exhaust, and,
also, the NAAQS applies to PM2 5 regardless of its composition.

   2.2.1.4.3 Ambient Exposures to Diesel Exhaust PM in Microenvironments

   One common microenvironment for ambient exposures to diesel exhaust PM is beside
freeways.  Although freeway locations are associated mostly with highway rather than nonroad
diesel enignes, there are many similarities between highway and nonroad diesel emissions, as
discussed in the Diesel HAD.  Also, similar spatial gradients in concentrations would be
expected where nonroad equipment is used.  The California Air Resources Board (California
ARB) has measured elemental carbon near the Long Beach Freeway in 1993.191 Levels
measured ranged from 0.4 to 4.0 jig/m3  (with one value as high as 7.5 |ig/m3) above background
levels. Microenvironments associated with nonroad engines would include construction zones.
PM and elemental carbon samples are being collected by NESCAUM in the immediate area of
the nonroad engine operations (such as at the edge or fence line of the construction zone).
Besides PM and elemental  carbon levels, various toxics such as benzene, 1,3-butadiene,
formaldehyde, and acetaldehyde will be sampled. The results should be especially useful since
they focus on microenvironments affected by nonroad diesel engines.

   Also, EPA is funding research in Fresno, California to measure indoor and outdoor PM
component concentrations in the homes of over 100 asthmatic children. Some of these homes
are located near agricultural, construction, and utility nonroad equipment operations. This work
will measure infiltration of elemental carbon and other PM components to indoor environments.
The project also evaluates lung function changes in the asthmatic children during fluctuations in
exposure concentrations and compositions.  This information may allow an evaluation of adverse
health effects associated with exposures to elemental carbon and other PM components from
on-road and nonroad sources.

2.2.2 Gaseous Air Toxics

   Nonroad diesel engine emissions contain several substances known or suspected as human or
animal carcinogens, or have noncancer health effects. These other compounds include benzene,
1,3-butadiene,  formaldehyde, acetaldehyde, acrolein, dioxin, and polycyclic organic matter
(POM). Mobile sources, including nonroad diesel engines, contribute significantly to total
emissions of these air toxics. All of these compounds were identified as national or regional
"risk" drivers in the 1996 NATA. That is, these compounds pose a significant portion of the
total inhalation cancer risk to a significant portion of the population.  As discussed later in this
section, this final rule will significantly reduce these emissions.

   Nonroad engines are major contributors to nationwide cancer risk from air toxic pollutants,
as indicated by the NATA  1996.192 In fact, this study and the National Toxics Inventory (NTI)
for 1996 are used throughout this section for toxics inventory information for nonroad sources.193

                                         2-75

-------
Final Regulatory Impact Analysis
Also, a supplemental paper provides more detail on nonroad diesel exhaust.194 In addition, a
paper published by the Society of Automotive Engineers gives future projections to 2007 for
these air toxics.195 These references form the basis for much of what will be discussed in this
section.

   Figure 2.2.2-1 summarizes the contribution of nonroad engines to average nationwide
lifetime upper bound cancer risk from outdoor sources in the 1996 NATA.  These data do not
include the cancer risk from diesel PM since EPA does not presently have a potency for diesel
particulate/exhaust.  Figure 2.2.2-2 depicts the nonroad engine contribution to average
nationwide inhalation exposure for benzene, 1,3-butadiene, formaldehyde, acetaldehyde, and
acrolein.  These compounds are all known or suspected human carcinogens, except for acrolein,
which has serious noncancer health effects.  All of these compounds were identified as national
or regional risk drivers in the 1996 NATA, and mobile sources contribute significantly to total
emissions in NATA. As indicated previously, NATA exposure and risk estimates are based on
air dispersion modeling using the ASPEN model. Comparisons of the predicted concentrations
from the model to monitor data indicate good agreement for benzene, where the ratio of median
modeled concentrations to monitor values is 0.92, and results are within a factor of two at almost
90 percent of monitors.196  Comparisons with aldehydes indicate significantly lower modeled
concentrations than monitor values. Comparisons with 1,3-butadiene have not been done.
Previously, extensive work was done on gaseous air toxic emissions including those from
nonroad diesel and reported in EPA's 1993 Motor Vehicle-Related Air Toxics Study.197 This
final rule will reduce these emissions. Dioxin and  some POM compounds have also been
identified as probable human carcinogens and are emitted by mobile sources, although nonroad
sources are less than 1% of total emissions for these compounds.
                                         2-76

-------
                                                           Air Quality,  Health, and Welfare Effects
                                            Figure 2.2.2-1
                                      1996 Risk Characterization
               DislfdMition at lifetime cancer risk lor the US population, based on 1996" exposure
                         la 2&carcmagerak aii pollutants Irom vaikxis source sectors
          0X1
                         C01
                                         01
                                                           i^r^mit p*r Million
                                                                       10
                                                                                      1CC
                                                                                                    1HM
f-K-a a-K ahcf
  x sj't- bMr«d an inftd/dltan *XOOM--IF* ID auldtou'
 or •!' lauctei ««r ri iL^Klinr, ^u'rm^! uad ^iwi
ciaim -:wir«' 'int rar ran«
caly, Jjllisuph llrcxa .-
     v we tr^vclca Jc l
     wi L
                                                                          lwilniy ndyc* Ihttm
                                                      2-77

-------
3
to
a
X
LU
HI
    c
    o

    .Q

    O
    O
                                                      Figure 2.2.2-2
        Contribution of Source Sectors to Average Annual Nationwide Inhalation Exposure to Air Toxics in 1996
        100%
         90%
         80%
         70%
         60%
      50%
         40%
         30%
         20%
          10%
          0%
                                                           Exposure
                                                           All Sources
                                                           = 0.96
                                                                                            Exposure
                                                                                            All Sources
                                                                                            = 0.10
                                    Background
                                  nArea
                                  rjMajor
                                  nNonroad Mobile
                                  rjOnroad Mobile
                    Benzene
                                    1 ,3-Butadiene
                                                  Formaldehyde
                                                    Pollutant
Acetaldehyde
                                                                                           Aero le in
Source: National Scale Air Toxics Assessment.

-------
                                                Air Quality, Health, and Welfare Effects
    2.2.2.1 Benzene

    Benzene is an aromatic hydrocarbon that is present as a gas in both exhaust and evaporative
emissions from mobile sources. Benzene accounts for one to two percent of the exhaust
hydrocarbons, expressed as a percentage of total organic gases (TOG), in diesel engines.198'199
For gasoline-powered highway vehicles, the benzene fraction of TOG varies depending on
control technology (e.g., type of catalyst) and the levels of benzene and other aromatics in the
fuel, but is generally higher than for diesel engines, about three to five percent.  The benzene
fraction of evaporative emissions from gasoline vehicles depends on control technology and fuel
composition and characteristics (e.g., benzene level and the evaporation rate) and is generally
about one percent.200

    Nonroad engines account for 28 percent of nationwide emissions of benzene with nonroad
diesel accounting for about 3 percent in 1996. Mobile sources as a whole account for 78 percent
of the total benzene emissions in the nation.  Nonroad sources as a whole account for an average
of about 17 percent of ambient benzene in urban areas and about 9 percent of ambient benzene in
rural areas across the U.S, in the 1996 NATA assessment.  Of ambient benzene levels due to
mobile sources, 5 percent in urban and 3 percent in rural areas come from nonroad diesel engines
(see Figure 2.2.2-3).

    The EPA's IRIS database lists benzene as a  known human carcinogen (causing leukemia) by
all routes of exposure.201 It is associated with additional health effects including chromosomal
changes in human and animal cells and increased proliferation of bone marrow cells in mice.202'
203  A number of adverse noncancer health effects including blood disorders, such as
preleukemia and aplastic anemia, have also been associated with long-term occupational
exposure to benzene.

    Inhalation is the major source of human exposure to benzene in the occupational  and non-
occupational setting.  At least half of this exposure is attributable to gasoline vapors and
automotive emissions. Long-term inhalation occupational exposure to benzene has been shown
to cause cancer of the hematopoetic (blood cell) system. Among these are acute nonlymphocytic
leukemia,1 chronic lymphocytic leukemia and possibly multiple myeloma
    leukemia is a blood disease in which the white blood cells are abnormal in type or number. Leukemia may be
divided into nonlymphocytic (granulocytic) leukemias and lymphocytic leukemias. Nonlymphocytic leukemia
generally involves the types of white blood cells (leukocytes) that are involved in engulfing, killing, and digesting
bacteria and other parasites (phagocytosis) as well as releasing chemicals involved in allergic and immune
responses. This type of leukemia may also involve erythroblastic cell types (immature red blood cells).
Lymphocytic leukemia involves the lymphocyte type of white bloods cell that are responsible for the immune
responses. Both nonlymphocytic and lymphocytic leukemia may, in turn, be separated into acute (rapid and fatal)
and chronic (lingering, lasting) forms.  For example; in acute myeloid leukemia (AML) there is diminished
production of normal red blood cells (erythrocytes), granulocytes, and platelets (control clotting), which leads to
death by anemia, infection, or hemorrhage. These events can be rapid.  In chronic myeloid leukemia (CML) the
leukemic cells retain the ability to differentiate (i.e., be responsive to stimulatory factors) and perform function; later
there is a loss of the ability to respond.

                                            2-79

-------
Final Regulatory Impact Analysis
                                      Figure 2.2.2-3
                      Contribution of Source Sectors to Total Average
                 Nationwide Mobile Source Ambient Concentrations in 1996
                                                                     L o c o m o live s 9
                                                                     Com m e re ia I M a rin e 8
                                                                       a ft?

                                                                    QHeavy Duty Diesel Vehicles3
                                                                    • L ig h t D uty D ie s e 12
                                                                    ^Gasoline h ig h w a y ve h ic Ie s 1
(primary malignant tumors in the bone marrow), although the evidence for the latter has
decreased with more recent studies.204'205 Leukemias, lymphomas, and other tumor types have
been observed in experimental animals exposed to benzene by inhalation or oral administration.
Exposure to benzene and/or its metabolites has also been linked with chromosomal changes in
humans and animals206 and increased proliferation of mouse bone marrow cells.207

   The latest assessment by EPA places the excess risk of developing acute nonlymphocytic
leukemia at 2.2 x 10"6 to 7.8 x 10"6 per |ig/m3. In other words, there is a risk of about two to
eight excess leukemia cases in one million people exposed to 1 i-ig/m3 over a lifetime (70
years).208 This range of unit risks are the maximum likelihood estimate (MLE) calculated from
different exposure assumptions and dose-response models that are linear at low doses. It should
be noted that not enough information is known to determine the slope of the dose-response curve
at environmental levels of exposure and to provide a sound scientific basis to choose any
particular extrapolation model to estimate human cancer risk at low doses.  In fact, recent data209
suggest that because genetic abnormalities occur at low exposure in humans, and the formation
of toxic metabolites plateaus above 25 ppm (80,000 i-ig/m3), the dose-response curve could be
supralinear below 25 ppm.  Thus, EPA believes the use of a linear extrapolation model as a
default approach is appropriate.
                                         2-80

-------
                                              Air Quality, Health, and Welfare Effects
   Based on average population exposures in the 1996 NATA Assessment, upper bound cancer
risk (using the upper end of the MLE range) from inhalation of benzene from ambient sources is
above 10 in a million across the entire United States.  These results are best interpreted as upper
estimates of risks to typical individuals (provided exposure estimates are not underestimated).
Thus most individuals are likely to have risks that are equal to or lower than these estimates, but
some individuals may have risks which are higher.  EPA projects a median nationwide reduction
in ambient concentrations of benzene from mobile sources of about 46 percent between 1996 and
2007, as a result of current and planned control programs based on the analysis referenced earlier
examining these pollutants in the  1996 to 2007 time frame based on the analysis of hazardous air
pollutants in the 1996 to 2007 time frame referenced earlier.

   A number of adverse noncancer health effects, blood disorders such as preleukemia and
aplastic anemia, have also been associated with long-term exposure to benzene.210'2n People
with long-term occupational exposure to benzene have experienced harmful effects on the blood-
forming tissues, especially in bone marrow. These effects can disrupt normal blood production
and suppress the production of important blood components, such as red and white blood cells
and blood platelets, leading to anemia (a reduction in the number of red blood cells), leukopenia
(a reduction in the number of white blood cells),  or thrombocytopenia (a reduction in the number
of blood platelets, thus reducing the ability of blood to clot). Chronic inhalation exposure to
benzene in humans and animals results in pancytopenia/ a condition characterized by decreased
numbers of circulating erythrocytes (red blood cells), leukocytes (white blood cells), and
thrombocytes (blood platelets).212'213 Individuals that develop pancytopenia and have continued
exposure to benzene may develop aplastic anemia,K whereas others exhibit both pancytopenia
and bone marrow hyperplasia (excessive cell formation), a condition that may indicate a
preleukemic state.214'215 It should be noted that these health effects occur in human and animal
studies at concentrations well  above those typically found in the ambient environment. The most
sensitive noncancer effect observed in humans, based on current data, is the depression of the
absolute lymphocyte count in blood.216 EPA's inhalation reference concentration (RfC, i.e., a
chronic exposure level presumed  to be "without appreciable risk" for noncancer effects) for
   JPancytopenia is the reduction in the number of all three major types of blood cells
(erythrocytes, or red blood cells, thrombocytes, or platelets, and leukocytes, or white blood
cells).  In adults, all three major types of blood cells are produced in the bone marrow of the
vertebra, sternum, ribs, and pelvis. The bone marrow contains immature cells, known as
multipotent myeloid stem cells, that later differentiate into the various mature blood cells.
Pancytopenia results from a reduction in the ability of the red bone marrow to produce adequate
numbers of these mature blood cells.

   KAplastic anemia is a more severe blood disease and occurs when the bone marrow ceases to
function, i.e.,these stem cells never reach maturity. The depression in bone marrow function
occurs in two stages - hyperplasia, or increased synthesis of blood cell elements, followed by
hypoplasia, or decreased synthesis.  As the disease progresses, the bone marrow decreases
functioning. This myeloplastic dysplasia (formation of abnormal tissue) without acute leukemias
known as preleukemia. The aplastic anemia can progress to AML (acute mylogenous leukemia).

                                          2-81

-------
Final Regulatory Impact Analysis
benzene is 30 i-ig/m3, based on suppressed absolute lymphocyte counts as seen in humans under
occupational exposure conditions.

   The average inhalation exposure concentration to benzene from ambient sources in the 1996
NATA assessment is 1.4 i-ig/m3, and the 95th percentile exposure concentration is about twice as
high (U. S. EPA, 2002).  However, the assessment does not account for localized hotspots.  In
these hot spots, such as in close proximity to roadways, inhalation exposures from ambient
sources are likely to be much higher.217'218'219'220'221'222  As mentioned above, nonroad diesel
engines are small but significant contributors to the ambient concentrations resulting in these
exposures.

   2.2.2.2  1,3-Butadiene

   1,3-Butadiene is formed in engine exhaust by the incomplete combustion of fuel.  It is not
present in engine evaporative emissions, because it is not present in any appreciable amount in
fuel. 1,3-butadiene accounts for less than one percent of total organic gas exhaust from mobile
sources.

   Nonroad engines account for 18 percent of nationwide emissions of 1,3-butadiene in 1996
with nonroad diesel  accounting for about 1.5 percent based on the NATA, NTI, and
supplemental information already discussed in the previous section. Mobile sources account for
63 percent of the total 1,3-butadiene emissions in the nation as a whole.  Nonroad sources as a
whole account for an average of about 21 percent of ambient butadiene in urban areas and about
13 percent of ambient 1,3-butadiene in rural areas across the United States. Of ambient
butadiene levels due to mobile sources, 4 percent in urban and 2 percent in rural areas come from
nonroad diesel (see Figure 2.2.2-3).

   EPA earlier identified 1,3-butadiene as a probable human carcinogen in its IRIS database.223
EPA characterized 1,3-butadiene as carcinogenic to humans by  inhalation.224'225'226  The specific
mechanisms of 1,3-butadiene-induced carcinogenesis are not fully characterized. However, the
data  strongly suggest that the carcinogenic effects are mediated by genotoxic metabolites of
1,3-butadiene. Animal data suggest that females may be more sensitive than males for cancer
effects; but more data are needed before reaching definitive conclusions  on potentially sensitive
subpopulations.

   The cancer unit risk estimate is 0.08/ppm or 3^10-5 per |ig/m3 (based primarily on linear
modeling and extrapolation of human data). In other words, it is estimated that approximately 30
persons in one million exposed to 1 i-ig/m3 1,3-butadiene continuously for their lifetime (70
years) would develop cancer as a result of this exposure. The human incremental lifetime unit
cancer risk (incidence) estimate is based on extrapolation from leukemias observed in an
occupational epidemiologic study.227 This estimate includes a twofold adjustment to the
epidemiologic-based unit cancer risk applied to reflect evidence from the rodent bioassays
suggesting that the epidemiologic-based estimate may underestimate total cancer risk from
1,3-butadiene exposure in the general population.  Based on average population exposure from
the 1996 NATA Assessment, upper bound lifetime cancer risk from inhalation of 1,3-butadiene

                                          2-82

-------
                                             Air Quality, Health, and Welfare Effects
is above 10 in a million across the entire United States. Most individuals are likely to have risks
that are equal to or lower than these estimates, but some individuals may have risks which are
higher. EPA projects a median nationwide reduction in ambient concentrations of butadiene
from mobile sources of about 46 percent between 1996 and 2007, as a result of current and
planned control programs.

   1,3-Butadiene also causes a variety of reproductive and developmental effects in mice; no
human data on these effects are available. The most sensitive effect was ovarian atrophy
observed in a lifetime bioassay of female mice.228 Based on this critical effect and the
benchmark concentration methodology, an RfC was calculated.  This RfC for chronic health
effects was 0.9 ppb, or about 2 |ig/m3.  The average inhalation exposure from outdoor sources in
the 1996 NATA assessment was 0.08 i-ig/m3, with a 95th percentile concentration of 0.2 |ig/m3
(U. S. EPA, 2002). As is the case with benzene, in some hot spots, such as in close proximity to
roadways, inhalation exposures from ambient sources are likely to be much higher. As
mentioned above, nonroad diesel engines are small but significant contributors to the ambient
concentrations resulting in these exposures.

   2.2.2.3 Formaldehyde

   Formaldehyde is the most prevalent aldehyde in engine exhaust. It is formed from
incomplete combustion of both gasoline and diesel fuel.  In a recent test program that measured
toxic emissions from several nonroad diesel engines, ranging from 50 to 480 horsepower,
formaldehyde consistently accounted for well over 10 percent of total exhaust hydrocarbon
emissions.229 Formaldehyde accounts for far less of total exhaust hydrocarbon emissions from
gasoline engines, although the amount can vary substantially by  duty cycle, emission control
system, and fuel composition.  It is not found in evaporative emissions.

   Nonroad engines account for 29 percent of nationwide emissions of formaldehyde in 1996,
with nonroad diesel accounting for about 22 percent based on the NATA, NTI, and supplemental
information already discussed. Mobile sources as a whole account for 56 percent of the total
formaldehyde emissions in  the nation.  Of ambient formaldehyde levels due to mobile sources,
37 percent in urban and 27 percent in rural  areas come from nonroad diesel. Nonroad sources as
a whole account for an average of about 41 percent of ambient formaldehyde in urban areas and
about 10 percent of ambient formaldehyde in rural areas across the U.S, in the 1996 NATA
assessment.  These figures are for tailpipe emissions of formaldehyde.  Formaldehyde in the
ambient air comes not only from tailpipe (of direct) emissions but is also formed from
photochemical reactions of hydrocarbons. Mobile sources are responsible for well over 50
percent of total formaldehyde including both the direct emissions and photochemically formed
formaldehyde in the ambient air, according to the NATA for 1996. EPA projects  a median
nationwide reduction in ambient concentrations of formaldehyde from mobile sources of about
43 percent between 1996 and 2007, as a result of current and planned control programs (Cook et
al., 2002).

   EPA has classified formaldehyde as a probable human carcinogen based on limited evidence
for carcinogenicity in humans and sufficient evidence of carcinogenicity in animal studies, rats,

                                         2-83

-------
Final Regulatory Impact Analysis
mice, hamsters, and monkeys.230'231 Epidemiological studies in occupationally exposed workers
suggest that long-term inhalation of formaldehyde may be associated with tumors of the
nasopharyngeal cavity (generally the area at the back of the mouth near the nose), nasal cavity,
and sinus.232 Studies in experimental animals provide sufficient evidence that long-term
inhalation exposure to formaldehyde causes an increase in the incidence of squamous (epithelial)
cell carcinomas (tumors) of the nasal cavity.233'234'235  The distribution of nasal tumors in rats
suggests that not only regional exposure but also local tissue susceptibility may be important for
the distribution of formaldehyde-induced tumors.236  Research has demonstrated that
formaldehyde produces mutagenic activity in cell cultures.237

   The agency is currently conducting a reassessment of risk from inhalation exposure to
formaldehyde based on new information including a study by the CUT Centers for Health
Research.238'239 The CUT information and other recent information, including recently published
epidemiological studies, are being reviewed and considered in the reassessment of the
formaldehyde unit risk estimate. The epidemiological studies examine the potential for
formaldehyde to cause cancer in organs other than those addressed by the CUT model.  We plan
to bring this reassessment to the Science Advisory Board in the summer of 2004.

   Formaldehyde exposure also causes a range of noncancer health effects. At low
concentrations (e.g. 60 - 2500 i-ig/m3), irritation of the eyes (tearing of the eyes and increased
blinking) and mucous membranes is the principal effect observed in humans.  At exposure to
1200-14,000 i-ig/m3, other human upper respiratory effects associated with acute formaldehyde
exposure include a dry or sore throat, and a tingling sensation of the nose. Sensitive individuals
may experience these effects at lower concentrations. Forty percent of formaldehyde-producing
factory workers reported nasal symptoms such as rhinitis (inflammation of the nasal membrane),
nasal obstruction, and nasal discharge following chronic exposure.240 In persons with bronchial
asthma, the upper respiratory irritation caused by formaldehyde can precipitate an acute
asthmatic attack, sometimes at concentrations below 6200 |ig/m3.241  Formaldehyde exposure
may also cause bronchial asthma-like symptoms in non-asthmatics.242'243

   Immune stimulation may occur following formaldehyde exposure, although conclusive
evidence is not available.  Also, little is known about formaldehyde's effect on the central
nervous system.  Several animal inhalation studies have been conducted to assess the
developmental toxicity of formaldehyde: The only exposure-related effect noted in these studies
was decreased maternal body weight gain at the high-exposure level. No adverse effects on
reproductive outcome of the fetuses that could be attributed to treatment were noted. An
inhalation reference concentration (RfC), below which long-term exposures would not pose
appreciable noncancer health risks, is not available for formaldehyde at this time. The Agency is
currently conducting a reassessment of risk from inhalation exposure to formaldehyde.
   Average inhalation exposure from outdoor sources in the 1996 NATA assessment was 0.9
I-ig/m3, with a 95th percentile concentration of 2.3 |ig/m3.
                                          2-84

-------
                                              Air Quality, Health, and Welfare Effects
   2.2.2.4 Acetaldehyde

   Acetaldehyde is a saturated aldehyde that is found in engine exhaust and is formed as a result
of incomplete combustion of both gasoline and diesel fuel.  In a recent test program that
measured toxic emissions from several nonroad diesel engines, ranging from 50 to 480
horsepower, acetaldehyde consistently accounted for over 5 percent of total exhaust hydrocarbon
emissions (Southwest Research, 2002). Acetaldehyde accounts for far less of total exhaust
hydrocarbon emissions from gasoline engines, although the amount can vary substantially by
duty cycle, emission control system, and fuel composition.  It is not a component of evaporative
emissions.

   Nonroad engines account for 43 percent of nationwide emissions of acetaldehyde with
nonroad diesel accounting for about 34 percent based on the NATA, NTI, and supplemental
information. Mobile sources as a whole account for 73 percent of the total acetaldehyde
emissions in the nation. Nonroad sources as a whole account for an average of about 36 percent
of ambient acetaldehyde in urban areas and about 21 percent of ambient acetaldehyde in rural
areas across the U.S, in the 1996 NATA assessment. Of ambient acetaldehyde levels due to
mobile sources, 24 percent in urban and 17 percent in rural  areas come from nonroad diesel..
Also, acetaldehyde can be formed photochemically in the atmosphere. Counting both direct
emissions and photochemically formed acetaldehyde, mobile sources are responsible for the
major portion of acetaldehyde in the ambient air according  to the NATA for 1996.

   Based primarily on nonhuman animal model studies, acetaldehyde is classified by EPA as a
probable human carcinogen.  Studies in experimental animals provide sufficient evidence that
long-term inhalation exposure to acetaldehyde causes an increase in the incidence of nasal
squamous cell carcinomas (epithelial tissue) and adenocarcinomas (glandular tissue)244'245'246'247'
248 The upper confidence limit estimate of a lifetime extra cancer risk from continuous
acetaldehyde exposure is about 2.2 x 10"6 per |ig/m3. In other words, it is estimated that about 2
persons in one million exposed to 1 i-ig/m3 acetaldehyde continuously for their lifetime (70 years)
would develop cancer  as a result of their exposure.  The Agency is currently conducting a
reassessment of risk from inhalation exposure to acetaldehyde. Based on the current unit risk
and average population exposure from the 1996 NATA Assessment, upper bound cancer risk
from inhalation of acetaldehyde from ambient sources is above one in a million for more than
one hundred million Americans. Most individuals are likely to have risks that are equal to or
lower than these estimates, but  some individuals may have  risks which are higher. EPA projects
a median nationwide reduction  in ambient concentrations of acetaldehyde from mobile  sources
of about 36 percent between 1996 and 2007, as a result of current and planned control programs

   EPA's IRIS database states  that noncancer effects in studies with rats and mice showed
acetaldehyde to be moderately toxic by the inhalation, oral, and intravenous  routes (EPA, 1988).
Similar conclusions have been made by the California Air Resources Board.249 The primary
acute effect of exposure to acetaldehyde vapors is irritation of the eyes, skin, and respiratory
tract. At
high concentrations, irritation and pulmonary effects can occur, which could facilitate the uptake
of other contaminants. Little research exists that addresses  the effects of inhalation of

                                         2-85

-------
Final Regulatory Impact Analysis
acetaldehyde on reproductive and developmental effects. Long-term exposures should be kept
below the reference concentration of 9 |ig/m3 to avoid appreciable risk of these noncancer health
effects (EPA, 1988).  The average inhalation exposure from outdoor sources in the 1996 NATA
assessment was 0.7 i-ig/m3, with a 95th percentile concentration of 1.8 |ig/m3 (U. S. EPA, 2002).
As is the case with other air toxic compounds emitted by mobile sources, in some hot spots, such
as in close proximity to roadways, inhalation exposures are likely to be much higher.  As
mentioned above, nonroad diesel engines are significant contributors to the ambient
concentrations resulting in these exposures.

   Acetaldehyde has been associated with lung function decrements in asthmatics.  In one
study, aerosolized acetaldehyde caused reductions in lung function and bronchoconstriction in
asthmatic subjects.250

   2.2.2.5 Acrolein

    In a recent test program that measured toxic emissions from several nonroad diesel engines,
ranging from 50 to 480 horsepower, acrolein accounted for about 0.5 to 2 percent of total
exhaust hydrocarbon emissions (Southwest Research, 2002). Acrolein accounts for far less  of
total exhaust hydrocarbon emissions from gasoline engines, although the amount can vary
substantially by duty cycle, emission control system, and fuel composition. It is not a
component of evaporative emissions.

   Nonroad engines account for 25 percent of nationwide emissions of acrolein in 1996 with
nonroad diesel accounting for about 17.5 percent based  on NATA, NTI, and the supplemental
information Mobile sources as a whole account for 43 percent of the total acrolein emissions in
the nation.  Of ambient acrolein levels due to mobile sources, 28 percent in urban and 18 percent
in rural areas come form nonroad diesel  according to NATA.

   Acrolein is intensely irritating to humans when inhaled, with acute exposure resulting in
substantial discomfort and sensory irritancy, mucus hypersecretion, and congestion. These
effects have been noted at acrolein levels ranging from 390 |ig/m3 to 990 |ig/m3.251 The intense
irritancy of this carbonyl has been demonstrated during  controlled tests in human subjects who
suffer intolerable eye and nasal mucosal sensory reactions within minutes of exposure.252 The
irritant nature of acrolein provides the basis for the OSHA Permissible Exposure Limit (PEL) for
the workplace of 0.1 ppm (230 |ig/m3) for an 8-hour exposure period. Acrolein has an odor
threshold of about 0.16 ppm (370 |ig/m3),253 and acute inhalation exposure of humans to 10  ppm
(23,000 |ag/m3) may result in death over a short period of time.254

   Acrolein is an extremely volatile vapor, and it possesses considerable water solubility.255 As
such, it readily absorbs into airway fluids in the respiratory tract when inhaled. Lesions to the
lungs and upper respiratory tract of rats, rabbits, and hamsters exposed to acrolein formed the
basis of the reference concentrations for inhalation (RfC) developed in 2003.256  The RfC of
acrolein is 0.02 |ig/m3.  Average population inhalation exposures from the 1996 NATA
assessment are between 0.02 |ig/m3 and 0.2 |ig/m3.  Thus, the hazard quotient (inhalation
exposure divided by the RfC) is greater than one for most of the U.S. population, indicating a

                                          2-86

-------
potential for adverse noncancer health effects.

   The toxicological data base demonstrating the highly irritating nature of this vapor has been
consistent regardless of test species.  Animal inhalation studies revealed early on that acrolein
induces damage throughout the respiratory tract at 0.7 ppm (1600 |ig/m3) 257 in concordance with
data showing similar vapor uptake along isolated upper and lower lung regions of animals.258 At
levels that humans may encounter incidentally, acrolein has been shown to alter breathing
mechanics259'260 and airway  structure in animals261 as well as to interfere with macrophage
function and to alter microbial infectivity.262'263'264 As with many other irritants, acrolein has the
potential to induce adaptation to its own irritancy with repeated exposures to low concentrations
(1260 |ig/m3)265 — a phenomenon consistent with the apparent human adaptation to the high
spikes of acrolein emanating in mainstream smoke from cigarettes.266  Hence,  sensory awareness
of exposure to low levels of acrolein may diminish the apparent acute discomfort, while
exposure and the potential for longer term impacts persist.  Prolonged exposure to acrolein has
been shown in animals to have an impact on pulmonary structure and function that can be
quantified.267 Over the range of 0.4 to 4.0 ppm (920 to 9200 |ig/m3) acrolein, distinct dose-
dependent changes in the degree of injury/disease are apparent, which have lung function
consequences. There are clear changes in the cell lining of the airways, including mucus cell
hyperplasia, as well as changes in the underlying supportive matrix of the airways. These
changes parallel changes in  airway hyperreactivity (sometimes referred to as "twitchiness").
Such changes are similar to  those observed with asthma. The structural changes in the larger
airways, likewise, are reminiscent of those associated with chronic exposure to tobacco smoke.

   Irritant effects in humans can be seen at levels encountered industrially that are below the
odor threshold and thus may be erroneously thought to be safe. Over time, these same
occupational levels of exposure in rats appear to alter airway structure and function.  As those in
the workplace generally  do not reflect the more sensitive groups of the public, the potential for
persistent, low level exposures eliciting health outcomes among susceptible groups, including
asthmatics who have sensitive airways is a concern.268

    EPA has concluded that the potential for carcinogen!city of acrolein cannot be determined
either for oral or inhalation routes of exposure.269

   2.2.2.6 Polycyclic Organic Matter

   POM is generally defined as a large class of chemicals consisting of organic compounds
having multiple benzene rings and a boiling point greater than 100 degrees C. Polycyclic
aromatic hydrocarbons (PAHs) are a chemical class that is a subset of POM. POM are naturally
occurring substances that are byproducts of the incomplete combustion of fossil fuels and plant
and animal biomass (e.g., forest fires). They occur as byproducts from steel and coke
productions and waste incineration. They also are a component of diesel PM emissions.  As
mentioned in Section 2.1.2.1.2, many of the compounds included in the class of compounds
known as POM are classified by EPA as probable human carcinogens based on animal data. In
particular, EPA obtained data on 7 of the POM compounds, which we analyzed separately as a
class in the NATA for 1996. Nonroad engines account for only 1 percent of these 7 POM
compounds with total mobile sources responsible for only 4 percent of the total; most of the 7
POMs come from area sources. For total POM compounds, mobile sources as a whole are
responsible for only 1 percent.  The mobile source emission numbers used to derive these

-------
Final Regulatory Impact Analysis
inventories are based only on particulate-phase POM and do not include the semi-volatile phase
POM levels. Were those additional POMs included (which is now being done in the NATA for
1999), these inventory numbers would be substantially higher. A study of indoor PAH found that
concentrations of indoor PAHs followed the a similar trend as outdoor motor traffic, and that
motor vehicle traffic was the largest outdoor source of PAH.270

   A recent study found that maternal exposures to polycyclic aromatic hydrocarbons (PAHs) in
a multiethnic population of pregnant women were associated with adverse birth outcomes,
including low birth weight, low birth length, and reduced head circumference.271

   2.2.2.7 Dioxins

   Exposure to dioxins are recognized by several authoritative bodies, including the
International Agency for Research on Cancer, the National Institute of Environmental Health
Sciences, the Agency for Toxic Substances and Disease Registry, EPA and some State health
and environmental agencies, to present a human health hazard for cancer and non-cancer effects.
Recent studies have confirmed that very small amounts of dioxins are formed by and emitted
from diesel engines (both heavy-duty diesel trucks and nonroad diesel engines). In an inventory
for dioxin sources in 1995, such emissions accounted for only about 1 percent of total dioxin
emissions.  These nonroad rules will have minimal impact on overall dioxin emissions since
these are a very small part of total emissions.

2.3 Ozone

   This section reviews health and welfare effects of ozone and describes the  air quality
information that forms the basis of our conclusion that ozone concentrations in many areas
across the country face a significant risk of exceeding the ozone standard into the year 2030.
Information on air quality was gathered from a variety of sources, including monitored ozone
concentrations from 1999-2001, air quality modeling forecasts conducted for  this rulemaking
and other state and local air quality information.

   Ground-level ozone, the main ingredient in smog,  is formed by the reaction of volatile
organic compounds (VOCs) and nitrogen oxides (NOx) in the atmosphere in the presence of heat
and sunlight. These pollutants, often referred to as ozone precursors, are emitted by many types
of pollution sources, including highway and nonroad motor vehicles and engines, power plants,
chemical plants, refineries, makers of consumer and commercial products, industrial facilities,
and smaller "area" sources.  VOCs are also emitted by natural sources such as vegetation.
Oxides of nitrogen are emitted largely from motor vehicles, off-highway equipment, power
plants,  and  other sources of combustion.

   The science of ozone formation, transport, and accumulation is complex.  Ground-level
ozone is produced and destroyed in a cyclical set of chemical reactions involving NOx, VOC,
heat, and sunlight. Many of the chemical reactions that are part of the ozone-forming cycle are
sensitive to temperature and sunlight. When ambient temperatures and  sunlight levels remain

-------
                                              Air Quality, Health, and Welfare Effects
high for several days and the air is relatively stagnant, ozone and its precursors can build up and
produce more ozone than typically would occur on a single high-temperature day. Further
complicating matters, ozone also can be transported into an area from pollution sources found
hundreds of miles upwind, resulting in elevated ozone levels even in areas with low VOC or
NOx emissions.  As a result, differences in NOx and VOC emissions and weather patterns
contribute to daily, seasonal, and yearly differences in ozone concentrations and differences from
city to city.

   These complexities also have implications for programs to reduce ozone. For example,
relatively small amounts of NOx enable ozone to form rapidly when VOC levels are relatively
high, but ozone production is quickly limited by removal of the NOx. Under these conditions,
NOx reductions are highly effective in reducing ozone while VOC reductions have little effect.
Such conditions are called "NOx-limited."  Because the contribution of VOC emissions from
biogenic (natural) sources to local ambient ozone  concentrations can be significant, even some
areas where man-made VOC emissions are relatively low can be NOx-limited.

   When NOx levels are relatively high and VOC levels relatively low, NOx forms inorganic
nitrates (i.e., particles) but relatively little ozone.  Such conditions are called "VOC-limited."
Under these conditions, VOC reductions are effective in reducing ozone, but NOx reductions can
actually increase local ozone under certain circumstances.  Even in VOC-limited urban areas,
NOx reductions are not expected to increase ozone levels if the NOx reductions are sufficiently
large. The highest levels of ozone are produced when both VOC and NOx emissions are present
in significant quantities on clear summer days.

   Rural areas are almost always NOx-limited, due to the relatively large amounts of biogenic
VOC emissions in such areas. Urban areas can be either VOC- or NOx-limited, or a mixture of
both, in which ozone levels exhibit moderate sensitivity to changes in either pollutant.

   Ozone concentrations in an area also can be lowered by the reaction of nitric oxide with
ozone, forming nitrogen dioxide (NO2); as the air  moves downwind and the cycle continues, the
NO2 forms additional ozone. The importance of this reaction depends, in part, on the relative
concentrations of NOx, VOC, and ozone, all of which change with time and location.

2.3.1 Health Effects of Ozone

   Exposure to ambient ozone contributes to a wide range of adverse health effects, which are
discussed in detail in the EPA Air Quality Criteria Document for Ozone.272 Effects include lung
function decrements, respiratory symptoms, aggravation of asthma, increased hospital and
emergency room visits, increased medication usage, inflammation of the lungs, as well  as a
variety of other respiratory effects.  People who are particularly at risk for high ozone exposures
inclue healthy  children and adults who are active outdoors. Susceptible subgroups include
children, people with respiratory disease, such as asthma, and people with unusual sensitivity to
ozone. More information on health  effects of ozone is also available at
http://www.epa.gOv/ttn/naaqs/standards/ozone/s 03 index.html.
                                          2-89

-------
Final Regulatory Impact Analysis
   Based on a large number of scientific studies, EPA has identified several key health effects
caused when people are exposed to levels of ozone found today in many areas of the country.
Short-term (1 to3 hours) and prolonged exposures (6 to 8 hours) to higher ambient ozone
concentrations have been linked to lung function decrements, respiratory symptoms, increased
hospital admissions and emergency room visits for respiratory problems.273' 274' 275' 276' 277' 278
Repeated exposure to ozone can make people more susceptible to respiratory infection and lung
inflammation and can aggravate preexisting respiratory diseases, such as asthma.279'280'281'282'283
It also can cause inflammation of the lung, impairment of lung  defense mechanisms, and
possibly irreversible changes in lung structure, which over time could lead to premature aging of
        s and/or chronic respiratory illnesses, such as emphysema and chronic bronchitis.284'285'
   Adults who are outdoors and active during the summer months, such as construction workers
and other outdoor workers, also are among those most at risk of elevated exposures.288 Thus, it
may be that children and outdoor workers are most at risk from ozone exposure because they
typically are active outside, playing and exercising, during the summer when ozone levels are
highest.289' 29° For example, summer camp studies in the Eastern United States and Southeastern
Canada have reported significant reductions in lung function in children who are active
outdoors.291' 292' 293' 294' 295' 296' 297' 298 Further, children are more at risk of experiencing health effects
than adults from ozone exposure because their respiratory systems are still developing.  These
individuals, as well as people with respiratory illnesses such as asthma, especially asthmatic
children, can experience reduced lung function and increased respiratory symptoms, such as
chest pain and cough, when exposed to relatively low ozone levels during prolonged periods of
moderate exertion.299'300'301'302

   The 8-hour NAAQS is based on well-documented science demonstrating that more people
are experiencing adverse health effects at lower levels of exertion, over longer periods, and at
lower ozone concentrations than addressed by the 1-hour ozone standard.303  Attaining the 8-hour
standard greatly limits ozone exposures of concern for the general population and populations
most at risk, including children active outdoors, outdoor workers, and individuals with pre-
existing respiratory disease, such as asthma.

   There has been new research that suggests additional serious health effects beyond those that
had been know when the 8-hour ozone standard was set. Since 1997, over 1,700 new health and
welfare studies have been published in peer-reviewed journals.304 Many of these studies have
investigated the impact of ozone exposure on such health effects as changes in lung structure and
biochemistry, inflammation of the lungs, exacerbation and causation of asthma, respiratory
illness-related school absence, hospital and emergency room visits for asthma and other
respiratory causes, and premature mortality. EPA is currently in the process of evaluating these
and other studies as part of the ongoing review of the air quality criteria and NAAQS for ozone.
A revised Air Quality Criteria Document for Ozone and Other Photochemical Oxidants will be
prepared in consultation with the EPA's Clean Air Scientific Advisory Committee (CAS AC).

   Key new health information falls into four general areas: development of new-onset  asthma,
hospital admissions for young children, school absence rate, and premature mortality. Examples

                                          2-90

-------
                                               Air Quality, Health, and Welfare Effects
of new studies in these areas are briefly discussed below.

   Aggravation of existing asthma resulting from short-term ambient ozone exposure was
reported prior to the 1997 decision and has been observed in studies published since.305'306 More
recent studies now suggest a relationship between long-term ambient ozone concentrations and
the incidence of new-onset asthma. In particular, such a relationship in adult males (but not in
females) was reported by McDonnell et al. (1999).307  Subsequently, McConnell et al. (2002)
reported that incidence of new diagnoses of asthma in children is associated with heavy exercise
in communities with high concentrations (i.e., mean 8-hour concentration of 59.6 ppb) of
ozone.308 This relationship was documented in children who played 3 or more sports and was
not statistically significant for those children who played one or two sports.L The larger effect of
high activity sports than low activity sports and an independent effect of time  spent outdoors also
in the higher ozone communities strengthened the inference that exposure to ozone may modify
the effect of sports on the development of asthma in some children.

   Previous studies have shown relationships between ozone and hospital admissions in the
general  population. A new study in Toronto reported a significant relationship between 1-hour
maximum ozone concentrations and respiratory hospital admissions in children under two.309
Given the relative vulnerability of children in this age category, we are particularly concerned
about the findings from the literature on ozone and hospital admissions.

   Increased respiratory disease that are serious enough to cause school absences has been
associated with 1-hour daily maximum and 8-hour average ozone concentrations in studies
conducted in Nevada in kindergarten to 6th grade310 and in Southern California in grades 4 to 6.311
These studies suggest that higher ambient ozone levels may result in increased school
absenteeism.

   The ambient air pollutant most clearly associated with premature mortality is PM, with
dozens of studies reporting such an association. However, repeated ozone exposure may be a
contributing factor for premature mortality, causing an inflammatory response in the lungs that
may  predispose elderly and other sensitive individuals to become more susceptible to the adverse
health effects of other air pollutants, such as PM.312'313 Although the findings in the past have
been mixed, the findings of three recent analyses suggests that ozone exposure is associated with
increased mortality. Although the National Morbidity, Mortality, and Air Pollution  Study
(NMMAPS) did not find an effect of ozone on total mortality across the full year, Samet  et al.
(2000),  who conducted the NMMAPS study, did report an effect after limiting the analysis to
summer when ozone levels are highest.314 Similarly, Thurston and Ito (1999)  have reported
associations between ozone and mortality.315 Toulomi et al., (1997) reported that 1-hour
maximum ozone levels were associated with daily numbers of deaths in 4 cities (London,
Athens, Barcelona, and Paris), and a quantitatively similar effect was found in a group of 4
additional cities (Amsterdam, Basel, Geneva, and Zurich).316
    In communities with mean 8-hour ozone concentration of 59.6 ppb, the relative risk of developing asthma in
children playing three or more sports was 3.3. (95% CI 1.9 - 5.8) compared with children playing no sports.

                                           2-91

-------
Final Regulatory Impact Analysis
   As discussed in Section 2.1 with respect to PM studies, the Health Effects Institute (HEI)
reported findings by health researchers that have raised concerns about aspects of the statistical
methodology used in a number of older time-series studies of short-term exposures to air
pollution and health effects.317

2.3.2 Attainment and Maintenance of the 1-Hour and 8-Hour Ozone NAAQS

   As shown earlier in Figure 2-1, unhealthy ozone concentrations (i.e., those exceeding the 8-
hour standard, which is requisite to protect public health with an adequate margin of safety)
occur over wide geographic areas, including most of the nation's major population centers.
These areas include much of the eastern half of the United States and large areas of California.
Nonroad engines contribute a substantial  fraction of ozone precursors in metropolitan areas.

   Emission reductions from this rule will assist nonattainment and maintenance areas in
reaching the standard by each area's respective attainment date and help maintaining the
standard in the future. We discuss both the 1-hour and the 8-hour NAAQS, which are based on
air quality measurements, called design values and other factors.

   An ozone design value is the concentration that determines whether a monitoring site meets
the NAAQS for ozone.  Because of the way they are defined, design values are determined based
on 3 consecutive-year monitoring periods. For example, an 8-hour design value is the fourth
highest daily maximum 8-hour average ozone concentration measured over a three-year period at
a given monitor.  The full details  of these determinations (including accounting  for missing
values and other  complexities) are given in Appendices H and I of 40 CFR Part 50. As discussed
in these appendices, design values are truncated to whole part per billion (ppb).  Due to the
precision with which the standards are expressed (0.08 parts per million (ppm) for the 8-hour), a
violation of the 8-hour standard is defined as a design value greater than or equal to 0.085 ppm.

   For a county, the  design value is the highest design value from among all the monitors with
valid design values within  that county. If a county does not contain an ozone monitor, it does
not have a design value. Thus, our analysis may underestimate the number of counties with
design values above the level of NAAQS. For the purposes of identifying areas likely to have an
ozone problem in the future, we used the  1999-2001 because these data were the most current at
the time we performed the  modeling (i.e,  2003 data were not yet available). In the recent
designations, the 2001-2003 data were used. The 1999-2001, the 2000-2002, and the 2001-2003
sets of design values  are listed in  the AQ  TSD, which is available in the docket to this rule.

   A number of States and local  areas in their public comments discussed their need for the rule
to reduce ozone levels.  The California Air Resources Board noted, "Adoption of the proposed
regulations outlined in the  NPRM by US  EPA is necessary for the  protection of public health in
California to comply with air quality standards." In addition, the South Coast Air Quality
Management District (SCAQMD) requested more federal reductions,  citing their need: "In 2010,
federal sources including non-road engines, ships, trains, aircraft, and 49-state vehicles would
contribute to 34% of the NOx emissions in the South Coast Air Basin (Basin).  Of this amount,

                                          2-92

-------
                                              Air Quality, Health, and Welfare Effects
non-road engines account for 14% or 108 tons per day of NOx in the Basin. ... without
aggressive regulations which would achieve substantial reductions by 2010 for non-road engines,
as well as other sources under federal jurisdiction, attainment of the federal 1-hour ozone and
PM2.5 standards could be seriously jeopardized. ...Where EPA has exclusive or nearly exclusive
jurisdiction, EPA must achieve the maximum feasible reductions to enable states to attain federal
standards. Therefore, it is incumbent upon EPA to craft its proposed regulation in a manner that
would provide maximum emissions benefit in the near term as well as on a long-term basis."

    The City of Houston commented that as the largest city with a severe 1-hour ozone
nonattainment area and a near-nonattainment area for PM that they had a need for "huge
emission reductions from all sectors in the 8-county area to reach attainment... While diesel
engines constitute less than 25% of the city's vehicle fleet, they account for over 40 percent of
our mobile source emissions and almost 35% of our overall emissions. The non-road portion of
our fleet alone produces 26% of our mobile source, and 21% of the city's overall emissions."

    Comments from Illinois Lieutenant Governor comments supported the  need for reductions in
ozone: "Working to relieve the affects of asthma is of particular importance in Illinois where the
mortality rate is the highest in the country  and is the number one reason for children missing
school."

    Similarly, New York State Department of Environmental Conservation "strongly supports
EPA's proposed rule to control emissions of air pollution from nonroad diesel engines and fuels.
We believe that these regulations, when  fully implemented, will provide substantial
environmental and public health benefits. ..Nonroad diesel equipment is a major source of NOx,
SOx and PM emissions and this proposal will help the state of New York attain and maintain the
NAAQS for ozone and PM."
2.3.2 Attainment and Maintenance of the 1-Hour and 8-Hour Ozone NAAQS

    As shown earlier in Figure 2-1, nonattainment with the ozone NAAQS occur over wide
geographic areas, including most of the nation's major population centers. These areas include
much of the eastern half of the United States, industrial midwest, and large areas of California.
Nonroad diesel engines contribute a substantial fraction of ozone precursors in metropolitan
areas.

    Emission reductions from this rule will assist nonattainment and maintenance areas in
reaching the standard by each area's respective attainment date and help maintaining the
standard in the future. We discuss both the 1-hour, an exceedance-based standard, and the 8-
hour NAAQS, which is based on air quality measurements, called design values, as well as other
factors.

    An ozone design value is a calculated ozone concentration that is used in determining
whether a monitoring site meets the NAAQS. Because of the way they are defined, design
values are determined based on 3 consecutive-year monitoring periods. For example, an 8-hour

                                          2-93

-------
Final Regulatory Impact Analysis
ozone design value is the average of the annual fourth highest daily maximum 8-hour average
ozone concentrations measured over a three-year period at a given monitor. Determination of
whether an area attains the 1-hour NAAQS is based on the number of "exceedances" of the
standard over a three year period. The full details of these determinations (including accounting
for missing values and other complexities) are given in Appendices H and I of 40 CFR Part 50.
As discussed in these appendices, design values are truncated to whole part per billion (ppb).
Due to the precision with which the standards are expressed (0.08 parts per million (ppm) for the
8-hour), a violation of the 8-hour standard is defined as a design value greater than or equal to
0.085 ppm.

   For a county, the design value is the highest design value from among all the monitors with
valid design values within that county. A nonattainment area may contain  counties both with
and without monitors. The highest design value of any county monitor representing the
nonattainment area would determine the design value for that nonattainment county. For the
purposes of identifying areas likely to have an ozone problem in the future, we performed
modeling and used the 1999-2001 air quality data as described below because these data were
the most current at the time we performed the modeling (i.e, 2003 data were not yet available).
In the 8-hour designations and classifications, we used the 2001-2003 data in addition to
considering other factors. The 1999-2001, the 2000-2002, and the 2001-2003 sets of design
values  are listed in the AQ TSD, which is available in the  docket to this rule.

   A number of States and local areas in their public comments discussed  their need for the rule
to reduce ozone levels. For example, the California Air Resources Board noted, "Adoption of
the proposed regulations outlined in the NPRM by US EPA is necessary for the protection of
public  health in California to comply with air quality standards."  In addition, the South Coast
Air Quality Management District (SCAQMD) requested more federal reductions, citing their
need: "In 2010, federal sources including non-road engines, ships, trains, aircraft, and 49-state
vehicles would contribute to 34% of the NOx emissions in the South Coast Air Basin (Basin).
Of this amount, non-road engines account for 14% or 108 tons per day of NOx in the Basin. ...
without aggressive regulations which would achieve substantial reductions by 2010 for non-road
engines, as well as other sources under federal jurisdiction, attainment of the federal 1-hour
ozone and PM25 standards could be seriously jeopardized. ...Where EPA has exclusive or nearly
exclusive jurisdiction, EPA must achieve the maximum feasible reductions to enable states to
attain federal standards.  Therefore, it is incumbent upon EPA to craft its proposed regulation in
a manner that would provide  maximum emissions benefit  in the near term as well as on a long-
term basis."

   The City of Houston commented that as the largest city with a severe 1-hour ozone
nonattainment area and a near-nonattainment area for PM that they had a need for "huge
emission reductions from all  sectors in the 8-county area to reach attainment... While diesel
engines constitute less than 25% of the city's vehicle fleet, they account for over 40 percent of
our mobile source emissions and almost 35% of our overall emissions. The non-road portion of
our fleet alone produces 26% of our mobile source, and 21% of the city's overall emissions."
                                          2-94

-------
                                              Air Quality, Health, and Welfare Effects
   Comments from Illinois Lieutenant Governor comments supported the need for reductions in
ozone: "Working to relieve the effects of asthma is of particular importance in Illinois where the
mortality rate is the highest in the country and is the number one reason for children missing
school."

   Similarly, New York State Department of Environmental Conservation "strongly supports
EPA's proposed rule to control emissions of air pollution from nonroad diesel engines and fuels.
We believe that these regulations, when fully implemented, will provide substantial
environmental and public health benefits. ..Nonroad diesel equipment is a major source of NOx,
SOx and PM emissions and this proposal will help the state of New York attain and maintain the
NAAQS for ozone and PM."
   2.3.2.1 1-Hour Ozone Nonattainment and Maintenance Areas and Concentrations

   Currently, there are 110 million people living in 53 1-hour ozone nonattainment areas
covering 219 counties.318 Of these areas, there are one extreme and 13 severe 1-hour ozone
nonattainment areas with a total affected population of 74 million as shown in Table 2.3-1. We
focus on these classifications of designated areas  because the timing of their attainment dates
relates to the timing of the new emission standards. Five severe 1-hour ozone nonattainment
areas have attainment dates of November 15, 2007. The Los Angeles South Coast Air Basin is
designated as an extreme nonattainment area and  has a compliance date of November 15, 2010.
While all of these areas are expected to be in attainment before the emission reductions from this
rule are fully realized,  these reductions will be important to assist these areas in achieving the
health and welfare protections of the standards and maintaining compliance with air quality
standards.
                                         2-95

-------
Final Regulatory Impact Analysis
                                          Table 2.3-1
                  1-Hour Ozone Extreme and Severe Nonattainment Areas
Nonattainment Area
Los Angeles South Coast Air Basin,
CAa
Chicago-Gary-Lake County, IL-IN
Houston-Galveston-Brazoria, TX
Milwaukee-Racine, WI
New York-New Jersey -Long Island,
NY-NJ-CT
Southeast Desert Modified AQMA, CA
Atlanta, GA
Baltimore, MD
Baton Rouge, LA
Philadelphia- Wilmington-Trenton, PA-
NJ-DE-MD
Sacramento, CA
San Joaquin Valley, CA
Ventura County, CA
Washington, DC-MD-VA
Total Population
Attainment
Date
November 15, 2010a
November 15, 2007
November 15, 2007
November 15, 2007
November 15, 2007
November 15, 2007
2005
2005
2005
2005
2005
2005
2005
2005
2000
Population
(millions)
14.6
8.8
4.7
1.8
19.2
1.0
3.7
0.8
0.6
6.3
2.0
3.2
0.7
4.5
2000-2002
Measured
Violation?
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
No
Yes
74million
a Extreme 1-Hour nonattainment areas.  All other areas are severe nonattainment areas.
Source: US EPA, Air Quality TSD 2004
    Many 1-hour ozone nonattainment areas continue to experience exceedances.
Approximately 53 million people are living in 73 counties with measured air quality violating
the 1-hour NAAQS in 2000-2002M See the AQ TSD for more details about the counties and
populations experiencing various levels of measured 1-hour ozone concentrations.
    MTypically, county design values (and thus exceedances) are consolidated where possible into design values for
consolidated metropolitan statistical areas (CMSA) or metropolitan statistical areas (MSA). Accordingly, the design
value for a metropolitan area is the highest design value among the included counties, and counties that are not in
metropolitan areas would be treated separately. However, for this section, we examined data on a county basis, not
consolidating into CMSA or MSA.  Designated nonattainment areas may contain more than one county, and some of
these counties have experienced recent exceedances, as indicated in the table. Further, the analysis is limited to areas
with ozone monitors.
                                             2-96

-------
                                              Air Quality, Health, and Welfare Effects
   The ability of states to maintain the ozone NAAQS once attainment is reached has proved
challenging, and the recent recurrence of violations of the NAAQS in some other areas increases
the Agency's concern about continuing maintenance of the standard. Recurrent nonattainment is
especially problematic for areas where high population growth rates lead to significant annual
increases in vehicle trips and VMT. Moreover, ozone modeling conducted for this rule predicted
exceedances in 2020 and 2030 (without additional controls), which adds to the Agency's
uncertainty about the prospect of continued attainment for these areas. The reductions from this
final rule will help  areas attain and maintain the 1-hour standards.

   2.3.2.2 8-Hour Ozone Levels: Current Nonattainment and Future Concentrations

   EPA has recently designated nonattainment areas for the 8-hour NAAQS by calculating air
quality design values (using 2001-2003 measurements) and considering other factors
(www.epa.gov/ozonedesignations).

   As described above in Section 2.3.1, the 8-hour NAAQS is based on well-documented
science demonstrating that more people are experiencing adverse health  effects at lower levels of
exertion, over longer periods, and at lower ozone concentrations than addressed by the 1-hour
ozone standard.319  The 8-hour standard greatly limits ozone exposures of concern for the general
population and sensitive populations.  This section describes the current  nonattainment with the
8-hour ozone NAAQS and describes our modeling to predict future 8-hour ozone concentrations,
which demonstrate a need for reductions in emissions from this final rule.

   2.3.2.2.1 Current 8-Hour Ozone Nonattainment

   All or part of 474 counties are in nonattainment, as shown in Figure 2-1, for either failing to
meet the 8-hour ozone NAAQS or for contributing to poor air quality in  a nearby area. About
159 million people live in the 126 areas that do not meet the 8-hour NAAQS.  Based upon the
measured data from years 2001-2003 and other factors, these areas were recently designated and
classified by EPA.). The nonattainment areas covered under subpart 1 will be required to attain
the standard no later than 5 years after designation and, in limited circumstances, they may
apply for an additional extension of up to 5 years (e.g., 2009 to  2014). The areas classified under
subpart 2 have attainment dates ranging from up to 3 years for marginal areas (2007) to up to 20
years for extreme areas (2024). .

Table 2.3-2 presents the areas, their design values for the 8-hour and 1-hour standards and their
category or classification.  The reductions from this rule will contribute to these areas' overall
strategy to attain and maintain the standards.
                                          2-97

-------
Final Regulatory Impact Analysis
EPA
Region Area Name
Table 2.3-2. 8-Hour Ozone Nonattainment Areas

      Design Value ppb (2001-2003 data)
             8-Hr  1-Hr  Category/Classification
2   Albany-Schenectady-Troy, NY
5   Allegan Co, MI
3   Allentown-Bethlehem-Easton, PA
3   Altoona, PA
9   Amador and Calaveras, CA(Central Mtn Co)
4   Atlanta, GA
3   Baltimore, MD
6   Baton Rouge, LA
6   Beaumont-Port Arthur, TX
5   Benton Harbor, MI
5   Benzie Co, MI                       88
3   Berkeley and Jefferson Counties, WV
4   Birmingham, AL
1   Boston-Lawrence-Worcester (E. MA), MA
1   Boston-Manchester-Portsmouth(SE),NH*
2   Buffalo-Niagara Falls, NY
5   Canton-Massillon, OH
5   Cass Co, MI
3   Charleston, WV
4   Charlotte-Gastonia-Rock Hill, NC-SC
4   Chattanooga, TN-GA
5   Chicago-Gary-Lake County, IL-IN
9   Chico, CA                          89
5,4 Cincinnati-Hamilton, OH-KY-IN
4   Clarksville-Hopkinsville, TN-KY
3   Clearfield and Indiana Cos, PA
5   Cleveland-Akron-Lorain, OH
4   Columbia, SC
5   Columbus, OH
6   Dallas-Fort Worth, TX
5   Dayton-Springfield, OH
8   Denver-Boulder-Greeley-Ft Collins-Love., CO
5   Detroit-Ann Arbor, MI
5   Door Co, WI
3   Erie, PA
2   Essex Co (Whiteface Mtn) NY
5   Evansville, IN
4   Fayetteville, NC
5   Flint, MI
87
97
91
85
91
91
103
86
91
91
116 <
86
87
95
95
99
90
93
86
100
88
101
102 <
96
85
90
103
89
95
100
90
' 87
97
94
92
91
85
87
90
115
115
114
107
117
125
143
131
129
117
•iubpe
105
113
124
124
116
109
124
107
129
113
134
•iubpz
118
99
106
128
108
117
135
117
114
127
113
114
113
106
108
103
                                    Subpart 1
                                    Subpart 1
                                    Subpart 1
                                    Subpart 1
                                    Subpart 1
                                    Subpart 2 Marginal
                                    Subpart 2 Moderate
                                    Subpart 2 Marginal
                                    Subpart 2 Marginal
                                    Subpart 1
                                   rtl
                                    EAC Subpart 1
                                    Subpart 1
                                    Subpart 2 Moderate
                                    Subpart 2 Moderate
                                    Subpart 1
                                    Subpart 1
                                    Subpart 2 Moderate
                                    Subpart 1
                                    Subpart 2 Moderate
                                    Subpart 1
                                    Subpart 2 Moderate
                                   rtl
                                    Subpart 1
                                   Subpart 1
                                    Subpart 1
                                    Subpart 2 Moderate
                                    EAC Subpart 1
                                    Subpart 1
                                    Subpart 2 Moderate
                                    Subpart 1
                                    EAC Subpart 1
                                    Subpart 2 Moderate
                                    Subpart 1
                                    Subpart 1
                                    Subpart 1
                                    Subpart 1
                                    EAC Subpart 1
                                    Subpart 1
                                        2-98

-------
                                           Air Quality, Health, and Welfare Effects
EPA
Region Area Name
Design Value ppb (2001-2003 data)
       8-Hr   1-Hr  Category/Classification
5   Fort Wayne, IN
3   Franklin Co, PA
3   Frederick Co, VA
3   Fredericksburg, VA*
5   Grand Rapids, MI
1   Greater Connecticut, CT
5   Greene Co, IN
3   Greene Co, PA
4   Greensboro-Winston Salem-High Point, NC
4   Greenville-Spartanburg-Anderson, SC
1   Hancock, Knox, Lincoln and Waldo Cos, ME
3   Harrisburg-Lebanon-Carlisle, PA
4   Haywood and Swain (Great Smoky NP), NC
4   Hickory-Morganton-Lenoir, NC
6   Houston-Galveston-Brazoria, TX
3,4 Huntington-Ashland, WV-KY
5   Huron Co, MI
9   Imperial Co, CA
5   Indianapolis, IN
5   Jackson Co, IN
2   Jamestown, NY
2   Jefferson Co, NY
4   Johnson City-Kingsport-Bristol, TN
3   Johnstown, PA
5   Kalamazoo-Battle Creek, MI          86
3   Kent and Queen Anne's Co, MD
9   Kern Co (Eastern Kern), CA
5   Kewaunee Co, WI
4   Knoxville, TN
5   La Porte Co, IN
3   Lancaster, PA
5   Lansing-East Lansing, MI
9   Las Vegas, NV
5   Lima,  OH                          89
9   Los Angeles South Coast Air Basin, CA
9   Los Angeles-San Bernardino (W Mojave),CA
4,5 Louisville, KY-IN
4   Macon, GA
    Madison and Page Cos (Shenandoah NP), VA
    Manitowoc Co, WI
    Mariposa and Tuolumne Cos, CA (S. Mtn Cos) 91
88
93
85
99
89
95
88
89
93
87
94
88
85
88
102
91
87
87
96
85
94
97
86
87
102 <
95
98
93
92
93
92
86
86
108 <
131
106
92
86
87
90
) 91
106
114
106
140
110
139
102
107
121
114
120
111
104
105
175
115
109
142
119
100
115
121
110
106
•iubpai
122
118
110
114
135
124
102
107
jubpai
180
138
120
113
104
110
113
                             Subpart 1
                             Subpart 1
                             EAC Subpart 1
                             Subpart 2 Moderate
                             Subpart 1
                             Subpart 2 Moderate
                             Subpart 1
                             Subpart 1
                             EAC Subpart 2 Moderate
                             EAC Subpart 1
                             Subpart 1
                             Subpart 1
                             Subpart 1
                             EAC Subpart 1
                              Subpart 2 Moderate
                             Subpart 1
                             Subpart 1
                             Subpart 2 Marginal
                             Subpart 1
                             Subpart 1
                             Subpart 1
                             Subpart 2 Moderate
                             EAC Subpart 1
                             Subpart 1
                             tl
                             Subpart 2 Moderate
                             Subpart 1
                             Subpart 1
                             Subpart 1
                             Subpart 2 Moderate
                             Subpart 2 Moderate
                             Subpart 1
                             Subpart 1
                             11
                              Subpart 2 Severe 17
                              Subpart 2 Moderate
                             Subpart 1
                             Subpart 1
                             Subpart 1
                             Subpart 1
                             Subpart 1
                                       2-99

-------
Final Regulatory Impact Analysis
EPA
Region Area Name
Design Value ppb (2001-2003 data)
       8-Hr   1-Hr  Category/Classification
5
4,6
5
5
4
5
4
9
2,1
3
3,5
3,2
9
O
1
2
1
4
O
O
9
O
2
4
9
6
9
9
9
O
5
5
1
7,5
O
5,3
9
5
3
5
9
Mason Co, MI
Memphis, TN-AR
Milwaukee-Racine, WI
Muncie, IN 88
Murray Co (Chattahoochee Nat Forest), GA
Muskegon, MI
Nashville, TN 86
Nevada Co, CA (Western Portion)
New York-N. N -Long Island,NY-NJ-CT
Norfolk- Virginia Beach-Newport News,VA
Parkersburg-Marietta, WV-OH
Philadelphia- Wilmin-Atl.City,PA-NJ-MD-DE
Phoenix-Mesa, AZ
Pittsburgh-Beaver Valley, PA
Portland, ME
Poughkeepsie, NY
Providence (All RI), RI
Raleigh-Durham-Chapel Hill, NC
Reading, PA
Richmond-Petersburg, VA
Riverside Co, (Coachella Valley), CA
Roanoke, VA
Rochester, NY
Rocky Mount, NC
Sacramento Metro, CA
San Antonio, TX
San Diego, CA
San Francisco Bay Area, CA
San Joaquin Valley, CA
Scranton-Wilkes-Barre, PA
Sheboygan, WI
South Bend-Elkhart, IN
Springfield (Western MA), MA
St Louis, MO-IL
State College, PA
Steubenville-Weirton, OH-WV
Sutler Co, CA (Sutler Buttes)
Terre Haute, IN
Tioga Co, PA
Toledo, OH
Ventura Co, CA
89
92
101
104
85
95
107
98
102
90
87
106
87
94
91
94
95
94
91
94
108
85
88
89
107
89
93
86
115
86
100
93
94
92
88
86
88
87
86
93
95
114
126
134
Subpai
103
121
EACS
116
146
121
113
133
111
120
126
126
130
118
116
131
133
107
110
106
143
119
118
123
151
108
124
116
132
122
109
113
113
108
102
112
124
                                                       Subpart 1
                                                       Subpart 2 Moderate
                                                        Subpart 2 Moderate
                                                       tl
                                                       Subpart 1
                                                       Subpart 2 Moderate
                                                       ubpart 1
                                                       Subpart 1
                                                       Subpart 2 Moderate
                                                       Subpart 2 Marginal
                                                       Subpart 1
                                                       Subpart 2 Moderate
                                                       Subpart 1
                                                       Subpart 1
                                                       Subpart 2 Marginal
                                                       Subpart 2 Moderate
                                                       Subpart 2 Moderate
                                                       Subpart 1
                                                       Subpart 1
                                                       Subpart 2 Moderate
                                                       Subpart 2 Serious
                                                       EAC Subpart 1
                                                       Subpart 1
                                                       Subpart 1
                                                       Subpart 2 Serious
                                                       EAC Subpart 1
                                                       Subpart 1
                                                       Subpart 2 Marginal
                                                       Subpart 2 Serious
                                                       Subpart 1
                                                        Subpart 2 Moderate
                                                       Subpart 1
                                                       Subpart 2 Moderate
                                                       Subpart 2 Moderate
                                                       Subpart 1
                                                       Subpart 1
                                                       Subpart 1
                                                       Subpart 1
                                                       Subpart 1
                                                       Subpart 1
                                                       Subpart 2 Moderate
                                       2-100

-------
                                             Air Quality, Health, and Welfare Effects
EPA
Region Area Name
Design Value ppb (2001-2003 data)
       8-Hr  1-Hr   Category/Classification
3   Washington Co (Hagerstown), MD
3   Washington, DC-MD-VA
3,5 Wheeling, WV-OH
3   York, PA
5,3 Youngstown-Warren-Sharon, OH-PA
                    86   109  EAC Subpart 1
                    99   140  Subpart 2 Moderate
                    87   111  Subpart 1
                    89   114  Subpart 1
                    95   118  Subpart 1
Boston-Manchester-Portsmouth(SE),NH has the same classification as Boston-Lawrence-
Worcester (E. MA), MA. Fredericksburg, VA has the same classification as Washington,
DC-MD-VA.

The level of the 8-hour ozone (O3) National Ambient Air Quality Standards (NAAQS) is 0.08
parts per million (ppm). The air quality design value for the 8-hour O3 NAAQS is the 3-year
average of the annual 4th highest daily maximum 8-hour average O3 concentration. The 8-hour
O3 NAAQS is not met when the 8-hour ozone design value is greater than 0.08 ppm (85 parts per
billion [ppb] rounds up).  Therefore, an area with a design value of 85 ppb does not meet the
NAAQS.

An area with a 1-hour design value of 120 ppb or lower is in a Subpart 1 category and must
attain the standard by up to 5 years after designation and they may apply for an extension of up
to 5 years.

Areas classified under Subpart 2 must attain the standards by the following attainment dates:

•  Marginal up to 3 years,
•  Moderate up to 6 years,
   Serious up to 9 years,
•  Severe up to 15 or 17 years,
•  Extreme up to 20 years.
                                        2-101

-------
Final Regulatory Impact Analysis
   2.3.2.2.2 Risk of Future 8-Hour Ozone Violations

   Our air quality modeling shows that there will continue to be a need for reductions in ozone
concentrations in the future without additional controls. In this section we describe the air
quality modeling including the non-emission inventory inputs. (See Chapter 3.6 summarizes the
emission inventory inputs.) We then discuss the results of the modeling for baseline conditions
absent additional control of nonroad diesel engines.

   We have also used our air quality modeling to estimate the change in future ozone levels that
would result from reductions in emissions from nonroad diesel engines. For this propose rule we
modeled a preliminary control scenario that illustrates the likely emission reductions. Because
of the substantial lead time to prepare the complex air quality modeling analyses, it was
necessary to develop a control options early in the proposal process based on our best judgment
at that time. Based on public comment and as additional  data regarding technical feasibility and
other factors became available, our judgment about the controls that are feasible has evolved.
Thus, the preliminary control option differs from what we are finalizing, as summarized in
Section 3.6 below.N It is important to note that these changes would not affect our estimates of
the baseline conditions without additional controls from nonroad diesel engines. This final rule
would produce nationwide air quality improvements in ozone levels, and we present the modeled
improvements in this section. Those interested in greater detail should review the AQ Modeling
TSD, which is available in the docket to this rule.

   2.3.2.2.3 Ozone Modeling Methodology, Domains and Simulation Periods

   In conjunction with this rulemaking, we performed a  series of ozone air quality modeling
simulations for the Eastern and Western United States using Comprehensive Air Quality Model
with Extension (CAMx).  The model simulations were performed for five emission scenarios: a
1996 baseline projection, a 2020 baseline projection and a 2020 projection with nonroad
controls, a 2030  baseline projection and a 2030 projection with nonroad controls.

   The model outputs from the 1996, 2020  and 2030 baselines, combined with current air
quality data, were used to identify areas expected to exceed the ozone NAAQS in 2020 and
2030. These  areas became candidates for being determined to be residual exceedance areas that
will require additional emission reductions to attain and maintain the ozone NAAQS. The
impacts of the new emission standards were determined by comparing the model results in the
future year control runs against the baseline simulations of the same year.  This modeling
supports the conclusion that there is a broad set of areas with predicted ozone concentrations at
or above 0.085 ppm between 1996 and 2030 in the baseline scenarios without additional
emission reductions.
   NBecause of the complexities and non-linear relationships in the air quality modeling, we are not attempting to
make any adjustments to the results.  Instead, we are presenting the results for the preliminary control option with
information about how the emission changes relate to what was modeled.

                                          2-102

-------
   The air quality modeling performed for this rule was based upon the same modeling system
as was used in the EPA's air quality assessment of the Clear Skies legislation with the addition
of updated inventory estimates for 1996, 2020 and 2030.  Further discussion of this modeling,
including evaluations of model performance relative to predicted future air quality, is provided in
the AQ Modeling TSD.

   CAMx was utilized to estimate base and future-year ozone concentrations over the Eastern
and Western United States for the various emission scenarios.  CAMx simulates the numerous
physical and chemical processes involved in the formation, transport, and destruction of ozone.
CAMx is a photochemical grid model that numerically simulates the effects of emissions,
advection, diffusion, chemistry, and surface removal processes on pollutant concentrations
within a three-dimensional grid.  This model is commonly used for purposes of determining
attainment/nonattainment as well as estimating the ozone reductions expected to occur from a
reduction in emitted pollutants.  The following sections provide an overview of the ozone
modeling completed as part of this rulemaking. More detailed information is included in the AQ
Modeling TSD, which is located in the docket for this rule.

   The regional ozone analyses used the modeling domains used previously for OTAG and the
highway passenger vehicle Tier 2 rulemaking.  The Eastern modeling domain encompasses the
area from the East coast to mid-Texas and consists of two grids with differing resolutions. The
model resolution was 36 km over the outer portions of the domain and 12 km in the  inner portion
of the grids.  The vertical height of the eastern modeling domain is 4,000 meters above ground
level with 9 vertical layers. The western modeling domain encompasses the area west of the 99th
degree longitude (which runs through North and South Dakota, Nebraska, Kansas, Oklahoma,
and Texas) and also consists of two grids with differing resolutions. The vertical height of the
western modeling domains is 4,800 meters above ground level with 11 vertical layers. As for the
Eastern United States, the model resolution was 36 km over the outer portions of the domain and
12 km in the inner portion of the grids.

   The simulation  periods modeled by CAMx included several multi-day periods when ambient
measurements were representative of ozone episodes over the Eastern and Western United
States. A simulation period, or episode, consists of meteorological data characterized over a
block of days that are used as inputs to the air quality model.  Three multi-day meteorological
scenarios during the summer of 1995 were used in the model simulations over the Eastern United
States: June 12-24,  July 5-15, and August 7-21. Two multi-day meteorological scenarios during
the summer of 1996 were used in the model simulations over the Western United States: July 5-
15 and July 18-31.  In general, these episodes do not represent extreme ozone events but, instead,
are generally  representative of ozone levels near local design values.  Each of the five emission
scenarios (1996 base year, 2020 base, 2020 control, 2030 baseline, 2030 control) were simulated
for the selected episodes.

   The meteorological data required for input into CAMx (wind, temperature, vertical mixing,
etc.) were developed by separate meteorological models.  For the Eastern United States, the
gridded meteorological data for the three historical 1995 episodes were developed using the
Regional Atmospheric Modeling System (RAMS), version 3b. This model provided needed data
at every grid cell on an hourly basis. For the Western United States, the gridded meteorological
data for the two historical 1996 episodes were developed using the Fifth-Generation National
Center for Atmospheric Research (NCAR) / Penn State Mesoscale Model (MM5).  These

-------
Final Regulatory Impact Analysis
meteorological modeling results were evaluated against observed weather conditions before
being input into CAMx and it was concluded that the model fields were adequate representations
of the historical meteorology. A more detailed description of the settings and assorted input files
employed in these applications is provided in the AQ Modeling TSD, which is located in the
docket for this rule.

   The modeling assumed background pollutant levels at the top and along the periphery of the
domain as in Tier 2. Additionally, initial conditions were assumed to be relatively clean as well.
Given the ramp-up days and the expansive domains, it is  expected that these assumptions will
not affect the modeling results, except in areas near the boundary (e.g., Dallas-Fort Worth TX).
The other non-emission CAMx inputs (land use, photolysis rates, etc.) were developed using
procedures  employed in the highway light duty Tier 2/OTAG regional modeling. The
development of model inputs is discussed in greater detail in the AQ Modeling TSD, which is
available in the docket for this rule.

   2.3.2.2.4 Model Performance Evaluation

   The purpose of the base year photochemical ozone modeling was to reproduce the
atmospheric processes resulting in the observed ozone concentrations over these domains and
episodes. One of the fundamental assumptions in air quality modeling is that a model that
adequately  replicates observed pollutant concentrations in the base year can be used to assess the
effects of future-year emission controls.

   A series of performance statistics was calculated for both model domains, the four quadrants
of the eastern domain, and multiple subregions in the eastern and western domains. Table 2.3-2
summarizes the performance statistics.  The model performance evaluation consisted solely of
comparisons against ambient surface ozone data. There was insufficient data available in terms
of ozone precursors or ozone aloft to allow for a more complete assessment of model
performance. Three primary statistical metrics were used to assess the overall accuracy of the
base year modeling simulations.

•  Mean normalized bias is defined as the average difference between the hourly model
   predictions and observations (paired in space and time) at each monitoring location,
   normalized by the magnitude  of the observations.

•  Mean normalized gross error is defined as the average absolute difference between the
   hourly model predictions and observations (paired in  space and time) at each monitoring
   location, normalized by the magnitude of the observations.

•  Average accuracy of the peak is defined as the average difference between peak daily model
   predictions and observations at each monitoring location, normalized by the magnitude of the
   observations.
                                         2-104

-------
                                              Air Quality, Health, and Welfare Effects
   In general, the model tends to underestimate observed ozone, especially in the modeling over
the Western United States, as shown in Table 2.3-3.  When all hourly observed ozone values
greater than a 60 ppb threshold are compared with their model counterparts for the 30 episode
modeling days in the eastern domain, the mean normalized bias is -1.1 percent and the mean
normalized gross error is 20.5  percent.  When the same statistics are calculated for the 19
episode days in the western domain, the bias is -21.4 percent and the error is 26.1 percent.

                                       Table 2.3-3
          Model Performance Statistics for the CAMx Ozone Predictions: Base Case
Region
Eastern U.S.
Western U.S.
Episode
June 1995
July 1995
August 1995
July 1996
Average Accuracy
of the Peak
-7.3
-3.3
9.6
-20.5
Mean Normalized
Bias
-8.8
-5.0
8.6
-21.4
Mean Normalized
Gross Error
19.6
19.1
623.3
26.1
   At present, there are no guidance criteria by which one can determine if a regional ozone
modeling exercise is exhibiting adequate model performance. These base case simulations were
determined to be acceptable based on comparisons to previously completed model rulemaking
analyses (e.g., Ozone Transport Assessment Group (OTAG), the light-duty passenger vehicle
Tier-2 standards, and on highway Heavy-Duty Diesel Engine 2007 standards). The modeling
completed for this rule exhibits less bias and error than any past regional ozone modeling
application done by EPA. Thus, the model  is  considered appropriate for use in projecting
changes in future year ozone concentrations and the resultant health and economic benefits due
to the anticipated emission reductions.

   2.3.2.2.5 Results of Photochemical Ozone  Modeling: Areas at Risk of Future 8-Hour
   Violations

   This section summarizes the results of our modeling of ozone air quality impact in the future
of reductions in nonroad diesel emissions.  Specifically, it provides information on our
calculations of the number of people estimated to live in counties in which ozone monitors are
predicted to exceed design values or to be within 10 percent of the design value in the future.
We also provide specific information about the number of people who would repeatedly
experience levels of ozone of potential concern over prolonged periods, i.e., over 0.085 ppm
ozone 8-hour concentrations over a number of days.

   The determination that an area is at risk of exceeding the ozone standard in the future was
made for all areas with current design values greater than or equal to 0.085 ppm (or within a 10
percent margin) and with modeling evidence that concentrations at and above this level will
persist into the future. The following sections provide background on methods for analysis of
                                         2-105

-------
attainment and maintenance. Those interested in greater detail should review the AQ TSD and
AQ Modeling TSD, both of which are available in the docket to this rule.

   The relative reduction factor method was used for interpreting the future-year modeling
results to determine where nonattainment is expected to occur in the 2020 and 2030 control
cases. The CAMx simulations were completed for base cases in 1996, 2020, and 2030
considering growth and expected emission controls that will affect future air quality.  The effects
of the nonroad engine reductions (control cases) were modeled for the two future years. As a
means of assessing the future levels of air quality with regard to the ozone NAAQS, future-year
estimates of ozone design values were calculated based on relative reduction factors (RRF)
between the various baselines and 1999-2001 ozone design values. The procedures for
determining the RRFs are similar to those in EPA's draft guidance for modeling for an 8-hour
ozone standard.320  Hourly model predictions were processed to determine daily maximum 8-
hour concentrations for each grid cell for each non-ramp-up day modeled. The RRF for a
monitoring site was determined by first calculating the multi-day  mean of the 8-hour daily
maximum predictions in the nine grid cells surrounding the site using only those predictions
greater than or equal to 70 ppb, as recommended in the guidance.°'321 This calculation was
performed for the base year scenario and each of the future-year baselines.  The RRF  for a site is
the ratio of the mean prediction in the future-year scenario to the mean prediction in the base
year scenario. RRFs were calculated on a site-by-site basis. The future-year design value
projections were then calculated by county, based on the highest resultant design values for a site
within that county from the RRF application.

   Based upon our air quality modeling for this rule, we anticipate that without emission
reductions beyond those already required under promulgated regulation and approved SIPs,
ozone nonattainment will likely persist into the future. With reductions from programs already
in place (but excluding the emission reductions from this rule), the number of counties violating
the ozone 8-hour standard is expected to decrease in 2020 to 30 counties where 43 million
people are projected to live.322 Thereafter, exposure to unhealthy levels of ozone is expected to
increase again.  In 2030 the  number of counties  violating the ozone 8-hour NAAQS, without
considering the emission reductions from this rule, is projected to increase to 32 counties where
47 million people are projected to live.

   EPA is still developing the implementation process for bringing the nation's air into
attainment with the ozone 8-hour NAAQS (see proposal, 68 FR 32702, June 2, 2003,  that was
recently finalized www.epa.gov/ozonedesignations) as described above.  Since the VOC and
NOx emission reductions expected from this final rule will go into effect during the period when
areas will need to attain the  8-hour ozone NAAQS, the projected reductions in nonroad diesel
emissions are expected to assist  States and local agencies in their effort to meet and maintain that
standard.  Many states mentioned this need in their public comments.  The following are sample
comments from states and state associations on the proposed rule, which corroborate that this
rule is  a critical element in States' NAAQS attainment efforts. Fuller information can be found
in the Summary and Analysis of Comments.
   °For the one-hour NAAQS we used a cut-off of 80 ppb. Please see the Highway Passenger
Vehicle Tier 2 Air Quality Modeling TSD for more details (EPA 1999b).

-------
   - "Unless emissions from nonroad diesels are sharply reduced, it is very likely that many
   areas of the country will be unable to attain and maintain health-based NAAQS for ozone
   and PM." (STAPPA/ALAPCO)
   - "Adoption of the proposed regulation ... is necessary for the protection of public health in
   California and to comply with air quality standards." (California Air Resources Board)
   - "Attainment of the NAAQS for ozone and PM25 is of immediate concern to the states in the
   northeast region....Thus, programs ... such as the proposed rule for nonroad diesel engines are
   essential." (NESCAUM)

       Furthermore, the inventories that underlie the ozone modeling conducted for this
rulemaking included emission reductions from all current or committed federal, State, and local
controls and, for the control case, including this rulemaking. There was no attempt to examine
the prospects of areas attaining or maintaining the ozone standard with possible future controls
(i.e., controls beyond current or committed federal, State, and local controls).  Tables 2.2-4 and
2.2-5 below should therefore be interpreted as indicating what counties are at risk of ozone
violations in 2020 or 2030 without additional federal or State measures that may be adopted  and
implemented after this rulemaking is finalized.  We expect many of the areas listed in Table
2.2-4 to adopt additional emission reduction programs, but we are unable to quantify or rely
upon future reductions from additional State programs since they have not yet been adopted.

   Since the emission reductions expected from this final rule begin in the same time period in
which areas will need reductions to attain by their attainment dates, the projected reductions in
nonroad emissions will be extremely important to States in meeting the new NAAQS. In public
comment, many States and local agencies commented that they will be relying on such nonroad
reductions to help them attain and maintain the 8-hour NAAQS.  Furthermore, since the nonroad
emission reductions will continue to grow in the years beyond 2014, they will also be important
for maintenance of the NAAQS for areas with attainment dates of 2014 and earlier.

   On a population-weighted basis, the average change in future year design values  would be a
decrease of 1.8  ppb in 2020, and 2.5 ppb in 2030. Within nonattainment areas, the population-
weighted average decrease would be somewhat higher: 1.9 ppb in 2020 and 3 ppb in 2030.p  In
terms of modeling accuracy, the count of modeled nonattaining counties is much less certain
than the average changes in air quality.  For example, actions by states to meet their SIP
obligations would not be expected to significantly change the overall concentration changes
induced by this final rule, but they could substantially change the number of counties in or out of
attainment. If state actions resulted in an increase in the number of areas that are very close to,
but still above, the NAAQS, then this rule might bring many of those counties down sufficiently
to change their attainment status. On the other hand, if state actions brought several  counties we
project to be very close to the standard in the future down sufficiently to reach attainment status,
then the air quality improvements from this rule might change the actual attainment status of
very few counties.  Bearing this limitation in mind, our modeling indicates that the nonroad
diesel emission reductions will decrease the net number of nonattainment counties by 2 in 2020
and by 4 in 2030, without consideration of new state or local programs.
   FThis is in spite of the fact that NOx reductions can at certain times in some areas cause ozone levels to
increase.  Such "disbenefits" are observed in our modeling, but these results make clear that the overall effect of this
final rule is positive.

-------
Final Regulatory Impact Analysis
    This air quality modeling suggests that without emission reductions beyond those already
required under promulgated regulations and approved SIPs, ozone nonattainment will likely
persist into the future. With reductions from programs already in place, the number of counties
violating the ozone 8-hour standard is expected to decrease from today's levels to 30 counties in
2020 where 43 million people are projected to live.323 Thereafter, exposure to unhealthy levels
of ozone is expected to begin to increase again.  In 2030 the number of counties violating the
ozone 8-hour NAAQS is projected to increase to 32 counties where 47 million people are
projected to live.  In addition, in 2030, 82 counties where 44 million people are projected to live
will be within 10 percent of violating the ozone 8-hour NAAQS. Specifically, counties
presented in Table 2.3-3 and 2.3-4 have monitored 1999-2001 air quality dataQ and our modeling
predicts violations of the 8-hour ozone NAAQS,  or predicts concentrations within 10 percent of
the standard, in 2020 or 2030. The base case indicates conditions predicted without the
reductions from this rule, and the control case represents a preliminary control option similar to
the final rule, as described in section 3.6 of the RIA.

    In Table 2.3-4 we  list the counties with 2020  and 2030 projected 8-hour ozone design values
(4th maximum concentration) that violate the 8-hour standard.  Counties are marked with an "V"
in the table if their projected design values are greater than or  equal to 85 ppb. The 1999-2001
average design values of these counties are also listed. Recall that we project future design
values only for counties that have 1999-2001 design values, so this list is limited to those
counties with ambient monitoring data sufficient to calculate these design values.
   QSince the air quality modeling and analyses performed at proposal used the 1999-2001 monitored data set, we
present these data rather than the 2000-2002 data for consistency.

                                          2-108

-------
                                                   Air Quality, Health, and Welfare Effects
           Table 2.3-4: Counties with 2020 and 2030 Projected Ozone Design Values
                           in Violation of the 8-Hour Ozone Standard."
State
CA
CA
CA
CA
CA
CA
CA
CT
CT
CT
GA
GA
GA
IL
IN
MD
MI
MI
NJ
NJ
NJ
NJ
NJ
NJ
NJ
NY
NY
NY
PA
PA
TX
TX
WI
County
Fresno
Kern
Los Angeles
Orange
Riverside
San Bernardino
Ventura
Fairfield
Middlesex
New Haven
Bibb
Fulton
Henry
Cook
Lake
Harford
Macomb
Wayne
Camden
Gloucester
Hudson
Hunterdon
Mercer
Middlesex
Ocean
Bronx
Richmond
Westchester
Bucks
Montgomery
Galveston
Harris
Kenosha
1999-2001
Design Value
(mb)
108
109
105
77
111
129
101
97
99
97
98
107
107
88
90
104
88
88
103
101
93
100
105
103
109
83
98
92
105
100
98
110
95
Number of Violating Counties
Population of Violating Counties'5
2020
Base
V
V
V
V
V
V
V
V
V
V
V
V
V
V

V

V
V
V
V
V
V
V
V

V
V
V
V
V
V
V
30
42.930.060
Control"
V
V
V
V
V
V
V
V
V
V

V

V



V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
28
43.532.490
2030
Base
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V

V
V
V
V
V
V
V
32
46.998.413
Control"
V
V
V
V
V
V
V
V
V
V



V


V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
28
46.038.489
Population
in 2000
799,407
661,645
9,519,338
2,846,289
1,545,387
1,709,434
753,197
882,567
155,071
824,008
153,887
816,006
119,341
5,376,741
484,564
218,590
788,149
2,061,162
508,932
254,673
608,975
121,989
350,761
750,162
510,916
1,332,650
443,728
923,459
597,635
750,097
250,158
3,400,578
149.577


a The projected emission reductions differ based on updated information (see Chapter 3.6); however, the base results
    presented here would not change, but we anticipate the control case improvements would generally be smaller.
b Populations are based on 2020 and 2030 estimates from the U.S. Census.
                                              2-109

-------
Final Regulatory Impact Analysis
   In Table 2.3-5 we present the counties with 1999-2001 design values and 2020 and 2030
projected 8-hour ozone design values that are within 10 percent of it in either base or control
scenarios. Counties are marked with an "X" in the table if their projected design values are
greater than or equal to 77 ppb, but less than 85 ppb. Counties are marked with a "V" in the
table if their projected design values are greater than or equal to 85 ppb.  This list is limited to
those counties with ambient monitoring data sufficient to calculate these design values, and the
1999-2001 average design values of these counties are also presented.  Most of these are
counties are not projected to violate the standard, but their future  values are project to be close to
the standard. Thus, the final rule will help ensure that these counties continue to meet the
standard.

                                       Table 2.3-5
                Counties with 2020 and 2030 Projected Ozone Design Values
                     within Ten Percent of the 8-Hour Ozone Standard.3
State
AR
AZ
CA
CA
CA
CO
CT
DC
DE
GA
GA
GA
GA
GA
GA
GA
GA
IL
IN
IN
LA
LA
LA
LA
LA
LA
County
Crittenden
Maricopa
Kings
Merced
Tulare
Jefferson
New London
Washington
New Castle
Bibb
Coweta
DeKalb
Douglas
Fayette
Fulton
Henry
Rockdale
McHenry
Lake
Porter
Ascension
Bossier
Calcasieu
East Baton Rouge
Iberville
Jefferson
1999-2001
Design Value
ftrob')
92
85
98
101
104
81
90
94
97
98
96
102
98
99
107
107
104
83
90
90
86
90
86
91
86
89
2020
Base
X
X
X
X
X
X
X
X
X
V
X
X
X
X
V
V
X
X
X
X
X
X
X
X
X
X
Control"
X
X
X
X
X
X

X
X
X
X
X


V
X
X

X
X
X
X
X
X

X
2030
Base
X
X
X
X
X
X
X
X
X
V
X
X
X
X
V
V
X
X
V
X
X
X
X
X
X
X
Control"
X
X
X
X
X
X

X
X
X
X
X


X
X
X

X
X
X
X
X
X

X
Population
in 2000
50,866
3,072,149
129,461
210,554
368,021
527,056
259,088
572,059
500,265
153,887
89,215
665,865
92,174
91,263
816,006
119,341
70,111
260,077
484,564
146,798
76,627
98,310
183,577
412,852
33,320
455,466
                                          2-110

-------
    Air Quality, Health, and Welfare Effects
State
LA
LA
LA
LA
LA
MA
MA
MD
MD
MD
MD
MD
MD
MI
MI
MI
MI
MI
MI
MO
MO
MS
MS
MS
NJ
NJ
NJ
NJ
NY
NY
NY
NY
NY
OH
OH
PA
PA
PA
PA
PA
County
Livingston
St Charles
St James
St John The Ba
West Baton Rou
Barnstable
Bristol
Anne Arundel
Baltimore
Cecil
Harford
Kent
Prince Georges
Benzie
Macomb
Mason
Muskegon
Oakland
St Clair
St Charles
St Louis
Hancock
Harrison
Jackson
Cumberland
Monmouth
Morris
Passaic
Bronx
Erie
Niagara
Putnam
Suffolk
Geauga
Lake
Allegheny
Delaware
Lancaster
Lehigh
Northampton
1999-2001
Design Value
frjob')
88
86
83
86
88
96
93
103
93
106
104
100
97
89
88
91
92
84
85
90
88
87
89
87
97
94
97
89
83
92
87
89
91
93
91
92
94
96
96
97
2020
Base
X
X

X
X
X
X
X
X
X
V
X
X
X
X
X
X
X



X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
Control"
X
X

X
X


X
X
X
X

X

X

X
X




X
X

X
X
X
V
X


X



X

X
X
2030
Base
X
X
X
X
X
X
X
X
X
X
V
X
X
X
V
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
Control"
X
X

X
X


X
X
X
X



V


X




X
X

X
X
X
V
X


X



X

X
X
Population
in 2000
91,814
48,072
21,216
43,044
21,601
222,230
534,678
489,656
754,292
85,951
218,590
19,197
801,515
15,998
788,149
28,274
170,200
1,194,156
164,235
283,883
1,016,315
42,967
189,601
131,420
146,438
615,301
470,212
489,049
1,332,650
950,265
219,846
95,745
1,419,369
90,895
227,511
1,281,666
550,864
470,658
312,090
267,066
2-111

-------
Final Regulatory Impact Analysis
State
PA
RI
RI
TN
TX
TX
TX
TX
TX
TX
TX
VA
VA
VA
WI
WI
WI
WI
WI
WI
WI
WI
County
Philadelphia
Kent
Washington
Shelby
Brazoria
Collin
Dallas
Denton
Jefferson
Montgomery
Tarrant
Alexandria City
Arlington
Fairfax
Door
Kewaunee
Manitowoc
Milwaukee
Ozaukee
Racine
Sheboygan
Waukesha
1999-2001
Design Value
frjob')
88
94
92
93
91
99
93
101
85
91
97
88
92
95
93
89
92
89
95
87
95
86
Number of Counties within 10%
Population of Counties within 10%b
2020
Base
X
X
X
X
X
X
X
X
X
X
X

X
X
X
X
X
X
X
X
X
X
79
40.465.492
Control"
X
X

X
X
X
X
X
X

X

X
X
X

X
X
X

X

58
33.888.031
2030
Base
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
82
44.013.587
Control"
X


X
X
X
X
X
X
X
X

X
X
X


X
X

X

54
35.631.215
Population
in 2000
1,517,550
167,090
123,546
897,472
241,767
491,675
2,218,899
432,976
252,051
293,768
1,446,219
128,283
189,453
969,749
27,961
20,187
82,887
940,164
82,317
188,831
112,646
360.767


a The projected emission reductions differ based on updated information (see Section 3.6); however, the base results
    presented here would not change, but we anticipate the control case improvements would generally be smaller.
b Populations are based on 2020 and 2030 estimates from the U.S. Census.
   Based on our modeling, we are also able to provide a quantitative prediction of the number of
people anticipated to reside in counties in which ozone concentrations are predicted to for 8-hour
periods in the range of 85 to 120 ppb and higher on multiple days.  Our analysis relies on
projected county-level population from the U.S. Department of Census for the period
representing each year analyzed.324

   For each of the counties analyzed, we determined the number of days for periods on which
the highest model-adjusted 8-hour concentration at any monitor in the county was predicted, for
example, to be equal to or above 85 ppb.  We then grouped the counties that had days with ozone
in this range according to the number of days this was predicted to happen and summed their
projected populations.
                                           2-112

-------
                                              Air Quality, Health, and Welfare Effects
   In the base case (i.e., before the application of emission reductions resulting from this rule),
we estimated in 2020 that 53 million people are predicted to live in counties with at least 2 days
with 8-hour average concentrations of 85 ppb or higher.  This baseline will increase in 2030 to
56 million people are predicted to live in counties with at least 2 days with 8-hour average
concentrations of 85 ppb or higher. About 30 million people live in counties with at least 7 days
of 8-hour ozone concentrations at or above 85 ppb in 2020 and 2030 without additional controls.
Approximately 15 million people are predicted to live in counties with at least 20 days of 8-hour
ozone concentrations at or above 85 ppb in 2020  and 2030 without additional controls.325 Thus,
reductions in ozone precursors from nonroad diesel engines are needed to assist States in
meeting the ozone NAAQS and to reduce ozone exposures.

   2.3.2.3 Potentially Counterproductive Impacts on Ozone Concentrations from NOx
   Emission Reductions

   While this final rule will reduce ozone levels  generally and provide significant ozone-related
health benefits, this is not always the case at the local level.  Due to the complex photochemistry
of ozone production, NOx emissions lead to both the formation and destruction of ozone,
depending on the relative quantities of NOx, VOC, and ozone catalysts such  as the OH and HO2
radicals.  In areas dominated by fresh emissions of NOx, ozone catalysts are  removed via the
production of nitric acid, which slows the ozone formation rate.  Because NOx is generally
depleted more rapidly than VOC, this effect is usually short-lived and the emitted NOx can lead
to ozone formation later and further downwind.  The terms "NOx disbenefits" or "ozone
disbenefits" refer to the ozone increases that can result from NOx emission reductions in these
localized areas. According to the NARSTO Ozone Assessment, these disbenefits are generally
limited to small regions within specific urban cores and are surrounded by larger regions in
which NOx control is beneficial.326

   In the context of ozone disbenefits, some have postulated that present-day weekend
conditions serve as a demonstration of the effects of future NOx reduction strategies because
NOx emissions decrease more than VOC emissions on weekends, due to a disproportionate
decrease in the activity of heavy-duty diesel trucks and other diesel equipment. Recent research
indicates that ambient ozone levels are higher in some metropolitan areas on  weekends than
weekdays.327'328  There are other hypotheses for the cause of the "weekend effect."329 For
instance, the role of ozone and ozone precursor carryover from previous  days is difficult to
evaluate because of limited ambient data, especially aloft. The role of the changed timing of
emissions is difficult to evaluate because of limited ambient and emission inventory information.
It is also important to note that in many areas with "weekend effects" (e.g., Los Angeles and San
Francisco) significant ozone reductions have been observed over the past 20  years for all days of
the week, during a period in which both NOx and VOC emissions have been greatly reduced.

   We received some public comments that in some cities, decreased motor  vehicle traffic
(particularly diesels) results in a higher VOC/NOx ratio which, in airsheds that are VOC-limited,
can result in higher ozone concentrations. EPA's air quality modeling predicts NOx disbenefits
in the areas identified by some studies as "VOC-limited" (e.g., Los Angeles). However, these

                                         2-113

-------
Final Regulatory Impact Analysis
areas represent a small minority of the area in the United States.  While some empirical studies
to date point to a weekend ozone effect related to NOx reduction, modeling conducted for this
rule predicts that this rule will result in net gains in benefits as a result of reduced ozone and
PM2 5 related to NOx.

    EPA maintains that the best available approach for determining the value of a particular
emission reduction strategy is the net air quality change projected to result from the rule,
evaluated on a nationwide basis and for all pollutants that are health and/or welfare concerns.
The primary tool for assessing the net impacts of this rule are the air quality simulation
models.330 Model scenarios of 2020 and 2030 with and without the emission controls from this
rulemaking are compared to determine the expected changes in future pollutant levels resulting
from the rule.  There are several factors related to the air quality modeling and inputs that should
be considered regarding the disbenefit issue.  First, our future year modeling does not contain
any local governmental actions beyond the controls in this rule. It is possible that significant
local controls of VOC and/or NOx could modify the conclusions regarding ozone changes in
some areas.  Second, the modeled NOx reductions are greater than those actually included in the
analysis to quantify the emission reductions resulting from the final rule (see Section 3.6 for
more detail).  This could lead to an exaggeration of the benefits and disbenefits expected to
result from the rule. Also, recent work by California ARB has indicated that model limitations
and uncertainties may lead to overestimates of ozone disbenefits attributed to NOx emission
reductions. While EPA maintains that the air quality simulations conducted for the rule
represent state-of-the-science analyses, any changes to the underlying chemical mechanisms,
grid resolution, and emissions/meteorological inputs could result in revised conclusions
regarding the strength and frequency of ozone disbenefits.

    A wide variety of ozone metrics were considered in assessing the emission reductions.  Three
of the most important assessments are: 1) the effect of the rule on projected future-year ozone
violations, 2) the effect of the rule in  assisting local areas in attainment and maintenance of the
NAAQS, and 3) an economic assessment of the rule benefits based on existing health studies.
Additional metrics for assessing the air quality effects are discussed in the TSD for the modeling.

    Based only on the reductions from this rule, our modeling predicts that periodic ozone
disbenefits will occur most frequently in New York City, Los Angeles, and Chicago.  Smaller
and less frequent disbenefits also occur in Boston, Detroit, and San Francisco.  As described
below, despite these localized increases, the net ozone impact of the rule nationally is positive
for the majority of the analysis metrics. Even within the few metropolitan areas that experience
periodic ozone increases, these disbenefits are infrequent relative to the benefits accrued at
ozone levels above the NAAQS. Furthermore, and most importantly, the overall air quality
impact of this final rule is  projected to be strongly positive due to the expected reductions in fine
PM.

    The projected net impact of the rule on 8-hour ozone violations in 2020 is that three counties
will no longer violate the NAAQS.331 Conversely, one county in the New York City CMSA
(Bronx County), which is currently not in violation of the NAAQS, is projected to violate the

                                          2-114

-------
                                               Air Quality, Health, and Welfare Effects
standard in 2020 as a result of the rule.  The net effect is a projected 1.4 percent increase in the
population living in violating counties.  It is important to note that ozone nonattainment
designations are historically based on larger geographical areas than counties (e.g., see public
comments from New York Department of Environmental Conservation requesting that EPA use
metropolitan areas instead of counties for its analyses for this reason).  Bronx County, NY is the
only county within the New York City CMSA in which increases are detected in 8-hour
violations in 2020. Considering a larger area, the modeling indicates that projected violations
over the entire New York City CMSA will be reduced by 6.8 percent.  Upon full turnover of the
fleet in 2030, the net impact of the rule on projected 8-hour ozone violations is a 2.0 percent
decrease in the population living in violating counties as two additional counties are no longer
projected to violate the NAAQS. The net impact of the rule on projected  1-hour ozone
violations is to eradicate projected violations from four counties (in both 2020 and 2030),
resulting in a 10.5 percent decrease in the population living in violating counties.

   Another way to assess the air quality impact of the rule is to calculate its effect on all
projected future year design values concentrations, as opposed to just those that cross the
threshold of the NAAQS. This metric helps assess the  degree to which the rule will assist local
areas in attaining and/or maintaining the NAAQS. Future year design values were calculated for
every location for which complete ambient monitoring data existed for the period 1999-2001.
These present-day design values were then projected by using the modeling projections (future
base vs. future control) in a relative sense. For the 1999-2001 monitoring period, there were
sites in 522 counties for which 8-hour design values could be calculated and sites in 510 counties
for which  1-hour design values could be calculated.

   Table 2.3.2-1 shows  the average change in future year eight-hour and one-hour ozone design
values.  Average changes are shown 1) for all counties with design values in 2001, 2) for
counties with design values that did not meet the standard in 1999-2001 ("violating" counties),
and 3) for counties that met the standard, but were within 10 percent of it in 1999-2001. This
last category is intended to reflect counties that meet the standard, but will likely benefit from
help in maintaining that status in the face of growth.  The average and population-weighted
average over all counties in Table 2.3.2-1 demonstrates a broad improvement in ozone air
quality. The average across violating counties shows that the rule will help bring these counties
into attainment.  The average over counties within ten percent of the standard shows that the rule
will also help those counties to maintain the standard. All of these metrics show a decrease in
2020 and a larger decrease in 2030 (due to fleet turnover), indicating in four different ways the
overall improvement in ozone air quality as measured by attainment of the NAAQS.
                                          2-115

-------
Final Regulatory Impact Analysis
                                          Table 2.3.2-1
                Average Change in Projected Future-Year Ozone Design Valuef
Design Value
8-Hour
1-Hour
Average*
All
All, population-weighted
Violating counties'5
Counties within 10
percent of the standard0
All
All, population-weighted
Violating counties'1
Counties within 10
percent of the standard6
Number of
Counties
522
522
289
130
510
510
73
130
2020 Control
minus Base (ppb)
-1.8
-1.6
-1.9
-1.7
-2.4
-2.3
-2.9
-2.4
2030 Control minus
Base (ppb)
-2.8
-2.6
-3
-2.6
-3.8
-3.6
-4.5
-3.8
a Averages are over counties with 2001 design values.
b Counties whose present-day design values exceeded the 8-hour standard (> 85 ppb).
0 Counties whose present-day design values were less than but within 10 percent of the 8-hour standard (77 125 ppb).
e Counties whose present-day design values were less than but within 10 percent of the 1-hour standard
    (112
-------
                                               Air Quality, Health, and Welfare Effects
                                       Table 2.3.2-2
                          Numbers of Counties Projected to Be in
        Different Design-Value Change Bins in 2020 and 2030 as a Result of the Rule3
Design value
change
> 2ppb increase
1 ppb increase
No change
1 ppb decrease
2-3 ppb decrease
4 ppb decrease
Total
2020
8-Hour
1
1
21
140
357
2
522
1-Hour
1
5
10
69
356
69
510
2030
8-Hour
1
3
10
42
333
133
522
1-Hour
1
2
5
22
193
287
510
1 The analysis in Chapter 3 differs based on updated information; however, we believe that the net results would
    approximate future emissions, although we anticipate the design value improvements would generally be slightly
    smaller.
   A third way to assess the impacts of the rule is an economic consideration of the economic
benefits. Benefits related to changes in ambient ozone are expected to be positive for the nation
as a whole. However, for certain health endpoints associated with longer ozone-averaging times,
such as minor restricted activity days related to 24-hour average ozone, the national impact may
be small or even negative. This is due to the forecasted increases in ozone for certain hours of
the day in some urban areas. Many of the increases occur during hours when baseline ozone
levels are low, but the benefits estimates rely on the changes in ozone along the full distribution
of baseline ozone levels, rather than changes occurring only above a particular threshold.  As
such, the benefits estimates are more sensitive to increases in ozone occurring due to the "NOx
disbenefits" effect described above. For more details on the economic effects of the rule, please
see Chapter 9: Public Health and Welfare Benefits.

   Historically, NOx reductions have been very successful at reducing regional and national
ozone levels. Consistent with that fact, the photochemical modeling completed for this rule
indicates that the projected emission reductions will significantly assist in the attainment and
maintenance of the ozone NAAQS at the national level.  Furthermore, NOx reductions also
result in reductions in PM and its associated health and welfare effects. This rule is one aspect
of overall emission reductions that States, local governments,  and Tribes need to reach their
clean air goals. It is expected that future state, local and national controls that decrease VOC,
CO,  and regional ozone will mitigate any localized disbenefits. EPA will continue to rely on
local attainment measures to ensure that the NAAQS  are not violated in the  future.  Many
organizations with an interest in improved air quality have supported the rule because they
believe the resulting NOx reductions will reduce both ozone and PM.332 EPA believes that a
                                          2-117

-------
Final Regulatory Impact Analysis
balanced air quality management approach that includes NOx emission reductions from nonroad
engines is needed as part of the nation's progress toward clean air.

2.3.3 Welfare Effects Associated with Ozone and its Precursors

    There are a number of significant welfare effects associated with the presence of ozone and
NOX in the ambient air.333 Because this rule will reduce ground-level ozone and nitrogen
deposition, benefits are expected to accrue to the welfare effects categories described in the
following paragraphs.

    2.3.3.1 Ozone-related welfare effects.

    The Ozone Criteria Document notes that "ozone affects vegetation throughout the United
States, impairing crops, native vegetation, and ecosystems more than any other air pollutant."334
Like carbon dioxide (CO2) and other gaseous substances, ozone enters plant tissues primarily
through apertures (stomata) in leaves in a process called "uptake". To a lesser extent, ozone can
also diffuse directly through surface layers to the plant's interior.335 Once ozone, a highly
reactive substance, reaches the interior of plant cells, it inhibits or damages essential cellular
components and functions, including enzyme activities, lipids, and cellular membranes,
disrupting the plant's osmotic (i.e., water) balance and energy utilization patterns.336'337 This
damage is commonly manifested as visible foliar injury such as chlorotic or necrotic spots,
increased leaf senescence (accelerated leaf aging) and/or as reduced photosynthesis. All these
effects reduce a plant's capacity to form carbohydrates, which are the primary form  of energy
used by plants.338 With fewer resources available, the plant reallocates existing resources away
from root growth and storage, above ground growth or yield, and reproductive processes, toward
leaf repair and maintenance.  Studies have shown that plants stressed in these ways may exhibit a
general loss of vigor, which can lead to secondary impacts that modify plants' responses to other
environmental factors.  Specifically, plants may become more sensitive to other air pollutants,
more susceptible to disease, insect attack, harsh weather (e.g., drought, frost) and other
environmental stresses (e.g., increasing CO2 concentrations). Furthermore, there is considerable
evidence that ozone can interfere with the formation of mycorrhiza, essential symbiotic fungi
associated with the roots of most terrestrial plants, by reducing the amount of carbon available
for transfer from the host to the symbiont.339

    Not all plants, however, are equally sensitive to ozone.  Much of the variation in sensitivity
between individual plants or whole species is related to the plant's ability to regulate the extent
of gas exchange via leaf stomata (e.g., avoidance of O3 uptake through closure of stomata).340'341'
342  Other resistance mechanisms may involve the intercellular production of detoxifying
substances. Several biochemical substances capable of detoxifying ozone have been reported to
occur in plants including the antioxidants ascorbate and glutathione.  After injuries have
occurred,  plants may be capable of repairing the damage to a limited extent.343 Because of the
differing sensitivities among plants to ozone, ozone pollution can also exert a selective pressure
that leads to changes in plant community composition.  Given the range of plant sensitivities and
the fact that numerous other environmental factors modify plant uptake and response to ozone, it

                                          2-118

-------
                                              Air Quality, Health, and Welfare Effects
is not possible to identify threshold values above which ozone is toxic for all plants. However,
in general, the science suggests that ozone concentrations of 0.10 ppm or greater can be
phytotoxic to a large number of plant species, and can produce acute foliar injury responses, crop
yield loss and reduced biomass production. Ozone concentrations below 0.10 ppm (0.05 to 0.09
ppm) can produce these effects in more sensitive plant species, and have the potential over a
longer duration of creating chronic stress on vegetation that can lead to effects of concern
associated with reduced carbohydrate production and decreased plant vigor.

    The economic value of some welfare losses due to ozone can be calculated, such as crop
yield loss from both reduced seed production (e.g., soybean) and visible injury to some leaf
crops (e.g., lettuce, spinach, tobacco) and visible injury to ornamental plants (i.e., grass, flowers,
shrubs), while other types of welfare loss may not be fully quantifiable in  economic terms (e.g.,
reduced aesthetic value of trees growing in Class I areas).

    Forests and Ecosystems. Ozone also has been shown conclusively to cause discernible
injury to forest trees.344'345 In terms of forest productivity and ecosystem diversity, ozone may
be the pollutant with the greatest potential for regional-scale forest impacts.346 Studies have
demonstrated repeatedly that ozone concentrations commonly observed in polluted areas can
have substantial impacts on plant function.347'348'349

    Because plants are at the center of the food web in many ecosystems, changes to the plant
community can affect associated organisms and ecosystems (including the suitability of habitats
that support threatened or endangered species and below ground organisms living in the root
zone). Ozone damages at the community and ecosystem-level vary widely depending upon
numerous factors, including concentration and temporal variation of tropospheric ozone, species
composition, soil properties and climatic factors.350 In most instances, responses to chronic or
recurrent exposure are subtle and not observable for many years. These injuries can cause stand-
level forest decline in sensitive ecosystems.351'352'353 It is not yet possible to predict ecosystem
responses to ozone with much certainty; however, considerable knowledge of potential
ecosystem responses has been acquired through long-term observations in highly damaged
forests in the United States.

    Given the scientific information establishing that ambient ozone levels cause visible injury to
foliage of some sensitive  forest species,354 there is a corresponding loss of public welfare from
reduced aesthetic properties of forests.355 However, present analytic tools and resources preclude
EPA from quantifying the benefits of improved forest aesthetics.

    Agriculture. Laboratory and field experiments have shown reductions in yields for
agronomic crops exposed to ozone, including vegetables (e.g., lettuce) and field crops (e.g.,
cotton and wheat).  The most extensive field experiments, conducted under the National Crop
Loss Assessment Network (NCLAN) examined 15 species and numerous  cultivars.  The
NCLAN results show that "several economically important crop species are sensitive to ozone
levels typical of those found in the Unites States."356  In addition, economic studies have shown a
relationship between observed ozone levels and crop yields.357 358 359

                                         2-119

-------
Final Regulatory Impact Analysis
   Urban Ornamentals. Urban ornamentals represent an additional vegetation category likely
to experience some degree of negative effects associated with exposure to ambient ozone levels
and likely to impact large economic sectors.  In the absence of adequate exposure-response
functions and economic damage functions for the potential range of effects relevant to these
types of vegetation, no direct quantitative analysis has been conducted.  It is estimated that more
than $20 billion (1990 dollars) are spent annually on landscaping using ornamentals, both by
private property owners/tenants and by governmental units responsible for public areas.360  This
is therefore a potentially important environmental effect. However, methods are not available to
allow for plausible estimates of the percentage of these expenditures that may be related to
impacts associated with ozone exposure.

   2.3.3.2 Nitrogen (NOx)-related welfare effects.

   Agriculture.  By reducing NOX emissions, this final rule will also reduce nitrogen deposition
on agricultural land and forests. There is  some evidence that nitrogen deposition may have
positive effects on agricultural output through passive fertilization. Holding all other factors
constant, farmers' and commercial tree growers use of purchased fertilizers or manure may
increase as deposited nitrogen  is reduced. Estimates of the potential value of this possible
increase in the use of purchased fertilizers are not available, but it is likely that the overall value
is very small relative to other health and welfare effects.  The share of nitrogen requirements
provided by this deposition is small, and the marginal cost of providing this nitrogen from
alternative sources is quite low. In some areas, agricultural lands suffer from nitrogen over-
saturation due to an abundance of on-farm nitrogen production, primarily from animal manure.
In these areas, reductions in atmospheric deposition of nitrogen represent additional agricultural
benefits.

   Forests and Ecosystems.  Information on the effects of changes in passive nitrogen
deposition on forests and other terrestrial  ecosystems is very limited. The multiplicity of factors
affecting forests, including  other potential stressors such as ozone,  and limiting factors such as
moisture and other nutrients, confound assessments of marginal changes in any one stressor or
nutrient in forest ecosystems. However, reductions in nitrogen deposition can have negative
effects on forest and vegetation growth in ecosystems where nitrogen is a limiting factor.361

   On the other hand, there is evidence that forest ecosystems in some areas of the United States
are already or are becoming nitrogen saturated.362 Once saturation is reached, adverse  effects of
additional nitrogen begin to occur, such as soil acidification, which can lead to leaching of
nutrients needed for plant growth and mobilization of harmful elements such as aluminum,
leading to reductions in tree growth or forest decline. Increased soil acidification is also linked
to higher amounts of acidic runoff to streams and lakes and leaching of harmful elements into
aquatic ecosystems, harming fish and other aquatic life.363

   The reductions in ground-level ozone  and nitrogen deposition that will result from  this rule
are expected to reduce the adverse impacts described above.  In particular, it is expected that
                                          2-120

-------
                                              Air Quality, Health, and Welfare Effects
economic impacts, such as those related to reduced crop yields and forest productivity, will be
reduced.

2.4 Carbon Monoxide

   This final rule will reduce levels of other pollutants for which NAAQS have been
established:  carbon monoxide (CO), nitrogen dioxide (NO2), and sulfur dioxide (SO2).
Currently every area in the United States has been designated to be in attainment with the NO2
NAAQS. As of August 27, 2003, there were 24 areas designated as nonattainment with the SO2
standard, and 11 designated CO nonattainment areas.  The rest of this section describes issues
related to CO.

   2.4.1  General Background

   Unlike many gases, CO is odorless, colorless, tasteless, and nonirritating. Carbon monoxide
results from incomplete combustion of fuel and is emitted directly from vehicle tailpipes.
Incomplete combustion is most likely to occur at low air-to-fuel ratios in the engine.  These
conditions are common during vehicle starting when air supply is restricted ("choked"), when
vehicles are not tuned properly, and at high altitude, where "thin" air effectively reduces the
amount of oxygen available for combustion (except in engines that are designed or adjusted to
compensate for altitude).  High concentrations of CO generally occur in areas with elevated
mobile-source emissions.  Carbon monoxide emissions increase dramatically in cold weather.
This is because engines need more fuel to start at cold temperatures and because some emission
control devices (such as oxygen sensors and catalytic converters) operate less efficiently when
they are cold.  Also, nighttime inversion conditions are more frequent in  the colder months of the
year.  This is due to the enhanced stability  in the atmospheric boundary layer, which inhibits
vertical mixing of emissions from the surface.

   As described in Chapter 3, nonroad diesel engines currently account for about one percent of
the national mobile source CO inventory.  EPA previously determined that the category of
nonroad diesel engines cause or contribute to ambient CO and  ozone in more than one
nonattainment area (65 FR 76790, December 7, 2000). In that action, EPA found that engines
subject to this final rule contribute to CO nonattainment in areas such as  Los Angeles, Phoenix,
Spokane, Anchorage, and Las Vegas. Nonroad land-based diesel engines emitted 1,004,600 tons
of CO in  1996 (1 percent of mobile source CO). Thus, nonroad diesel engines contribute to CO
nonattainment in more than one of these areas.

   Although nonroad diesel engines have relatively low per-engine CO emissions, they  can be a
significant source  of ambient CO levels in  CO nonattainment areas. Thus, the emission benefits
from this final rule will help areas to attain and maintain the CO NAAQS.
                                         2-121

-------
Final Regulatory Impact Analysis
   2.4.2 Health Effects of CO

   Carbon monoxide enters the bloodstream through the lungs and forms carboxyhemoglobin
(COHb), a compound that inhibits the blood's capacity to carry oxygen to organs and tissues.364'
365 Carbon monoxide has long been known to have substantial adverse effects on human health,
including toxic effects on blood and tissues, and effects on organ functions. Although there are
effective compensatory increases in blood flow to the brain, at some concentrations of COHb,
somewhere above 20 percent, these compensations fail to maintain sufficient oxygen delivery,
and metabolism declines.366 The subsequent hypoxia in brain tissue then produces behavioral
effects, including decrements in continuous performance and reaction time.367

   Carbon monoxide has been linked to increased risk for people with heart disease, reduced
visual perception, cognitive functions and aerobic capacity, and possible fetal effects.368  Persons
with heart disease are especially sensitive to carbon monoxide poisoning and may experience
chest pain if they breathe the gas while exercising.369  Infants, elderly persons, and individuals
with respiratory diseases are also particularly sensitive. Carbon monoxide can affect healthy
individuals, impairing exercise capacity, visual perception, manual dexterity, learning functions,
and ability to perform complex tasks.370

   Several recent epidemiological studies have shown a link between CO and premature
morbidity (including angina,  congestive heart failure, and other cardiovascular diseases.  Several
studies in the United States and Canada have also reported an association of ambient CO
exposures with frequency of cardiovascular hospital admissions, especially for congestive heart
failure (CHF). An association of ambient CO exposure with mortality has also been reported in
epidemiological studies, though not as consistently or specifically as with CHF admissions.
EPA reviewed these studies as part of the Criteria Document review process.371

   2.4.3 CO Nonattainment

   The current primary NAAQS for CO are 35 parts  per million for the one-hour average and 9
parts per million for the eight-hour average. These values are not to be exceeded more than once
per year. Air quality carbon monoxide value is estimated using EPA guidance for calculating
design values. Over 19 million people currently live in the 11  nonattainment areas for the CO
NAAQS.

   Nationally, significant progress has been made over the last decade to reduce CO emissions
and ambient CO concentrations. Total CO emissions from all  sources have decreased 16 percent
from 1989 to  1998, and ambient CO concentrations decreased by 39 percent. During that time,
while the mobile source CO contribution of the inventory remained steady at about 77 percent,
the highway portion decreased from 62 percent of total CO emissions to 56 percent while the
nonroad portion increased from 17 percent to 22 percent.372 Over the next decade, we expect
there to be a minor decreasing trend from the highway segment due primarily to the more
stringent standards for certain light-duty trucks and gasoline nonroad engines.373  CO standards
                                         2-122

-------
                                             Air Quality, Health, and Welfare Effects
for passenger cars and other light-duty trucks and heavy-duty vehicles did not change as a result
of other recent rulemakings.

    As noted above, CO has been linked to numerous health effects; however, we are unable to
quantify the CO-related health or environmental effects of the Nonroad Diesel Engine rule at this
time.  However, nonroad diesel engines do contribute to nonattainment in some areas. Thus, the
emission benefits from this rule will help areas to attain and maintain the CO NAAQS.
                                         2-123

-------
Final Regulatory Impact Analysis
Chapter 2 References

1.  U.S. EPA (1996) Air Quality Criteria for Particulate Matter - Volumes I, II, and III, EPA600-
P-95-OOlaF, EPA600-P-95-001bF, EPA600-P-95-001cF. Docket No. A-99-06. Document Nos.
II-A-18to20,  and U.S. EPA (2003). Air Quality Criteria for Particulate Matter - Volumes I
and II (Fourth External Review Draft,, EPA/600/P-99/002aD June 2003.  Air Quality Criteria
for Particulate Matter - Revised Chapters 7 and 8, U.S. EPA (2003). This material is available
electronically at http://cfpub.epa.gov/ncea/cfm/partmatt.cfm).

2.  U.S. EPA (2002). Health Assessment Document for Diesel Engine Exhaust.  EPA600-8-90-
057F Office of Research and Development, Washington DC. This  document is available
electronically at http://cfpub.epa.gov/ncea/cfm/recordisplay.cfm?deid=29060.

3.  Schwartz, J.; Morris, R. (1995) Air pollution and hospital admissions for cardiovascular
disease in Detroit, Michigan. Am. J. Epidemiol. 142: 23-35.

4.  Lippmann, M.; Ito, K.; Nadas, A.; et al. (2000) Association of particulate matter components
with daily mortality and morbidity in urban populations.  Res Rep Health Effects Inst 95.

5.  Thurston, G. D.; Ito, K.; Hayes, C. G.; Bates, D. V.; Lippmann,  M. (1994) Respiratory
hospital admissions and summertime haze air pollution in Toronto, Ontario: consideration of the
role of acid aerosols. Environ. Res.65: 271-290.

6.  Schwartz, J. (1995) Short term fluctuations in air pollution and hospital admissions of the
elderly for respiratory disease. Thorax 50: 531-538.

7.  Schwartz, J.; Spix, C.; Touloumi, G.; Bacharova, L.; Barumamdzadeh, T.; le Tertre, A.;
Piekarksi, T.; Ponce de Leon, A.; Ponka, A.;  Rossi, G.; Saez, M.; Schouten, J. P. (1996b)
Methodological issues in studies of air pollution and daily counts of deaths or hospital
admissions. In: St Leger, S., ed. The APHEA project.  Short term effects of air pollution on
health: a European approach using epidemiological time series data. J. Epidemiol. Community
Health 50(suppl. 1): S3-S11.

8.  Schwartz, J. (1996) Air pollution and hospital admissions for respiratory disease.
Epidemiology 7(l):20-8.

9.  Schwartz J.  (1994) Air pollution and hospital admissions for the elderly in Detroit, Michigan.
Am J Respir Crit Care Med 150(3):648-55.

10. Schwartz, J. (1994) PM10, ozone, and hospital admissions for the elderly in Minneapolis-St.
Paul, Minnesota. Arch Environ Health 49(5):366-74.

11. Schwartz, J. (1994) What are people dying of on high air pollution days? Environ Res
64(l):26-35.
                                         2-124

-------
                                              Air Quality, Health, and Welfare Effects
12. Schwartz, 1; Dockery, D. W.; Neas, L. M.; Wypij, D.; Ware, J. H.; Spengler, J. D.;
Koutrakis, P.; Speizer, F. E.; Ferris, B. G., Jr. (1994) Acute effects of summer air pollution on
respiratory symptom reporting in children.  Am. J. Respir. Crit. Care Med. 150: 1234-1242.

13. Pope, C. A., III. (1991) Respiratory hospital admissions associated with PM10 pollution in
Utah, Salt Lake, and Cache Valleys. Arch. Environ. Health 46: 90-97.

14. Pope, C.A.  III. and Dockery, D.W. (1992) Acute health effects of PM10 pollution on
symptomatic and asymptomatic children. Am Rev Respir Dis 145(5): 1123-8.

15. Schwartz, J.; Dockery, D. W.; Neas, L. M. (1996) Is daily mortality associated specifically
with fine particles?  J. Air Waste Manage. Assoc. 46: 927-939.

16. Pope, C. A., Ill; Schwartz, J.; Ransom, M. R. (1992) Daily mortality and PM10 pollution in
Utah valley. Arch. Environ. Health 47: 211-217.

17. Dockery, D. W.; Schwartz, J.; Spengler, J. D. (1992) Air pollution and daily mortality:
associations with particulates and acid aerosols. Environ. Res. 59: 362-373.

18. Schwartz, J. (1993) Air pollution and daily mortality in Birmingham, Alabama. Am. J.
Epidemiol. 137: 1136-1147.

19. Samet, J.M.; Dominici, F; Zeger, S.L.; et al. (2000) The National Morbidity, Mortality, and
Air Pollution  Study. Part I: methods and methodologic issues.  Res Rep Health Eff Inst 94, Part
I.  Docket A-2000-01. Document No. IV-A-205.

20. Samet, J.M.; Zeger, S.L.; Dominici, F; et al.  (2000) The National Morbidity, Mortality,
and Air Pollution Study. Part II: morbidity and mortality from air pollution in the United States.
Res Rep Health Eff Inst Number 94, Part II. Docket A-2000-01.  Document No. IV-A-206.

21. Dominici, F; McDermott, A.; Zeger S.L.; et al. (2002) On the use of generalized additive
models in time-series studies of air pollution and health.  Am J Epidemiol  156(3):193-203.

22. Laden F;  Neas LM; Dockery DW; et al. (2000). Association of fine particulate matter from
different  sources with daily mortality in six U.S. cities.  Environ Health Perspectives
108(10):941-947.

23. Schwartz J; Laden F; Zanobetti A. (2002). The concentration-response relation between
PM(2.5) and daily deaths.  Environ Health Perspect 110(10):  1025-1029.

24. Janssen NA; Schwartz J; Zanobetti A.; et al.  (2002). Air conditioning and source-specific
particles as modifiers of the effect of PM10 on hospital admissions for heart and lung disease.
Environ Health  Perspect 110(l):43-49.

25.Health Effects Institute. (2003a) Revised analyses of time-series studies of
air pollution and health. Available:  http://www.healtheffects.org/Pubs/TimeSeries.pdff21 Nov,

                                         2-125

-------
Final Regulatory Impact Analysis
2003].

26.Kunzli, N.; Medina, S.; Kaiser, R.; et al. (2001) Assessment of deaths attributable to air
pollution: should we use risk estimates based on time series or on cohort studies? Am J
Epidemiol 153(11): 1050-1055.

27.Gauderman, W.J.; McConnell, R.; Gilliland, F.; et al. (2000) Association between air
pollution and lung function growth in southern California children.  Am J Respir Crit Care Med
162(4 Pt 1): 1383-90.

28.Gauderman, W.J.; Gilliland, G.F.; Vora, H.; et al. (2002) Association between air pollution
and lung function growth in southern California children:  results from a second cohort. Am J
Respir Crit Care Med 166(1): 76-84.

29.Peters, J.M.; Avol, E.; Navidi, W.; et al. (1999) A study of twelve southern California
communities with differing levels and types of air pollution:  I. Prevalence of respiratory
morbidity. Am J Respir Crit Care Med 159(3): 760-7.

SO.Hoek, G; Brunekreef, B; Goldbohm, S; et al. (2002). Association between mortality and
indicators of traffic-related air pollution in the Netherlands: a cohort study. Lancet
360(9341):1203-1209.

Sl.Finkelstein, M.M.; Jerrett, M.; Deluca, P.; et al. (2003) Relation between income, air
pollution, and mortality: a cohort study. Canadian Med Assoc J  169(5): 397-402.

32. Dockery, DW; Pope, CA, III; Xu, X; et al. (1993). An association between air pollution and
mortality in six U.S. cities. N Engl J Med 329:1753-1759.-75.

33. Pope, CA, III; Thun, MJ; Namboordiri, MM;  et al. (1995). Particulate air pollution as a
predictor of mortality in a prospective study of U.S. adults. Am J Respir Crit Care Med
151:669-674. and Pope, CA, III; Burnett, RT; Thun,  MJ; Calle, EE; et al. (2002) Lung cancer,
cardiopulmonary mortality, and long-term exposure to fine particulate air pollution. JAMA J.
Am. Med. Assoc. 287: 1132-1141.

34. Health Effects Institute Report, "Reanalysis of the Harvard Six Cities Study and the
American Cancer Society Study of Particulate Air Pollution and Mortality" Docket A-99-06.
Document No. IV-G-75. and Pope, CA, III; Burnett, RT; Thun, MJ; Calle, EE;  et al. (2002)
Lung cancer, cardiopulmonary mortality, and long-term exposure to fine particulate air pollution.
JAMA J. Am. Med. Assoc. 287: 1132-1141.

35.Abbey, D. E.; Nishino, N.; McDonnell, W. F.; Burchette, R. J.; Knutsen, S. F.; Beeson, W. L.;
Yang, J. X. (1999) Long-term inhalable particles and other air pollutants related to mortality in
nonsmokers. Am. J. Respir.  Crit. Care Med. 159: 373-382.
                                         2-126

-------
                                              Air Quality, Health, and Welfare Effects
36.McDonnell, W.F.; Nishino-Ishikawa, N.; Peterson, F.F.; et al. (2000) Relationships of
mortality with the fine and coarse fractions of long-term ambient PM10 concentrations in
nonsmokers.  J Exposure Anal Environ Epidemiol 10: 427-436.

37.Lipfert, F. W.; Perry, H. M., Jr.; Miller, J. P.; Baty, J. D.; Wyzga, R. E.; Carmody, S. E.
(2000b) The Washington University-EPRI veterans' cohort mortality study: preliminary results.
In: Grant, L. D., ed. PM2000: particulate matter and health. Inhalation Toxicol. 12(suppl. 4):
41-73.

38.Hoek, G; Brunekreef, B; Goldbohm, S; et al. (2002). Association between mortality and
indicators of traffic-related air pollution in the Netherlands: a cohort study.  Lancet
360(9341):1203-1209.

39.Hoek, G; Fischer, P.; van den Brandt, P.; Goldbohm, S.; and Brunekreef, B. (2001)
Estimation of long-term average exposure to outdoor air pollution for a
cohort study on mortality. J Expo Anal Environ Epidemiol 11: 459-69.

40.Finkelstein, M.M.; Jerrett, M.; Deluca, P.; et al. (2003) Relation between income, air
pollution, and mortality: a  cohort study. Canadian Med Assoc J 169(5):  397-402.

41.Finkelstein, M.M.; Jerrett, M.; Deluca, P.; et al. (2003) Relation between income, air
pollution, and mortality: a  cohort study. Canadian Med Assoc J 169(5):  397-402.

42.Churg, A and Brauer, M. (1997) Human lung parenchyma retains PM2.5. Am J Respir Crit
Care Med 155(6):2109-11.

43.Churg, A.; Brauer, M.; del Carmen Avila-Casado, M.; et al. (2003) Chonic exposure to high
levels of particulate air pollution and small  airway remodeling.  Environ Health Perspect 111(5):
714-718.

44.Calderon-Garciduenas,  L.; Mora-Tiscareno, A.; Fordham, L.A.; et al. (2001) Canines as
sentinel species for assessing chronic exposures to air pollutants: part 2. Respiratory pathology.
Toxicol Sci 61(2): 342-355.

45. Calderon-Garciduenas, L.; Gambling, T.M.; Acuna, H.; et al. (2001) Canines as sentinel
species for assessing chronic exposures to air pollutants: part 2. Cardiac pathology. Toxicol Sci
61(2): 356-67.

46.Bunn, H.J.; Dinsdale, D.;  Smith, T.; et al. (2001) Ultrafine particles in alveolar macrophages
from normal children.  Thorax 56(12):932-4.

47. Liao, D.; Creason, J.; Shy, C.; et al. (1999) Daily variation of particulate air pollution and
poor autonomic control in the elderly. Environ Health Perspect 107(7):521-525.United States

48. Creason, J.; Neas, L.; Walsh, D; et al. (2001) Particulate matter and heart rate variability
among elderly retirees:  the Baltimore 1998  PM study. J Exposure Anal Environ Epidemiol

                                         2-127

-------
Final Regulatory Impact Analysis
11:116-122.

49. Magari SR, Hauser R, Schwartz J; et al. (2001). Association of heart rate variability with
occupational and environmental exposure to particulate air pollution.  Circulation
104(9):986-991.

50. Pope, C.A. Ill; Dockery, D.W.; Kanner, R.E.; et al. (1999) Oxygen saturation, pulse rate,
and particulate air pollution. Am J Respir Crit Care Med 159: 356-372.

51. Pope, C.A. Ill; Verrier, R.L.; Lovett, E.G.; et al. (1999) Heart rate variability associated with
particulate air pollution.  Am Heart J 138: 890-899.

52. Gold, D.R.; Litonjua, A; Schwartz, J; et al. (2000) Ambient pollution and heart rate
variability. Circulation 101: 1267-1273.

53. Liao, D.; Cai, J.; Rosamond W.D.; et al. (1997)  Cardiac autonomic function and incident
coronary heart disease: a population-based case-cohort study. The ARIC Study. Atherosclerosis
Risk in Communities Study. Am J Epidemiol 145(8):696-706.

54. Dekker, J.M., Crow, R.S., Folsom, A.R.; et al. (2000) Low heart rate variability in a
2-minute rhythm strip predicts risk of coronary heart disease and mortality from several causes:
the ARIC Study. Atherosclerosis Risk In Communities. Circulation 102(11): 1239-44.

55. La Rovere, M.T.; Pinna G.D.; Maestri R.; et al.  (2003) Short-term heart rate variability
strongly predicts sudden cardiac death in chronic heart failure patients.
Circulationl07(4):565-70.

56. Kennon, S., Price, C.P., Mills, P.G.; et al. (2003) Cumulative risk assessment in unstable
angina: clinical, electrocardiographic, autonomic, and biochemical markers. Heart 89(1):36-41.

57. Salvi et al. (1999) Acute inflammatory responses in the airways and peripheral blood after
short-term exposure to diesel exhaust in healthy human volunteers. Am J Respir Crit Care Med
159: 702-709.

58. Salvi et al. (2000) Acute exposure to diesel exhaust increases IL-8 and GRO-a production  in
healthy human airways. Am J Respir Crit Care Med 161:  550-557.

59. Holgate et al. (2003) Health effects of acute exposure to air pollution. Part I: healthy and
asthmatic subjects exposed to diesel exhaust. Res Rep Health Eff Inst  1 12.

60. Ghio, A.J.; Kim, C.; and Devlin R.B. (2000) Concentrated ambient air particles induce mild
pulmonary inflammation in healthy human volunteers.  Am J Respir Crit Care Med 162(3 Pt
61. Seaton et al. (1999) Particulate air pollution and the blood. Thorax 54: 1027-1032.
                                          2-128

-------
                                              Air Quality, Health, and Welfare Effects
62. Peters et al. (200la) Particulate air pollution is associated with an acute phase response in
men; results from the MONICA-Augsburg study. Eur Heart J 22(14): 1198-1204.

63. Tan et al. (2000) The human bone marrow response to acute air pollution caused by forest
fires.  Am J Respir Crit Care Med 161: 1213-1217.

64. Peters et al. (1997) Increased plasma viscosity during and air pollution episode: a link to
mortality? Lancet 349: 1582-87.

65. Zimmerman, M.A.; Selzman, C.H.; Cothren, C.; et al. (2003) Diagnostic implications of
C-reactive protein.  Arch Surg 138(2):220-4.

66. Engstrom, G,; Lind, P.; Hedblad, B.; et al. (2002)  Effects of cholesterol and
inflammation-sensitive plasma proteins on incidence of myocardial infarction and stroke in men.
Circulation 105(22):2632-7.

67. Suwa, T.; Hogg, J.C.; Quinlan, K.B.; et al. (2002) Particulate air pollution induces
progression of atherosclerosis.  J Am Coll Cardiol 39(6): 935-942.

68. Calderon-Garciduenas, L.; Gambling, T.M.; Acuna, H.; et al. (2001) Canines as sentinel
species for assessing chronic exposures to air pollutants: part 2. Cardiac pathology.  Toxicol Sci
61(2): 356-67.

69. Peters, A.; Liu, E.; Verrier, R.L.; et al. (2000) Air pollution and incidence of cardiac
arrhythmia.  Epidemiology 11: 11-17.

70. Peters, A.; Dockery, D.W.; Muller, I.E.; et al. (2001) Increased particulate air pollution and
the triggering of myocardial infarction.  Circulation 103(23): 2810-2815.

71. U.S. EPA (2002). Health assessment document for diesel engine exhaust.
EPA/600/8-90/057F Office of Research and Development, Washington DC.  This document is
available electronically at http://cfpub.epa.gov/ncea/cfm/recordisplay.cfm?deid=29060.

72. U.S. EPA (1985). Size specific total particulate emission factor for mobile sources.
EPA460-3-85-005. Office of Mobile Sources, Ann Arbor, MI.

73.Hoek, G; Fischer, P.; van den Brandt, P.; Goldbohm, S.; and Brunekreef, B. (2001)
Estimation of long-term average exposure to outdoor air pollution for a
cohort study on mortality. J Expo Anal Environ Epidemiol 11: 459-69.

74.Maheswaran, R. and Elliott, P. (2003) Stroke mortality associated with living near main roads
in England and Wales. A geographical study.  Stroke.  Published online November 13, 2003.
Available: http://stroke.ahajournals.org/strokeasap.shtml.

75.Garshick, E.; Laden, F.; Hart, I.E.; et al. (2003) Residence near a major road and respiratory
symptoms in U.S. veterans. Epidemiology 14: 728-736.

                                          2-129

-------
Final Regulatory Impact Analysis
76. Brauer, M; Hoek, G; Van Vliet, P.; et al. (2002) Air pollution from traffic and the
development of respiratory infections and asthmatic and allergic symptoms in children. Am J
Respir Crit Care Med 166(8): 1092-8.

77.English, P.; Neutra, R.; Scalf, R.; et al. (1999). Examining associations between childhood
asthma and traffic flow using a geographic information system.  Environ Health Perspect
107:761-767.

78.Pershagen, G.; Rylander, E.; Norberg, S.; et al. (1995) Air pollution involving nitrogen
dioxide exposure and wheezing bronchitis in children. Int J Epidemiol 24:1147-1153.

79.Weiland, S.K.; Mundt, K.A.; Ruckmann, A.; et al. (1994) Self-reported wheezing and allergic
rhinitis in children and traffic density on street of residence. Ann Epidemiol 4:243-247.

SO.Duhme, H.; Weiland, S.K.; Keil, U.; et al. (1996) The association between self-reported
symptoms of asthma and allergic rhinitis and self-reported traffic density on street of residence
in adolescents. Epidemiology 7:578-582.

81.van Vliet, P.; Knape, M.; de Hartog, J; et al. (1997) Motor vehicle exhaust and chronic
respiratory symptoms in children living near freeways. Environ Res 74:122-132.

82.Waldron, G; Pottle, B; and Dod, J. (1995) Asthma and the motorways — one district's
experience. J Public Health Med 17:85-89.

83. Delfino RJ. (2002). Epidemiologic evidence for asthma and exposure to air toxics: linkages
between occupational, indoor, and community air pollution research.  Env Health Perspect Suppl
110(4): 573-589.

84. Brunekreef, B; Janssen NA; de Hartog, J; et al.  (1997). Air pollution from traffic and lung
function in children living near motor ways.  Epidemiology (8): 298-303.

85. Wilhelm, M. and Ritz, B. (2003) Residential proximity to traffic and adverse birth outcomes
in Los Angeles County, California, 1994-1996. Environ Health Perspect 111(2): 207-216.

86. Bunn, H.J.; Dinsdale, D.; Smith, T.; et al. (2001) Ultrafme particles in alveolar macrophages
from normal children.  Thorax 56(12):932-4.

87. Zhu, Y.; Hinds, W.C.; Kim, S.; et al.  (2002) Concentration  and size distribution of ultrafine
particles near a major highway. J Air Waste Manage Assoc 52: 1032-1042.

88. Zhu, Y.; Hinds, W.C.; Kim, S.; et al.  (2002) Study of ultafme particles near a major highway
with heavy-duty diesel traffic. Atmos Environ 36:4323-4335.

89. Kittleson, D.B.; Watts, W.F.; and Johnson, J.P. (2001) Fine particle (nanoparticle) emissions
on Minnesota highways. Minnesota Department of Transportation Report No. MN/RC-2001-12.
                                         2-130

-------
                                             Air Quality, Health, and Welfare Effects
90. Koman memorandum to the Docket.  One-hour Ozone and PM10 Nonattainment Status and
Air Quality Data Update. August 11, 2003.

91. Rao, Venkatesh; Frank, N.; Rush, A.; and Dimmick, F. (November 13-15, 2002). Chemical
speciation of PM25 in urban and rural areas (November 13-15, 2002) In the Proceedings of the
Air & Waste Management Association Symposium on Air Quality Measurement Methods and
Technology, San Francisco Meeting.

92. EPA (2002) Latest Finds on National Air Quality, EPA454-K-02-001.

93. Mansell (2000).  User's Instructions for the Phase 2 REMSAD Preprocessors, Environ
International.  Novato, CA.  2000.

94. IMPROVE (2000).  Spatial and Seasonal Patterns and Temporal Variability of Haze and its
Constitutents in the United States.  Report III. Cooperative Institute for Research in the
Atmosphere, ISSN: 0737-5352-47.

95. California Air Resources Board and New York State Department of Environmental
Conservation (April 9, 2002). Letter to EPA Administrator Christine Todd Whitman.

96. State and Territorial Air Pollution Program Administrators (STAPPA) and Association of
Local Air Pollution Control Officials (ALAPCO) (December 17, 2002). Letter to EPA Assistant
Administrator Jeffrey R. Holmstead.

97. Western Regional Air Partnership (WRAP) January 28, 2003), Letter to Governor Christine
Todd Whitman.

98.National Research Council,  1993. Protecting Visibility in National Parks and Wilderness
Areas. National Academy of Sciences  Committee on Haze in National Parks and Wilderness
Areas. National Academy Press, Washington, DC. This document is available on the internet at
http://www.nap. edu/books/0309048443/html/.
U.S. EPA (1996). "Air Quality Criteria for Paniculate Matter (PM)" Vol I - III.
EPA600-P-99-002a;  Docket No. A-99-06. Document Nos. II-A-18 to 20.
U.S. EPA (1996). Review of the National Ambient Air Quality Standards for Particulate Matter:
Policy Assessment of Scientific and Technical Information OAQPS Staff Paper. EPA452-R-96-
013, 1996.  Docket Number A-99-06, Documents No. II-A-23. The particulate matter air quality
criteria documents are also available at http://www.epa.gov/ncea/partmatt.htm.  Also, U.S. EPA.
Review of the National Ambient Air Quality Standards for Particulate Matter: Policy
Assessment of Scientific and Technical Information, OAQPS Staff Paper.  Preliminary  Draft.
June 2001.  Docket A-2000-01, Document IV-A-199.

99. Council on Environmental Quality, 1978.  Visibility Protection for Class I Areas, the
Technical Basis. Washington DC. Cited in U.S. EPA, Review of the National Ambient Air
Quality Standards for Particulate Matter: Policy Assessment of Scientific and Technical
Information. OAQPS Staff Paper.  EPA452-R-96-013.  This document is available in Docket


                                        2-131

-------
Final Regulatory Impact Analysis
A-99-06, Document II-A-23.

100.U.S. EPA Trends Report 2001. This document is available on the internet at
http ://www. epa. gov/airtrends/.

101. Sisler, James F.  Spatial and Seasonal Patterns and Long Term Variability of the
Composition of Haze in the United States: An Analysis of Data from the IMPROVE Network.
1996. A copy of the relevant pages of this document can be found in Docket A-99-06,
Document No. II-B-21.

102.U.S. EPA Criteria for Particulate Matter, 8-3; U.S. EPA Review of the National Ambient
Air Quality Standards for Particulate Matter: Policy Assessment of Scientific and Technical
Information OAQPS Staff Paper. EPA452-R-96-013.  1996.  Docket Number A-99-06,
Documents Nos. II-A-18, 19, 20, and 23. The particulate matter air quality criteria documents
are also available at http ://www. epa. gov/ncea/partmatt.htm. Also, U.S. EPA. Review of the
National Ambient Air Quality Standards for Particulate Matter: Policy Assessment of Scientific
and Technical Information, OAQPS Staff Paper.  Preliminary Draft.  June 2001. Docket A-
2000-01, Document IV-A-199.

103. National Research Council, 1993 (Ibid). This document is available on the internet at
http://www.nap.edu/books/0309048443/html/.

104.National Research Council, 1993 (Ibid). This document is available on the internet at
http://www.nap.edu/books/0309048443/html/.

105. National Acid Precipitation Assessment Program (NAPAP), 1991.  Office of the Director.
Acid Deposition: State of Science and Technology. Report 24, Visibility: Existing and
Historical Conditions - Causes and Effects. Washington, DC.  Cited in U.S. EPA, Review of the
National Ambient Air Quality Standards for Particulate Matter: Policy Assessment of Scientific
and Technical Information. OAQPS Staff Paper. EPA452-R-96-013. This document is
available in Docket A-99-06, Document II-A-23. Also, U.S. EPA. Review of the National
Ambient Air Quality Standards for Particulate Matter: Policy  Assessment of Scientific and
Technical Information, OAQPS Staff Paper. Preliminary Draft. June 2001. Docket A-2000-01,
Document IV-A-199.

106.U.S. EPA. (2003). Air Quality Technical Support Document for the proposed Nonroad
Diesel rulemaking.  OAQPS.  April 2003.

107.U.S. EPA (1996). Review of the National Ambient Air Quality  Standards for Particulate
Matter: Policy Assessment of Scientific and Technical Information OAQPS Staff Paper.
EPA452-R-96-013.  1996. Docket Number A-99-06, Documents No. II-A-23. The paniculate
matter air quality criteria documents are also available at http://www.epa.gov/ncea/partmatt.htm.

108. U.S. EPA (1996). Review of the National Ambient Air Quality Standards for Particulate
Matter: Policy Assessment for Scientific and Technical Information, OAQPS Staff Paper,


                                        2-132

-------
                                            Air Quality, Health, and Welfare Effects
EPA452-R-96-013, July, 1996, at IV-7. This document is available from Docket A-99-06,
Document II-A-23.

109.US EPA Trends Report 2002

110. See 64 FR 35722, July 1, 1999.

111. Technical Memorandum, EPA Air Docket A-99-06, Eric O. Ginsburg, Senior Program
Advisor, Emissions Monitoring and Analysis Division, OAQPS, Summary of Absolute Modeled
and Model-Adjusted Estimates of Fine Particulate Matter for Selected Years, December 6, 2000,
Table P-2. Docket Number 2000-01, Document Number II-B-14.

112.Western Regional Air Partnership (WRAP) letter dated Jan 28, 2003 to Administrator
Christine Todd Whitman.

113. U.S. EPA. (1993). Effects of the 1990 Clean Air Act Amendments on Visibility in Class I
Areas: An EPA Report to Congress. EPA452-R-93-014, Docket A-2000-01, Document IV-A-
220. And see also 64 FR 35722, July  1, 1999.

114.This goal was recently upheld by the U.S. Court of Appeals. American Corn Growers
Association v. EPA, 291F.3d 1 (D.C .Cir 2002).  A copy of this decision can be found in Docket
A-2000-01, Document IV-A-113.

115.U.S. EPA. (1993). Effects of the 1990 Clean Air Act Amendments on Visibility in Class I
Areas: An EPA Report to Congress. EPA452-R-93-014, Docket A-2000-01, Document IV-A-
20.U.S. EPA Trends Report 2002.

116. For more information and the IMPROVE data, see
http://vista.cira.colostate.edu/improve/data/IMPROVE/improve_data.htm.

117.National Park Service.  Air Quality in the National Parks, Second edition.  NFS, Air
Resources Division. D 2266.  September 2002.

118.U.S. EPA Trends Report 2002.

119. Chestnut, L.G., andR.D. Rowe.  1990a. Preservation for Visibility Protection at the
National Parks: Draft Final Report. Prepared for Office of Air Quality Planning and Standards,
U.S. Environmental Protection Agency, and Air Quality Management Division, National Park
Service; Chestnut, L.G., and R.D. This document is available from Docket A-97-10, Document
II-A-33 Rowe. 1990b.  A New National Park Visibility Value Estimates. In Visibility and Fine
Particles., Transactions of an AWMA/EPA International Speciality Conference. C.V. Mathai,
ed., Air and Waste Management Association, Pittsburg. Docket A-2000-01, IV-A-2000.

120.Much of the information in this subsection was excerpted from the EPA document, Human
Health Benefits from Sulfate Reduction, written under Title IV of the 1990 Clean Air Act
Amendments, U.S. EPA, Office of Air and Radiation, Acid Rain Division, Washington, DC

                                        2-133

-------
Final Regulatory Impact Analysis
20460, November 1995.

121.Acid Rain: Emissions Trends and Effects in the Eastern United States, U.S. General
Accounting Office, March, 2000 (GOA/RCED-00-47).

122. Acid Deposition Standard Feasibility Study: Report to Congress, EPA430-R-95-001a,
October, 1995.

123. Deposition of Air Pollutants to the Great Waters, Third Report to Congress, June, 2000.

124.Deposition of Air Pollutants to the Great Waters, Third Report to Congress, June, 2000.
Great Waters are defined as the Great Lakes, the Chesapeake Bay, Lake Champlain, and coastal
waters.  The first report to Congress was delivered in May,  1994; the second report to Congress
in June, 1997.

125. Bricker, Suzanne B., et al.,  NationalEstuarine Eutrophication Assessment, Effects of
Nutrient Enrichment in the Nation's Estuaries, National Ocean Service, National Oceanic and
Atmospheric Administration, September, 1999.

126. Deposition of Air Pollutants to the Great Waters, Third Report to Congress, June, 2000.

127. Valigura, Richard, et al., Airsheds and Watersheds II: A Shared Resources Workshop, Air
Subcommittee of the Chesapeake Bay Program, March, 1997.

128. The Impact of Atmospheric Nitrogen Deposition on Long Island Sound, The Long Island
Sound Study, September, 1997.

129. Dennis, Robin L., Using the Regional Acid Deposition Model to Determine the Nitrogen
Deposition Airshed of the Chesapeake Bay Watershed, SET AC Technical Publications Series,
1997.

130. Dennis, Robin L., Using the Regional Acid Deposition Model to Determine the Nitrogen
Deposition Airshed of the Chesapeake Bay Watershed, SET AC Technical Publications Series,
1997.

131. Much of this information was taken from the following EPA document: Deposition of Air
Pollutants to the Great Waters-Second Report to Congress, Office of Air Quality Planning and
Standards, June 1997, EPA453-R-97-011.  Refer to that document for a more detailed
discussion.

132. The 1996 National Toxics Inventory, Office of Air Quality Planning and Standards,
October 1999.

133. U.S. EPA. Control of Emissions of Hazardous Air Pollutants from Mobile Sources; Final
Rule (66 FR 17230-17273, March 29, 2001).
                                        2-134

-------
                                              Air Quality, Health, and Welfare Effects
134. U.S. EPA. (1999). Guidelines for Carcinogen Risk Assessment. Review Draft.
NCEA-F-0644, July. Risk Assessment Forum, Washington, DC.
http://www.epa.gov/ncea/raf/cancer.htm.

135. U.S. EPA. (1986) .Guidelines for carcinogen risk assessment. Federal Register
51(185):33992-34003.

136. National Institute for Occupational Safety and Health (NIOSH). (1988).  Carcinogenic
effects of  exposure to diesel exhaust. NIOSH Current Intelligence Bulletin 50. DHHS (NIOSH)
Publication No. 88-116. Atlanta, GA: Centers for Disease Control.

137. International Agency for Research on Cancer - IARC. (1989). Monographs on the
evaluation of carcinogenic risks to humans. Vol. 46. Diesel and Gasoline Engines Exhausts and
some Nitroarenes. Lyon, France: IARC, pp. 362-375.

138. National Institute for Occupational Safety and Health (NIOSH). (1988). Carcinogenic
effects of  exposure to diesel exhaust.. NIOSH Current Intelligence Bulletin 50. DHHS (NIOSH)
Publication No. 88-116. Atlanta, GA: Centers for Disease Control.

139. World Health Organization International Program on Chemical Safety (1996).
Environmental Health Criteria 171. Diesel fuel and exhaust emissions. Geneva: World Health
Organization, pp.172-176.

140. California Environmental Protection Agency. (Cal EPA, OEHHA) (1998). Health risk
assessment for diesel  exhaust.  Public and Scientific Review Draft.

141. National Toxicology Program (NTP). (2000). 9th report on carcinogens.  Public Health
Service, U.S. Department of Health and Human Services, Research Triangle Park, NC.
Available from: http://ntp-server.niehs.nih.gov.

142. Health Effects Institute (HEI). (1995). Diesel exhaust: a critical analysis of emissions,
exposure, and health effects. Cambridge,  MA.

143. Health Effects Institute (HEI) (1999). Diesel emissions and lung cancer: epidemiology
and quantitative risk  assessment.  A special  report of the Institute's Diesel Epidemiology
Expert Panel.  Cambridge,  MA.

144. Health Effects Institute (HEI).  (2002). Research directions to improve estimates of human
exposure and risk assessment.  A special report of the Institute's Diesel Epidemiology Working
Group, Cambridge, MA.

145. Ishinishi, N.,  Kuwabara, N., Takaki, Y., et al. (1988). Long-term inhalation experiments
on diesel exhaust. In:  Diesel exhaust and health risks. Results of the HERP studies. Ibaraki,
Japan: Research Committee for HERP Studies; pp. 11-84.
                                         2-135

-------
Final Regulatory Impact Analysis
146. Lewtas, J. (1983). Evaluation of the mutagenicity and carcinogenicity of motor vehicle
emissions in short-term bioassays.  Environ Health Perspect 47:141-152/.

147. Garshick, E., Schenker, M., Munoz, A, et al. (1987). A case-control study of lung cancer
and diesel exhaust exposure in railroad workers. Am Rev Respir Dis 135:1242-1248.

148. Garshick, E., Schenker, M., Munoz, A, et al. (1988). A retrospective cohort study of
lung cancer and diesel exhaust exposure in railroad workers. Am Rev Respir Dis 137:820-825.

149. Woskie, SR; Smith, TJ; Hammond, SK; et al. (1988). Estimation of the diesel exhaust
exposures of railroad workers. I. Current exposures. Am J Ind Med 13:381-394.

150. Steenland, K., Silverman,  D, Hornung, R. (1990).  Case-control study of lung cancer and
truck driving in the Teamsters Union. Am J Public Health 80:670-674.

151. Steenland, K., Deddens, J., Stayner,  L. (1998). Diesel exhaust and lung cancer in the
trucking industry: exposure-response analyses and risk assessment. Am J Ind Med 34:220-228.

152. Zaebst, DD; Clapp, DE; Blake, LM; et al. (1991). Quantitative determination of trucking
industry workers' exposures to diesel exhaust particles. Am Ind Hyg Assoc J 52:529-541.

153. Saverin, R. (1999). German potash miners: cancer mortality. Health Effects Institute
Number 7. March 7-9, Stone Mountain, GA, pp. 220-229.

154. Friones, JR; Hinds, WC; Duffy, RM; Lafuente, EJ; Liu, WV. (1987). Exposure of
firefighters to diesel emissions in fire stations. Am Ind Hyg Assoc J 48:202-207.

155. Bruske-Hohlfeld, I, Mohner, M., Ahrens, W., et al. (1999). Lung cancer risk in male
workers occupationally exposed to diesel motor emissions in Germany. Am J Ind Med
36:405-414.

156. Wong, O; Morgan, RW; Kheifets, L; et al. (1985). Mortality among members of a heavy
construction equipment operators union with potential exposure to diesel exhaust emissions. Br J
Ind Med 42:435-448. U.S. Environmental Protection Agency.

157. Bhatia, R., Lopipero, P., Smith, A. (1998). Diesel exhaust exposure and lung cancer.
Epidemiology 9(1):84-91.

158. Lipsett, M: Campleman, S.; (1999). Occupational exposure to diesel exhaust and lung
cancer: a meta-analysis. Am J Public Health 80(7): 1009-1017.

159. U.S. EPA (2002), National-Scale Air Toxics Assessment for  1996.  This material is
available electronically at http://www.epa.gov/ttn/atw/nata/.

160. Ishinishi, N; Kuwabara, N; Takaki, Y; et al. (1988) Long-term inhalation experiments on
diesel exhaust. In: Diesel exhaust and health risks. Results of the HERP studies. Ibaraki, Japan:

                                         2-136

-------
                                             Air Quality, Health, and Welfare Effects
Research Committee for HERP Studies; pp.
11-84.

161. Heinrich, U; Fuhst, R; Rittinghausen, S; et al. (1995) Chronic inhalation exposure of
Wistar rats and two different strains of mice to diesel engine exhaust, carbon black, and titanium
dioxide. Inhal Toxicol 7:553-556.

162. Mauderly, JL; Jones, RK; Griffith, WC; et al. (1987) Diesel exhaust is a pulmonary
carcinogen in rats exposed chronically by inhalation. Fundam Appl Toxicol 9:208-221.

163. Nikula, KJ; Snipes, MB; Barr, EB; et al. (1995) Comparative pulmonary toxicities and
carcinogenicities of chronically inhaled diesel exhaust and carbon black in F344 rats. Fundam
Appl Toxicol 25:80-94.

164. Reger, R; Hancock, J; Hankinson, J; et al. (1982) Coal miners exposed to diesel exhaust
emissions. Ann Occup Hyg 26:799-815.

165. Attfield, MD. (1978) The effect of exposure to silica and diesel exhaust in underground
metal and nonmetal miners. In: Industrial hygiene for mining and tunneling: proceedings of a
topical symposium; November; Denver, CO. Kelley, WD, ed. Cincinnati, OH: The American
Conference of Governmental Industrial Hygienists, Inc.; pp.
129-135.

166. El  Batawi, MA; Noweir, MH. (1966) Health problems resulting from prolonged exposure
to air pollution in diesel bus garages. Ind Health 4:1-10.

167. Wade, JF, III; Newman, LS. (1993) Diesel asthma: reactive airways disease following
overexposure to locomotive exhaust. J Occup Med 35:149-154

168.Delfino, R.J. (2002) Epidemiologic evidence for asthma and exposure to air toxics: linkages
between occupational, indoor, and community air pollution research. Environ Health Perspect
HO(SuppU): 573-589.

169. U.S. EPA (1995). User's Guide for the Industrial  Source Complex (ISC3) Dispersion
Models.  Office of Air Quality Planning and Standards,  Research Triangle Park, NC. Report No.
EPA454-B-95-003b.

170. U.S. EPA.  (2002). Example Application of Modeling Toxic Air Pollutants in Urban Areas.
Report No. EPA454-R-02-003. Office of Air Quality Planning and Standards, Research Triangle
Park, North Carolina.

171. U.S. EPA. (2000). Regulatory Impact Analysis: Heavy Duty Engine and Vehicle
Standards and Highway Diesel Fuel Sulfur Control Requirements. Office of Transportation and
Air Quality.  Report No. EPA420-R-00-026. (December, 2000). Docket No. A-99-06.
Document No. V-B-01.
                                        2-137

-------
Final Regulatory Impact Analysis
172. U.S. EPA. (2002). Diesel PM Model-to-measurement Comparison. Prepared by ICF
Consulting for EPA, Office of Transportation and Air Quality. Report No. EPA420-D-02-004.
EPA.  2002.

173. Zheng, M., Cass, G. R., Schauer, J. J., and Edgerton, E. S. (2002).  Source Apportionment
of PM2.5 in the Southeastern United States Using Solvent-Extractable Organic Compounds as
Tracers. Environmental Science and Technology. In press.

174. Ramadan, Z., Song, X-H, and Hopke, P. K. (2000). Identification of Sources of Phoenix
Aerosol by Positive Matrix Factorization. J. Air & Waste Manage. Assoc. 50, pp. 1308-1320.

175. Schauer, J. J., Rogge, W. F., Hildemann, L. M., Mazurek, M. A., Cass, G. R., and
Simoneit, B. R. T. (1996). Source Apportionment of Airborne PM Using Organic Compounds as
Tracers. Atmospheric Environment. Vol 30, No. 22, pp. 3837 -3855.

176. Schauer, J. J., and Cass, G. R. (2000). Source  Apportionment of Wintertime Gas-Phase and
Particle Phase Air Pollutants Using Organic Compounds as Tracers. Environmental Science and
Technology. Vol 34, No. 9, pp. 1821 -1832.

177. Watson, J. G, Fujita, E., Chow, J.  G, Zielinska, B., Richards, L. W., Neff, W., and
Dietrich, D. (1998). Northern Front Range Air Quality Study Final Report. Desert Research
Institute. 6580-685-8750.1F2.

178. Air Improvement Resources. (1997). Contribution of Gasoline Powered Vehicles to
Ambient Levels of Fine Particulate Matter. CRC Project A-18.

179. Cass, G. R. (1997). Contribution of Vehicle Emissions to Ambient  Carbonaceous
Particulate Matter: A Review and Synthesis of the Available Data in the South Coast Air Basin.
CRC Project A-18.

180. Zheng, M; Cass, GR; Schauer, JJ; et al.  (2002) Source apportionment of PM25 in the
Southeastern United States using solvent-extractable organic compounds as tracers.  Environ Sci
Technol 36: 2361-2371.

181. Schauer, JJ; Rogge, WF; Hildemann, LM; et al.  (1996). Source  apportionment of airborne
particulate matter using organic compounds as tracers. Atmos Environ 30(22): 3837-3855.

182. Watson, JG;  Fujita, EM; Chow, JC; et al. (1998). Northern Front Range Air Quality Study
final report. Prepared by Desert Research Institute for Colorado State University, Cooperative
Institute for Research in the Atmosphere, 1998.

183. Schauer, JJ and Cass, GR.(1999). Source apportionment of wintertime gas-phase and
particle-phase air pollutants using organic compounds as tracers. Environ Sci Technol

184. Schauer, JJ; Fraser, MP; Cass, GR; et al. (2002). Source reconciliation of atmospheric gas-
phase and particle-phase pollutants during a severe  photochemical smog episode. Environ Sci

                                         2-138

-------
                                             Air Quality, Health, and Welfare Effects
Technol 36: 3806-3814.

185. Cal-EPA. (1998) Measuring concentrations of selected air pollutants inside California
vehicles. Final report.

186. Whittaker, LS; Macintosh, DL; Williams, PL. (1999). Employee Exposure to Diesel
Exhaust in the Electric Utility Industry. Am Ind Hyg Assoc J 60:635-640.

187. Groves, J;  Cain, JR. (2000). A Survey of Exposure to Diesel Engine Exhaust Emissions in
the Workplace. Ann Occ Hyg 44(6):435-447.

188. Blute, NA; Woskie, SR; Greenspan, CA. (1999). Exposure Characterization for Highway
Construction Part 1: Cut and Cover and Tunnel Finish Stages. Applied Occ Envir Hyg
14(9):632-641.

189.Verma, D.K.; Kurtz, L.A.; Sahai, D.; et al. (2003) Current chemical exposures among
Ontario construction workers.  Appl Occup Environ Hygiene 18: 1031-1047.

190. U.S. EPA ( 2002). Diesel PM model-to-measurement comparison. Prepared by ICF
Consulting for EPA, Office of Transportation and Air Quality. Report No. EPA420-D-02-004.

191. California EPA.  (1998).  Proposed Identification of Diesel Exhaust as a Toxic Air
Contaminant. Appendix III, Part A: Exposure Assessment. California Environmental Protection
Agency.  California Air Resources Board, April 22, 1998. Available at
http://www.arb.ca.gov/toxics/diesel/diesel.htm.

192. U.S. EPA (2002). National-Scale Air Toxics Assessment.  This material is available
electronically at http://www.epa.gov/ttn/atw/nata/.

193. U.S. EPA (2001). 1996 National Toxics Inventory. This material is available
electronically at http://www.epa.gov/ttn/chief/nti/.

194. CookR., M.  Strum, J. Touma and R. Mason. (2002). Contribution of Highway and
Nonroad Mobile source Categories to Ambient Concentrations of 20 Hazardous Air Pollutants in
1996. SAE Technical Paper No. 2002-01-0650.

195. Cook, R., M.  Strum, J. Touma, W. Battye, and R. Mason (2002). Trends in Mobile Source-
Related Ambient Concentrations of Hazardous Air Pollutants, 1996 to 2007.  SAE Technical
Paper No. 2002-01-1274.

196. U.S. EPA. (2002). Comparison of ASPEN Modeling System Results to Monitored
Concentrations,  http://www.epa.gov/ttn/atw/nata/draft6.html#SecI.

197. U.S. EPA (1993). Motor Vehicle-Related Air Toxics Study, U.S. Environmental Protection
Agency, Office of Mobile Sources, Ann Arbor, MI, EPA Report No. EPA 420-R-93-005, April
1993. http://www.epa.gov/otaq/toxics.htm.

                                         2-139

-------
Final Regulatory Impact Analysis
198. Eastern Research Group. (2000). Documentation for the 1996 Base Year National Toxics
Inventory for Onroad Sources. Prepared for U.S. EPA, Emission Factor and Inventory Group,
Office of Air Quality Planning and Standards, June 2, 2000. http://www.epa.gov/ttn/chief/nti/.

199. Cook, R. and E. Glover (2002). Technical Description of the Toxics Module for
MOBILE6.2 and Guidance on Its Use for Emission Inventory Preparation. U.S. EPA, Office of
Transportation and Air Quality, Ann Arbor, MI. Report No. EPA420-R-02-011.
http ://www. epa.gov/otaq/m6. htm.

200. U.S. EPA. (1999). Analysis of the Impacts of Control Programs on Motor Vehicle Toxic
Emissions and Exposure in Urban Areas and Nationwide: Volume I.  Prepared for EPA by Sierra
Research, Inc.  and Radian International Corporation/Eastern Research Group, November 30,
1999. Report No. EPA420-R-99-029. http://www.epa.gov/otaq/toxics.htm.

201. U.S. EPA (2000). Integrated Risk Information System File for Benzene. This material is
available electronically at http://www.epa.gov/iris/subst/0276.htm.

202. International Agency for Research on Cancer, IARC. (1982). Monographs on the
evaluation of carcinogenic risk of chemicals to humans, Volume 29, Some industrial chemicals
and dyestuffs, International Agency for Research on Cancer, World Health Organization, Lyon,
France, p. 345-389.

203. Irons,  R.D., W.S. Stillman, D.B. Colagiovanni, and V.A. Henry. (1992) Synergistic action
of the benzene metabolite hydroquinone on myelopoietic stimulating activity of
granulocyte/macrophage colony-stimulating factor in vitro, Proc. Natl. Acad. Sci. 89:3691-3695.

204. U.S. EPA (1985). Environmental Protection Agency, Interim quantitative cancer unit risk
estimates due to inhalation of benzene, prepared by the Office of Health and Environmental
Assessment, Carcinogen Assessment Group, Washington, DC. for the Office of Air Quality
Planning and Standards, Washington, DC., 1985.

205. Clement Associates, Inc. (1991). Motor vehicle air toxics health information, for U.S. EPA
Office of Mobile Sources, Ann Arbor, MI, September 1991.

206. International Agency for Research on Cancer (IARC) (1982). IARC monographs on the
evaluation of carcinogenic risk of chemicals to humans, Volume 29, Some industrial chemicals
and dyestuffs, International Agency for Research on Cancer, World Health Organization, Lyon,
France, p. 345-389.

207. Irons,  R.D., W.S. Stillman, D.B. Colagiovanni, and V.A. Henry (1992). Synergistic action
of the benzene metabolite hydroquinone on myelopoietic stimulating activity of
granulocyte/macrophage colony-stimulating factor in vitro, Proc. Natl. Acad. Sci. 89:3691-3695.

208. U.S. EPA (1998).  Environmental Protection Agency, Carcinogenic Effects of Benzene: An
Update, National Center for Environmental Assessment, Washington, DC. 1998. EPA600-P-97-
                                         2-140

-------
                                             Air Quality, Health, and Welfare Effects
001F. http://www.epa.gov/ncepihom/Catalog/EPA600P97001F.html.

209. Hayes, R. B., Yin, S. N., Dosemici, M. S., et al. (1997). Benzene and the dose-related
incidence of hematological neoplasms in China. J. Nat. Cancer Inst. 89:1065-1071.

210. Aksoy, M. (1989). Hematotoxicity and carcinogenicity of benzene. Environ. Health
Perspect. 82: 193-197.

211. Goldstein, B.D. (1988). Benzene toxicity. Occupational medicine. State of the Art
Reviews. 3: 541-554.

212. Aksoy, M (1991).  Hematotoxicity, leukemogenicity and carcinogenicity of chronic
exposure to benzene. In: Arinc, E.; Schenkman, J.B.; Hodgson, E., Eds.  Molecular Aspects of
Monooxygenases and Bioactivation of Toxic Compounds. New York: Plenum Press, pp. 415-
434.

213. Goldstein, B.D. (1988). Benzene toxi city. Occupational medicine. State of the Art
Reviews. 3: 541-554.

214. Aksoy, M., S. Erdem, and G. Dincol. (1974). Leukemia in shoe-workers exposed
chronically to benzene.  Blood 44:837.

215. Aksoy, M. and K. Erdem. (1978).  A follow-up study on the mortality and the development
of leukemia in 44 pancytopenic patients associated with long-term exposure to benzene. Blood
52:  285-292.

216. Rothman, N., G.L. Li, M. Dosemeci, W.E. Bechtold, G.E. Marti, Y.Z. Wang, M. Linet,
L.Q. Xi, W. Lu, M.T. Smith, N. Titenko-Holland, L.P. Zhang, W. Blot, S.N. Yin, and R.B.
Hayes (1996). Hematotoxicity among Chinese workers heavily exposed to benzene. Am. J. Ind.
Med. 29: 236-246.

217. Kinnee, E., A. Beidler, J. S. Touma, M. Strum, C. R. Bailey, and R. Cook. (2004).
Allocation of Onroad Mobile Emissions to Road Segments for Air Toxics Modeling in Harris
County, Texas. Transportation Research Part D: Transport and Environment 9(2): 139-150.

218. Cohen, J., R. Cook, C. R. Bailey, and E. Carr.  (2004).  Relationship Between Motor
Vehicle Emissions of Hazardous Pollutants, Roadway Proximity, and Ambient Concentrations in
Portland Oregon. Environmental Modelling and Software, in press.

219. Sapkota, A. and T.  J. Buckley. (2003).  The mobile source effect on curbside 1,3-butadiene,
benzene, and particle-bound polycyclic aromatic hydrocarbons assessed at a tollbooth. J Air
Waste Manage Assoc 53: 740-748.

220. Janssen, N.A.H.; P. H. N. van Vliet, F. Aarts, et al. (2000) Assessment of exposure to traffic
related air pollution of children attending schools near motorways.  Atmos Environ 35: 3875-
3884.

                                        2-141

-------
Final Regulatory Impact Analysis
221. Skov, H.; A. B. Hansen, G. Lorenzen, et al. (2001) Benzene exposure and the effect of
traffic pollution in Copenhagen, Denmark. Atmos Environ 35: 2463-2471.

222. Payne-Sturges, D., T. A. Burke, P. Breysse, et al. (2004) Personal exposure meets risk
assessment: a comparison of measured and modeled exposures and risks in an urban community.
Environ Health Perspect doi:10.1289/ehp.6496 [Online at http://dx.doi.org/]

223. U.S. EPA (1987).  Integrated Risk Information System File of Butadiene. This material is
available electronically at http://www.epa.gov/iris/subst/0139.htm

224. U.S. EPA. (2002).  Health Assessment of 1,3-Butadiene. Office of Research and
Development, National Center for Environmental Assessment, Washington Office, Washington,
DC. Report No. EPA600-P-98-001F.

225. U.S. EPA (2002). Health Assessment of Butadiene, This material is available electronically
at http://cfpub.epa.gov/ncea/cfm/recordisplay.cfm?deid=54499.

226. U.S. EPA (1998).  A Science Advisory Board  Report: Review of the Health Risk
Assessment of 1,3-Butadiene. EPA-SAB-EHC-98.

227. Delzell, E; Sathiakumar, N; Macaluso, M.; et al. (1995) A follow-up study of synthetic
rubber workers. Submitted to the International Institute of Synthetic Rubber Producers.
University of Alabama at Birmingham.  October 2, 1995.

228. Bevan, C; Stadler, JC; Elliot, GS; et al. (1996) Subchronic toxicity of 4-vinylcyclohexene in
rats and mice by inhalation. Fundam. Appl. Toxicol. 32:1-10.

229. Southwest Research Institute.  (2002). Nonroad Duty Cycle Testing for  Toxic Emissions.
Prepared for the U.S. Environmental Protection Agency, Office of Transportation and Air
Quality, September 2002.  Report No. SwRI 08.5004.05.

230. U.S. EPA (1987).  Environmental Protection Agency, Assessment of health risks to
garment workers and certain home residents from exposure to formaldehyde, Office of Pesticides
and Toxic Substances, April 1987.

231. U.S. EPA (1991). Integrated Risk Information System File of Formaldehyde.  This material
is available electronically at http://www.epa.gov/iris/subst/0419.htm.

232. Blair, A., P.A. Stewart, R.N. Hoover, et al. (1986). Mortality among industrial workers
exposed to formaldehyde. J. Natl. Cancer Inst. 76(6): 1071-1084.

233. Kerns, W.D., K.L. Pavkov, D.J. Donofrio, EJ. Gralla and J.A. Swenberg. (1983).
Carcinogenicity of formaldehyde in  rats and mice after long-term inhalation exposure. Cancer
Res. 43:4382-4392.
                                         2-142

-------
                                             Air Quality, Health, and Welfare Effects
234. Albert, R.E., A.R. Sellakumar, S. Laskin, M. Kuschner, N. Nelson and C.A. Snyder.
Gaseous formaldehyde and hydrogen chloride induction of nasal cancer in the rat. J. Natl.
Cancer Inst. 68(4): 597-603.

235. Tobe, M., T. Kaneko, Y. Uchida, et al. (1985) Studies of the inhalation toxicity of
formaldehyde. National Sanitary and Medical Laboratory Service (Japan), p. 1-94.

236. Clement Associates, Inc. (1991). Motor vehicle air toxics health information, for U.S. EPA
Office of Mobile Sources, Ann Arbor, MI, September 1991.

237. Ulsamer, A. G., J. R. Beall, H. K. Kang, et al. (1984). Overview of health effects of
formaldehyde. In: Saxsena, J. (ed.) Hazard Assessment of Chemicals - Current Developments.
NY: Academic Press, Inc. 3:337-400.

238. Chemical Industry Institute of Toxicology (1999). Formaldehyde: Hazard Characterization
and Dose-Response Assessment for Carcinogenicity by the Route of Inhalation.

239. Blair, A., P. Stewart, P. A. Hoover, et al. (1987).  Cancers of the nasopharynx and
oropharynx and formaldehyde exposure. J. Natl. Cancer Inst. 78(1): 191-193.

240. Wilhelmsson, B. and M. Holmstrom. (1987). Positive formaldehyde PAST after prolonged
formaldehyde exposure by inhalation. The Lancet: 164.

241. Burge, P.S., M.G. Harries, W.K. Lam, I.M. O'Brien, and P.A. Patchett. (1985).
Occupational  asthma due to formaldehyde. Thorax 40:225-260.

242. Hendrick, D.J., RJ. Rando, DJ. Lane, and MJ. Morris (1982). Formaldehyde asthma:
Challenge exposure levels and fate after five years. J. Occup. Med. 893-897.

243. Nordman, H., H. Keskinen, and M. Tuppurainen. (1985). Formaldehyde asthma - rare or
overlooked? J. Allergy Clin. Immunol. 75:91-99.

244. U.S. EPA (1988).  Integrated Risk Information System File of Acetaldehyde. This material
is available electronically at http://www.epa.gov/iris/subst/0290.htm.

245. Feron, VJ. (1979). Effects of exposure to acetaldehyde in Syrian hamsters simultaneously
treated with benzo(a)pyrene or diethylnitrosamine. Prog. Exp. Tumor Res. 24: 162-176.

246. Feron, V.J., A. Kruysse and R.A. Woutersen. (1982). Respiratory tract tumors in hamsters
exposed to acetaldehyde vapour alone or simultaneously to benzo(a)pyrene or
diethylnitrosamine. Eur. J. Cancer Clin. Oncol. 18: 13-31.

247. Woutersen, R.A. and L.M. Appelman. (1984). Lifespan inhalation carcinogenicity study of
acetaldehyde in rats. III. Recovery after 52 weeks of exposure. Report No. V84.288/190172.
CIVO-Institutes TNO, The Netherlands.
                                         2-143

-------
Final Regulatory Impact Analysis
248. Wouterson, R., A. Van Garderen-Hoetmer and L.M. Appelman. 1985. Lifespan (27
months) inhalation carcinogenicity study of acetaldehyde in rats. Report No. V85.145/190172.
CIVO-Institutes TNO, The Netherlands.

249. California Air Resources Board (1992). Preliminary Draft: Proposed identification of
acetaldehyde as a toxic air contaminant, Part B Health assessment, California Air Resources
Board, Stationary Source Division, August, 1992.

250. Myou, S.; Fujimura, M.; Nishi, K.; et al. (1993) Aerosolized acetaldehyde induces
histamine-mediated bronchoconstriction in asthmatics.  Am Rev Respir Dis 148(4 Pt 1): 940-3.

251. Agency for Toxic Substances and Disease Registry (ATSDR). Toxicological Profile for
Acrolein. Public Health Service, U.S. Department of Health and Human Services, Atlanta, GA.
1990.

252. Sim VM, Pattle RE. Effect of possible smog irritants on human subjects JAMA165: 1980-
2010, 1957.

253. U.S. Environmental Protection Agency. Integrated Risk Information System (IRIS) on
Acrolein. National Center for Environmental  Assessment, Office of Research and Development,
Washington, D.C. 2003.

254. Agency for Toxic Substances and Disease Registry (ATSDR). Toxicological Profile for
Acrolein. Public Health Service, U.S. Department of Health and Human Services, Atlanta, GA.
1990.

255.Agency for Toxic Substances and Disease Registry (ATSDR). Toxicological Profile for
Acrolein. Public Health Service, U.S. Department of Health and Human Services, Atlanta, GA.
1990.

256.U.S. Environmental Protection Agency. Integrated Risk Information System (IRIS) on
Acrolein. National Center for Environmental  Assessment, Office of Research and Development,
Washington, D.C. 2003.

257.Lyon J, Jenkins L, Jones R, Coon R, siegel J. Repeated and continuous exposure of
laboratory animals to acrolein. Toxicol. Appl. Pharmacol. 17:726-732, 1970.

258.Egle JL. Retention of inhaled formaldehyde, proprionaldehyde, and acrolein in the dosg.
Arch. Environ. Health. 25; 119-124, 1972.

259.Kane LE, alarie Y. Sensory irritation to formaldehyde and acrolein during single and
repeated exposures to mice. Am Ind. Assoc. J. 38: 509-517, 1977.
Perhaps the most significant exposure humans have to acrolein results from mainstream tobacco
smoke where acrolein concentrations can peak at 90 ppm per puff (see Newsome et al. 1965).
Much of the toxicology of acrolein has been associated with tobacco smoke or linked to potential
                                         2-144

-------
                                              Air Quality, Health, and Welfare Effects
industrial accidental exposures, though there have been more prolonged studies of the irritant as
might be encountered in the work place or urban setting with ambient smog.

260.Murphy SD, Klingshirn DA, Ulrich CE. Respiratory response of guinea pigs during acrolein
inhlation and its modification by drugs. Pharmacol. Exp. Ther. 141: 79-83,  1963.

261.Feron VJ, Kruysee A, Til HP, Immel HR. Repeated exposure to acrolein vapour: subacute
studies in hamsters, rats, and rabbits. Toxicol.  9: 47-57, 1978.

262.Astey CL, Jakab GJ. The effects of acrolein exposure on antibacterial defenses. Toxicol.
Appl. Pharmacol. 67: 49-54, 1983.

263. Jakab GJ. The toxicologic interactions resulting from inhalation of carbon black and acrolein
on pulmonary antibacterial and anitviral defenses. Toxicol. Appl. Pharmacol.  121(2): 167-175,
1993.

264. Jakab GJ. Adverse effect of a cigarette smoke componenet, acrolein, on pulmonary
antibacterial defences and viral-bacterial interactions in the lung. Am. Rev.  Resp. Dis.  115: 33-
38, 1977.

265.Bouley G, Dubreuil A, Godin J, Boissel M, Boudene C. Phenomena of adaptation  in rats
continuously exposed to low concentrations of acrolein. Ann.  Occup. Hyg.  19: 27-32, 1976.

266.Newsome JR, Norma V, Keither CH. Vapor phase anlysis of tobacco smoke. Tobacco Sci.
9: 102-110, 1965.

267.Costa DL, Kutzman RS, Lehmann JR, Drew RT. Altered  lung function and structure in the
rat after subchronic exposure to acrolein. Am.  Rev. Resp. dis.  133: 286-291, 1986.

268.Kutzman R, Wehner R, Haber S (1984) Selected responses of hypertension-sensitive and
resistant rats to inhaled acrolein. Toxicology, 31:53-65.

269. U.S. EPA (2003). Integrated Risk Information System File of Acrolein. This material is
available electronically at http://www.epa.gov/iris/subst/0364.htm.

270. Dubowsky,  S.D.; Wallace, L.A.; and Buckley, T.J. (1999) The contribution of traffic to
indoor concentrations of poly cyclic aromatic hydrocarbons. J Expo Anal Environ Epidemiol
9(4):312-21.

271. Perera, P.P.; Rauh, V.; Tsai, W.Y.; et al. (2003) Effects of transplacental exposure to
environmental pollutants on birth outcomes in a multiethnic population. Environ Health
Perspect 111(2): 201-205.

272.U.S. EPA (1996).  Air Quality Criteria for Ozone and Related Photochemical Oxidants,
EPA600-P-93-004aF.  Docket No. A-99-06. Document Nos. II-A-15 to 17. More information
on health effects of ozone is also available at

                                         2-145

-------
Final Regulatory Impact Analysis
http ://www. epa. gov/ttn/naaqs/standards/ozone/s_03_index.html.

273.Bates, D.V.; Baker-Anderson, M.; Sizto, R. (1990) Asthma attack periodicity: a study of
hospital emergency visits in Vancouver. Environ. Res. 51: 51-70.

274.Thurston, G.D.; Ito, K.; Kinney, P.L.; Lippmann, M. (1992) A multi-year study of air
pollution and respiratory hospital admissions in three New York State metropolitan areas:
results for 1988 and 1989 summers.  J. Exposure Anal. Environ. Epidemiol. 2:429-450.

275.Thurston, G.D.; Ito, K.; Hayes, C.G.; Bates, D.V.; Lippmann, M. (1994) Respiratory
hospital admissions and summertime haze air pollution in Toronto, Ontario: consideration of the
role of acid aerosols. Environ. Res. 65: 271-290.
276.Lipfert, F.W.; Hammerstrom, T. (1992) Temporal patterns in air pollution and hospital
admissions. Environ. Res. 59: 374-399.

277.Burnett, R.T.; Dales, R.E.; Raizenne, M.E.; Krewski, D.; Summers, P.W.; Roberts, G.R.;
Raad-Young, M.; Dann,T.; Brook, J. (1994) Effects of low ambient levels of ozone and sulfates
on the frequency of respiratory admissions to Ontario hospitals. Environ.  Res. 65: 172-194.

278. U.S. EPA (1996). Air Quality Criteria for Ozone and Related Photochemical Oxidants,
EPA600-P-93-004aF. Docket No. A-99-06. Document Nos. II-A-15 to 17. (See page 9-33).

279. U.S. EPA (1996). Air Quality Criteria for Ozone and Related Photochemical Oxidants,
EPA600-P-93-004aF. Docket No. A-99-06. Document Nos. II-A-15 to 17. (See page 7-167).

280.Devlin, R. B.; McDonnell, W. F.; Mann, R.; Becker, S.; House, D. E.; Schreinemachers, D.;
Koren, H. S. (1991) Exposure of humans to ambient levels of ozone for 6.6 hours causes
cellullar and biochemical changes in the lung.  Am. J. Respir. Cell Mol. Biol. 4: 72-81.

281.Koren, H.  S.; Devlin, R. B.; Becker, S.; Perez, R.; McDonnell, W. F.  (1991) Time-
dependent changes of markers associated with inflammation in the lungs of humans exposed to
ambient levels of ozone. Toxicol. Pathol. 19: 406-411.

282.Koren, H.  S.; Devlin, R. B.; Graham, D. E.; Mann, R.; McGee, M. P.; Horstman, D. H.;
Kozumbo, W. J.; Becker, S.; House, D. E.; McDonnell, W. F.; Bromberg, P. A. (1989a) Ozone-
induced inflammation in the lower airways of human subjects. Am. Rev. Respir. Dis. 139: 407-
415.

283.Schelegle, E.S.; Siefkin, A.D.; McDonald, RJ. (1991) Time course of ozone-induced
neutrophilia in normal humans.  Am. Rev. Respir. Dis.  143:1353-1358.

284.U.S. EPA (1996). Air Quality Criteria for Ozone and Related Photochemical Oxidants,
EPA600-P-93-004aF. Docket No. A-99-06. Document Nos. II-A-15 to 17. (See page 7-171)
                                        2-146

-------
                                             Air Quality, Health, and Welfare Effects
285.Hodgkin, I.E.; Abbey, D.E.; Euler, G.L.; Magie, A.R. (1984) COPD prevalence in
nonsmokers in high and low photochemical air pollution areas. Chest 86: 830-838.

286.Euler, G.L.; Abbey, D.E.; Hodgkin, I.E.; Magie, A.R. (1988) Chronic obstructive
pulmonary disease symptom effects of long-term cumulative exposure to ambient levels of total
oxidants and nitrogen dioxide in California Seventh-day Adventist residents. Arch. Environ.
Health 43: 279-285.

287.Abbey, D.E.; Petersen, F.; Mills, P.K.; Beeson, W.L. (1993) Long-term ambient
concentrations of total suspended particulates, ozone, and sulfur dioxide and respiratory
symptoms in a nonsmoking population. Arch. Environ. Health 48: 33-46.

288.U.S. EPA. (1996). Review of National Ambient Air Quality Standards for Ozone,
Assessment of Scientific and Technical Information, OAQPS Staff Paper, EPA452-R-96-007.
Docket No. A-99-06. Document No. II-A-22.

289.U.S. EPA (1996). Air Quality  Criteria for Ozone and Related Photochemical Oxidants,
EPA600-P-93-004aF. Docket No. A-99-06. Document Nos. II-A-15 to 17.

290.U.S. EPA. (1996). Review of National Ambient Air Quality Standards for Ozone,
Assessment of Scientific and Technical Information, OAQPS Staff Paper, EPA452-R-96-007.
Docket No. A-99-06. Document No. II-A-22.

291.U.S. EPA (1996). Air Quality  Criteria for Ozone and Related Photochemical Oxidants,
EPA600-P-93-004aF. Docket No. A-99-06. Document Nos. II-A-15 to 17. (See page 7-170)

292.Avol, E. L.; Trim, S. C.; Little, D. E.; Spier, C. E.;  Smith,  M. N.; Peng, R.-C.; Linn, W. S.;
Hackney, J. D.; Gross, K. B.; D'Arcy, J. B.; Gibbons, D.; Higgins, I. T. T. (1990) Ozone
exposure and lung function in children attending a southern California summer camp.  Presented
at: 83rd annual meeting and exhibition of the Air & Waste Management Association; June;
Pittsburgh, PA. Pittsburgh, PA: Air & Waste Management Association; paper no. 90-150.3.

293.Higgins, I. T. T.; D'Arcy, J. B.; Gibbons, D. L; Avol, E. L.; Gross, K. B. (1990) Effect of
exposures to ambient ozone on ventilatory lung function in children. Am. Rev. Respir. Dis. 141:
1136-1146.

294.Raizenne, M. E.; Burnett, R. T.; Stern, B.; Franklin, C. A.; Spengler, J. D. (1989)  Acute lung
function responses to ambient acid aerosol exposures in children. Environ. Health Perspect. 79:
179-185.

295.Raizenne, M.; Stern, B.; Burnett, R.; Spengler, J. (1987) Acute respiratory function and
transported air pollutants:  observational studies. Presented at: 80th annual meeting of the Air
Pollution Control Association; June; New York, NY. Pittsburgh, PA: Air Pollution Control
Association; paper no. 87-32.6.
                                         2-147

-------
Final Regulatory Impact Analysis
296.Spektor, D. M.; Lippmann, M. (1991) Health effects of ambient ozone on healthy children at
a summer camp.  In: Berglund, R. L.; Lawson, D. R.; McKee, D. J., eds. Tropospheric ozone and
the environment:  papers from an international conference; March 1990; Los Angeles, CA.
Pittsburgh, PA: Air & Waste Management Association; pp. 83-89. (A&WMA transaction series
no. TR-19).

297.Spektor, D. M.; Thurston, G. D.; Mao, J.; He, D.; Hayes, C.; Lippmann, M. (1991) Effects
of single- and multiday ozone exposures on respiratory function in active normal children.
Environ.  Res. 55: 107-122.

298.Spektor, D. M.; Lippman, M.; Lioy, P. J.; Thurston, G. D.;s Citak, K.; James, D. J.; Bock,
N.; Speizer, F. E.; Hayes, C. (1988a) Effects of ambient ozone on respiratory function in active,
normal children. Am. Rev. Respir. Dis. 137: 313-320.

299.U.S.  EPA (1996). Air Quality Criteria for Ozone and Related Photochemical Oxidants,
EPA600-P-93-004aF.  Docket No. A-99-06. Document Nos. II-A-15 to 17. (See pages 7-160 to
7-165)

SOO.Hazucha, M. J.; Folinsbee, L. J.; Seal, E., Jr. (1992) Effects of steady-state and variable
ozone concentration profiles on pulmonary function. Am. Rev. Respir. Dis. 146: 1487-1493.

SOl.Horstman, D.H.; Ball, B.A.; Folinsbee, L.J.; Brown, J.; Gerrity, T. (1995) Comparison of
pulmonary responses of asthmatic and nonasthmatic subjects performing light exercise while
exposed to a low level of ozone. Toxicol.  Ind. Health.

302.Horstman, D.H.; Folinsbee, L.J.; Ives, P.J.; Abdul-Salaam, S.; McDonnell, W.F. (1990)
Ozone concentration and pulmonary response relationships for 6.6-hour exposures with five
hours of moderate exercise to 0.08, 0.10, and 0.12 ppm. Am. Rev. Respir. Dis. 142: 1158-1163.

303.U.S.  EPA. (1996). Review of National Ambient Air Quality Standards for Ozone,
Assessment of Scientific and Technical Information, OAQPS  Staff Paper, EPA452-R-96-007.
Docket No. A-99-06.  Document No. II-A-22.

304.  New Ozone Health and Environmental Effects References, Published Since Completion of
the Previous Ozone AQCD, National Center for Environmental Assessment, Office of Research
and Development, U.S. Environmental Protection Agency, Research Triangle Park, NC 27711,
July 2002. Docket No. A-2001-11, Document No. IV-A-19.

305. Thurston, G.D., M.L. Lippman, M.B. Scott, and J.M. Fine. 1997. Summertime Haze Air
Pollution and Children with Asthma. American Journal of Respiratory Critical Care Medicine,
155: 654-660. Ostro et al., 2001)

306. Ostro, B, M. Lipsett, J. Mann, H. Braxton-Owens, and M. White (2001) Air pollution and
exacerbation of asthma in African-American children in Los Angeles. Epidemiology 12(2): 200-
208.
                                        2-148

-------
                                             Air Quality, Health, and Welfare Effects
307.McDonnell, W.F., D.E. Abbey, N. Nishino and M.D. Lebowitz. 1999. "Long-term ambient
ozone concentration and the incidence of asthma in nonsmoking adults: the ahsmog study."
Environmental Research. 80(2 Pt 1): 110-121.

308.McConnell, R.; Berhane, K.; Gilliland, F.; London, S. J.; Islam, T.; Gauderman, W. J.; Avol,
E.; Margolis, H. G.; Peters, J. M. (2002) Asthma in exercising children exposed to ozone: a
cohort study. Lancet 359: 386-391.

309.Burnett, R. T.; Smith_Doiron, M.; Stieb, D.; Raizenne, M. E.; Brook, J. R.; Dales, R. E.;
Leech, J. A.; Cakmak, S.; Krewski, D. (2001) Association between ozone and hospitalization for
acute respiratory diseases in children less than 2 years of age. Am. J. Epidemiol. 153: 444-452.

310. Chen, L.; Jennison, B. L.; Yang, W.; Omaye, S. T. (2000) Elementary school absenteeism
and air pollution. Inhalation Toxicol. 12: 997-1016.

311. Gilliland, FD, K  Berhane, EB Rappaport, DC Thomas, E Avol, WJ Gauderman, SJ London,
HG Margolis, R McConnell, KT Islam, JM Peters (2001) The effects of ambient air pollution on
school absenteeism due to respiratory illnesses Epidemiology 12:43-54.

312.Devlin, R. B.; Folinsbee, L. J.; Biscardi, F.; Hatch, G.; Becker, S.; Madden, M. C.; Robbins,
M.; Koren, H. S. (1997) Inflammation and cell damage induced by repeated exposure of humans
to ozone. Inhalation Toxicol. 9: 211-235.

313.Koren HS, Devlin RB, Graham DE, Mann R, McGee MP, Horstman DH, Kozumbo WJ,
Becker S, House DE,  McDonnell SF, Bromberg, PA.  1989. Ozone-induced inflammation in the
lower airways of human subjects.  Am. Rev. Respir. Dies.  139: 407-415.

314.Samet JM, Zeger SL, Dominici F, Curriero F, Coursac I, Dockery  DW, Schwartz J,
Zanobetti A. 2000. The National Morbidity, Mortality and Air Pollution Study: Part II:
Morbidity, Mortality and Air Pollution in the United States. Research  Report No. 94, Part II.
Health Effects Institute, Cambridge MA, June 2000. (Docket Number A-2000-01, Document
Nos. IV-A-208 and 209)

315.Thurston, G. D.; Ito, K. (2001) Epidemiological studies of acute ozone exposures and
mortality. J. Exposure Anal. Environ. Epidemiol. 11: 286-294.

316.Touloumi, G.; Katsouyanni, K.; Zmirou, D.; Schwartz, J.; Spix, C.; Ponce de Leon, A.;
Tobias, A.; Quennel, P.; Rabczenko, D.; Bacharova, L.; Bisanti, L.; Vonk, J. M.; Ponka, A.
(1997) Short-term effects of ambient oxidant exposure on mortality: a  combined analysis within
the APHEA project. Am. J. Epidemiol. 146: 177-185.

317. Greenbaum, D.  Letter to colleagues dated May 30, 2002. [Available at
www.healtheffects.org]. Letter from L.D. Grant, Ph.D. to Dr. P. Hopke re: external review of
EPA's Air Quality Criteria for Particulate Matter, with copy of 05/30/02 letter from Health
Effects Institute re: re-analysis of National Morbidity, Mortality and Air Pollution Study data
                                        2-149

-------
Final Regulatory Impact Analysis
attached. Docket No. A-2000-01. Document No. IV-A-145.

318. U.S. EPA (2004). Technical Support Document for Nonroad Diesel Engine and Fuel
Rulemaking. Office of Air Quality Planning and Standards.  April 2004.

319.U.S. EPA (1996). Review of National Ambient Air Quality Standards for Ozone,
Assessment of Scientific and Technical Information, OAQPS Staff Paper, EPA452R-96-007.
Docket No. A-99-06. Document No. II-A-22.

320. U.S. EPA (1999). Draft Guidance on the Use of Models and Other Analyses in Attainment
Demonstrations for the 8-Hour Ozone NAAQS, Office of Air Quality Planning and Standards,
Research Triangle Park, NC. http://www.epa.gov/scram001/guidance/guide/drafto3.pdf

321.U.S. EPA (1999). "Technical Support Document for Tier 2/Gasoline Sulfur Ozone
Modeling Analyses" [memo from Pat Dolwick, OAQPS].  December 16, 1999. Docket No. A-
99-06.  Docket No. II-A-30.

322.U.S. EPA (2003). Technical Support Document for Nonroad Diesel Proposed Rulemaking.

323.U.S. EPA (2003). Technical Support Document for Nonroad Diesel Proposed Rulemaking

324.U.S. EPA (2003). Technical Support Document for Nonroad Diesel Proposed Rulemaking

325.U.S. EPA (2003). Technical Support Document for Nonroad Diesel Proposed Rulemaking.

326. NARSTO Synthesis Team (2000).  An Assessment of Tropospheric Ozone Pollution: A
North American Perspective.

327. Fujita, E.M., W.R. Stockwell, D.E.  Campbell, R.E. Keislar, and D.R.  Lawson (2003).
Evolution of the Magnitude and Spatial Extent of the Weekend Ozone Effect in California's
South Coast Air Basin from 1981 to 2000, Submitted to the J. Air &  Waste Manage. Assoc.

328. Marr, L.C. and R.A. Harley (2002). Modeling the Effect of Weekday-Weekend
Differences in Motor Vehicle Emissions  on Photochemical Air Pollution in Central California,
Environ. Sci. Technol, 36, 4099-4106.

329. Larsen, L.C. (2003). The Ozone Weekend Effect in California: Evidence Supporting NOx
Emissions Reductions, Submitted to the J. Air & Waste Manage. Assoc.

330. U.S. EPA (2003). Air Quality Technical Support Document for the proposed Nonroad
Diesel rulemaking.

331. Two counties in the Atlanta CMSA and one in the Baltimore-Washington CMSA.

332. For example, see letters in the Air Docket for this rule from American Lung Association,
Clean Air Trust, California Environmental Protection Agency, New York State Department of

                                       2-150

-------
                                            Air Quality, Health, and Welfare Effects
Environmental Conservation, Texas Commission on Environmental Quality (TCEQ, formerly
Texas Natural Resources Conservation Commission), State and Territorial Air Pollution
Program Administrators and the Association of Local Air Pollution Control Officials
(STAPPA/ALAPCO), Natural Resources Defense Council, Sierra Club, and Union of Concerned
Scientists.)

333. U.S. Environmental Protection Agency, 1999. The Benefits and Costs of the Clean Air
Act, 1990-2010. Prepared for U.S. Congress by U.S. EPA, Office of Air and Radiation, Office
of Policy Analysis and Review, Washington, DC, November; EPA report no. EPA410-R-99-001.

334. U.S. EPA (1996).  Air Quality Criteria for Ozone and Related Photochemical Oxidants,
EPA600-P-93-004aF. Docket No. A-99-06. Document Nos. II-A-15 to 17.

335. Winner, W.E., and CJ. Atkinson. 1986. Absorption of air pollution by plants, and
consequences for growth. Trends  in Ecology and Evolution 1:15-18.

336. U.S. EPA (1996).  Air Quality Criteria for Ozone and Related Photochemical Oxidants,
EPA600-P-93-004aF. Docket No. A-99-06. Document Nos. II-A-15 to 17.

337. Tingey, D.T., and Taylor, G.E. 1982. Variation in plant response to ozone: a conceptual
model of physiological events.  In: Effects of Gaseous Air Pollution in Agriculture and
Horticulture (Unsworth, M.H., Omrod, D.P., eds.) London, UK: Butterworth Scientific, pp. 113-
138.

338. U.S. EPA (1996).  Air Quality Criteria for Ozone and Related Photochemical Oxidants,
EPA600-P-93-004aF. Docket No. A-99-06. Document Nos. II-A-15 to 17.

339. U.S. EPA (1996).  Air Quality Criteria for Ozone and Related Photochemical Oxidants,
EPA600-P-93-004aF. Docket No. A-99-06. Document Nos. II-A-15 to 17.

340. U.S. EPA (1996).  Air Quality Criteria for Ozone and Related Photochemical Oxidants,
EPA600-P-93-004aF. Docket No. A-99-06. Document Nos. II-A-15 to 17.

341. Ollinger, S.V., J.D. Aber and P.B. Reich. 1997. Simulating ozone effects on forest
productivity: interactions between leaf canopy and stand level processes. Ecological
Applications 7:1237-1251.

342. Winner, W.E., 1994. Mechanistic analysis of plant responses to air pollution. Ecological
Applications, 4(4):651-661.

343. U.S. EPA (1996).  Air Quality Criteria for Ozone and Related Photochemical Oxidants,
EPA600-P-93-004aF. Docket No. A-99-06. Document Nos. II-A-15 to 17.

344. U.S. EPA (1996).  Air Quality Criteria for Ozone and Related Photochemical Oxidants,
EPA600P-93-004aF. Docket No. A-99-06. Document Nos. II-A-15 to 17.
                                        2-151

-------
Final Regulatory Impact Analysis
345. Fox, S., and R. A. Mickler, eds.. 1996.  Impact of Air Pollutants on Southern Pine Forests.
Springer-Verlag, NY, Ecol. Studies, Vol. 118, 513 pp.

346. National Acid Precipitation Assessment Program (NAPAP), 1991. National Acid
Precipitation Assessment Program. 1990 Integrated Assessment Report. National Acid
Precipitation Program. Office of the Director, Washington DC.

347. U.S. EPA (1996). Air Quality Criteria for Ozone and Related Photochemical Oxidants,
EPA600-P-93-004aF. Docket No. A-99-06.  Document Nos. II-A-15 to 17.

348. De Steiguer, J., J. Pye, C. Love.  1990. Air pollution Damage to U.S. forests. Journal of
Forestry, Vol 88(8) pp. 17-22.

349. Pye, J.M. Impact of ozone on the growth and yield  of trees: A review. Journal of
Environmental Quality 17 pp.347-360.,  1988.

350. U.S. EPA (1996). Air Quality Criteria for Ozone and Related Photochemical Oxidants,
EPA600-P-93-004aF. Docket No. A-99-06.  Document Nos. II-A-15 to 17.

351. U.S. EPA (1996). Air Quality Criteria for Ozone and Related Photochemical Oxidants,
EPA600-P-93-004aF. Docket No. A-99-06.  Document Nos. II-A-15 to 17.

352. McBride, J.R., P.R. Miller, and R.D. Laven. 1985. Effects of oxidant air pollutants on
forest succession in the mixed conifer forest type of southern California. In:  Air Pollutants
Effects On Forest Ecosystems, Symposium Proceedings,  St. P, 1985, p. 157-167.

353. Miller, P.R., O.C. Taylor, R.G. Wilhour. 1982. Oxidant air pollution effects on a western
coniferous forest ecosystem. Corvallis, OR: U.S. Environmental Protection Agency,
Environmental Research Laboratory (EPA600-D-82-276).

354. U.S. EPA (1996). Air Quality Criteria for Ozone and Related Photochemical Oxidants,
EPA600-P-93-004aF. Docket No. A-99-06.  Document Nos. II-A-15 to 17.

355. Hardner, J., A. VanGeel, K. Stockhammer, J. Neumann, and S. Ollinger.  1999.
Characterizing the Commercial Timber Benefits from Tropospheric Ozone Reduction
Attributable to the 1990 Clean Air Act Amendments, 1990-2010. Prepared for Office of Air
Quality Planning and Standards, U.S. Environmental Protection Agency.

356. U.S. EPA (1996). Air Quality Criteria for Ozone and Related Photochemical Oxidants,
EPA600-P-93-004aF. Docket No. A-99-06.  Document Nos. II-A-15 to 17.

357. Kopp, R. J.; Vaughn, W. J.; Hazilla, M.; Carson, R.  1985.  Implications of environmental
policy for U.S. agriculture: the case of ambient ozone standards. J. Environ. Manage. 20:321-
331.
                                        2-152

-------
                                            Air Quality, Health, and Welfare Effects
358. Adams, R. M.; Hamilton, S. A.; McCarl, B. A.  1986.  The benefits of pollution control: the
case of ozone and U.S. agriculture.  Am. J. Agric. Econ.  34: 3-19.

359. Adams, R. M.; Glyer, J. D.; Johnson, S. L.; McCarl, B. A.  1989.  A reassessment of the
economic effects of ozone on U.S. agriculture.  JAPCA 39:960-968.

360. Abt Associates, Inc.  1995.  Urban ornamental plants: sensitivity to ozone and potential
economic losses. U.S. EPA,  Office of Air Quality Planning and Standards, Research Triangle
Park.  Under contract to RADIAN Corporation, contract no. 68-D3-0033, WA no. 6.  pp. 9-10.

361. U.S. EPA (1993). Air Quality  Criteria for Oxides of Nitrogen, EPA600-8-91-049aF.
Docket No. A-2000-01. Document  Nos. II-A-89.

362.  U.S. EPA (1993). Air Quality Criteria for Oxides of Nitrogen, EPA600-8-91-049aF.
Docket No. A-2000-01. Document  Nos. II-A-89.

363. Hardner, J., A. VanGeel, K. Stockhammer, J. Neumann, and S. Ollinger.  1999.
Characterizing the Commercial Timber Benefits from Tropospheric Ozone Reduction
Attributable to the 1990 Clean Air Act Amendments, 1990-2010.  Prepared for Office of Air
Quality Planning and Standards, U.S. Environmental Protection Agency.

364. U.S. EPA (2000). Air Quality  Criteria for Carbon Monoxide. EPA600-P-99-001F. June 1,
2000. U.S. Environmental Protection Agency, Office of Research and Development, National
Center for Environmental Assessment, Washington, D.C.
http://www.epa.gov/ncea/pdfs/coaqcd.pdf (Docket A-2000-01, Document II-A-29).

365. Coburn, R.F. (1979) Mechanisms of carbon monoxide toxicity. Prev. Med. 8:310-322.

366. Helfaer, M.A., and Traystman, R.J. (1996) Cerebrovascular effects of carbon monoxide.
In:  Carbon Monoxide (Penney, D.G., ed). Boca Raton, CRC Press, 69-86.

367. Benignus, V.A.  (1994) Behavioral effects of carbon monoxide: meta analyses and
extrapolations. J. Appl. Physiol.  76:1310-1316. Docket A-2000-01, Document IV-A-127.

368. U.S. EPA (2000). Air Quality  Criteria for Carbon Monoxide. EPA600-P-99-001F. June 1,
2000. U.S. Environmental Protection Agency, Office of Research and Development, National
Center for Environmental Assessment, Washington, D.C.
http://www.epa.gov/ncea/pdfs/coaqcd.pdf (Docket A-2000-01, Document II-A-29).

369. U.S. EPA (2000). Air Quality  Criteria for Carbon Monoxide. EPA600-P-99-001F. June 1,
2000. U.S. Environmental Protection Agency, Office of Research and Development, National
Center for Environmental Assessment, Washington, DC.
http://www.epa.gov/ncea/pdfs/coaqcd.pdf (Docket A-2000-01, Document II-A-29).

370. U.S. EPA (2000). Air Quality  Criteria for Carbon Monoxide. EPA600-P-99-001F. June 1,
2000. U.S. Environmental Protection Agency, Office of Research and Development, National

                                        2-153

-------
Final Regulatory Impact Analysis
Center for Environmental Assessment, Washington, DC.
http://www.epa.gov/ncea/pdfs/coaqcd.pdf (Docket A-2000-01, Document II-A-29).

371. U.S. EPA (2000). Air Quality Criteria for Carbon Monoxide. EPA600-P-99-001F. June 1,
2000. U.S. Environmental Protection Agency, Office of Research and Development, National
Center for Environmental Assessment, Washington, DC.
http://www.epa.gov/ncea/pdfs/coaqcd.pdf (Docket A-2000-01, Document II-A-29).

372. National Air Quality and Emissions Trends Report,  1998, March, 2000; this document is
available at http://www.epa.gov/oar/aqtrnd98  National Air Pollutant Emission Trends, 1900-
1998 (EPA454-R-00-002), March 2000. These documents are available at Docket No. A-2000-
01, Document No. II-A-72. See also Air Quality Criteria for Carbon Monoxide, U.S. EPA,
EPA600-P-99-001F, June 2000, at 3-10. Air Docket A-2001-11.  This document is also
available at http://www.epa.gov/ncea/coabstract.htm.

373. Ref for Tier 2  and Large SI rules
                                        2-154

-------
CHAPTER 3: Emission Inventory
   3.1 Nonroad Diesel Baseline Emission Inventory Development  	3-2
      3.1.1 Land-Based Nonroad Diesel Engines—PM25, NOX, SO2, VOC, and CO Emissions
           	'	3-2
          3.1.1.1 Overview	3-2
          3.1.1.2NONROAD'sMajorInputs	3-3
          3.1.1.3 Emission Estimation Process	3-8
          3.1.1.4 Estimation of VOC Emissions	3-9
          3.1.1.5 Estimation of SO2 Emissions	3-10
          3.1.1.6 Estimation of PM25 Emissions	3-10
          3.1.1.7 Estimation of Fuel Consumption	3-11
          3.1.1.8 Changes from Draft NONROAD2002 to Draft NONROAD2004  	3-11
          3.1.1.9 Baseline Inventory	3-12
      3.1.2 Land-Based Nonroad Diesel Engines—Air Toxics Emissions 	3-15
      3.1.3 Commercial Marine Vessels and Locomotives  	3-16
      3.1.4 Recreational Marine Engines  	3-21
      3.1.5 Fuel Consumption for Nonroad Diesel Engines	3-24
   3.2 Contribution of Nonroad Diesel Engines to National Emission Inventories	3-26
      3.2.1 Baseline Emission Inventory Development	3-26
      3.2.2 PM25 Emissions  	3-28
      3.2.3 NOX Emissions  	3-28
      3.2.4 SO2 Emissions	3-29
      3.2.5 VOC Emissions	3-29
      3.2.6 CO Emissions	3-29
   3.3 Contribution of Nonroad Diesel Engines to Selected Local Emission Inventories . .  . 3-37
      3.3.1 PM25 Emissions  	3-37
      3.3.2NOXEmissions  	3-41
   3.4 Nonroad Diesel Controlled Emission Inventory Development	3-43
      3.4.1 Land-Based Diesel Engines—PM25, NOX, SO2, VOC, and CO Emissions 	3-43
          3.4.1.1 Standards and Zero-Hour Emission Factors  	3-44
          3.4.1.2 Transient Adjustment Factors  	3-44
          3.4.1.3 Deterioration Rates  	3-47
          3.4.1.4 In-Use Sulfur Levels, Certification Sulfur Levels, and Sulfur Conversion
             Factors 	3-47
          3.4.1.5 Controlled Inventory 	3-49
      3.4.2 Land-Based Diesel Engines—Air Toxics Emissions	3-52
      3.4.3 Commercial Marine Vessels and Locomotives  	3-53
      3.4.4 Recreational Marine Engines  	3-55
   3.5 Projected Emission Reductions from the Final Rule	3-58
      3.5.1 PM25 Reductions	3-58
      3.5.2 NOX Reductions	3-66
      3.5.3 SO2 Reductions	3-68
      3.5.4 VOC and Air Toxics Reductions  	3-75
      3.5.5 CO Reductions  	3-78
      3.5.6 PM25 and SO2 Reductions from the 15 ppm Locomotive and Marine (LM) Fuel
          Program  	3-79
      3.5.7 SO2 and Sulfate PM Reductions from Other Nonhighway Fuel	3-81
   3.6 Emission Inventories Used for Air Quality Modeling	3-86

-------
Final Regulatory Impact Analysis
                   CHAPTER 3:  Emission Inventory
   This chapter presents our analysis of the emission impact of the final rule for the four
categories of nonroad diesel engines affected: land-based diesel engines, commercial marine
diesel vessels, locomotives, and recreational marine diesel  engines. New engine controls are
being adopted for the land-based diesel engine category. For the other three nonroad diesel
categories, the final rule includes no new engine controls; however, the diesel fuel sulfur
requirements will decrease emissions of particulate matter  smaller than 2.5 microns (PM25) and
sulfur dioxide (SO2) for these categories.

   Section 3.1 presents an overview of the methodology used to generate the baseline
inventories.  The baseline inventories represent current and future  emissions with only the
existing standards. Sections 3.2 and 3.3 then describe the contribution of nonroad diesel engines
to national and selected local baseline inventories, respectively. Section 3.4 describes the
development of the controlled inventories, specifically the  changes made to the baseline inputs to
incorporate the new standards and fuel sulfur requirements. Section 3.5 follows with the
projected emission reductions resulting from the final rule. Section 3.6 concludes the chapter by
describing the changes  in the inputs and resulting emission inventories between the preliminary
baseline and control scenarios used for the air quality modeling and the updated baseline and
control scenarios in this final rule.

   The controlled inventory estimates do not include the potential uses of the averaging,
banking, and trading (ABT) program or the transition provisions for engine manufacturers, since
these are flexibilities that would be difficult to predict and  model.  More information regarding
these provisions can be found in Section III of the preamble.

   The estimates of baseline emissions and emission reductions for nonroad land-based,
recreational marine, locomotive, and commercial marine vessel diesel engines are reported for
both 48-state and  50-state inventories. The 48-state inventories are used for the air quality
modeling that EPA uses to analyze regional ozone and PM air quality,  of which Alaska and
Hawaii are not a part. In addition, 50-state emission estimates for  other sources (such as
stationary and area sources) are not available.  As a result,  in cases where nonroad diesel sources
are compared with other emission  sources, the 48-state  emission inventory estimates are used.

   Inventories are presented for the following pollutants: PM2 5, PM10, oxides of nitrogen (NOX),
SO2, volatile organic compounds (VOC), carbon monoxide (CO), and air toxics.  The specific air
toxics are  benzene, formaldeyde, acetaldehyde, 1,3-butadiene, and acrolein.  The PM inventories
include directly emitted PM only, although secondary sulfates are  taken into account in the air
quality modeling.
                                           5-2

-------
                                                                  Emission Inventory
3.1 Nonroad Diesel Baseline Emission Inventory Development

   This section describes how the baseline emission inventories were developed for the four
categories of nonroad diesel engines affected by this final rule: land-based diesel engines,
commercial marine diesel vessels, locomotives, and recreational marine diesel engines.  For
land-based diesel engines, there is a section that discusses inventory development for PM2 5,
NOX, SO2, VOC, and CO, followed by a section for air toxics.

3.1.1 Land-Based Nonroad Diesel Engines—PM2 5, NOX, SO2, VOC, and CO Emissions

   The baseline emission inventories for land-based diesel engines were generated using the
draft NONROAD2004 model. The baseline inventories account for the effect of existing federal
emission standards that establish three tiers of emission standards (Tier 1 through Tier 3).
Section 3.1.1.1 provides an overview of the draft NONROAD2004 model and a description of
the methodology used in the model to estimate emissions.  Details of the baseline modeling
inputs (e.g., populations, activity, and emission factors) for land-based diesel engines can be
found in the technical reports documenting the model. The single scenario option variable that
affects diesel emissions is the in-use fuel sulfur level.  The in-use diesel fuel sulfur level inputs
used for the baseline scenarios are given in Section 3.1.1.2.3.

   For the proposed rule, the draft NONROAD2002 model was used.  Section 3.1.1.8 describes
the changes made to the model for the  final rule.

   3.1.1.1 Overview

   The draft NONROAD2004 model  estimates emission inventories of important air emissions
from diverse nonroad equipment. The model's scope includes all nonroad sources with the
exception of locomotives, aircraft and  commercial marine vessels. Users can construct
inventories for criteria pollutants including carbon monoxide (CO), oxides of nitrogen (NOX),
oxides of sulfur (SO2), and particulate  matter (PM), as well as other emissions including total
hydrocarbon (THC) and carbon dioxide (CO2). As a related feature, the model estimates fuel
consumption. The model can distinguish emissions on the basis of equipment type,  size and
technology group. A central feature of the model is projection of future or past emissions
between 1970 and 2050.

   The draft NONROAD2004 model  contains three major components: (1) the core model, a
FORTRAN program that performs model calculations, (2) the reporting utility, a Microsoft
Access application that compiles and presents results, and (3) the graphic user interface (GUI), a
Visual-Basic application that allows users to easily construct scenarios for submission to the core
model. The following discussion will describe processes performed by the core model in the
calculation of emission  inventories.

-------
Final Regulatory Impact Analysis
   This section describes how the draft NONROAD2004 model estimates emissions particularly
relevant to this analysis, including parti culate matter (PM), oxides of nitrogen (NOX), oxides of
sulfur (SO2), carbon monoxide (CO) and volatile organic compounds (VOC). As appropriate,
we will focus on estimation of emissions of these pollutants by diesel engines. The model
estimates emissions from approximately 80 types of diesel equipment. As with other engine
classes, the model defines engine or equipment "size" in terms of the rated power (horsepower)
of the engine. For diesel engines, the regulations also classify engines on the basis of rated
power.

   The first four chemical species are exhaust emissions, i.e., pollutants emitted directly as
exhaust from combustion of diesel fuel in the engine. However, the last emission, VOC, includes
both exhaust and evaporative components. The exhaust component represents hydrocarbons
emitted as products of combustion; the evaporative component includes compounds emitted
from unburned fuel during operation, i.e., "crankcase emissions." For VOC, we will first
describe estimation of total hydrocarbon exhaust emissions, in conjunction with the description
for the other exhaust emissions. We discuss subsequent estimation of associated VOC emissions
in Section 3. 1.1. 4.

   3.1.1.2 NONROAD's Major Inputs

   The draft NONROAD2004 model uses three major sets of inputs in estimation of exhaust
emission inventories: (1) emission calculation variables, (2) projection variables, and (3)
scenario option variables.

   3.1.1.2.1 Emission Calculation Variables

   The draft NONROAD2004 model estimates exhaust emissions using the equation

                                    = E^-A-L-P-N
where each term is defined as follows:
   /exh = the exhaust emission inventory (gram/year, gram/day),
   Eexh = exhaust emission factor (gram/hp-hr),
   A = equipment activity (operating hours/year),
   L = Load factor (average proportion of rated power used during operation (percent)),
   P = average rated power (hp)
   N = Equipment population (units).

Emissions are then converted and reported as tons/year or tons/day.

   For diesel engines, each of the inputs applies to sub-populations of equipment, as classified
by type (dozer, tractor, backhoe, etc.), rated power class (50-100 hp, 100-300 hp, etc.) and
regulatory tier (tier 1, tier 2, etc.).
                                           5-4

-------
                                                                   Emission Inventory
   Exhaust Emission Factor.  The emission factor in a given simulation year consists of three
components, a "zero-hour" emission level (ZHL) , a transient adjustment factor (TAP) and a
deterioration factor (DF). The ZHL represents the emission rate for recently manufactured
engines, i.e., engines with few operating hours, and is typically derived directly from laboratory
measurements on new or nearly new engines on several commonly used duty cycles, hence the
term "zero-hour."

   Because most emission data have been collected under steady-state conditions (constant
engine speed and load), and because most real-world operation involves transient conditions
(variable speed and load), we attempt to adjust for the difference between laboratory
measurements and real-world operation through the use of transient adjustment factors (TAFs).
The TAF is a ratio representing the difference in the emission rate between transient and steady-
state operation. The TAFs are estimated by collecting emission measurements on specific
engines using both transient and steady-state cycles, and calculating the ratio
                                            FF
                                   r-p A 77 _     transient
                                        ~ EF
                                             steady-state
where EFtransient is the measurement for a given engine on a specific transient cycle, and EFsteady.state
is the corresponding measurement for the same engine on  a selected steady-state cycle.
Data from seven transient cycles were used to develop seven TAFs for each of the four
pollutants.  The seven cycle TAFs were then binned into two categories, based on the cycle load
factors. TAFs were then assigned to each equipment type represented in the model  on the basis
of engineering judgment. If steady-state operation was typical of an equipment type, no
adjustment was made (i.e., TAF = 1.0).

   Emission factors in the model input file represent the product (ZHL-TAF) for each
combination of equipment type, size class and regulatory tier represented by the model. We refer
to this product as the "baseline emission factor." For more detail on the derivation and
application of EFs and TAFs, refer to the model documentation on diesel emission factors.1

   During a model run,  the model applies emission deterioration to the baseline emission factor,
based on the age distribution of the equipment type in the  year simulated. Deterioration
expresses an assumption that emissions increase with equipment age and is expressed as a
multiplicative deterioration factor (DF). Thus, the final emission factor applied in the simulation
year is the product ZHL-TAF-DF. Deterioration factors vary from year to year; we  describe their
calculation in more detail in Section 3.1.1.2.2 below.

   The model estimates fuel consumption by substituting brake-specific fuel consumption
(BSFC, Ib/hp-hr) for the emission factor in the  equation above. We apply a TAF to the BSFC but
assume that BSFC does not  deteriorate with equipment age.

   In estimation of PM emissions, we apply an additional adjustment to the emission factor to
account for the in-use  sulfur level of diesel fuel.1 Based on user-specified diesel sulfur levels for
                                           5-5

-------
Final Regulatory Impact Analysis
a given scenario, NONROAD adjusts the PM emission factor by the margin SPMadj (g/hp-hr)
calculated as
                      SPMadj =  BSFC • m^ • mPHS • 0.01 • (Sbase - Sm_use )
where:    BSFC = brake-specific fuel consumption (g fuel/hp-hr),
       mso4,s = a constant, representing the sulfate fraction of total particulate sulfur, equal to 7.0
       g PM SO4/g PM S,
       mpu,s = a constant, representing the fraction of fuel sulfur converted to particulate sulfur,
       equal to 0.02247 g PM S/g fuel S,
       0.01 = conversion factor from wt% to wt fraction
       Sbase = base sulfur level in NONROAD (0.33 wt%, 3300 ppm for pre-control and Tier 1
       engines; 0.20 wt%, 2000 ppm for Tier 2-3 engines),
       Sm-use = in-use diesel  sulfur level as specified by user (wt%).

   Equipment Activity . Activity represents the usage of equipment, expressed in operating
hours per year. Activity estimates are specific to equipment types and remain constant in any
given simulation year. Activity estimates for diesel equipment have been adopted from the
Partslink model, a commercial  source developed and maintained by Power Systems
Research/Compass International, Inc. For discussion of activity estimates for specific equipment
types, refer to the technical documentation for the model.2

   Load Factor.  This parameter represents the average fraction  of rated power that equipment
uses during operation. Load factors are assigned by equipment type, and remain constant in any
simulation year. For use in draft NONROAD2004, we derived  load factors from the results of a
project designed to develop transient engine test cycles. During the course of the project, seven
cycles were developed,  designed to represent the operation of specific common equipment types.

   Specific load factors for the cycles fell into two broad groups, which we designated as "high"
and "low." We calculated an average for each group, with the high group containing four cycles
and the low group three; resulting load factors were 0.59 for the high group and 0.21 for the low
group. Then, we assigned one of these two factors to each equipment type for which we believed
engineering judgment was sufficient to make an assignment. For  remaining equipment types, for
which we considered engineering judgment insufficient to make an assignment, we assigned a
'steady-state' load factor, calculated as the average of load factors for all seven transient cycles
(0.43). Of NONROAD' s 90 diesel applications, half were assigned 'high' or 'low' load factors,
with the remainder assigned 'steady-state'  load factors. For more detail on the derivation of load
factors and assignment to specific equipment types, refer to the appropriate technical report2.

   Rated Power.  This parameter represents the average rated power for equipment, as assigned
to each combination of equipment type and rated-power class represented by the model. Values
assigned to a given type/power combination represents the sales-weighted average of engines for
that equipment type  in that rated-power class.3 Rated-power assignments remain constant in any
given simulation year. For use in draft NONROAD2004, we obtained estimates from the
Partslink database, maintained by Power Systems Research/Compass International, Inc. The
                                           5-6

-------
                                                                   Emission Inventory
product of load factor and rated power (LP) represents actual power output during equipment
operation.

   Equipment Population. As the name implies, this model input represents populations of
equipment pieces. For diesel engines, the model generates separate sub-populations for
individual combinations of equipment type and rated-power class. However, unlike activity and
load factor, populations do not remain constant from year to year. Projection of future or past
populations is the means through which the draft NONROAD2004 model projects future or past
emissions. As a reference point, the input file contains populations in the model's base year 2000
(updated from  1998 in draft NONROAD2002). We generated populations in the base year using
a simple attrition model that calculated base-year populations as a function of equipment sales,
scrappage, activity and load factor. Equipment sales by model year were obtained from the
commercially available Partslink database, developed and maintained by Power Systems
Research/Compass International, Inc. (PSR). This database contains sales estimates for nonroad
equipment for model years 1973 through 2000. Base-year population development is discussed
in the technical documentation.3

   3.1.1.2.2 Projection Variables

   The model uses three variables to project emissions over time: the annual population growth
rate, the equipment median life, and the relative deterioration rate. Collectively, these variables
represent population growth, changes in the equipment age distribution, and emission
deterioration.

   Annual Population Growth Rate (percent/year).  The population growth rate represents the
percentage increase in the equipment population for a given equipment type over successive
years. The growth rate is linear for diesel equipment,  and is applied to the entire population,
including all rated-power classes and tiers.4 Diesel growth rates vary by sector (e.g.,
agricultural, construction).

   Equipment Median Life (hours @full load). This variable represents the period of time over
which 50 percent of the engines in a given "model-year cohort" are scrapped.  A "model-year
cohort" represents a sub-population of engines represented as entering the population in a given
year.  The input value assumes that (1) engines are run at full load until failure, and (2)
equipment scrappage follows the model's scrappage curve. During a simulation, the model uses
the "annualized median life," which represents the actual service life of equipment in years,
depending on how much and how hard the equipment is used. Annualized median life is
calculated as median life in hours (lh\ divided by the product of activity and load factor (ly =
lf/AL). Engines persist in the equipment population over two median lives (2ly); during the first
median life,  50 percent of the engines are scrapped, and over the second, the remaining 50
percent are scrapped. For a more detailed description of median life, see the model
documentation.2
                                           5-7

-------
Final Regulatory Impact Analysis
   Relative Deterioration Rate (percent increase in emission factor/percent median life
expended).  This variable plays a key role in calculation of the deterioration factor.  Values of the
relative deterioration rate are assigned based on pollutant, rated-power class, and tier. Using the
relative deterioration rate (J), the annualized median life (ly) and the equipment age, draft
NONROAD2004 calculates the deterioration factor as
                         -L-'J- n,-,11ii+fin+ +iat- -tranr     *- '  *-^i
                            pollutant,tier,year ~       pollutant,tier    7
where:
   DFpollutantyea]. = the deterioration factor for a given pollutant for a model-year cohort in the
                 simulation year
   d     =  the relative deterioration rate for a given pollutant (percent increase in emission
              factor /percent useful life expended) and regulatory tier
   age   =  the age of a specific model-year group of engines in the simulation year
   ly     =  the annualized median life of the given model-year cohort (years)

   The deterioration factor adjusts the exhaust emission factor for engines in a given model-year
cohort in relation to the proportion of median life expended.  The model calculates the
deterioration linearly over one median life for a given model-year cohort (represented as a
fraction of the entire population). Following the first median life, the deteriorated emission
factor is held constant over the remaining life for engines in the cohort.  The model's
deterioration calculations are discussed in greater detail in the technical documentation.1

   3.1.1.2.3 Scenario Option Variables

   These inputs apply to entire model runs or scenarios, rather than to equipment. Scenario
options describe fuel characteristics and ambient weather conditions.  The option that applies to
inventories for diesel equipment is the in-use  diesel sulfur level (wt%).

   The in-use diesel fuel sulfur level inputs used for land-based diesel engines for the baseline
scenarios are provided in Table 3.1-1. The fuel sulfur levels account for spillover use of
highway fuel and are discussed in more detail in Chapter 7.  The in-use sulfur levels in Table
3.1-1 used for modeling differ slightly from those presented in Chapter 7, since minor revisions
were made subsequent to the modeling.

-------
                                                                    Emission Inventory
                                       Table 3.1-1
                    Modeled Baseline In-Use Diesel Fuel Sulfur Content
                          for Land-Based Nonroad Diesel Engines
Calendar Year
through 2005
2006
2007-2009
2010
2011+
48-State Fuel Sulfur
(ppm)
2283
2249
2224
2167
2126
50-State Fuel Sulfur
(ppm)
2284
2242
2212
2155
2114
   3.1.1.3 Emission Estimation Process

   To project emissions in a given year, the draft NONROAD2004 model performs a series of
steps (not necessarily in the order described).

   Equipment Population.  The model projects the equipment population for the user-specified
simulation year.  The current year's population (jVyear) is projected as a function of the base-year
population (Nb,J as
where g is the annual growth rate and n is the number of years between the simulation year and
the base year. For diesel equipment, population projection follows a linear trend as in the
equation above. Diesel growth rates in the model vary only by sector (e.g., agricultural,
construction). The sector-specific growth rates are applied to all equipment types and hp
categories within each sector.

   Equipment Age Distribution.  The model assigns an age distribution for each sub-population
calculated in the previous step. This calculation divides the total population into a series of
model -year cohorts of decreasing size, with the number of cohorts equal to twice the annualized
median life for the rated-power class under consideration (2ly).  Each model-year cohort is
estimated as a fraction of the total population, using fractions derived from NONROAD's
scrappage curve, scaled to the useful life of the given rated-power class, also equal to 2/r5

   Emission and Deterioration Factors. Because the previous steps were performed for engines
of a given rated-power class, the model assigns emission factors to different model year cohorts
simply by relating equipment age to regulatory tier. Similarly, the model calculates deterioration
factors for each cohort. The algorithm identifies the appropriate relative deterioration rate in
relation to tier and rated-power class, calculates the age of the cohort, and supplies these inputs
to the deterioration factor equation.
                                           5-9

-------
Final Regulatory Impact Analysis
   Activity and Load Factor.  The model obtains the appropriate activity, load factor and rated
power estimates. Activity and load factor are defined on the basis of equipment type alone; they
are constant for all model-year cohorts, and rated power is determined on the basis of equipment
type and rated power class.

   Emission Calculation. For a given pollutant, the calculations described above are performed
and the resulting inputs multiplied in the exhaust emission equation. The steps are repeated for
each rated-power class within an equipment type to obtain total emissions for that type. The
resulting subtotals for equipment types are then summed to obtain total emissions from all
equipment types included in the simulation. These processes are repeated for each pollutant
requested for the simulation. Using summation notation, the process may be summarized as
exh,poll
                            = 1
                                       sum over all equipment types
                                        sum over all rated-power classes
                                        within an equipment type

                                         sum over all model-year cohorts
                                         within a rated-power class
   3.1.1.4 Estimation of VOC Emissions

   Volatile organic compounds are a class of hydrocarbons considered to be of regulatory
interest. For purposes of inventory modeling, we define VOC as total hydrocarbon (THC) plus
reactive oxygenated species, represented by aldehydes (RCHO) and alcohols (RCOH), less
nonreactive species represented by methane and ethane (CH4 and CH3CH3), as follows:

                 VOC = THC + (RCHO + RCOH) - (CH4 + CH3CH3)

The NONROAD model estimates VOC in relation to THC, where THC is defined as those
hydrocarbons measured by a flame ionization detector (FID) calibrated to propane. Total
hydrocarbon has exhaust and evaporative components, where the evaporative THC represents
'crankcase emissions.' Crankcase emissions are hydrocarbons that escape from the cylinder
through the piston rings into the crankcase.  The draft NONROAD2004 model assumes that all
diesel engines have open crankcases, allowing that gases in the crankcase to escape to the
atmosphere.

   For diesel engines, the emission factor for crankcase emissions (EFcrank) is estimated as a
fraction of the exhaust emission factor (EFexh), as
                            FF          = 002-FF
                            ^x crank,HC,year   v.v± J^l exh,HC,year
                                          3-10

-------
                                                                   Emission Inventory
Note that the model adjusts crankcase emissions for deterioration. In a given simulation year,
the crankcase emission factor is calculated from the deteriorated exhaust emission factor for that
year, i.e., EFexhyear = ZHL-TAF -DFyear.

   The model estimates exhaust and crankcase VOC as a fraction of exhaust and crankcase
THC, respectively.
                 VOCexh = 1.053 • THCexh,   VOCcrank = 1.053 • THCcrank

Note the fraction is greater than one, reflecting the addition of oxygenated species to THC. For
additional discussion of the model's estimation of crankcase  and VOC emissions, refer to the
model documentation.1'6

   3.1.1.5 Estimation of SO2 Emissions

   To estimate SO2 emissions, the draft NONROAD2004 model does not use an explicit
emission factor. Rather, the model estimates a SO2 emission  factor EFS02 on the basis of brake-
specific fuel consumption, the user-defined diesel sulfur level, and the emission factor for THC.

                   EFso2  = [BSFC-(l-wPMS)-EFTHC]-Sm_use-ws02S

where:
       BSFC = brake-specific fuel consumption (g/hp-hr),
       mpu,s = a constant, representing the fraction of fuel sulfur converted to particulate sulfur,
       equal to 0.02247 g PM S/g fuel S,
       EFTHC = the in-use adjusted THC emission factor (g/hp-hr),
       Sm-use = the user-specified scenario-specific sulfur content of diesel fuel (weight fraction),
       and
       mso2s = a constant, representing fraction of fuel sulfur converted to SO2, equal to 2.0 g
       S02/g S.

   This equation includes corrections for the fraction of sulfur that is converted to PM (wPMiS)
and for the sulfur remaining in the unburned fuel (EF^c). The correction for unburned fuel, as
indicated by THC emissions, is more significant for gasoline emissions, but insubstantial for
diesel emissions.

Having estimated EFS02, the model estimates SO2 emissions  as it does other exhaust emissions.

   3.1.1.6 Estimation of PM25 Emissions

   The model estimates emissions of diesel PM25 as a multiple of PM10 emissions. PM25 is
estimated to compose 97 percent of PM10 emissions.  This is  an updated estimate, based on an
analysis of size distribution data for diesel  engines.7
                                          3-11

-------
Final Regulatory Impact Analysis
   3.1.1.7 Estimation of Fuel Consumption

   The draft NONROAD2004 model estimates fuel consumption using the equation

                                      BSFC-A-L-P-N
                                F =	D	

where:
       F = fuel consumption (gallons/year)
       BSFC = brake-specific fuel consumption (Ib/hp-hr)
       A = equipment activity (operating hours/year)
       L = load factor (average proportion of rated power used during operation (percent))
       P = average rated power (hp)
       N = equipment population (units)
       D = fuel  density (Ib/gal); diesel fuel density = 7.1 Ib/gal

   The fuel consumption estimates for land-based diesel and recreational marine diesel engines
are given in Section 3.1.5.

   3.1.1.8 Changes from Draft NONROAD2002 to Draft NONROAD2004

   For the final rule, we have updated the model to incorporate the following changes:

1) Draft NONROAD2004 contains more horsepower bins in order to model the final standards.
   Specifically, the 50-100 hp bin was split into 50-75 hp and 75-100 hp bins. Also, the 1000-
   1500 hp bin was split into 1000-1200 hp and 1200-1500 hp bins.

2) Draft NONROAD2004 eliminates the Tier 3 NOx and PM transient adjustment factors
   (TAFs) for steady-state applications, which were mistakenly included in draft
   NONROAD2002.

3) The base year populations in draft NONROAD2004 were updated from 1998 to 2000, based
   on newer sales data.

4) The PM25 fraction of PM10 was revised from 0.92 to 0.97, based on an updated analysis of
   size distribution data for diesel engines.

5) The recreational marine populations, median life, and deterioration factors for HC and NOX
   were revised to match what was used in the 2002 final rulemaking that covers large spark
   ignition engines (>25 hp), recreational equipment, and recreational marine diesel engines
   (>50 hp).8 The exhaust emission factors for these three categories were also revised in draft
   NONROAD2004 to reflect the final standards.

6) The output label was changed  from 'SOX'  to 'SO2' to avoid confusion, since SO2 emissions
   are calculated by the model.

                                         3-12

-------
                                                                   Emission Inventory
   For land-based diesel nonroad engines, the net effect of these changes is generally within 3
percent, with the direction and variation of the change dependent on the calendar year and
pollutant of interest.

   3.1.1.9 Baseline Inventory

   Tables 3.1-2a and 3.1-2b present the PM10, PM25, NOX, SO2, VOC, and CO baseline
emissions for land-based nonroad engines in 1996 and 2000-2040, for the 48-state and 50-state
inventories, respectively.
                                          3-13

-------
Final Regulatory Impact Analysis
                                   Table3.1-2a
      Baseline (48-State) Emissions for Land-Based Nonroad Diesel Engines (short tons)
Year
1996
2000
2001
2002
2003
2004
2005
2006
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2031
2032
2033
2034
2035
2036
2037
2038
2039
2040
PM10
192,275
176,056
170,451
165,017
159,268
153,932
148,720
143,840
139,990
137,366
135,097
132,712
130,964
130,091
129,779
129,700
129,831
130,128
130,606
131,211
131,993
133,049
134,251
135,491
136,799
138,136
139,555
141,007
142,429
143,901
145,385
146,891
148,452
150,035
151,640
153,253
154,851
156,499
158,171
160,204
162,240
164.275
PM25
186,507
170,774
165,338
160,067
154,490
149,314
144,259
139,525
135,791
133,245
131,044
128,730
127,035
126,189
125,885
125,809
125,936
126,224
126,688
127,275
128,034
129,058
130,223
131,426
132,695
133,992
135,369
136,777
138,156
139,584
141,023
142,484
143,999
145,534
147,091
148,655
150,205
151,804
153,426
155,398
157,373
159.346
NOX
1,564,904
1,550,355
1,537,890
1,526,119
1,505,435
1,486,335
1,467,547
1,435,181
1,399,787
1,359,661
1,317,995
1,278,038
1,242,159
1,211,982
1,188,162
1,168,310
1,152,199
1,139,969
1,130,663
1,124,057
1,120,529
1,119,481
1,120,802
1,124,159
1,129,090
1,135,338
1,142,889
1,151,480
1,160,868
1,170,868
1,181,457
1,192,833
1,205,007
1,217,535
1,230,337
1,243,467
1,256,924
1,270,722
1,284,718
1,299,415
1,314,296
1.329.330
S02
143,572
161,977
166,644
171,309
175,971
180,630
185,287
187,085
189,511
194,019
198,526
197,829
198,415
202,740
207,062
211,382
215,699
219,971
224,241
228,510
232,777
237,044
241,309
245,573
249,836
254,099
258,360
262,591
266,822
271,052
275,282
279,511
283,740
287,969
292,198
296,426
300,654
304,882
309,110
313,337
317,564
321.792
voc
220,971
199,887
191,472
183,525
176,383
169,873
163,663
156,952
150,357
143,306
136,426
129,711
123,573
118,363
114,022
110,284
107,084
104,426
102,252
100,383
98,766
97,513
96,566
95,837
95,344
95,061
94,975
95,043
95,234
95,529
95,906
96,374
96,942
97,568
98,241
98,967
99,747
100,591
101,473
102,472
103,495
104.543
CO
1,004,586
916,507
880,129
845,435
813,886
787,559
763,062
741,436
724,449
710,202
697,893
687,234
678,980
674,285
672,732
672,819
674,296
677,095
681,156
685,866
691,194
697,630
704,932
712,591
720,565
729,001
737,967
747,219
756,611
766,274
776,141
786,181
796,408
806,761
817,199
827,712
838,224
848,884
859,588
870,258
880,968
891.684
                                       3-14

-------
                                                         Emission Inventory
                              Table3.1-2b
Baseline (50-State) Emissions for Land-Based Nonroad Diesel Engines (short tons)
Year
1996
2000
2001
2002
2003
2004
2005
2006
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2031
2032
2033
2034
2035
2036
2037
2038
2039
2040
PM10
193,166
176,881
171,256
165,801
160,030
154,670
149,434
144,479
140,579
137,945
135,668
133,274
131,521
130,648
130,337
130,260
130,394
130,695
131,178
131,788
132,575
133,637
134,844
136,091
137,406
138,750
140,177
141,637
143,067
144,547
146,038
147,552
149,123
150,715
152,329
153,950
155,557
157,214
158,896
160,938
162,984
165.028
PM25
187,371
171,575
166,118
160,827
155,229
150,030
144,951
140,145
136,362
133,807
131,598
129,276
127,576
126,729
126,426
126,352
126,482
126,774
127,243
127,835
128,598
129,628
130,799
132,008
133,284
134,587
135,972
137,388
138,775
140,211
141,657
143,126
144,649
146,193
147,759
149,332
150,891
152,498
154,129
156,110
158,095
160.077
NOX
1,573,083
1,558,392
1,545,852
1,534,007
1,513,203
1,493,989
1,475,092
1,442,534
1,406,936
1,366,584
1,324,685
1,284,510
1,248,440
1,218,098
1,194,153
1,174,204
1,158,023
1,145,751
1,136,425
1,129,817
1,126,301
1,125,276
1,126,633
1,130,034
1,135,015
1,141,319
1,148,929
1,157,584
1,167,040
1,177,111
1,187,773
1,199,225
1,211,478
1,224,086
1,236,969
1,250,181
1,263,722
1,277,605
1,291,688
1,306,473
1,321,443
1.336.566
SO2
144,409
162,920
167,615
172,307
176,996
181,683
186,368
187,508
189,505
194,013
198,521
197,795
198,360
202,685
207,006
211,325
215,641
219,912
224,181
228,449
232,716
236,982
241,246
245,509
249,772
254,033
258,294
262,525
266,754
270,984
275,213
279,442
283,670
287,898
292,126
296,354
300,581
304,808
309,035
313,262
317,489
321.715
voc
222,084
200,903
192,447
184,462
177,287
170,744
164,505
157,762
151,134
144,049
137,135
130,388
124,220
118,984
114,621
110,863
107,647
104,977
102,793
100,917
99,294
98,037
97,086
96,355
95,860
95,575
95,490
95,558
95,752
96,049
96,429
96,900
97,472
98,102
98,779
99,511
100,296
101,146
102,033
103,038
104,068
105.122
CO
1,009,804
921,226
884,645
849,756
818,037
791,568
766,944
745,216
728,159
713,862
701,516
690,829
682,563
677,865
676,320
676,420
677,918
680,746
684,843
689,593
694,964
701,445
708,795
716,502
724,528
733,017
742,039
751,348
760,798
770,520
780,446
790,547
800,835
811,250
821,751
832,326
842,901
853,624
864,392
875,126
885,901
896.682
                                 3-15

-------
Final Regulatory Impact Analysis
3.1.2 Land-Based Nonroad Diesel Engines—Air Toxics Emissions

   EPA focused on five major air toxics pollutants for this rule: benzene, formaldehyde,
acetaldehyde, 1,3-butadiene, and acrolein. These pollutants are VOCs and are included in the
total land-based nonroad diesel VOC emission estimate.  EPA developed the baseline inventory
estimates for these pollutants by multiplying the baseline VOC emissions from the draft
NONROAD2004 model for a given year by the constant fractional amount that each air toxic
pollutant contributes to VOC emissions. Table 3.1-3 shows the fractions that EPA used for each
air toxics pollutant.  EPA developed these nonroad air toxics pollutant fractions for the National
Emission Inventory.9

                                      Table 3.1-3
                               Air Toxics Fractions of VOC
Benzene
0.020
Formaldehyde
0.118
Acetaldehyde
0.053
1,3 -Butadiene
0.002
Acrolein |
0.003 1
   Tables 3.1-4a and 3.1-4b show our 48-state and 50-state estimates of national baseline
emissions for five selected major air toxic pollutants (benzene, formaldehyde, acetaldehyde,
1,3-butadiene, and acrolein) for 1996, as well as for selected years from 2005 to 2030, modeled
with the existing Tier 1-3 standards.  Toxics emissions decrease over time until 2025 as engines
meeting the Tier 1-3 standards are introduced into the fleet. Beyond 2025, the growth in
population overtakes the effect of the existing emission standards. Chapter 2 discusses the
health effects of these pollutants.

                                      Table3.1-4a
                         Baseline (48-State) Air Toxics Emissions
                    for Land-Based Nonroad Diesel Engines (short tons)
Year
1996
2000
2005
2007
2010
2015
2020
2025
2030
Benzene
4,419
3,998
3,273
3,007
2,594
2,142
1,950
1,900
1,927
Formaldehyde
26,075
23,587
19,312
17,742
15,306
12,636
11,507
11,207
11,372
Acetaldehyde
11,711
10,594
8,674
7,969
6,875
5,675
5,168
5,034
5,108
1,3 -Butadiene
442
400
327
301
259
214
195
190
193
Acrolein
663
600
491
451
389
321
293
285
289
                                          3-16

-------
                                                                  Emission Inventory
                                     Table3.1-4b
                        Baseline (50-State) Air Toxics Emissions
                    for Land-Based Nonroad Diesel Engines (short tons)
Year
1996
2000
2005
2007
2010
2015
2020
2025
2030
Benzene
4,442
4,018
3,290
3,023
2,608
2,153
1,961
1,910
1,938
Formaldehyde
26,206
23,707
19,412
17,834
15,386
12,702
11,568
11,268
11,434
Acetaldehyde
11,770
10,648
8,719
8,010
6,911
5,705
5,196
5,061
5,136
1,3 -Butadiene
444
402
329
302
261
215
196
191
194
Acrolein
666
603
494
453
391
323
294
286
291
3.1.3 Commercial Marine Vessels and Locomotives

   Though no new engine controls are being proposed for diesel commercial marine and
locomotive engines, these engines use diesel fuel and the effects of the fuel changes in the final
rule need to be modeled.  This section addresses the modeling of the baseline case for these
engines, which includes effects of certain other rules such as (a) the April 1998 final rule for
locomotives, (b) the December 1999 final rule  for Category 1 and 2 commercial marine diesel
engines, (c) the January 2003 final rule for Category 3 commercial marine residual engines, and
(c) the January 2001 heavy  duty highway diesel fuel rule that takes effect in June 2006.

   Since the draft NONROAD2004 model does not generate emission estimates for these
applications, the emission inventories were calculated using the following methodology. VOC,
CO, and NOX emissions for 1996, 2020, and 2030 (the years chosen for air quality modeling) for
commercial marine diesel engines were taken from the rulemaking documentation. For
locomotives, the fuel-specific emission factors from the rulemaking documentation were
multiplied by the updated fuel consumption annual estimates described in Chapter 7 to obtain the
emission estimates. The VOC, CO, and NOx emission estimates for commercial marine diesel
engines and locomotives  are presented in Table 3.1-5. VOC emissions were calculated by
multiplying THC emissions by a factor of 1.053, which is also the factor used for land-based
diesel engines.
                                         3-17

-------
Final Regulatory Impact Analysis
                                      Table 3.1-5
                     Baseline (48-State) NOX, VOC, and CO Emissions
             for Locomotives and Commercial Marine Diesel Vessels (short tons)
Year
1996
2020
2030
NOX
Locomotives
934,070
508,084
481,077
CMV
639,630
587,115
602,967
VOC
Locomotives
38,035
30,125
28,580
CMV
21,540
24,005
26,169
CO
Locomotives
92,496
99,227
107,780
CMV
93,638
114,397
123,436
   Tables 3.1-6a and 3.1-6b provide the 48-state and 50-state baseline fuel volumes, fuel sulfur
levels, PM sulfate, PM25, and SO2 emissions. The fuel sulfur levels account for "spillover" of
low-sulfur highway diesel fuel into use by nonroad applications. The slight decrease in average
sulfur level in 2006 is due to the introduction of highway diesel fuel meeting the 2007 15 ppm
standard, and the "spillover" of this highway fuel into the nonroad fuel pool.  The derivation of
the fuel volumes and sulfur levels is discussed in more detail in Chapter 7. The marine fuel
volumes reported in Chapter 7 include both commercial and recreational marine usage.  The fuel
consumption specific to commercial marine in Tables 3.1-6a and 3.1-6b was  calculated by
subtracting the recreational marine fuel consumption as generated by the draft NONROAD2004
model.
                                         3-18

-------
                            Table3.1-6a
Baseline (48-State) Fuel Sulfur Levels, SO2, Sulfate PM, and PM2 5 Emissions
         for Locomotives and Commercial Marine Diesel Vessels
Year
1996
2000
2001
2002
2003
2004
2005
2006
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2031
2032
2033
2034
2035
2036
2037
2038
2039
2040
Locomotiv
e Usage
(109gal/yr)
3.065
2.687
2.772
2.692
2.722
2.741
2.762
2.818
2.868
2.900
2.939
2.986
3.043
3.073
3.097
3.121
3.148
3.181
3.210
3.234
3.266
3.288
3.305
3.335
3.364
3.393
3.426
3.455
3.483
3.513
3.542
3.572
3.602
3.632
3.662
3.693
3.724
3.755
3.786
3.818
3.850
3.882
Commercial
Marine
Usage
(109gal/yr)
1.644
1.556
1.533
1.493
1.507
1.518
1.522
1.556
1.575
1.594
1.609
1.625
1.646
1.663
1.674
1.691
1.706
1.718
1.733
1.757
1.786
1.804
1.823
1.852
1.870
1.893
1.912
1.935
1.958
1.981
2.005
2.030
2.055
2.080
2.106
2.132
2.158
2.185
2.213
2.240
2.269
2.298
Base
Sulfur
Level
(ppm)
2641
2641
2637
2638
2638
2639
2639
2616
2599
2599
2599
2444
2334
2334
2334
2335
2335
2335
2335
2336
2337
2338
2339
2340
2340
2341
2341
2341
2342
2343
2343
2344
2345
2345
2346
2346
2347
2348
2348
2349
2349
2350
Base
SO2
Loco
(tons/yr)
56,193
49,268
50,737
49,291
49,843
50,205
50,583
51,170
51,736
52,317
53,021
50,658
49,278
49,779
50,176
50,581
51,011
51,551
52,028
52,437
52,973
53,352
53,646
54,148
54,635
55,123
55,659
56,140
56,624
57,113
57,606
58,103
58,605
59,111
59,621
60,136
60,655
61,179
61,707
62,240
62,777
63.319
CMV
(tons/yr)
30,136
28,523
28,065
27,339
27,598
27,793
27,867
28,252
28,416
28,749
29,019
27,565
26,655
26,947
27,118
27,395
27,645
27,837
28,093
28,495
28,972
29,268
29,593
30,072
30,364
30,745
31,062
31,440
31,825
32,216
32,615
33,020
33,433
33,852
34,279
34,713
35,154
35,603
36,059
36,523
36,995
37.475
Sulfate PM
Loco
(tons/yr)
4,521
3,964
4,082
3,966
4,010
4,039
4,070
4,117
4,162
4,209
4,266
4,076
3,965
4,005
4,037
4,069
4,104
4,147
4,186
4,219
4,262
4,292
4,316
4,356
4,396
4,435
4,478
4,517
4,556
4,595
4,635
4,675
4,715
4,756
4,797
4,838
4,880
4,922
4,964
5,007
5,051
5.094
CMV
(tons/yr)
2,424
2,295
2,258
2,199
2,220
2,236
2,242
2,273
2,286
2,313
2,335
2,218
2,144
2,168
2,182
2,204
2,224
2,240
2,260
2,292
2,331
2,355
2,381
2,419
2,443
2,473
2,499
2,529
2,560
2,592
2,624
2,657
2,690
2,723
2,758
2,793
2,828
2,864
2,901
2,938
2,976
3.015
Total PM2 5
Loco
(tons/yr)
22,266
19,522
20,137
19,554
19,772
19,913
19,474
19,270
18,998
18,588
18,526
18,183
18,527
18,384
18,198
18,007
17,821
17,671
17,490
17,619
17,444
17,213
16,947
16,743
16,891
16,675
16,469
16,238
16,374
16,136
15,892
16,025
15,775
15,519
15,649
15,385
15,514
15,644
15,370
15,499
15,218
15.345
CMV
(tons/yr)
17,782
18,542
18,723
18,905
19,090
19,019
18,915
18,808
18,671
18,533
18,394
18,259
18,125
17,996
17,871
17,752
17,640
17,575
17,541
17,538
17,588
17,665
17,765
17,890
18,032
18,188
18,356
18,533
18,720
18,906
19,098
19,294
19,497
19,701
19,903
20,108
20,315
20,523
20,733
20,945
21,158
21.372
                                3-19

-------
                             Table3.1-6b
Baseline (50-State) Fuel Sulfur Levels, SO2, Sulfate PM, and PM2 5 Emissions
          for Locomotives and Commercial Marine Diesel Vessels
Year
1996
2000
2001
2002
2003
2004
2005
2006
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2031
2032
2033
2034
2035
2036
2037
2038
2039
2040
Locomotiv
e Usage
(109gal/yr)
3.072
2.691
2.776
2.696
2.726
2.745
2.766
2.823
2.873
2.904
2.944
2.990
3.047
3.077
3.102
3.126
3.152
3.186
3.215
3.239
3.271
3.293
3.310
3.339
3.369
3.398
3.431
3.460
3.489
3.518
3.547
3.577
3.607
3.637
3.668
3.698
3.729
3.760
3.792
3.824
3.856
3.888
Commercial
Marine
Usage
(109gal/yr)
1.724
1.634
1.610
1.569
1.584
1.595
1.599
1.636
1.656
1.675
1.691
1.708
1.731
1.749
1.761
1.778
1.794
1.807
1.824
1.849
1.879
1.898
1.919
1.949
1.968
1.992
2.012
2.037
2.061
2.086
2.111
2.137
2.163
2.190
2.217
2.244
2.272
2.301
2.330
2.359
2.389
2.420
Base
Sulfur
Level
(ppm)
2640
2640
2635
2637
2637
2637
2637
2588
2552
2552
2552
2400
2292
2292
2292
2293
2293
2293
2293
2294
2295
2295
2296
2297
2297
2298
2298
2298
2299
2299
2300
2300
2301
2301
2302
2302
2303
2303
2304
2304
2305
2305
Base
SO2
Loco
(tons/yr)
56,287
49,305
50,778
49,330
49,882
50,244
50,622
50,693
50,877
51,447
52,140
49,822
48,471
48,962
49,351
49,748
50,169
50,701
51,170
51,567
52,091
52,462
52,747
53,236
53,714
54,191
54,717
55,187
55,661
56,139
56,621
57,107
57,597
58,092
58,591
59,094
59,601
60,113
60,629
61,150
61,675
62.205
CMV
(tons/yr)
31,587
29,926
29,454
28,702
28,978
29,186
29,269
29,374
29,330
29,676
29,958
28,464
27,529
27,832
28,012
28,299
28,559
28,761
29,028
29,442
29,934
30,240
30,576
31,069
31,372
31,766
32,095
32,486
32,884
33,288
33,699
34,118
34,543
34,976
35,416
35,864
36,319
36,782
37,252
37,731
38,217
38.711
Sulfate PM
Loco
(tons/yr)
4,528
3,967
4,085
3,969
4,013
4,042
4,073
4,078
4,093
4,139
4,195
4,008
3,900
3,939
3,970
4,002
4,036
4,079
4,117
4,149
4,191
4,221
4,244
4,283
4,321
4,360
4,402
4,440
4,478
4,517
4,555
4,594
4,634
4,674
4,714
4,754
4,795
4,836
4,878
4,920
4,962
5.005
CMV
(tons/yr)
2,541
2,408
2,370
2,309
2,331
2,348
2,355
2,363
2,360
2,388
2,410
2,290
2,215
2,239
2,254
2,277
2,298
2,314
2,335
2,369
2,408
2,433
2,460
2,500
2,524
2,556
2,582
2,614
2,646
2,678
2,711
2,745
2,779
2,814
2,849
2,885
2,922
2,959
2,997
3,036
3,075
3.114
Total PM2 5
Loco
(tons/yr)
22,319
19,551
20,167
19,583
19,801
19,943
19,502
19,298
19,026
18,616
18,553
18,210
18,554
18,411
18,225
18,034
17,847
17,697
17,516
17,645
17,469
17,238
16,972
16,767
16,916
16,699
16,493
16,262
16,398
16,159
15,916
16,049
15,798
15,542
15,672
15,408
15,537
15,667
15,393
15,522
15,240
15 368
CMV
(tons/yr)
18,717
19,518
19,708
19,900
20,095
20,020
19,911
19,798
19,653
19,508
19,363
19,220
19,079
18,943
18,811
18,686
18,568
18,500
18,464
18,461
18,514
18,595
18,700
18,831
18,981
19,146
19,322
19,509
19,705
19,901
20,104
20,309
20,523
20,738
20,951
21,166
21,384
21,603
21,825
22,047
22,271
22.497
                                3-20

-------
                                                                    Emission Inventory
   Annual SO2 emission estimates for locomotives and commercial marine vessels were
calculated by multiplying the gallons of fuel use by the fuel density, the fuel sulfur content, and
the molecular weight ratio of SO2 to sulfur.  This is then reduced by the fraction of fuel sulfur
that is converted to sulfate PM (2.247 percent on average for engines without aftertreatment).1
Following is an example of the calculation for the case when fuel sulfur content is 2300 ppm.

   SO2 tons =   gallons x 7.1 Ib/gallon x 0.0023 S wt. Fraction x (1-0.02247 S fraction converted to SO2) x 64/32
              SO2 to S M. W. ratio / 2000 Ib/ton

   Unlike the equation used in the draft NONROAD2004 model for land-based diesel and
recreational marine diesel  engines (described in  Section 3.1.1.5), this equation does not include a
correction for the sulfur remaining in the unburned fuel. The correction for unburned fuel, as
indicated by THC emissions is insubstantial for diesel emissions.

   Annual sulfate PM emission estimates for locomotives and commercial marine vessels were
calculated by multiplying the gallons of fuel use by the fuel density, the fuel sulfur content, the
molecular weight ratio of hydrated  sulfate to sulfur, and the fraction of fuel  sulfur converted to
sulfate on average. Following is an example of the  calculation for the case when fuel sulfur
content is 2300 ppm.

Sulfate tons = gallons x 7.1 Ib/gallon x 0.0023  S wt. Fraction x 0.02247 fraction of S converted
              to sulfate x 224/32 sulfate to S M.W. ratio / 2000  Ib/ton

The baseline sulfate PM estimates are not used to generate baseline PM10 emission estimates, but
are needed in order to calculate the PM benefits  of reductions in fuel sulfur levels with the final
rule.

   Annual total  PM10 emission estimates for locomotives were calculated by multiplying the
gallons of fuel use by the gram per gallon PM emission factor from the 1998 locomotive final
rule Regulatory Support Document. Following is an example calculation:

PM10 tons =   gallons x g/gal EF / 454g/lb / 2000 Ibs/ton

   Annual PM10 emission estimates for commercial marine vessels were derived from the
rulemaking documentation.

   PM10 is assumed to be equivalent to total PM, and PM2 5 is estimated by multiplying PM10
emissions by a factor of 0.97. This is the factor used for all nonroad diesel engines; the basis is
described in Section 3.1.1.6.
                                          3-21

-------
Final Regulatory Impact Analysis
3.1.4 Recreational Marine Engines

   Diesel recreational marine engines consist mainly of inboard engines used in larger power
boats and sailboats, but there are also a small number of outboard diesel engines in use.
Emission estimates for this category were generated using the draft NONROAD2004 model.
Details of the modeling inputs (e.g., populations, activity, and emission factors) for these engines
can be found in the technical  reports documenting the draft NONROAD2004 model.  The
emission inventory numbers presented here assume that recreational marine applications will use
diesel fuel with the same sulfur content and sulfur-to-sulfate conversion rate as locomotives and
commercial marine vessels.

   It should be noted that, unlike the previous version of the NONROAD model, these
inventory values generated with the draft NONROAD2004 model now account for the newest
standards promulgated in September 2002, which take effect in 2006-2009, for diesel
recreational marine engines greater than 37 kw (50 hp).  Although those standards provide
substantial benefits for the affected engines (e.g., 25 to 37 percent reductions of PM, NOX, and
HC in 2030), the impact of this on the total nonroad diesel inventory is quite small, representing
less than 1 percent of the baseline nonroad diesel inventory (without locomotives or commercial
marine) for PM, NOX, and HC in 2030.

   Tables 3.1-7a and 3.1-7b  present the PM10, PM25, NOX, SO2, VOC, and CO emissions for
recreational marine engines in 1996 and 2000-2040 for the 48-state and 50-state inventories,
respectively.
                                          3-22

-------
                                                         Emission Inventory
                              Table3.1-7a
Baseline (48-State) Emissions for Recreational Marine Diesel Engines (short tons)
Year
1996
2000
2001
2002
2003
2004
2005
2006
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2031
2032
2033
2034
2035
2036
2037
2038
2039
2040
PM10
951
1,070
1,099
1,130
1,160
1,190
1,220
1,233
1,247
1,262
1,275
1,257
1,245
1,254
1,261
1,269
1,275
1,280
1,285
1,290
1,295
1,300
1,304
1,309
1,314
1,320
1,330
1,344
1,359
1,376
1,394
1,413
1,432
1,451
1,471
1,491
1,511
1,531
1,552
1,573
1,593
1.615
PM25
97.3
1,038
1,066
1,096
1,125
1,154
1,183
1,196
1,210
1,225
1,237
1,219
1,208
1,216
1,223
1,230
1,236
1,242
1,247
1,251
1,256
1,261
1,265
1,270
1,275
1,281
1,290
1,303
1,319
1,335
1,352
1,371
1,389
1,408
1,427
1,446
1,466
1,485
1,505
1,526
1,546
1.566
NOX
33 679
37,943
39,071
40,198
41,325
42,452
43,578
44,105
44,602
45,066
45,415
45,729
46,022
46,282
46,528
46,765
46,969
47,168
47,362
47,525
47,687
47,847
48,003
48,182
48,363
48,593
48,961
49,501
50,092
50,716
51,392
52,085
52,790
53,510
54,228
54,959
55,702
56,444
57,197
57,963
58,729
59.506
SO2
4 7.86
4,831
4,968
5,114
5,259
5,406
5,551
5,647
5,754
5,897
6,041
5,816
5,682
5,811
5,939
6,070
6,198
6,327
6,455
6,587
6,718
6,850
6,982
7,114
7,243
7,375
7,504
7,633
7,765
7,897
8,026
8,158
8,290
8,419
8,552
8,681
8,814
8,946
9,075
9,208
9,338
9.471
VOC
1 797
1,455
1,494
1,533
1,571
1,609
1,647
1,657
1,664
1,670
1,670
1,668
1,665
1,660
1,655
1,649
1,642
1,634
1,627
1,618
1,611
1,604
1,597
1,592
1,586
1,583
1,587
1,599
1,614
1,630
1,649
1,669
1,689
1,710
1,731
1,753
1,775
1,798
1,820
1,844
1,868
1.892
CO
5 474
6,098
6,271
6,444
6,615
6,787
6,958
7,128
7,298
7,467
7,636
7,804
7,971
8,137
8,303
8,469
8,635
8,802
8,969
9,137
9,308
9,482
9,655
9,829
10,004
10,178
10,354
10,529
10,704
10,880
11,056
11,232
11,409
11,585
11,762
11,938
12,115
12,292
12,469
12,646
12,823
13.001
                                 3-23

-------
Final Regulatory Impact Analysis
                                   Table3.1-7b
      Baseline (50-State) Emissions for Recreational Marine Diesel Engines (short tons)
Year
1996
2000
2001
2002
2003
2004
2005
2006
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2031
2032
2033
2034
2035
2036
2037
2038
2039
2040
PM10
957
1,076
1,106
1,137
1,167
1,197
1,227
1,236
1,246
1,262
1,274
1,256
1,245
1,253
1,261
1,268
1,273
1,279
1,284
1,288
1,293
1,298
1,302
1,307
1,312
1,317
1,327
1,341
1,357
1,373
1,391
1,410
1,429
1,448
1,467
1,487
1,507
1,527
1,548
1,568
1,589
1.610
PM25
929
1,044
1,073
1,103
1,132
1,161
1,190
1,199
1,209
1,224
1,236
1,219
1,208
1,215
1,223
1,230
1,235
1,241
1,245
1,250
1,254
1,259
1,263
1,268
1,272
1,278
1,287
1,301
1,316
1,332
1,349
1,367
1,386
1,404
1,423
1,442
1,462
1,481
1,501
1,521
1,542
1.562
NOX
33,891
38,182
39,317
40,452
41,586
42,719
43,852
44,383
44,883
45,350
45,701
46,018
46,312
46,573
46,821
47,060
47,265
47,465
47,660
47,825
47,987
48,148
48,305
48,485
48,667
48,899
49,269
49,813
50,408
51,036
51,716
52,413
53,123
53,847
54,570
55,305
56,053
56,799
57,558
58,329
59,099
59.881
S02
4,312
4,859
4,995
5,145
5,290
5,436
5,582
5,622
5,685
5,827
5,969
5,747
5,615
5,742
5,869
5,998
6,125
6,252
6,379
6,509
6,639
6,766
6,897
7,027
7,155
7,285
7,412
7,540
7,670
7,797
7,928
8,055
8,186
8,313
8,444
8,572
8,703
8,830
8,961
9,089
9,220
9.348
VOC
1,305
1,464
1,503
1,542
1,581
1,619
1,658
1,667
1,674
1,680
1,680
1,678
1,675
1,671
1,665
1,660
1,652
1,645
1,637
1,629
1,621
1,614
1,607
1,602
1,596
1,593
1,597
1,609
1,624
1,640
1,659
1,679
1,700
1,721
1,742
1,764
1,786
1,809
1,832
1,856
1,879
1.904
CO
5,458
6,137
6,311
6,484
6,657
6,829
7,001
7,173
7,344
7,514
7,684
7,853
8,021
8,189
8,356
8,523
8,690
8,857
9,025
9,195
9,367
9,541
9,716
9,891
10,067
10,243
10,419
10,595
10,772
10,949
11,126
11,303
11,481
11,658
11,836
12,013
12,191
12,369
12,547
12,726
12,904
13.082
                                       3-24

-------
                                                                  Emission Inventory
3.1.5 Fuel Consumption for Nonroad Diesel Engines

   Table 3.1-8 presents the fuel consumption estimates for the land-based, recreational marine,
locomotive, and commercial marine nonroad diesel categories. Fuel consumption estimates are
provided for 1996 and 2000-2040 for the 48-state and 50-state inventories.

   The fuel consumption estimates for land-based diesel and recreational marine diesel engines
were obtained using the draft NONROAD2004 model.  The methodology is described in Section
3.1.1.7. The derivation of the fuel consumption estimates for locomotives and commercial
marine vessels is described in Section 3.1.3.

   For the final rule,  the draft NONROAD2004 estimates for fuel consumption are the basis for
both inventory generation and for the cost analyses.  The land-based diesel fuel estimates in
Chapter 7 differ from those presented in Table 3.1-8 by less than 1 percent, due to simple
rounding error.

   Although the locomotive diesel demand volumes in this chapter are identical to those
described in Chapter 7, the  marine diesel volumes are slightly different.  In Chapter 7, the marine
end-use category is a  combination of both commercial and recreational marine end uses. In this
chapter, recreational marine demand is estimated separately with the draft NONROAD2004
model for each calendar year, and subtracted from the respective combined marine end use
volume to produce the commercial marine estimate.
                                         3-25

-------
Final Regulatory Impact Analysis
                                   Table 3.1-8
                    Fuel Consumption for Nonroad Diesel Engines
Year
1996
2000
2001
2002
2003
2004
2005
2006
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2031
2032
2033
2034
2035
2036
2037
2038
2039
2040
Fuel Consumption (106 gal/year)
Land-Based Diesel
48-State
9,120
10,276
10,568
10,861
11,153
11,445
11,737
12,028
12,318
12,608
12,898
13,188
13,480
13,772
14,063
14,355
14,647
14,936
15,224
15,513
15,802
16,091
16,380
16,668
16,957
17,246
17,535
17,821
18,108
18,395
18,682
18,968
19,255
19,542
19,829
20,116
20,402
20,689
20,976
21,263
21,549
21.836
50-State
9,169
10,331
10,625
10,919
11,213
11,507
11,801
12,092
12,384
12,676
12,968
13,259
13,553
13,846
14,139
14,433
14,726
15,016
15,307
15,597
15,887
16,178
16,468
16,759
17,049
17,339
17,630
17,918
18,206
18,495
18,783
19,071
19,360
19,648
19,936
20,225
20,513
20,801
21,090
21,378
21,666
21.955
Recreational Marine
48-State
234
264
272
280
288
296
303
311
319
327
335
343
351
359
367
375
383
391
399
407
415
423
431
438
446
454
462
470
478
486
494
502
510
518
526
534
542
549
557
565
573
581
50-State
236
266
274
282
289
297
305
313
321
329
337
345
353
361
369
377
385
393
401
409
417
425
433
441
449
457
465
473
481
489
497
505
513
521
529
537
545
553
561
569
577
585
Locomotives
48-State
3,065
2,687
2,772
2,692
2,722
2,741
2,762
2,818
2,868
2,900
2,939
2,986
3,043
3,073
3,097
3,121
3,148
3,181
3,210
3,234
3,266
3,288
3,305
3,335
3,364
3,393
3,426
3,455
3,483
3,513
3,542
3,572
3,602
3,632
3,662
3,693
3,724
3,755
3,786
3,818
3,850
3.882
50-State
3,072
2,691
2,776
2,696
2,726
2,745
2,766
2,823
2,873
2,904
2,944
2,990
3,047
3,077
3,102
3,126
3,152
3,186
3,215
3,239
3,271
3,293
3,310
3,339
3,369
3,398
3,431
3,460
3,489
3,518
3,547
3,577
3,607
3,637
3,668
3,698
3,729
3,760
3,792
3,824
3,856
3.888
Commercial Marine
48-State
1,644
1,556
1,533
1,493
1,507
1,518
1,522
1,556
1,575
1,594
1,609
1,625
1,646
1,663
1,674
1,691
1,706
1,718
1,733
1,757
1,786
1,804
1,823
1,852
1,870
1,893
1,912
1,935
1,958
1,981
2,005
2,030
2,055
2,080
2,106
2,132
2,158
2,185
2,213
2,240
2,269
2.298
50-State
1,724
1,634
1,610
1,569
1,584
1,595
1,599
1,636
1,656
1,675
1,691
1,708
1,731
1,749
1,761
1,778
1,794
1,807
1,824
1,849
1,879
1,898
1,919
1,949
1,968
1,992
2,012
2,037
2,061
2,086
2,111
2,137
2,163
2,190
2,217
2,244
2,272
2,301
2,330
2,359
2,389
2.420
                                      3-26

-------
                                                                  Emission Inventory
3.2 Contribution of Nonroad Diesel Engines to National Emission
Inventories

   This section provides the contribution of nonroad diesel engines to national baseline
emission inventories in 1996, 2020, and 2030. The emission inventories are based on 48-state
inventories that exclude Alaska and Hawaii to be consistent with the air quality modeling region.
The baseline cases represent current and future emissions only with the existing standards. For
the final rule, these baseline inventories now incorporate recent standards that cover large spark-
ignition engines (>25 hp), recreational equipment, and recreational marine diesel engines (>50
hp).10

   The calendar years correspond to those chosen for the air quality modeling. Pollutants
discussed include PM2 5, NOX, SO2, VOC, and CO.  VOC includes both exhaust and evaporative
emissions.

   Of interest are the contributions of emissions from nonroad diesel  sources affected by the
final rule.  For PM2 5 and SO2, this includes emissions from all nonroad diesel sources. For NOX,
VOC, and CO, this includes emissions from land-based nonroad diesel engines. Contributions to
both total mobile source emissions and total emissions from all sources are presented. For PM2 5,
contributions of nonroad diesel  engines to both total diesel PM25 and total manmade PM25 are
also presented.

   The development of the 1996, 2020, and 2030 baseline emission inventories for the nonroad
sector and for the sectors not affected by this rule are briefly described, followed by discussions
for each pollutant of the contribution of nonroad diesel engines to national baseline inventories.

3.2.1 Baseline Emission Inventory Development

   For 1996, 2020, and 2030, county-level emission estimates were developed by Pechan under
contract to EPA. These were used as input for the air quality modeling.  These inventories
account for county-level differences in parameters such as fuel characteristics and temperature.
The draft NONROAD2002 model was used to generate the county-level emission estimates for
all nonroad sources, with the exception of commercial marine  engines, locomotives, and aircraft.
The methodology has been documented elsewhere.11

   The highway estimates are based on the MOBILESb model, but with some further
adjustments to reflect MOBILE6 emission factors.  The highway inventories are similar to those
prepared for HD2007 rulemaking, with the exception of adjustments to NOX and VOC for
California counties, based on county-level estimates from the California Air Resources Board.12

   The stationary point and area source estimates are also based on the HD2007 rulemaking,
with the exception of adjustments to NOX and VOC for California counties,  based on county -
                                         3-27

-------
Final Regulatory Impact Analysis
level estimates from the California Air Resources Board. There were also some stack parameter
corrections made to the point source estimates.

   The inventories developed by Pechan were used in this section for the following categories:
recreational marine spark-ignition engines, commercial marine vessels fueled with gasoline and
coal, aircraft, and stationary point and area sources. For the remaining categories, updated
national estimates were substituted that reflect recent rulemakings and/or updated model inputs,
fuel parameters and usage. The basis for the updated estimates for the remaining categories is
described below.

   The model inputs for the nonroad diesel sources have been described in detail in Section 3.1.
The emission estimates for the land-based diesel and recreational  marine diesel categories were
based on national level runs with the draft NONROAD2004 model.  This was done for two
reasons.  First, the baseline inventories for 2020 and 2030 were revised since the county-level
estimates were developed (specifically, PM2 5 and SO2  emissions were changed to reflect revised
diesel fuel sulfur inputs, standards affecting recreational marine diesel engines were
promulgated, and model inputs such as base year populations were updated). It was not possible
to develop revised county-level estimates for these categories due to resource and time
constraints.  Second, county-level estimates were developed only  for 2020 and 2030. Estimates
for interim years are also needed to fully evaluate the anticipated emission benefits  of the final
rule. Interim year estimates are generated using national level model runs.  To be consistent with
other sections of the Final RIA in which interim year estimates from  1996 to 2030 are presented,
the inventory estimates presented here for the land-based diesel and recreational marine diesel
categories are based on national level model runs. Model results for national level runs are
similar to those based on an aggregation of county-level runs.

   For nonroad spark-ignition engines, the emission estimates were based on national level runs
with the draft NONROAD2004 model, in order to account for the recent rulemaking that affects
large spark-ignition engines. The draft NONROAD2004 model accounts for the exhaust
provisions of the rule.  Additional adjustments were made to the VOC model output to account
for the evaporative provisions of the rule, since the draft NONROAD2004 model does not yet
incorporate the evaporative provisions of the rulemaking.

   The commercial marine category has been divided  into three subcategories: commercial
marine diesel, commercial marine residual, and commercial marine other. The commercial
marine diesel category includes compression-ignition engines using diesel fuel (generally
includes  Category 1 and 2 engines). The commercial marine residual category includes
compression-ignition engines using residual fuel (includes Category 3 engines). The commercial
marine other category includes commercial marine engines using  gasoline or coal.  The emission
estimates for the commercial marine diesel and residual categories were updated to  reflect the
1999 and 2003 rulemakings affecting commercial marine compression-ignition engines. In
addition, the SO2 estimates for commercial marine diesel vessels  are based  on the updated fuel
sulfur levels and fuel consumption estimates provided in Section 3.1.
                                          3-28

-------
                                                                   Emission Inventory
   Emission estimates for the locomotive category were revised to reflect the updated fuel sulfur
levels and fuel consumption estimates provided in Section 3.1.  Finally, the motorcycle portions
of the highway estimates were revised to incorporate updated estimates contained in the recent
rulemaking affecting motorcycles.

3.2.2 PM7. Emissions
   Table 3.2-1 provides the contribution of land-based diesel engines and other source
categories to total diesel PM2 5 emissions.

   PM25 emissions from land-based nonroad diesel engines are 46 percent of the total diesel
PM2 5 emissions in 1996, and this percentage increases to 72 percent by 2030. Emissions from
land-based nonroad diesel engines actually decrease from 186,507 tons in 1996 to 129,058 tons
in 2020 due to the existing emission standards. From 2020 to 2030, however, emissions increase
to 142,484 tons, as growth in this sector offsets the effect of the existing emission standards.

   PM25 emissions from recreational marine diesel engines,  commercial marine diesel engines,
and locomotives will also be affected by this rule due to the fuel sulfur requirements. For all
nonroad diesel sources affected by this rule, the contribution to total diesel PM2 5 emissions
increases from 56 percent in 1996 to 91 percent in 2030.

   Table 3.2-2 provides the contribution of land-based diesel engines and other source
categories to total manmade PM2 5 emissions. PM2 5 emissions from land-based nonroad diesel
engines are 8 percent of the total manmade PM25 emissions in 1996, and this percentage drops
slightly to 6 percent in 2020 and 2030. The contribution of land-based diesel engines to total
mobile source PM25 emissions is 33 percent in 1996, rising slightly to 35 percent by 2030. For
all nonroad diesel sources, the contribution to total manmade PM2 5 emissions is 10 percent in
1996, and this percentage drops slightly to 8 percent in 2020  and 2030.

3.2.3 NOV Emissions
        'x
   Table 3.2-3 provides the contribution of land-based diesel engines and other source
categories to total NOX emissions.

   NOX emissions from land-based nonroad diesel engines are 6 percent of the total emissions in
1996, and this percentage increases to 8 percent by 2030. The contribution of land-based diesel
engines to total mobile source NOX emissions is 12 percent in 1996, rising to 24 percent by 2030.
Emissions from land-based nonroad diesel engines actually decrease from 1,564,904 tons in
1996 to 1,119,481 tons in 2020 due to the existing emission standards.  From 2020 to 2030,
however, emissions increase to 1,192,833 tons, as growth in this sector offsets the effect of the
existing emission standards.
                                          3-29

-------
Final Regulatory Impact Analysis
   NOX emissions from recreational marine diesel engines, commercial marine diesel engines,
and locomotives will not be affected by this rule.  For these categories combined, the
contribution to total NOX emissions remains stable at 7-8 percent from 1996 to 2030.

3.2.4 SO2 Emissions

   Table 3.2-4 provides the contribution of land-based diesel engines and other source
categories to total SO2 emissions.

   SO2 emissions from land-based nonroad diesel engines are 1 percent of the total emissions in
1996, and this percentage increases to 2 percent by 2030.  The contribution of land-based diesel
engines to total mobile source SO2 emissions is 20 percent in 1996, rising to 33 percent by 2030,
due to continued growth in this sector.

   SO2 emissions from recreational marine diesel engines, commercial  marine diesel engines,
and locomotives will also be affected by this rule  due to the fuel sulfur requirements. For all
nonroad diesel sources affected by this rule, the contribution to total SO2 emissions remains
relatively stable at 1 percent.

3.2.5 VOC Emissions

   Table 3.2-5 provides the contribution of land-based diesel engines and other source
categories to total VOC emissions.  VOC includes both exhaust and evaporative emissions.
VOC is an ozone precursor; therefore, VOC inventories are required for air quality modeling.

   VOC emissions from land-based nonroad diesel engines are 1 percent of the total emissions
in 1996, and this percentage increases to 2 percent by 2030. The contribution of land-based
diesel engines to total mobile source VOC emissions is 3 percent in 1996, decreasing slightly to
2 percent by 2030. Emissions from land-based nonroad diesel engines actually decrease from
220,971 tons in 1996 to 97,513 tons in 2020 due to the existing emission standards.  From 2020
to 2030, however, emissions increase to 96,374 tons, as growth in this sector offsets the effect of
the existing emission standards.

   VOC emissions from recreational marine diesel engines, commercial marine diesel engines,
and locomotives will not be affected by this rule.  For these categories combined, the
contribution to total VOC emissions is less than 1 percent.

3.2.6 CO Emissions

   Table 3.2-6 provides the contribution of land-based diesel engines and other source
categories to total CO emissions.

   CO emissions from land-based nonroad diesel engines are 1 percent of the total emissions in
1996, and this percentage remains stable at 1 percent by 2030.  The contribution of land-based
diesel engines to total mobile source CO emissions is also 1 percent in 1996, remaining at 1

                                          3-30

-------
                                                                   Emission Inventory
percent by 2030. Emissions from land-based nonroad diesel engines actually decrease from
1,004,586 tons in 1996 to 697,630 tons in 2020 due to the existing emission standards. From
2020 to 2030, however, emissions increase to 786,181 tons, as growth in this  sector offsets the
effect of the existing emission standards.

   CO emissions from recreational marine diesel engines, commercial marine diesel engines,
and locomotives will not be affected by this rule. For these categories combined, the
contribution to total CO emissions is less than 1  percent in 1996 and 2030.
                                         3-31

-------
     Final Regulatory Impact Analysis
                                               Table 3.2-1
       Annual Diesel PM2 s Baseline Emission Levels for Mobile and Other Source Categories3
Category
Land-Based Nonroad
Diesel
Recreational Marine
Diesel <50 hp
Recreational Marine
Diesel >50 hp
Commercial Marine
Diesel b
Locomotive
Total Nonroad Diesel
Total Highway Diesel
Total Mobile Source
Diesel
Stationary Point and
Area Source Diesel °
Total Man-Made
Diesel Sources
Mobile Source
Percent of Total
1996
short tons
186,507
56
867
17,782
22,266
227,478
167,384
394,862
12,199
407,061
97%
%of
mobile
source
47.2%
0.0%
0.2%
4.5%
5.6%
58%
42%
100%
—
—
—
%of
total
45.8%
0.0%
0.2%
4.4%
5.5%
56%
41%
97%
3%


2020
short
tons
129,058
46
1,214
17,665
17,213
165,196
18,426
183,622
4,010
187,632
98%
%of
mobile
source
70.3%
0.0%
0.7%
9.6%
9.4%
90%
10%
100%
—
—
—
%of
total
68.8%
0.0%
0.6%
9.4%
9.2%
88%
10%
98%
2%


2030
short
tons
142,484
50
1,321
19,294
16,025
179,173
13,948
193,121
4,231
197,352
98%
%of
mobile
source
73.8%
0.0%
0.7%
10.0%
8.3%
93%
7%
100%
—
—
—
%of
total
72.2%
0.0%
0.7%
9.8%
8.1%
91%
7%
98%
2%


a These are 48-state inventories. They do not include Alaska and Hawaii.
b This category includes compression-ignition (CI) vessels using diesel fuel. It does not include CI vessels using residual fuel.
°This category includes point sources burning either diesel, distillate oil (diesel), or diesel/kerosene fuel.

-------
                                                                            Emission Inventory
                                            Table 3.2-2
         Annual PM2 s Baseline Emission Levels for Mobile and Other Source Categories a'b
Category
Land-Based Nonroad
Diesel
Recreational Marine
Diesel <50 hp
Recreational Marine
Diesel >50 hp
Recreational
Marine SI
Nonroad SI <25 hp
Nonroad SI >25hp
Recreational SI
Commercial Marine
Diesel °
Commercial Marine
Residual c
Commercial Marine
Other0
Locomotive
Aircraft
Total Nonroad
Total Highway
Total Mobile Sources
Stationary Point and
Area Sources
Total Man-Made
Sources
Mobile Source
Percent of Total
1996
short tons
186,507
56
867
35,147
24,309
1,374
7,968
17,782
16,126
1,370
22,266
27,891
341,663
230,684
572,346
1,653,392
2,225,738
26%
%of
mobile
source
32.6%
0.0%
0.2%
6.1%
4.2%
0.2%
1.4%
3.1%
2.8%
0.2%
3.9%
4.9%
60%
40%
100%
—
—
—
%of
total
8.4%
0.0%
0.0%
1.6%
1.1%
0.1%
0.4%
0.8%
0.7%
0.1%
1.0%
1.3%
15%
10%
26%
74%


2020
short tons
129,058
46
1,214
26,110
30,151
2,302
9,963
17,665
34,532
1,326
17,213
30,024
299,603
72,377
371,980
1,712,004
2,083,984
18%
%of
mobile
source
34.7%
0.0%
0.3%
7.0%
8.1%
0.6%
2.7%
4.7%
9.3%
0.4%
4.6%
8.1%
81%
19%
100%
—
—
—
%of
total
6.2%
0.0%
0.1%
1.3%
1.4%
0.1%
0.5%
0.8%
1.7%
0.1%
0.8%
1.4%
14%
4%
18%
82%


2030
short tons
142,484
50
1,321
27,223
34,598
2,692
9,460
19,294
51,026
1,427
16,025
30,606
336,206
75,825
412,030
1,824,609
2,236,639
18%
%of
mobile
source
34.6%
0.0%
0.3%
6.6%
8.4%
0.7%
2.3%
4.7%
12.4%
0.3%
3.9%
7.4%
82%
18%
100%
—
—
—
%of
total
6.4%
0.0%
0.1%
1.2%
1.5%
0.1%
0.4%
0.9%
2.3%
0.1%
0.7%
1.4%
15%
3%
18%
82%


a These are 48-state inventories. They do not include Alaska and Hawaii.
b Excludes natural and miscellaneous sources.

-------
     Final Regulatory Impact Analysis
0 Commercial marine diesel includes Category 1 and 2 compression-ignition (CI) engines using diesel fuel. The residual

    category includes Category 3 CI engines using residual fuel.  The other category includes engines using gasoline and

    steamships fueled with coal.
                                                    > o A
                                                    5-34

-------
                                                                        Emission Inventory
                                         Table 3.2-3
         Annual NOX Baseline Emission Levels for Mobile and Other Source Categories
Category
Land-Based Nonroad
Diesel
Recreational Marine
Diesel <50 hp
Recreational Marine
Diesel >50 hp
Recreational
Marine SI
Nonroad SI <25 hp
Nonroad SI >25hp
Recreational SI
Commercial Marine
Diesel b
Commercial Marine
Residual b
Commercial
Marine Other b
Locomotive
Aircraft
Total Nonroad
Total Highway
Total Mobile
Sources
Stationary Point and
Area Sources c
Total Man-Made
Sources
Mobile Source
Percent of Total
1996
short tons
1,564,904
438
33,241
33,304
63,120
273,082
4,297
639,630
184,275
5,979
934,070
165,018
3,901,357
9,060,923
12,962,279
11,449,752
24,412,031
53%
%of
mobile
source
12.1%
0.0%
0.3%
0.3%
0.5%
2.1%
0.0%
4.9%
1.4%
0.0%
7.2%
1.3%
30%
70%
100%
—
—
—
%of
total
6.4%
0.0%
0.1%
0.1%
0.3%
1.1%
0.0%
2.6%
0.8%
0.0%
3.8%
0.7%
16%
37%
53%
47%


2020
short tons
1,119,481
491
47,356
61,749
98,584
43,315
17,129
587,115
356,445
4,207
508,084
228,851
3,072,808
1,975,312
5,048,120
10,050,213
15,098,333
33%
%of
mobile
source
22.2%
0.0%
0.9%
1.2%
2.0%
0.9%
0.3%
11.6%
7.1%
0.1%
10.1%
4.5%
61%
39%
100%
—
—
—
%of
total
7.4%
0.0%
0.3%
0.4%
0.7%
0.3%
0.1%
3.9%
2.4%
0.0%
3.4%
1.5%
20%
13%
33%
67%


2030
short tons
1,192,833
554
51,531
67,893
114,447
43,527
19,389
602,967
514,881
4,020
481,077
258,102
3,351,220
1,566,902
4,918,123
10,320,361
15,238,484
32%
%of
mobile
source
24.3%
0.0%
1.0%
1.4%
2.3%
0.9%
0.4%
12.3%
10.5%
0.1%
9.8%
5.2%
68%
32%
100%
—
—
—
%of
total
7.8%
0.0%
0.3%
0.4%
0.8%
0.3%
0.1%
4.0%
3.4%
0.0%
3.2%
1.7%
22%
10%
32%
68%


' These are 48-state inventories. They do not include Alaska and Hawaii.

-------
     Final Regulatory Impact Analysis
b Commercial marine diesel includes Category 1 and 2 compression-ignition (CI) engines using diesel fuel. The residual
    category includes Category 3 CI engines using residual fuel. The other category includes engines using gasoline and
    steamships fueled with coal.
0 Does not include effects of the proposed Rule to Reduce Interstate Transport of Fine Particulate Matter and Ozone (Interstate
    Air Quality Rule). 69 FR 4566 (January 30, 2004). See http://www.epa.gov/interstateairquality/rule.html.
                                                Table 3.2-4
           Annual SO2 Baseline Emission Levels  for Mobile and Other Source Categories
Category
Land-Based Nonroad
Diesel
Recreational Marine
Diesel <50 hp
Recreational Marine
Diesel >50 hp
Recreational
Marine SI
Nonroad SI <25 hp
Nonroad SI >25hp
Recreational SI
Commercial Marine
Diesel b
Commercial Marine
Residual b
Commercial
Marine Other b
Locomotive
Aircraft
Total Nonroad
Total Highway
Total Mobile
Sources
Stationary Point and
Area Sources °
Total Man-Made
Sources
Mobile Source
Percent of Total
1996
short tons
143,572
53
4,234
2,170
6,803
890
949
30,136
151,559
9,266
56,193
11,305
417,128
302,938
720,066
17,636,602
18,356,668
4%
%of
mobile
source
19.9%
0.0%
0.6%
0.3%
0.9%
0.1%
0.1%
4.2%
21.0%
1.3%
7.8%
1.6%
58%
42%
100%
—
—
—
%of
total
0.8%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.2%
0.8%
0.1%
0.3%
0.1%
2%
2%
4%
96%


2020
short tons
237,044
85
6,766
2,522
8,623
879
2,561
29,268
263,076
9,677
53,352
15,267
629,118
35,311
664,429
14,510,426
15,174,855
4%
%of
mobile
source
35.7%
0.0%
1.0%
0.4%
1.3%
0.1%
0.4%
4.4%
39.6%
1.5%
8.0%
2.3%
95%
5%
100%
—
—
—
%of
total
1.6%
0.0%
0.0%
0.0%
0.1%
0.0%
0.0%
0.2%
1.7%
0.1%
0.4%
0.1%
4%
0%
4%
96%


2030
short tons
279,511
101
8,057
2,698
10,007
998
2,691
33,020
387,754
10,366
58,103
16,813
810,119
40,788
850,907
14,782,220
15,633,127
5%
%of
mobile
source
32.8%
0.0%
0.9%
0.3%
1.2%
0.1%
0.3%
3.9%
45.6%
1.2%
6.8%
2.0%
95%
5%
100%
—
—
—
%of
total
1.8%
0.0%
0.1%
0.0%
0.1%
0.0%
0.0%
0.2%
2.5%
0.1%
0.4%
0.1%
5%
0%
5%
95%


' These are 48-state inventories. They do not include Alaska and Hawaii.

-------
                                                                                        Emission Inventory
b Commercial marine diesel includes Category 1 and 2 compression-ignition (CI) engines using diesel fuel. The residual
    category includes Category 3 CI engines using residual fuel. The other category includes engines using gasoline and
    steamships fueled with coal.
0 Does not include effects of the proposed Rule to Reduce Interstate Transport of Fine Particulate Matter and Ozone (Interstate
    Air Quality Rule).  69 FR 4566 (January 30, 2004). See http://www.epa.gov/interstateairquality/rule.html.
                                                      J-J

-------
     Final Regulatory Impact Analysis
                                              Table 3.2-5
          Annual VOC Baseline Emission Levels for Mobile and Other Source Categories a
Category
Land-Based Nonroad
Diesel
Recreational Marine
Diesel <50 hp
Recreational Marine
Diesel >50 hp
Recreational
Marine SI
Nonroad SI <25 hp
Nonroad SI >25hp
Recreational SI
Commercial Marine
Diesel b
Commercial Marine
Residual b
Commercial
Marine Other b
Locomotive
Aircraft
Total Nonroad
Total Highway
Total Mobile
Sources
Stationary Point and
Area Sources
Total Man-Made
Sources
Mobile Source
Percent of Total
1996
short tons
220,971
106
1,191
804,488
1,332,392
88,526
322,766
21,540
7,446
892
38,035
176,394
3,014,747
5,291,388
8,306,135
10,249,136
18,555,271
45%
%of
mobile
source
2.7%
0.0%
0.0%
9.7%
16.0%
1.1%
3.9%
0.3%
0.1%
0.0%
0.5%
2.1%
36%
64%
100%
—
—
—
%of
total
1.2%
0.0%
0.0%
4.3%
7.2%
0.5%
1.7%
0.1%
0.0%
0.0%
0.2%
1.0%
16%
29%
45%
55%


2020
short tons
97,513
52
1,552
380,891
656,845
10,629
345,649
24,005
17,584
925
30,125
239,654
1,805,424
2,071,456
3,876,880
9,648,376
13,525,256
29%
%of
mobile
source
2.5%
0.0%
0.0%
9.8%
16.9%
0.3%
8.9%
0.6%
0.5%
0.0%
0.8%
6.2%
47%
53%
100%
—
—
—
%of
total
0.7%
0.0%
0.0%
2.8%
4.9%
0.1%
2.6%
0.2%
0.1%
0.0%
0.2%
1.8%
13%
15%
29%
71%


2030
short tons
96,374
50
1,619
372,970
758,512
9,664
327,403
26,169
26,711
1,001
28,580
265,561
1,914,614
2,312,561
4,227,175
10,751,134
14,978,309
28%
%of
mobile
source
2.3%
0.0%
0.0%
8.8%
17.9%
0.2%
7.7%
0.6%
0.6%
0.0%
0.7%
6.3%
45%
55%
100%
—
—
—
%of
total
0.6%
0.0%
0.0%
2.5%
5.1%
0.1%
2.2%
0.2%
0.2%
0.0%
0.2%
1.8%
13%
15%
28%
72%


a These are 48-state inventories.  They do not include Alaska and Hawaii.
b Commercial marine diesel includes Category 1 and 2 compression-ignition (CI) engines using diesel fuel. The residual
    category includes Category  3 CI engines using residual fuel. The other category includes engines using gasoline and
    steamships fueled with coal.

-------
                                                                                 Emission Inventory
                                              Table 3.2-6
           Annual CO Baseline Emission Levels for Mobile and Other Source Categories a
Category
Land-Based Nonroad
Diesel
Recreational Marine
Diesel <50 hp
Recreational Marine
Diesel >50 hp
Recreational
Marine SI
Nonroad SI <25 hp
Nonroad SI >25hp
Recreational SI
Commercial Marine
Diesel b
Commercial Marine
Residual b
Commercial
Marine Other b
Locomotive
Aircraft
Total Nonroad
Total Highway
Total Mobile
Sources
Stationary Point and
Area Sources
Total Man-Made
Sources
Mobile Source
Percent of Total
1996
short tons
1,004,586
304
5,120
1,995,907
18,013,533
1,614,394
921,345
93,638
15,245
5,869
92,496
949,313
24,711,750
53,685,026
78,396,776
16,318,451
94,715,227
83%
%of
mobile
source
1.3%
0.0%
0.0%
2.5%
23.0%
2.1%
1.2%
0.1%
0.0%
0.0%
0.1%
1.2%
32%
68%
100%
—
—
—
%of
total
1.1%
0.0%
0.0%
2.1%
19.0%
1.7%
1.0%
0.1%
0.0%
0.0%
0.1%
1.0%
26%
57%
83%
17%


2020
short tons
697,630
243
9,239
1,977,403
26,372,980
275,647
1,820,865
114,397
36,165
6,542
99,227
1,387,178
32,797,515
48,529,203
81,326,718
15,648,555
96,975,273
84%
%of
mobile
sources
0.9%
0.0%
0.0%
2.4%
32.4%
0.3%
2.2%
0.1%
0.0%
0.0%
0.1%
1.7%
40%
60%
100%
—
—
—
%of
total
0.7%
0.0%
0.0%
2.0%
27.2%
0.3%
1.9%
0.1%
0.0%
0.0%
0.1%
1.4%
34%
50%
84%
16%


2030
short tons
786,181
259
10,973
2,075,666
30,611,599
264,047
1,836,350
123,436
54,924
7,058
107,780
1,502,265
37,380,538
55,847,203
93,227,742
16,325,306
109,553,048
85%
%of
mobile
source
0.8%
0.0%
0.0%
2.2%
32.8%
0.3%
2.0%
0.1%
0.1%
0.0%
0.1%
1.6%
40%
60%
100%
—
—
—
%of
total
0.7%
0.0%
0.0%
1.9%
27.9%
0.2%
1.7%
0.1%
0.1%
0.0%
0.1%
1.4%
34%
51%
85%
15%


a These are 48-state inventories. They do not include Alaska and Hawaii.
b Commercial marine diesel includes Category 1 and 2 compression-ignition (CI) engines using diesel fuel.  The residual
    category includes Category 3 CI engines using residual fuel. The other category includes engines using gasoline and
    steamships fueled with coal.
                                                     3-39

-------
Final Regulatory Impact Analysis
3.3 Contribution of Nonroad Diesel Engines to Selected Local Emission
Inventories

   The contribution of land-based nonroad compression-ignition (CI) engines to PM25 and NOX
emission inventories in many U.S. cities can be significantly greater than that reflected by
national average values.A This is not surprising given the high density of these engines one would
expect to be operating in urban areas. EPA selected a collection of typical cities spread across the
United States to compare projected urban inventories with national average ones for 1996,  2020,
and 2030. The results of this analysis are shown below.

3.3.1 PM2 5 Emissions

   As illustrated in Tables 3.3-1, 3.3-2, and 3.3-3, EPA's city-specific analysis of selected
metropolitan areas for 1996, 2020, and 2030 show that land-based nonroad diesel engine engines
are a significant contributor to total PM2 5 emissions from all man-made sources.
       Construction, industrial, and commercial nonroad diesel equipment comprise most of the land-based nonroad emission
       inventory. These types of equipment are more concentrated in urban areas where construction projects, manufacturing,
       and commercial operations are prevalent.

                                           3-40

-------
                                                                        Emission Inventory
                                          Table 3.3-1
                          Land-Based Nonroad Percent Contribution
                     to PM2 s Inventories in Selected Urban Areas in 1996a-b
MSA, CMSA / State
Atlanta, GA
Boston, MA
Chicago, IL
Dallas-Fort Worth, TX
Indianapolis, IN
Minneapolis, MN
New York, NY
Orlando, FL
Sacramento, CA
San Diego, CA
Denver, CO
El Paso, TX
Las Vegas, NV-AZ
Phoenix -Mesa, AZ
Seattle, WA
Land-Based
Diesel
(short tons)
1,650
4,265
3,374
1,826
1,040
1,484
2,991
764
529
879
1,125
252
1,155
1,549
1,119
Mobile
Sources
(short tons)
7,308
9,539
10,106
5,606
3,126
4,238
6,757
2,559
2,140
3,715
3,199
822
2,700
4,994
4,259
Total Man-
Made Sources
(short tons)
22,190
23,254
40,339
13,667
7,083
15,499
23,380
5,436
7,103
9,631
10,107
1,637
7,511
10,100
15,187
Land-Based
Diesel as % of
Total
7%
18%
8%
13%
15%
10%
13%
14%
7%
9%
11%
15%
15%
15%
7%
Land-Based
Diesel as % of
Mobile Sources
23%
45%
33%
33%
33%
35%
44%
30%
25%
24%
35%
31%
43%
31%
26%
a Includes only direct exhaust emissions; see Chapter 2 for a discussion of secondary fine PM levels.
b Based on inventories developed for the proposed rule.
                                             3-41

-------
Final Regulatory Impact Analysis
                                         Table 3.3-2
                      Annual Land-Based Nonroad Diesel Contributions
                     to PM2 s Inventories in Selected Urban Areas in 202Ob
MSA, CMSA / State
Atlanta, GA
Boston, MA
Chicago, IL
Dallas-Fort Worth, TX
Indianapolis, IN
Minneapolis, MN
New York, NY
Orlando, FL
Sacramento, CA
San Diego, CA
Denver, CO
El Paso, TX
Las Vegas, NV-AZ
Phoenix -Mesa, AZ
Seattle, WA
Land-Based
Diesel
(short tons)
1,429
3,580
2,824
1,499
794
1,188
2,573
652
391
678
923
212
961
1,299
946
Mobile
Sources
(short tons)
4,506
6,720
6,984
3,544
1,779
2,509
4,549
1,743
1,301
2,478
2,149
478
2,080
3,512
3,043
Total Man-
Made Sources
(short tons)
22,846
20,365
42,211
15,202
6,238
15,096
21,566
5,627
5,505
9,135
10,954
1,140
7,804
10,768
13,094
Land-Based
Diesel as % of
Total
6%
18%
7%
10%
13%
8%
12%
12%
7%
7%
8%
19%
12%
12%
7%
Land-Based
Diesel as % of
Mobile Sources
32%
53%
40%
42%
45%
47%
57%
37%
30%
27%
43%
44%
46%
37%
31%
a Includes only direct exhaust emissions; see Chapter 2 for a discussion of secondary fine PM levels.
b Based on inventories developed for the proposed rule.
                                            3-42

-------
                                                                        Emission Inventory
                                          Table 3.3-3
                          Land-Based Nonroad Percent Contribution
                     to PM2 s Inventories in Selected Urban Areas in 203Ob
MSA, CMSA / State
Atlanta, GA
Boston, MA
Chicago, IL
Dallas-Fort Worth, TX
Indianapolis, IN
Minneapolis, MN
New York, NY
Orlando, FL
Sacramento, CA
San Diego, CA
Denver, CO
El Paso, TX
Las Vegas, NV-AZ
Phoenix -Mesa, AZ
Seattle, WA
Land-Based
Diesel
(short tons)
1,647
4,132
3,236
1,721
902
1,354
2,953
752
447
111
1,060
244
1,113
1,499
1,084
Mobile
Sources
(short tons)
4,937
7,529
7,735
3,919
1,934
2,769
5,064
1,957
1,445
2,770
2,379
524
2,307
3,870
3,357
Total Man-
Made Sources
(short tons)
24,880
21,846
45,975
16,622
6,753
16,586
22,891
6,084
5,890
10,096
12,117
1,243
8,512
11,989
14,148
Land-Based
Diesel as % of
Total
7%
19%
7%
10%
13%
8%
13%
12%
8%
8%
9%
20%
13%
13%
8%
Land-Based
Diesel as % of
Mobile Sources
33%
55%
42%
44%
47%
49%
58%
38%
31%
28%
45%
47%
48%
39%
32%
a Includes only direct exhaust emissions; see Chapter 2 for a discussion of secondary fine PM levels.
b Based on inventories developed for the proposed rule.
                                             3-43

-------
3.3.2 NOV Emissions
    As presented in Tables 3.3-4, 3.3-5, and 3.3-6, EPA's city-specific analysis of selected
metropolitan areas for 1996, 2020, and 2030 show that land-based nonroad diesel engine engines
are a significant contributor to total NOX emissions from all man-made sources.

                                        Table 3.3-4
                         Land-Based Nonroad Percent Contribution
                     to NOV Inventories in Selected Urban Areas in 1996a
MSA, CMSA / State
Atlanta, GA
Boston, MA
Chicago, IL
Dallas-Fort Worth, TX
Indianapolis, IN
Minneapolis, MN
New York, NY
Orlando, FL
Sacramento, CA
San Diego, CA
Denver, CO
El Paso, TX
Las Vegas, NV-AZ
Phoenix -Mesa, AZ
Seattle, WA
Land-Based
Diesel
(short tons)
16,238
43,362
32,276
17,852
9,487
13,843
29,543
7,493
5,666
9,460
11,080
2,498
11,788
15,145
11,227
Mobile
Sources
(short tons)
205,465
232,444
296,710
152,878
89,291
124,437
184,384
61,667
55,144
99,325
86,329
24,382
50,724
115,544
115,264
Total Man-
Made Sources
(short tons)
298,361
311,045
509,853
186.824
113,300
224,817
262,021
75,714
58,757
107,024
146,807
30,160
108,875
161,606
133,840
Land-Based
Diesel as % of
Total
5%
14%
6%
10%
8%
6%
11%
10%
10%
9%
8%
8%
11%
9%
8%
Land-Based
Diesel as % of
Mobile Sources
8%
19%
11%
12%
11%
11%
16%
12%
10%
10%
13%
10%
23%
13%
10%
' Based on inventories developed for the proposed rule.

-------
                                        Table 3.3-5
                      Annual Land-Based Nonroad Diesel Contributions
                     to NOV Inventories in Selected Urban Areas in 202O
MSA, CMSA / State
Atlanta, GA
Boston, MA
Chicago, IL
Dallas-Fort Worth, TX
Indianapolis, IN
Minneapolis, MN
New York, NY
Orlando, FL
Sacramento, CA
San Diego, CA
Denver, CO
El Paso, TX
Las Vegas, NV-AZ
Phoenix -Mesa, AZ
Seattle, WA
Land-Based
Diesel
(short tons)
12,650
31,282
24,732
13,334
6,982
10,376
22,456
5,837
4,297
7,464
8,251
1,847
8,501
11,560
8,283
Mobile
Sources
(short tons)
69,816
93,308
123,823
60,745
36,283
47,375
67,083
28,653
18,870
46,005
38,435
10,105
26,840
48,348
51,252
Total Man-
Made Sources
(short tons)
193,456
167,572
333,945
101,453
60,059
165,775
112,960
45,362
23,111
51,909
103,533
12,452
72,829
105,185
76,161
Land-Based
Diesel as % of
Total
7%
19%
7%
13%
12%
6%
20%
13%
19%
14%
8%
15%
12%
11%
11%
Land-Based
Diesel as % of
Mobile Sources
18%
34%
20%
22%
19%
22%
33%
20%
23%
16%
21%
18%
32%
24%
16%
' Based on inventories developed for the proposed rule.
                                            3-45

-------
                                       Table 3.3-6
                        Land-Based Nonroad Percent Contribution
                     to NOV Inventories in Selected Urban Areas in 203O
MSA, CMSA / State
Atlanta, GA
Boston, MA
Chicago, IL
Dallas-Fort Worth, TX
Indianapolis, IN
Minneapolis, MN
New York, NY
Orlando, FL
Sacramento, CA
San Diego, CA
Denver, CO
El Paso, TX
Las Vegas, NV-AZ
Phoenix -Mesa, AZ
Seattle, WA
Land-Based
Diesel
(short tons)
14,190
35,039
27,525
14,839
7,641
11,444
25,064
6,551
4,806
8,401
9,185
2,062
9,544
12,952
9,247
Mobile
Sources
(short tons)
65,746
92,537
120,694
56,907
34,442
45,326
67,163
28,365
17,498
43,930
37,105
9,422
26,349
46,280
49,258
Total Man-
Made Sources
(short tons)
191,932
168,422
334,334
100,721
58,793
167,154
108,215
45,267
21,952
50,296
104,217
11,905
72,926
106,061
77,133
Land-Based
Diesel as % of
Total
7%
21%
8%
15%
13%
7%
23%
14%
22%
17%
9%
17%
13%
12%
12%
Land-Based
Diesel as % of
Mobile Sources
22%
38%
23%
26%
22%
25%
37%
23%
27%
19%
25%
22%
36%
28%
19%
' Based on inventories developed for the proposed rule.
3.4 Nonroad Diesel Controlled Emission Inventory Development

   This section describes how the controlled emission inventories were developed for the four
categories of nonroad diesel engines affected by this rule: land-based diesel engines, commercial
marine diesel vessels, locomotives, and recreational marine diesel engines. For land-based diesel
engines, there are separate sections for criteria (i.e., PM2 5, NOX, SO2, VOC, and CO)  and air
toxics emission development.

3.4.1 Land-Based Diesel Engines—PM2 5, NOX, SO2, VOC, and CO Emissions

   The emission inventory estimates used in this rule were generated using the draft
NONROAD2004 model with certain input modifications to account for the in-use diesel fuel
sulfur reductions and the engine controls associated with the new emission standards.  This
section will describe only these modifications to the model inputs, since the other aspects of the
model, including inputs for earlier engines, are covered in detail in the technical reports that
document the draft NONROAD2004 model.
                                          3-46

-------
                                                                  Emission Inventory
   3.4.1.1 Standards and Zero-Hour Emission Factors

   The new emission standards are summarized in Table 3.4-1. The modeled emission factors
corresponding to the new emission standards are shown in Table 3.4-2.  These emission factors
are derived from the standards by applying an assumed 8  percent compliance margin to the
standard. This compliance margin was derived from data for highway diesel vehicles and used in
the HD2007 rulemaking.

   Besides exhaust emissions, the final rule includes changes in crankcase hydrocarbon
emissions. Crankcase losses before Tier 4 have been modeled as 2.0 percent of exhaust HC, and
any crankcase emissions of other pollutants have been considered negligible.  For all Tier 4
engines, including those using transitional controls without particulate traps, our modeling now
assumes zero crankcase emissions.

   3.4.1.2 Transient Adjustment Factors

   The supplemental nonroad transient test will apply to  a nonroad diesel engine when that
engine must first show compliance with the Tier 4 PM and NOX+NMHC emissions standards
which are based on the performance of the advanced post-combustion emissions control systems
(e.g., catalyzed-diesel particulate filters and NOX adsorbers).  This is 2011 for engines at or above
175 hp, 2012 for 75-175 hp engines, and 2013 for engines under 75 hp.  Details regarding the
transient testing requirements and manufacturer options are provided in Section III of the
preamble. More broadly though, transient emissions control is expected to be an integral part of
all Tier 4 engine design considerations, including engines under 75 hp meeting either the 0.22
g/hp-hr or 0.30 g/hp-hr Tier 4 PM standards in 2008. Thus, there was no Transient Adjustment
Factor (TAP) applied to the emission factors for Tier 4  engines (i.e., the model applies a TAF of
1.0); the zero-hour emission factor was modeled simply as the value of the standard minus an
assumed 8 percent compliance margin.
                                          3-47

-------
Final Regulatory Impact Analysis
                                               Table 3.4-1
                                  Tier 4 Emission Standards Modeled
Engine
Power
kW<19
(hp <25)
19 < kW<56
(25 < hp < 75)
56 < kW<130
(75 < hp < 175)
130 < kW<560
(175 < hp < 750)
kW > 560
(hp > 750)
except Generator sets
Generator sets
560 < kW < 895
(750 < hp < 1200)
Generator sets
kW > 895
(hp > 1200)
Emission Standard
(g/hp-hr)
transitional
or final
final
transitional
final
transitional
final
transitional
final
transitional
final
transitional
final
transitional
final
PM
0.30
0.22
0.02
0.01
0.01
0.01
0.01
0.075
0.03
0.075
0.02
0.075
0.02
NO/
NMHCa
5.6 b'c
5.6/3. 5 b'c
3.5b
0.30
(50%)
0.30
0.30
(50%)
0.30
2.6
2.6
2.6
0.50
0.50
0.50
0.14
(50%)
0.14
0.14
(50%)
0.14
0.30
0.14
0.30
0.14
0.30
0.14
cod
6.0/4.9 c
4.1/3.7 c
4.1/3.7°
3.7 c
3.7 c
2.6 c
2.6 c
2.6 c
2.6 c
2.6 c
2.6 c
2.6 c
2.6 c
Model
Year(s)
2008
2008-2012
2013
2012-2013
2014
2011-2013
2014
2011-2014
2015
2011-2014
2015
2011-2014
2015
a Percentages are model year sales fractions required to comply with the indicated NOX and NMHC standards, for model
years where less than 100 percent is required.  For a complete description of manufacturer options and alternative
standards, refer to Section II of the preamble.
b This is a combined NMHC + NOX standard.
0 This emission standard level is unchanged from the level that applies in the previous model year. For 25-75 hp engines,
    the transitional NMHC + NOX standard is 5.6 g/hp-hr for engines below 50 hp and 3.5 g/hp-hr for engines at or above
    50 hp. For engines under 75 hp, the CO standard is 6.0 g/hp-hr for engines below 11 hp, 4.9 g/hp-hr for engines 11 to
    under 25 hp, 4.1 g/hp-hr for engines 25 to below 50 hp and 3.7 g/hp-hr for engines at  or above 50 hp.
d There are no Tier 4 CO standards. The CO emission standard level is unchanged from the level that applies in the
previous model year.
                                                   3-48

-------
                                                                                            Emission  Inventory
                                                      Table 3.4-2
                              NONROAD Model EF Inputs for Tier 4 Engines
Engine
Power
hp< 11
1 1< hp < 25
25 < hp < 50
50 < hp < 75
75 750
except Generator
sets
Generator sets
750 < hp < 1200
Generator sets
hp > 1200
Emission Factor Modeling Inputs, g/hp-hr a
Type of
standard
final
final
transitional
final
transitional
final
transitional
final
transitional
final
transitional
final
transitional
final
transitional
final
transitional
final
transitional
final
transitional
final
PM
0.28
0.28
0.20
0.018
0.20
0.018
0.01
0.01
0.01
0.01
0.01
0.01
0.01
0.01
0.01
0.01
0.069
0.028
0.069
0.018
0.069
0.018
N0xb-c
4.
4.
4.
3
3
3
3.0 (50%)
0.
2.5 (75%)
0.
2.5 (50%)
0.
2.5 (50%)
0.
2.5 (50%)
0.
2.
2.
2.
0.
0.
0.
30
44
73
.0
.0
.0
0.28 (50%)
28
0.28 (25%)
28
0.28 (50%)
28
0.28 (50%)
28
0.28 (50%)
28
39
39
39
46
46
46
THC c'd
0.55
0.44
0.28
0.13
0.18
0.13
0.13
0.13
0.13
0.13
0.13
0.13
0.13
0.13
0.13
0.13
0.28
0.13
0.28
0.13
0.28
0.13
coe
4.11
2.16
1.53
0.15
2.4
0.24
0.24
0.24
0.087
0.087
0.075
0.075
0.084
0.084
0.13
0.13
0.076
0.076
0.076
0.076
0.076
0.076
Model
Year(s)
2008
2008
2008
2013
2008
2013
2012-2013
2014
2012-2014
2015
2011-2013
2014
2011-2013
2014
2011-2013
2014
2011-2014
2015
2011-2014
2015
2011-2014
2015
a Transient emission control is assumed for Tier 4 engines, so Transient Adjustment Factors are not applied to the emission factors shown here.
b Percentages are model-year sales fractions required to comply with the indicated standard.
0 NMHC + NOX is a combined standard, so for modeling purposes the NOX and HC are separated using a NOX/HC ratio that approximates the
     results found in prior test programs, as described in technical report NR-009b.
d HC Standards are in terms  of NMHC, but the model expects inputs as THC, so a conversion factor of 1.02 is applied to the NMHC value to get
     the THC model input.
                                                          3-49

-------
Final Regulatory Impact Analysis
e CO emissions from Tier 4 engines are assumed to decrease by 90% from its prior levels in any cases where particulate traps are expected for PM
   control.
   3.4.1.3 Deterioration Rates

   The deterioration rates (d) used for the modeling of Tier 4 engines are the same as used for
Tier 3 engines for all affected pollutants (PM, NOX, HC, and CO). These are listed in Table 3.4-3
below and  are fully documented in technical report NR

                                         Table 3.4-3
                        Deterioration Rates for Nonroad Diesel Engines
Pollutant
HC
CO
NOX
PM
Relative Deterioration Rate (percent increase per percent useful life expended)3
Base/Tier 0
0.047
0.185
0.024
0.473
Tierl
0.036
0.101
0.024
0.473
Tier 2
0.034
0.101
0.009
0.473
Tier 3
0.027
0.151
0.008
0.473
Tier 4
0.027
0.151
0.008
0.473
' At the median life point, the Deterioration Factor = 1 + relative deterioration rate.
    3.4.1.4 In-Use Sulfur Levels, Certification Sulfur Levels, and Sulfur Conversion Factors

    Tables 3.4-4 and 3.4-5 show the certification and in-use fuel sulfur levels by calendar year
and engine power range that were assumed for modeling the engines regulated under this rule.
The certification sulfur levels are the default fuel sulfur levels used to calculate the zero mile PM
and SO2 emission factors in the model (referred to as Sbase in Section 3.1.1.2.1). The in-use fuel
sulfur level is the episodic fuel sulfur level (referred to as Sin.use in Section 3.1.1.2.1).
Adjustments to PM and SO2 for in-use fuel sulfur levels are made relative to the certification
sulfur levels in the model. As described above for the baseline inventory development, the in-use
fuel sulfur content, fuel consumption, sulfate conversion factor, and exhaust HC emission factor
(unburned fuel) determine the SO2 emissions,  and a fraction of the fuel sulfur is also converted to
sulfate PM. The changes  for modeling of the control case are (a) lower sulfur content for in-use
and certification fuel per this rule, and (b) the use of a higher sulfur-to-sulfate conversion factor
for engines that are expected to use a particulate trap/filter to achieve the PM standards of 0.01 or
0.02 g/hp-hr (30 percent conversion instead of 2.247 percent that is used for all earlier nontrap-
equipped engines).

    The in-use sulfur levels  account for the 500 ppm standard beginning in 2007, the 15 ppm
standard for land-based engines beginning in 2010, and the 15 ppm standard for marine engines
                                            3-50

-------
                                                                         Emission Inventory
and locomotives beginning in 2012. The derivation of the annual fuel sulfur levels is described in
detail in Chapter 7.  The in-use sulfur levels in Table 3.4-5 used for modeling differ slightly from
those presented in Chapter 7, since minor revisions were made subsequent to the modeling.

                                          Table 3.4-4
                        Modeled Certification Diesel Fuel Sulfur Content
Engine
Power
kW<56
(hp <75)
56 < kW < 75
(75 < hp < 100)
75 < kW<130
(100 < hp < 175)
130 < kW<560
(175 < hp < 750)
kW > 560
(hp > 750)
Standards
Tier 2
transitional
final
Tier 3 transitional a
final
Tier 3
final
Tier 3
final
Tier 2
final
Modeled Certification Fuel
Sulfur Content, PPM
2000
500
15
500
15
2000
15
2000
15
2000
15
Model
Year(s)
through 2007
2008-2012
2013
2008-2011
2012
2007-2011
2012
2006-2010
2011
2006-2010
2011
' The emission standard here is still Tier 3 as in the Baseline case, but since the Tier 3 standard begins in 2008 for 50-100
    hp engines it is assumed that this new technology introduction will allow manufacturers to take advantage of the
    availability of 500 ppm fuel that year.
                                              3-51

-------
Final Regulatory Impact Analysis
                                      Table 3.4-5
   Modeled 48-State & 50-State In-Use Diesel Fuel Sulfur Content for Controlled Inventories
Applications
Land-based,
all power ranges
Recreational Marine,
Commercial Marine, and
Locomotives
Calendar
Year(s)
through 2005
2006
2007
2008-2009
2010
2011-2013
2014
2015+
through 2000
2001
2002-2003
2004-2005
2006
2007
2008-2009
2010
2011
2012
2013
2014
2015-2017
2018-2038
2039-2040
Modeled In-Use Fuel Sulfur Content, ppm
48-State
2283
2249
1140
348
163
31
19
11
2641
2637
2638
2639
2616
1328
408
307
234
123
43
51
56
56
55
50-State
2284
2242
1139
351
165
32
20
11
2640
2635
2637
2637
2588
1332
435
319
236
124
44
52
56
55
55
   3.4.1.5 Controlled Inventory

   Tables 3.4-6a and 3.4-6b present the PM10, PM25, NOX, SO2, VOC, and CO controlled
emissions for land-based nonroad diesel engines in 1996 and 2000-2040, for the 48-state and 50-
state inventories, respectively.
                                         3-52

-------
                                                          Emission Inventory
                               Table3.4-6a
Controlled (48-State) Emissions for Land-Based Nonroad Diesel Engines (short tons)
Year
1996
2000
2001
2002
2003
2004
2005
2006
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2031
2032
2033
2034
2035
2036
2037
2038
2039
2040
PM,n PM,, NOT SO, VOC CO
192,275
176,056
170,451
165,017
159,268
153,932
148,720
143,840
132,534
123,646
120,512
116,263
110,940
104,319
97,187
89,522
81,780
74,718
68,079
61,986
56,496
51,613
47,285
43,376
39,837
36,548
33,508
30,735
28,234
26,125
24,177
22,369
20,873
19,492
18,188
16,970
15,877
14,930
14,053
13,577
13,194
12,852
186,507
170,774
165,338
160,067
154,490
149,314
144,259
139,525
128,558
119,936
116,896
112,775
107,612
101,189
94,271
86,837
79,326
72,476
66,036
60,127
54,801
50,065
45,866
42,074
38,642
35,452
32,503
29,813
27,387
25,341
23,452
21,698
20,247
18,907
17,643
16,461
15,401
14,482
13,631
13,169
12,798
12,467
1,564,904
1,550,355
1,537,890
1,526,119
1,505,435
1,486,335
1,467,547
1,435,181
1,399,787
1,359,631
1,317,925
1,277,888
1,224,329
1,165,155
1,108,560
1,031,680
958,769
890,935
828,178
772,291
722,094
677,420
639,156
606,068
576,872
551,570
529,260
510,126
493,869
479,930
467,852
458,649
451,478
445,218
439,984
435,620
432,306
429,867
428,058
427,438
427,591
428,084
143,572
161,977
166,644
171,309
175,971
180,630
185,287
187,085
97,142
30,359
31,064
14,881
2,853
2,850
2,832
1,724
992
987
984
983
984
986
991
996
1,003
1,011
1,019
1,028
1,039
1,050
1,062
1,074
1,087
1,100
1,113
1,126
1,140
1,155
1,169
1,183
1,198
1,213
220,971
199,887
191,472
183,525
176,383
169,873
163,663
156,952
150,357
143,138
136,085
129,186
122,434
115,877
109,726
104,160
98,766
93,976
89,760
85,896
82,398
79,372
76,813
74,680
72,854
71,291
69,973
68,878
68,008
67,319
66,761
66,344
66,118
65,979
65,904
65,909
66,004
66,186
66,418
66,781
67,195
67,645
1,004,586
916,507
880,129
845,435
813,886
787,559
763,062
741,436
724,449
707,098
691,627
677,599
650,276
609,685
563,695
518,729
475,349
435,137
398,578
365,813
336,094
309,593
286,679
266,071
247,738
231,324
216,510
203,435
192,100
182,716
174,448
167,014
161,116
155,882
151,053
146,747
143,229
140,378
137,840
135,517
133,748
132,256
                                  3-53

-------
Final Regulatory Impact Analysis
                                    Table3.4-6b
     Controlled (50-State) Emissions for Land-Based Nonroad Diesel Engines (short tons)
Year
1996
2000
2001
2002
2003
2004
2005
2006
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2031
2032
2033
2034
2035
2036
2037
2038
2039
2040
PM,n PM9S MX SO, VOC CO
193,166
176,881
171,256
165,801
160,030
154,670
149,434
144,479
133,159
124,257
121,113
116,841
111,492
104,846
97,687
89,993
82,171
75,070
68,395
62,269
56,750
51,840
47,489
43,560
40,006
36,703
33,651
30,866
28,355
26,237
24,280
22,464
20,963
19,577
18,269
17,047
15,951
15,000
14,120
13,642
13,257
12.915
187,371
171,575
166,118
160,827
155,229
150,030
144,951
140,145
129,165
120,529
117,479
113,336
108,147
101,700
94,757
87,293
79,706
72,818
66,343
60,401
55,047
50,285
46,064
42,254
38,806
35,602
32,641
29,940
27,504
25,450
23,552
21,790
20,334
18,990
17,721
16,536
15,472
14,550
13,696
13,233
12,859
12.527
1,573,083
1,558,392
1,545,852
1,534,007
1,513,203
1,493,989
1,475,092
1,442,534
1,406,936
1,366,553
1,324,613
1,284,357
1,230,489
1,170,969
1,114,051
1,036,731
963,408
895,198
832,101
775,920
725,464
680,563
642,114
608,874
579,551
554,147
531,753
512,553
496,243
482,261
470,147
460,918
453,730
447,458
442,218
437,851
434,539
432,104
430,302
429,692
429,857
430 365
144,409
162,920
167,615
172,307
176,996
181,683
186,368
187,508
97,580
30,786
31,501
15,145
2,961
2,957
2,939
1,825
997
992
989
988
989
991
996
1,001
1,008
1,016
1,024
1,034
1,044
1,056
1,068
1,080
1,093
1,106
1,119
1,133
1,147
1,161
1,175
1,190
1,204
1.219
222,084
200,903
192,447
184,462
177,287
170,744
164,505
157,762
151,134
143,880
136,792
129,859
123,074
116,483
110,299
104,704
99,281
94,464
90,227
86,343
82,828
79,786
77,214
75,070
73,234
71,662
70,338
69,237
68,363
67,671
67,110
66,690
66,464
66,324
66,250
66,256
66,352
66,535
66,769
67,135
67,551
68.004
1,009,804
921,226
884,645
849,756
818,037
791,568
766,944
745,216
728,159
710,743
695,221
681,150
653,692
612,882
566,639
521,423
477,800
437,357
400,587
367,637
337,757
311,112
288,075
267,360
248,939
232,449
217,569
204,437
193,052
183,622
175,312
167,841
161,916
156,659
151,810
147,486
143,953
141,089
138,541
136,210
134,435
132.940
                                       3-54

-------
                                                                   Emission Inventory
3.4.2 Land-Based Diesel Engines—Air Toxics Emissions

   Since air toxics emissions are part of the VOC emission inventory, NMHC standards in this
rule will also affect air toxics emissions. Tables 3.4-7a and 3.4-7b show 48-state and 50-state
estimated emissions for five major air toxics, benzene, formaldehyde, acetaldehyde,
1,3-butadiene, and acrolein, resulting from the final rule. EPA uses the same fractions used to
calculate the base air toxic emissions without the final rule (see Section 3.1.2), along with the
estimated VOC emissions resulting from the final rule, to calculate the air toxics emissions
resulting from the final rule.

                                       Table3.4-7a
 Controlled (48-State) Air Toxic Emissions for Land-Based Nonroad Diesel Engines (short tons)
Year
2000
2005
2007
2010
2015
2020
2025
2030
Benzene
3,998
3,273
3,007
2,584
1,975
1,587
1,399
1,327
Formaldehyde
23,587
19,312
17,742
15,244
11,654
9,366
8,257
7,829
Acetaldehyde
10,594
8,674
7,969
6,847
5,235
4,207
3,709
3,516
1,3 -butadiene
400
327
301
258
198
159
140
133
Acrolein
600
491
451
388
296
238
210
199
                                       Table3.4-7b
 Controlled (50-State) Air Toxic Emissions for Land-Based Nonroad Diesel Engines (short tons)
Year
2000
2005
2007
2010
2015
2020
2025
2030
Benzene
4,018
3,290
3,023
2,597
1,986
1,596
1,407
1,334
Formaldehyde
23,707
19,412
17,834
15,323
11,715
9,415
8,300
7,869
Acetaldehyde
10,648
8,719
8,010
6,883
5,262
4,229
3,728
3,535
1,3 -butadiene
402
329
302
260
199
160
141
133
Acrolein
603
494
453
390
298
239
211
200
                                          3-55

-------
Final Regulatory Impact Analysis
3.4.3 Commercial Marine Vessels and Locomotives

   The control case locomotive and commercial marine inventories for VOC, CO, and NOX are
identical to the base case inventories, since no new controls apply for these engines.  However,
due to the new requirements to reduce sulfur levels in diesel fuel, decreases are expected in PM
and SO2 inventories for these engines.

   The method used for estimating PM and SO2 emissions in the control case is nearly almost
identical to that described in Section 3.1.3 for the base case, but the fuel sulfur levels in the
equations are changed to reflect the control case sulfur. The control case PM and SO2 emission
inventory estimates presented here assume that locomotive and commercial marine applications
will use diesel fuel meeting a 500 ppm sulfur standard beginning in June 2007 and a 15 ppm
sulfur standard beginning in June 2012. Additional sulfur adjustments were made to account for
the "spillover" of low-sulfur highway fuel meeting a 15 ppm standard  in the applicable years
before the start of the 15 ppm nonroad fuel standard.

   As in the base case, the same sulfur-to-sulfate conversion rate was used as for land-based
diesel applications before they started using aftertreatment technologies (2.247 percent).  The
slight decrease in average sulfur level in 2006 is due to the introduction of highway diesel fuel
meeting the 2007 15 ppm standard, and the "spillover" of this highway fuel into the nonroad fuel
pool. Note that there are transition years in which the control sulfur level begins in June,  in which
case the annual average sulfur level shown reflects an interpolation of five months at the  higher
sulfur level of the prior year plus seven months at the new lower sulfur level.  The derivation of
these sulfur levels are described  in more detail in Chapter 7.

   The control case locomotive and commercial marine PM inventories were calculated by
subtracting the sulfate PM benefits (from decreased fuel sulfur content) described above from the
base case locomotive and commercial marine PM inventories.  The 48-state and 50-state  control
case locomotive and commercial marine PM25 and SO2 inventories are given in Tables 3.4-8a and
3.4-8b, respectively.
                                          3-56

-------
                                                         Emission Inventory
                              Table3.4-8a
                Controlled (48-State) Fuel Sulfur Levels, SO2
Sulfate PM, and PM2 s Emissions for Locomotives and Commercial Marine Vessels
Year
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2031
2032
2033
2034
2035
2036
2037
2038
2039
2040
Control
Sulfur Level
(ppm)
1,328
408
408
307
234
123
43
51
56
56
56
56
56
56
56
56
56
56
56
56
56
56
56
56
56
56
56
56
56
56
56
56
55
55
Control
SO2
Loco
(tons/yr)
26,430
8,210
8,321
6,352
4,944
2,614
921
1,099
1,231
1,244
1,255
1,263
1,274
1,282
1,288
1,298
1,309
1,319
1,332
1,342
1,352
1,363
1,373
1,384
1,394
1,405
1,416
1,427
1,438
1,449
1,460
1,471
1,482
L494
CMV
(tons/yr)
14,517
4,512
4,554
3,457
2,675
1,415
498
595
667
672
678
687
697
703
710
721
727
736
743
751
760
769
111
786
795
805
814
824
833
843
853
863
874
884
Sulfate PM
Loco
(tons/yr)
2,126
661
669
511
398
210
74
88
99
100
101
102
103
103
104
104
105
106
107
108
109
110
110
111
112
113
114
115
116
117
117
118
119
120
CMV
(tons/yr)
1,168
363
366
278
215
114
40
48
54
54
55
55
56
57
57
58
59
59
60
60
61
62
63
63
64
65
65
66
67
68
69
69
70
71
Total PM2 5
Loco (tons/yr)
17,023
15,146
15,038
14,725
15,067
14,703
14,354
14,146
13,936
13,745
13,527
13,626
13,409
13,149
12,861
12,618
12,729
12,476
12,229
11,962
12,060
11,785
11,504
11,599
11,310
11,016
11,107
10,804
10,893
10,983
10,669
10,757
10,434
10.520
CMV (tons/yr)
17,586
16,641
16,485
16,377
16,254
16,003
15,793
15,660
15,534
15,455
15,402
15,367
15,382
15,436
15,511
15,599
15,719
15,846
15,990
16,138
16,295
16,452
16,614
16,778
16,950
17,122
17,292
17,463
17,636
17,811
17,986
18,162
18,339
18.517
                                 3-57

-------
Final Regulatory Impact Analysis
                                      Table3.4-8b
                       Controlled (50-State) Fuel Sulfur Levels, SO2
       Sulfate PM, and PM2 s Emissions for Locomotives and Commercial Marine Vessels
Year
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2031
2032
2033
2034
2035
2036
2037
2038
2039
2040
Control
Sulfur Level
(ppm)
1,332
435
435
319
236
124
44
52
56
56
56
55
55
55
55
55
55
55
55
55
55
55
55
55
55
55
55
55
55
55
55
55
55
55
Control
SO2
Loco
(tons/yr)
26,548
8,764
8,881
6,615
4,990
2,646
943
1,133
1,215
1,228
1,239
1,247
1,258
1,266
1,271
1,281
1,291
1,302
1,314
1,324
1,334
1,344
1,355
1,365
1,375
1,386
1,397
1,407
1,418
1,429
1,440
1,451
1,462
1.473
CMV
(tons/yr)
15,305
5,055
5,103
3,779
2,834
1,504
535
645
692
697
703
712
723
729
737
747
754
763
771
779
788
797
806
815
825
834
844
854
864
874
885
895
906
917
Sulfate PM
Loco
(tons/yr)
2,136
705
715
532
401
213
76
91
98
99
100
100
101
102
102
103
104
105
106
107
107
108
109
110
111
112
112
113
114
115
116
117
118
119
CMV
(tons/yr)
1,231
407
411
304
228
121
43
52
56
56
57
57
58
59
59
60
61
61
62
63
63
64
65
66
66
67
68
69
70
70
71
72
73
74
Total PM2 5
Loco (tons/yr)
17,127
15,285
15,177
14,838
15,161
14,796
14,447
14,240
14,027
13,836
13,619
13,719
13,502
13,243
12,955
12,713
12,825
12,572
12,326
12,058
12,158
11,883
11,603
11,699
11,411
11,116
11,208
10,906
10,996
11,087
10,774
10,863
10,541
10.628
CMV (tons/yr)
18,559
17,587
17,423
17,293
17,152
16,888
16,667
16,528
16,393
16,310
16,254
16,219
16,234
16,292
16,372
16,465
16,591
16,726
16,878
17,034
17,200
17,366
17,537
17,710
17,892
18,073
18,253
18,434
18,617
18,801
18,987
19,173
19,359
19.548
3.4.4 Recreational Marine Engines

   Even though this final rule does not include any emission standards for marine engines, there
are PM and SO2 benefits associated with these engines due to the fuel sulfur standards. The
emission inventory estimates presented in Tables 3.4-9a and 3.4-9b assume that recreational
                                         3-58

-------
                                                                 Emission Inventory
marine applications will use diesel fuel meeting the same standards as locomotive and
commercial marine diesel fuel, as shown in Table 3.4-5.

                                      Table3.4-9a
      Controlled (48-State) Emissions for Recreational Marine Diesel Engines (short tons)
Year
1996
2000
2001
2002
2003
2004
2005
2006
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2031
2032
2033
2034
2035
2036
2037
2038
2039
2040
PM10
951
1,070
1,099
1,130
1,160
1,190
1,220
1,233
1,020
862
865
847
833
810
792
790
787
783
778
772
767
761
756
750
745
740
740
744
749
756
763
772
781
790
799
808
818
828
838
849
859
870
PM25
923
1,038
1,066
1,096
1,125
1,154
1,183
1,196
990
836
839
822
808
786
768
767
764
759
755
749
744
738
733
728
722
718
717
721
727
733
741
749
757
766
775
784
794
803
813
823
833
844
NOX
33,679
37,943
39,071
40,198
41,325
42,452
43,578
44,105
44,602
45,066
45,415
45,729
46,022
46,282
46,528
46,765
46,969
47,168
47,362
47,525
47,687
47,847
48,003
48,182
48,363
48,593
48,961
49,501
50,092
50,716
51,392
52,085
52,790
53,510
54,228
54,959
55,702
56,444
57,197
57,963
58,729
59 506
SO2
4,286
4,831
4,968
5,114
5,259
5,406
5,551
5,647
2,940
926
948
731
570
306
109
133
149
152
155
158
161
164
167
170
173
176
180
183
186
189
192
195
198
201
204
207
210
213
216
220
219
222
voc
1,297
1,455
1,494
1,533
1,571
1,609
1,647
1,657
1,664
1,670
1,670
1,668
1,665
1,660
1,655
1,649
1,642
1,634
1,627
1,618
1,611
1,604
1,597
1,592
1,586
1,583
1,587
1,599
1,614
1,630
1,649
1,669
1,689
1,710
1,731
1,753
1,775
1,798
1,820
1,844
1,868
1.892
CO
5,424
6,098
6,271
6,444
6,615
6,787
6,958
7,128
7,298
7,467
7,636
7,804
7,971
8,137
8,303
8,469
8,635
8,802
8,969
9,137
9,308
9,482
9,655
9,829
10,004
10,178
10,354
10,529
10,704
10,880
11,056
11,232
11,409
11,585
11,762
11,938
12,115
12,292
12,469
12,646
12,823
13.001
                                         3-59

-------
Final Regulatory Impact Analysis
                                     Table3.4-9b
      Controlled (50-State) Emissions for Recreational Marine Diesel Engines (short tons)
Year
1996
2000
2001
2002
2003
2004
2005
2006
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2031
2032
2033
2034
2035
2036
2037
2038
2039
2040
PM,n
957
1,076
1,106
1,137
1,167
1,197
1,227
1,236
1,027
872
875
855
839
816
797
795
792
788
783
111
111
766
760
755
749
745
744
748
754
760
768
776
785
794
804
813
823
833
843
854
865
876
PM,,
929
1,044
1,073
1,103
1,132
1,161
1,190
1,199
997
846
849
829
814
791
773
772
768
764
759
753
748
743
737
732
727
722
722
726
731
737
745
753
762
771
779
789
798
808
818
828
839
849
MX
33,891
38,182
39,317
40,452
41,586
42,719
43,852
44,383
44,883
45,350
45,701
46,018
46,312
46,573
46,821
47,060
47,265
47,465
47,660
47,825
47,987
48,148
48,305
48,485
48,667
48,899
49,269
49,813
50,408
51,036
51,716
52,413
53,123
53,847
54,570
55,305
56,053
56,799
57,558
58,329
59,099
59,881
SO,
4,312
4,859
4,995
5,145
5,290
5,436
5,582
5,622
2,967
993
1,017
764
578
311
113
136
150
153
156
156
159
162
165
168
171
174
177
180
183
187
190
193
196
199
202
205
208
211
214
217
220
223
voc
1,305
1,464
1,503
1,542
1,581
1,619
1,658
1,667
1,674
1,680
1,680
1,678
1,675
1,671
1,665
1,660
1,652
1,645
1,637
1,629
1,621
1,614
1,607
1,602
1,596
1,593
1,597
1,609
1,624
1,640
1,659
1,679
1,700
1,721
1,742
1,764
1,786
1,809
1,832
1,856
1,879
1,904
CO
5,458
6,137
6,311
6,484
6,657
6,829
7,001
7,173
7,344
7,514
7,684
7,853
8,021
8,189
8,356
8,523
8,690
8,857
9,025
9,195
9,367
9,541
9,716
9,891
10,067
10,243
10,419
10,595
10,772
10,949
11,126
11,303
11,481
11,658
11,836
12,013
12,191
12,369
12,547
12,726
12,904
13,082
                                        3-60

-------
                                                                   Emission Inventory
3.5 Projected Emission Reductions from the Final Rule

   Emissions from nonroad diesel engines will continue to be a significant part of the emission
inventory in the coming years. In the absence of new emission standards, we expect overall
emissions from nonroad diesel engines to generally decline across the nation for the next 10 to 15
years, depending on the pollutant.  Although nonroad diesel engine emissions decline during this
period, this trend will not be enough to adequately reduce the large amount of emissions that
these engines contribute. In addition, after the  2010 to 2015 time period we project that this trend
reverses and emissions rise into the future in the absence of additional regulation of these engines.
The initial downward trend occurs as the nonroad fleet becomes increasingly dominated over time
by engines that comply with existing emission  regulations.  The upturn in emissions beginning
around 2015 results  as growth in the nonroad sector overtakes the effect of the existing emission
standards.

   The engine and fuel standards in this rule will affect fine particulate matter (PM2 5), oxides of
nitrogen (NOX), sulfur oxides (SO2), volatile organic hydrocarbons (VOC), air toxics, and carbon
monoxide (CO).  For engines used in locomotives, commercial  marine vessels,  and recreational
marine vessels, the requirements for low-sulfur fuel will affect PM2 5 and SO2.

   This section discusses the projected emission reductions associated with this final rule.  The
baseline case represents future emissions with current standards. The controlled case estimates
the future emissions of these engines based on  the new emission standards and fuel requirements.
Both 48-state and 50-state results are presented. Tables 3.5-la and 3.5-lb present a summary of
the total 48-state and 50-state emission reductions for each pollutant.

3.5.1 PM2 5 Reductions

   48-State and 50-state emissions of PM25 from land-based nonroad diesel engines are shown in
Tables 3.5-2a and 3.5-2b, respectively, along with estimates of the reductions from this final rule.
PM2 5 will be  reduced as a result of the  new PM emission standards and changes in the sulfur level
in nonroad diesel fuel. The exhaust emission standards begin in 2008 for engines less than  75 hp,
and are completely phased in for all hp categories by 2015.  Nonroad diesel fuel sulfur is reduced
to a 500 ppm standard in June of 2007, and further reduced to a 15 ppm standard (11 ppm in-use)
in June of 2010.  The 15 ppm standard  is fully phased in starting in 2011.

   Tables 3.5-2a and 3.5-2b present results for five-year increments from 2000 to 2030.
Individual years from 2007 to 2011 are also included, since fuel sulfur levels are changing during
this period. Emissions are projected to 2030 to reflect close to complete turnover of the fleet to
engines meeting the  new emission  standards. For comparison purposes, emission reductions are
also shown from reducing the diesel fuel sulfur level to 500 ppm in 2007 and to 15 ppm in 2010,
without any new emission standards.
                                          3-61

-------
                     Table 3.5-la
Total Emission Reductions (48-State) from the Final Rule
Year
2000
2005
2007
2008
2009
2010
2011
2012
2015
2020
2025
2030
PM25
0
0
10,511
19,031
19,943
21,692
25,154
31,103
53,072
85,808
110,043
128,350
NOX
0
0
0
30
70
149
17,830
46,827
193,431
442,061
613,629
734,184
SO2
0
0
134,388
236,976
241,719
256,447
268,989
278,092
297,513
323,378
349,312
375,354
voc
0
0
0
168
341
525
1,139
2,486
8,318
18,141
25,002
30,030
CO
0
0
0
3,104
6,266
9,634
28,704
64,599
198,947
388,037
521,457
619,167
Benzene
0
0
0
3
7
11
23
50
166
363
500
601
Formaldehyde
0
0
0
20
40
62
134
293
981
2,141
2,950
3,544
Acetaldehyde
0
0
0
9
18
28
60
132
441
961
1,325
1,592
1,3 -butadiene
0
0
0
0
1
1
2
5
17
36
50
60
Acrolein
0
0
0
1
1
2
3
7
25
54
75
90
                    Table3.5-lb
Total Emission Reductions (50-State) from the Final Rule
Year
2000
2005
2007
2008
2009
2010
2011
2012
2015
2020
2025
2030
PM25
0
0
10,403
18,908
19,821
21,627
25,142
31,122
53,238
86,157
110,508
128,899
NOX
0
0
0
31
72
153
17,951
47,129
194,615
444,714
617,176
738,307
SO2
0
0
132,998
235,366
240,084
255,525
268,613
277,804
297,440
323,302
349,233
375,269
VOC
0
0
0
169
343
529
1,146
2,501
8,367
18,251
25,152
30,210
CO
0
0
0
3,119
6,296
9,680
28,871
64,983
200,118
390,333
524,471
622,706
Benzene
0
0
0
3
7
11
23
50
167
365
503
604
Formaldehyde
0
0
0
20
41
62
135
295
987
2,154
2,968
3,565
Acetaldehyde
0
0
0
9
18
28
61
133
443
967
1,333
1,601
1,3 -butadiene
0
0
0
0
1
1
2
5
17
37
50
60
Acrolein
0
0
0
1
1
2
3
8
25
55
75
91

-------
                                                                 Emission Inventory
                                     Table3.5-2a
                           Estimated National (48-State) PM2 5
            Emissions and Reductions From Nonroad Land-Based Diesel Engines3
Year
2000
2005
2007
2008
2009
2010
2011
2015
2020
2025
2030
PM2 5 Emissions [short tons]
Without
Rule
170,774
144,259
135,791
133,245
131,044
128,730
127,035
125,936
129,058
135,369
142,484
With fuel
sulfur
reduced to
500 ppm in
2007;
No Tier 4
standards
170,774
144,259
128,558
120,434
117,938
115,273
113,243
110,950
112,595
117,428
123,076
With fuel
sulfur further
reduced to
1 5 ppm in
2010; No
Tier 4
standards
170,774
144,259
128,558
120,434
117,938
114,416
111,739
109,157
110,625
115,281
120,754
With Rule
(Fuel sulfur
reduced to 15
ppm in 2010;
Tier 4
standards)
170,774
144,259
128,558
119,936
116,896
112,775
107,612
79,326
50,065
32,503
21,698
PM2 5 Reductions [short tons]
With fuel
sulfur
reduced to
500 ppm in
2007;
No Tier 4
standards
0
0
7,232
12,811
13,106
13,458
13,792
14,986
16,463
17,941
19,408
With fuel
sulfur
further
reduced to
1 5 ppm in
2010; No
Tier 4
standards
0
0
7,232
12,811
13,106
14,315
15,296
16,779
18,433
20,087
21,730
With
Rule
0
0
7,232
13,309
14,148
15,955
19,423
46,610
78,993
102,866
120,786
' PM25 represents 97 percent of PM10 emissions.
                                         3-63

-------
Final Regulatory Impact Analysis
                                      Table3.5-2b
                           Estimated National (50-State) PM2 5
            Emissions and Reductions From Nonroad Land-Based Diesel Engines3
Year
2000
2005
2007
2008
2009
2010
2011
2015
2020
2025
2030
PM2 5 Emissions [short tons]
Without
Rule
171,575
144,951
136,362
133,807
131,598
129,276
127,576
126,482
129,628
135,972
143,126
With fuel
sulfur
reduced to
500 ppm in
2007;
No Tier 4
standards
171,575
144,951
129,165
121,030
118,526
115,846
113,797
111,511
113,181
118,049
123,737
With fuel
sulfur further
reduced to
1 5 ppm in
2010; No
Tier 4
standards
171,575
144,951
129,165
121,030
118,526
114,984
112,292
109,708
111,200
115,891
121,402
With Rule
(Fuel sulfur
reduced to 15
ppm in 2010;
Tier 4
standards)
171,575
144,951
129,165
120,529
117,479
113,336
108,147
79,706
50,285
32,641
21,790
PM25 Reductions [short tons]
With fuel
sulfur
reduced to
500 ppm in
2007;
No Tier 4
standards
0
0
7,197
12,777
13,071
13,430
13,778
14,971
16,447
17,923
19,389
With fuel
sulfur
further
reduced to
1 5 ppm in
2010; No
Tier 4
standards
0
0
7,197
12,777
13,071
14,292
15,283
16,774
18,428
20,081
21,724
With
Rule
0
0
7,197
13,277
14,118
15,940
19,428
46,777
79,343
103,331
121,336
' PM25 represents 97 percent of PM10 emissions.
   The benefits in the early years of the program (i.e., pre-2010) are primarily from reducing the
diesel fuel sulfur level to 500 ppm. As the standards phase in and fleet turnover occurs, PM2 5
emissions are impacted more significantly from the requirements of the final rule. PM25
emissions from land-based diesel engines are projected to decrease by roughly 120,000 tons by
2030 as a result of this rule.

   Figure 3.5-1 shows EPA's estimate of 50-state PM25 emissions from land-based diesel
engines for 2000 to 2030 with and without the new PM emission standards. We estimate that
PM2 5 emissions from this source would decrease by 85 percent in 2030.
                                          3-64

-------
                                                                  Emission Inventory
             Figure 3.5-1: Estimated Reductions in PM2.5 Emissions
                From Land-Based Nonroad Engines (tons/year)

             200,000  -
             180,000  -
             160,000  -
             140,000  -
             120,000  -
             100,000  -
              80,000  -
              60,000  -
              40,000  -
              20,000  -
                   0
                   2000  2005  2010  2015  2020  2025  2030
•Base 50-State
•Control 50-State
   Nonroad diesel engines used in locomotives, commercial marine vessels, and recreational
marine vessels are not affected by the emission standards in this rule.  PM2 5 emissions from these
engines will be reduced as a result of the lower fuel sulfur levels from a current in-use average of
about 2640 ppm to about 55 ppm by 2015. The estimated 48-state and 50-state reductions in
PM2 5 emissions from these engines based on the diesel fuel-sulfur requirements are given in
Tables 3.5-3a and 3.5-3b, respectively.  Total PM25 reductions reach roughly 7,500 tons in 2030
for these engine categories.

   Tables 3.5-4a and 3.5-4b present the PM25 emissions and reductions for all nonroad diesel
categories combined. The 50-state results are also presented graphically in Figure 3.5-2.  For all
nonroad diesel categories combined, the estimated reductions in PM2 5 emissions are 86,000 tons
in 2020, increasing to 128,000 tons in 2030. Simply reducing the fuel sulfur level to 500 ppm in
2007 will lead to projected PM25 reductions of 23,000 tons in 2020 and 26,000 tons in 2030.
Reducing the fuel  sulfur level further to 15 ppm (in 2010 for land-based diesel engines and in
2012 for marine engines and locomotives) in the absence of Tier 4 standards (i.e., a fuel only
program) will lead to projected PM25 reductions of 25,000 tons in 2020 and 29,000 tons in 2030.
                                          3-65

-------
Final Regulatory Impact Analysis
                                    Table3.5-3a
                     Estimated National (48-State) PM2 5 Reductions
       From Locomotives, Commercial Marine, and Recreational Marine Diesel Engines
Year
2000
2005
2007
2008
2009
2010
2011
2015
2020
2025
2030
PM2 5 Reductions with Rule [short tons]
Locomotives
0
0
1,975
3,442
3,488
3,458
3,460
3,885
4,063
4,240
4,426
Commerical
Marine Diesel
0
0
1,085
1,891
1,909
1,882
1,871
2,105
2,229
2,366
2,516
Recreational
Marine Diesel
0
0
220
389
398
397
400
473
522
572
622
Total PM2 5
Reductions
0
0
3,279
5,722
5,796
5,737
5,731
6,463
6,815
7,178
7,564
                                    Table3.5-3b
                     Estimated National (50-State) PM2 5 Reductions
       From Locomotives, Commercial Marine, and Recreational Marine Diesel Engines
Year
2000
2005
2007
2008
2009
2010
2011
2015
2020
2025
2030
PM2 5 Reductions with Rule [short tons]
Locomotives
0
0
1,899
3,331
3,376
3,372
3,393
3,820
3,995
4,168
4,350
Commerical
Marine Diesel
0
0
1,095
1,921
1,940
1,927
1,927
2,175
2,303
2,445
2,599
Recreational
Marine Diesel
0
0
212
378
387
390
394
467
516
565
614
Total PM2 5
Reductions
0
0
3,206
5,630
5,702
5,689
5,714
6,462
6,814
7,177
7,563
                                       3-66

-------
                                                           Emission Inventory
                                 Table3.5-4a
         Estimated National (48-State) PM2 5 Emissions and Reductions from
Land-Based Nonroad, Locomotive, Commercial Marine, and Recreational Marine Vessels
Year
2000
2005
2007
2008
2009
2010
2011
2012
2015
2020
2025
2030
PM25 Emissions [short tons]
Without
Rule
209,876
183,831
174,668
171,591
169,201
166,391
164,894
163,784
162,633
165,196
171,484
179,173
With fuel
sulfur
reduced to
500 ppm in
2007; No
Tier 4
standards
209,876
183,831
164,157
153,058
150,300
147,235
145,438
143,965
141,757
142,522
147,002
152,873
With fuel
sulfur further
reduced to
1 5 ppm in
2010/2012;
No Tier 4
standards
209,876
183,831
164,157
153,058
150,300
146,340
143,868
142,054
139,391
139,948
144,219
149,880
With Rule
(Fuel sulfur
further
reduced to 15
ppm in
2010/2012;
Tier 4
standards)
209,876
183,831
164,157
152,560
149,258
144,699
139,741
132,681
109,560
79,388
61,440
50,824
PM2 5 Reductions [short tons]
With fuel
sulfur
reduced to
500 ppm in
2007; No
Tier 4
standards
0
0
10,511
18,533
18,901
19,156
19,457
19,819
20,876
22,674
24,482
26,300
With fuel
sulfur
further
reduced to
1 5 ppm in
2010/2012;
No Tier 4
standards
0
0
10,511
18,533
18,901
20,051
21,027
21,730
23,241
25,248
27,265
29,293
With Rule
(Fuel sulfur
further
reduced to 15
ppm in
2010/2012;
Tier 4
standards)
0
0
10,511
19,031
19,943
21,692
25,154
31,103
53,072
85,808
110,043
128,350
                                    3-67

-------
Final Regulatory Impact Analysis
                                    Table3.5-4b
            Estimated National (50-State) PM2 5 Emissions and Reductions from
   Land-Based Nonroad, Locomotive, Commercial Marine, and Recreational Marine Vessels
Year
2000
2005
2007
2008
2009
2010
2011
2012
2015
2020
2025
2030
PM2 5 Emissions [short tons]
Without
Rule
211,688
185,555
176,250
173,154
170,750
167,923
166,416
165,298
164,133
166,719
173,075
180,851
With fuel
sulfur
reduced to
500 ppm in
2007; No
Tier 4
standards
211,688
185,555
165,847
154,747
151,976
148,844
146,990
145,510
143,289
144,080
148,630
154,591
With fuel
sulfur further
reduced to
15 ppm in
2010/2012;
No Tier 4
standards
211,688
185,555
165,847
154,747
151,976
147,944
145,419
143,591
140,897
141,477
145,816
151,565
With Rule
(Fuel sulfur
further
reduced to 1 5
ppm in
2010/2012;
Tier 4
standards)
211,688
185,555
165,847
154,247
150,929
146,296
141,274
134,176
110,894
80,562
62,567
51,953
PM2 5 Reductions [short tons]
With fuel
sulfur
reduced to
500 ppm in
2007; No
Tier 4
standards
0
0
10,403
18,407
18,774
19,079
19,426
19,788
20,843
22,639
24,445
26,260
With fuel
sulfur
further
reduced to
1 5 ppm in
2010/2012;
No Tier 4
standards
0
0
10,403
18,407
18,774
19,979
20,997
21,707
23,236
25,242
27,259
29,287
With Rule
(Fuel sulfur
further
reduced to 1 5
ppm in
2010/2012;
Tier 4
standards)
0
0
10,403
18,908
19,821
21,627
25,142
31,122
53,238
86,157
110,508
128,899
                                       3-68

-------
                                                                   Emission Inventory
                Figure 3.5-2: Estimated Reductions in
               Emissions From Land-Based Nonroad Engines,
                CMVs, RMVs, and Locomotives (tons/year)
             250,000 -,
             200,000 -
                                                           •Base 50-State
                                                           •Control 50-State
                  0
                   2000  2005  2010 2015 2020 2025 2030
3.5.2 NOX Reductions

   Tables 3.5-5a and 3.5-5b show the estimated 48-state and 50-state NOX emissions in five-year
increments from 2000 to 2030 with and without this rule.  The 50-state results are shown
graphically in Figure 3.5-3.  We estimate that NOX emissions from these engines will be reduced
by 62 percent in 2030.

   We note that the magnitude of NOx reductions determined in the final rule analysis is
somewhat less than what was reported in the proposal's draft RIA, especially in the later years
when the fleet has mostly turned over to Tier 4 designs. The greater part of this is  due to the fact
that we have deferred setting a long-term NOx standard for mobile machinery over 750 hp to a
later action. When this future  action is completed, we would expect roughly  equivalent
reductions between the proposal and the overall final program, though there are some other
effects reflected in the differing NOx reductions as well, due to updated modeling assumptions
and the adjusted NOx standards levels for engines over 750 hp. Preamble Section  II. A.4 contains
a detailed discussion of the NOx standards we are adopting for engines over 750 hp,  and the basis
for those standards.

   NOX emissions from locomotives, commercial marine diesel vessels, and  recreational marine
diesel vessels are not affected  by this rule.
                                          3-69

-------
Final Regulatory Impact Analysis
                                    Table3.5-5a
                      Estimated National (48-State) NOX Emissions
                 and Reductions From Nonroad Land-Based Diesel Engines
Year
2000
2005
2010
2015
2020
2030
NOX Emissions Without
Rule [short tons]
1,550,355
1,467,547
1,278,038
1,152,199
1,119,481
1,192,833
NOX Emissions With
Rule
1,550,355
1,467,547
1,277,888
958,769
677,420
458,649
NOX Reductions With
Rule
0
0
149
193,431
442,061
734,184
                                    Table3.5-5b
                      Estimated National (50-State) NOX Emissions
                 and Reductions From Nonroad Land-Based Diesel Engines
Year
2000
2005
2010
2015
2020
2030
NOX Emissions Without
Rule [short tons]
1,558,392
1,475,092
1,284,510
1,158,023
1,125,276
1,199,225
NOX Emissions With
Rule
1,558,392
1,475,092
1,284,357
963,408
680,563
460,918
NOX Reductions With
Rule
0
0
153
194,615
444,714
738,307
                                        3-70

-------
                                                                   Emission Inventory
                 Figure 3.5-3: Estimated Reductions in NOx Emissions
                    From Land-Based Nonroad Engines (tons/year)
              1,800,000 n
              1,600,000 -
              1,400,000 -
              1,200,000 -
              1,000,000 -
               800,000 -
               600,000 -
               400,000 -
               200,000 -
• Base 50-State
•Control 50-State
                     2000 2005 2010 2015 2020 2025  2030
3.5.3 SO2 Reductions

   As part of this final rule, sulfur levels in fuel will be significantly reduced, leading to large
reductions in nonroad diesel SO2 emissions. By 2007, the sulfur in diesel fuel used by all nonroad
diesel engines will be reduced to 500 ppm.  By 2010, the sulfur in diesel fuel used by nonroad
land-based engines will be further reduced to 15 ppm.  By 2012, the sulfur in diesel fuel used by
marine engines and locomotives will also be reduced to 15 ppm.

   48-State and 50-state emissions of SO2 from land-based nonroad diesel engines are shown in
Tables 3.5-6a and 3.5-6b, respectively, along with estimates of the emission reductions resulting
from this final rule.  Results are presented for five-year increments from 2000 to 2030. Individual
years from 2007 to 2011 are also included,  since fuel sulfur levels are changing during this
period.  SO2 will be reduced due to the changes in the sulfur level in nonroad diesel fuel. For
comparison purposes, emission reductions are also shown from reducing the diesel fuel sulfur
level to  500 ppm beginning in June of 2007, without any new emission standards or any
additional sulfur level reductions.
                                          3-71

-------
Final Regulatory Impact Analysis
                                    Table3.5-6a
                          Estimated National (48-State)
            Emissions and Reductions From Nonroad Land-
SO2
Based Diesel Engines
Year
2000
2005
2007
2008
2009
2010
2011
2015
2020
2025
2030
SO2 Emissions [short tons]
Without
Rule
161,977
185,287
189,511
194,019
198,526
197,829
198,415
215,699
237,044
258,360
279,511
With fuel sulfur
reduced to 500 ppm in
2007
161,977
185,287
97,142
30,359
31,064
25,835
22,119
24,045
26,425
28,801
31,159
With Rule
(Fuel sulfur reduced
to 15 ppm in 2010)
161,977
185,287
97,142
30,359
31,064
14,881
2,853
992
986
1,019
1,074
SO2 Reductions [short tons]
With fuel sulfur
reduced to 500 ppm in
2007
0
0
92,370
163,660
167,462
171,993
176,296
191,654
210,619
229,559
248,352
With
Rule
0
0
92,370
163,660
167,461
182,948
195,562
214,707
236,057
257,341
278,437
                                       3-72

-------
                                                                   Emission Inventory
                                       Table3.5-6b
                            Estimated National (50-State)
             Emissions and Reductions From Nonroad Land-
SO2
Based Diesel Engines
Year
2000
2005
2007
2008
2009
2010
2011
2015
2020
2025
2030
SO2 Emissions [short tons]
Without
Rule
162,920
186,368
189,505
194,013
198,521
197,795
198,360
215,641
236,982
258,294
279,442
With fuel sulfur
reduced to 500 ppm in
2007
162,920
186,368
97,580
30,786
31,501
26,159
22,238
24,175
26,568
28,957
31,328
With Rule
(Fuel sulfur reduced
to 15 ppm in 2010)
162,920
186,368
97,580
30,786
31,501
15,145
2,961
997
991
1,024
1,080
SO2 Reductions [short tons]
With fuel sulfur
reduced to 500 ppm in
2007
0
0
91,926
163,227
167,019
171,637
176,122
191,466
210,414
229,337
248,114
With
Rule
0
0
91,926
163,227
167,019
182,651
195,400
214,644
235,990
257,270
278,362
   The benefits in the early years of the program (i.e., pre-2010) are from reducing the diesel fuel
sulfur level to 500 ppm.  Reducing the diesel fuel sulfur level to 15 ppm in June of 2010
proportionately reduces SO2 further.  Total 50-state SO2 emissions are projected to decrease by
278,000 tons in 2030 as a result of this final rule. Note that SO2 emissions continue to increase
over time due to the growth in the nonroad sector.

   Nonroad diesel engines used in locomotives, commercial marine vessels, and recreational
marine vessels are also affected by the new fuel sulfur requirements. The estimated 48-state and
50-state reductions in SO2 emissions from these engines based on the new requirements for diesel
fuel are given in Tables 3.5-7a and 3.5-7b, respectively. Total 50-state SO2 reductions reach
96,000 tons in 2030 for these nonroad diesel engine categories.

   Tables 3.5-8a and 3.5-8b present the SO2 emissions and reductions for all nonroad diesel
categories combined. The 50-state results are also presented graphically in Figure 3.5-4.  For all
nonroad diesel categories combined, the estimated 50-state reductions in SO2 emissions resulting
from the final rule are 323,000 tons in 2020, increasing to 375,000 tons in 2030. Simply reducing
the fuel sulfur level to 500 ppm in 2007 will result in SO2 reductions of 289,000 tons in 2020 and
336,000 tons in 2030.
                                          3-73

-------
Final Regulatory Impact Analysis
                                   Table3.5-7a
                     Estimated National (48-State) SO2 Reductions
       From Locomotives, Commercial Marine, and Recreational Marine Diesel Engines
Year
2000
2005
2007
2008
2009
2010
2011
2015
2020
2025
2030
SO2 Reductions with Rule [short tons]
Locomotives
0
0
25,305
44,107
44,700
44,306
44,334
49,779
52,070
54,328
56,720
Commerical Marine
Diesel Vessels
0
0
13,899
24,238
24,465
24,108
23,980
26,977
28,564
30,319
32,234
Recreational
Marine Diesel
Vessels
0
0
2,814
4,972
5,093
5,085
5,112
6,049
6,686
7,324
7,963
Total SO2
Reductions
0
0
42,018
73,316
74,257
73,499
73,426
82,806
87,320
91,971
96,917
                                       3-74

-------
                                                       Emission Inventory
                             Table3.5-7b
               Estimated National (50-State) SO2 Reductions
From Locomotives, Commercial Marine, and Recreational Marine Diesel Engines
Year
2000
2005
2007
2008
2009
2010
2011
2015
2020
2025
2030
SO2 Reductions with Rule [short tons]
Locomotives
0
0
24,329
42,683
43,258
43,207
43,481
48,954
51,196
53,404
55,742
Commerical Marine
Diesel Vessels
0
0
14,025
24,621
24,855
24,685
24,695
27,867
29,511
31,325
33,302
Recreational
Marine Diesel
Vessels
0
0
2,718
4,834
4,952
4,983
5,037
5,975
6,604
7,235
7,863
Total SO2
Reductions
0
0
41,072
72,139
73,065
72,875
73,213
82,797
87,311
91,963
96,907
                                3-75

-------
Final Regulatory Impact Analysis
                                    Table3.5-8a
             Estimated National (48-State) SO2 Emissions and Reductions from
   Land-Based Nonroad, Locomotive, Commercial Marine, and Recreational Marine Vessels
Year
2000
2005
2007
2008
2009
2010
2011
2012
2015
2020
2025
2030
SO2 Emissions [short tons]
Without
Rule
244,599
269,288
275,416
280,983
286,606
281,867
280,031
285,277
300,552
326,514
352,585
378,793
With fuel sulfur
reduced to 500
ppm in 2007
244,599
269,288
141,029
44,007
44,887
36,860
31,152
31,735
33,434
36,322
39,218
42,128
With fuel sulfur
further reduced to 1 5
ppm in 20 10/20 12
244,599
269,288
141,029
44,007
44,888
25,420
11,041
7,185
3,039
3,136
3,273
3,439
SO2 Reductions [short tons]
With fuel sulfur
reduced to 500
ppm in 2007
0
0
134,388
236,976
241,719
245,007
248,879
253,542
267,118
290,192
313,367
336,665
With fuel sulfur
further reduced to 15
ppm in 20 10/20 12
0
0
134,388
236,976
241,719
256,447
268,989
278,092
297,513
323,378
349,312
375,354
                                       3-76

-------
                                                           Emission Inventory
                                 Table3.5-8b
          Estimated National (50-State) SO2 Emissions and Reductions from
Land-Based Nonroad, Locomotive, Commercial Marine, and Recreational Marine Vessels
Year
2000
2005
2007
2008
2009
2010
2011
2012
2015
2020
2025
2030
SO2 Emissions [short tons]
Without
Rule
247,010
271,841
275,397
280,964
286,588
281,828
279,976
285,221
300,494
326,450
352,519
378,722
With fuel sulfur
reduced to 500
ppm in 2007
247,010
271,841
142,399
45,598
46,503
37,802
31,486
32,075
33,788
36,701
39,625
42,565
With fuel sulfur
further reduced to 1 5
ppm in 20 10/20 12
247,010
271,841
142,399
45,598
46,503
26,303
11,363
7,418
3,054
3,149
3,286
3,453
SO2 Reductions [short tons]
With fuel sulfur
reduced to 500
ppm in 2007
0
0
132,998
235,366
240,085
244,026
248,490
253,147
266,706
289,749
312,894
336,157
With fuel sulfur
further reduced to 1 5
ppm in 20 10/20 12
0
0
132,998
235,366
240,084
255,525
268,613
277,804
297,440
323,302
349,233
375,269
                                    3-77

-------
Final Regulatory Impact Analysis
                Figure 3.5-4: Estimated Reductions in SO2 Benefits
                From Reducing Fuel Sulfur for Land-Based Nonroad
               Engines, CMVs, RMVs, and Locomotives (tons/year)
                                                             • Base 50-State
                                                             • Control 50-State
                      2000 2005  2010  2015 2020  2025 2030
3.5.4 VOC and Air Toxics Reductions

   Tables 3.5-9a and 3.5-9b show our projection of the 48-state and 50-state reductions in VOC
emissions expected from implementing the new NMHC emission standards.

   Although this final rule does not include specific standards for air toxics, these pollutants
decrease as manufacturers take steps to meet the NMHC emission standards. Tables 3.5-10a and
3.5-10b show our estimate of reduced emissions of benzene, formaldehyde, acetaldehyde,
1,3-butadiene, and acrolein.  We base these numbers on the assumption that air toxic emissions
are a constant fraction of hydrocarbon exhaust emissions.
                                         3-78

-------
                                                   Emission Inventory
                        Table3.5-9a
VOC Reductions (48-State) from Land-Based Nonroad Diesel Engines
Calendar Year
2000
2005
2010
2015
2020
2025
2030
VOC
Without Rule
[short tons]
199,887
163,663
129,711
107,084
97,513
94,975
96,374
VOC
With Rule
[short tons]
199,887
163,663
129,186
98,766
79,372
69,973
66,344
VOC Reductions
With Rule
[short tons]
0
0
525
8,318
18,141
25,002
30,030
                        Table3.5-9b
VOC Reductions (50-State) from Land-Based Nonroad Diesel Engines
Calendar Year
2000
2005
2010
2015
2020
2025
2030
VOC
Without Rule
[short tons]
200,903
164,505
130,388
107,647
98,037
95,490
96,900
VOC
With Rule
[short tons]
200,903
164,505
129,859
99,281
79,786
70,338
66,690
VOC Reductions
With Rule
[short tons]
0
0
529
8,367
18,251
25,152
30,210
                            3-79

-------
Final Regulatory Impact Analysis
                                   Table3.5-10a
                       Air Toxic Reductions (48-State) (tons/year)
Year
2000


2005


2007


2010


2015


2020


2025


2030



Base
Control
Reduction
Base
Control
Reduction
Base
Control
Reduction
Base
Control
Reduction
Base
Control
Reduction
Base
Control
Reduction
Base
Control
Reduction
Base
Control
Reduction
Benzene
3,998
3,998
0
3,273
3,273
0
3,007
3,007
0
2,594
2,584
11
2,142
1,975
166
1,950
1,587
363
1,900
1,399
500
1,927
1,327
601
Formaldehyde
23,587
23,587
0
19,312
19,312
0
17,742
17,742
0
15,306
15,244
62
12,636
11,654
981
11,507
9,366
2,141
11,207
8,257
2,950
11,372
7,829
3,544
Acetaldehyde
10,594
10,594
0
8,674
8,674
0
7,969
7,969
0
6,875
6,847
28
5,675
5,235
441
5,168
4,207
961
5,034
3,709
1,325
5,108
3,516
1,592
1,3 -butadiene
400
400
0
327
327
0
301
301
0
259
258
1
214
198
17
195
159
36
190
140
50
193
133
60
Acrolein
600
600
0
491
491
0
451
451
0
389
388
2
321
296
25
293
238
54
285
210
75
289
199
90
                                       3-80

-------
                                                                   Emission Inventory
                                      Table3.5-10b
                        Air Toxic Reductions (50-State) (tons/year)
Year
2000


2005


2007


2010


2015


2020


2025


2030



Base
Control
Reduction
Base
Control
Reduction
Base
Control
Reduction
Base
Control
Reduction
Base
Control
Reduction
Base
Control
Reduction
Base
Control
Reduction
Base
Control
Reduction
Benzene
4,018
4,018
0
3,290
3,290
0
3,023
3,023
0
2,608
2,597
11
2,153
1,986
167
1,961
1,596
365
1,910
1,407
503
1,938
1,334
604
Formaldehyde
23,707
23,707
0
19,412
19,412
0
17,834
17,834
0
15,386
15,323
62
12,702
11,715
987
11,568
9,415
2,154
11,268
8,300
2,968
11,434
7,869
3,565
Acetaldehyde
10,648
10,648
0
8,719
8,719
0
8,010
8,010
0
6,911
6,883
28
5,705
5,262
443
5,196
4,229
967
5,061
3,728
1,333
5,136
3,535
1,601
1,3 -butadiene
402
402
0
329
329
0
302
302
0
261
260
1
215
199
17
196
160
37
191
141
50
194
133
60
Acrolein
603
603
0
494
494
0
453
453
0
391
390
2
323
298
25
294
239
55
286
211
75
291
200
91
3.5.5 CO Reductions

   Tables 3.5-1 la and 3.5-1 Ib show the estimated 48-state and 50-state emissions of CO from
land-based diesel engines in five-year increments from 2000 to 2030 with and without the final
rule. Although there are no Tier 4 CO standards, CO is estimated to decrease by 90 percent with
the advent of trap-equipped engines (corresponding to the start of 0.02 or 0.01 g/hp-hr PM
standards). We estimate that 50-state CO emissions from these engines will decrease by 623,000
tons in 2030.

   CO emissions from locomotives, commercial marine  diesel vessels, and recreational marine
diesel vessels are not affected by this rule.
                                          3-81

-------
Final Regulatory Impact Analysis
                                     Table 3.5-1 la
                            Estimated National (48-State) CO
            Emissions and Reductions From Nonroad Land-Based Diesel Engines
Year
2000
2005
2010
2015
2020
2030
CO Emissions
Without Rule
[short tons]
916,507
763,062
687,234
674,296
697,630
786,181
CO Emissions
With Rule
[short tons]
916,507
763,062
677,599
475,349
309,593
167,014
CO Reductions
With Rule
[short tons]
0
0
9,634
198,947
388,037
619,167
                                     Table 3.5-1 lb
                            Estimated National (50-State) CO
            Emissions and Reductions From Nonroad Land-Based Diesel Engines
Year
2000
2005
2010
2015
2020
2030
CO Emissions
Without Rule
[short tons]
921,226
766,944
690,829
677,918
701,445
790,547
CO Emissions
With Rule
[short tons]
921,226
766,944
681,150
477,800
311,112
167,841
CO Reductions
With Rule
[short tons]
0
0
9,680
200,118
390,333
622,706
3.5.6 PM2 5 and SO2 Reductions from the 15 ppm Locomotive and Marine (LM) Fuel
Program

   Tables 3.5-12a and 3.5-12b provide the 48-state and 50-state PM25 and SO2 emissions and
reductions from reducing locomotive and marine fuel sulfur from 500 ppm to 15 ppm in 2012.
This is referred to as the 15 ppm LM fuel program. The reductions are shown relative to the full
engine and fuel program for land-based diesel engines, and locomotive and marine fuel sulfur
control to 500 ppm starting in 2007.  To model the reductions for this program, the in-use fuel
sulfur levels in Chapter 7 were used. The 15 ppm LM fuel program provides additional PM25
reductions of approximately 400 tons by 2030, and additional SO2 reductions of approximately
5,300 tons by 2030.
                                         3-82

-------
                                                  Emission Inventory
                        Table3.5-12a
Estimated National (48-State) PM2 5 and SO2 Emissions and Reductions
     from a 15 ppm Locomotive and Marine (LM) Fuel Program
Year
2000
2005
2010
2012
2015
2020
2030
Emissions (short tons)
Land-based full engine and fuel
program; LM fuel sulfur reduced
to 500 ppm in 2007
PM25
209,876
183,831
144,667
133,144
110,027
79,870
51,296
SO2
244,599
269,288
24,864
11,639
8,285
8,517
8,925
Land-based full engine and fuel
program; LM fuel sulfur further
reduced to 15 ppm in 2012
PM25
209,876
183,831
144,667
132,755
109,613
79,450
50,882
SO2
244,599
269,288
24,864
7,269
2,977
3,139
3,621
Reductions (short tons)
LM fuel sulfur reduced from
500 ppm to 1 5 ppm in 20 1 2
PM25
0
0
0
389
414
420
414
S02
0
0
0
4,370
5,308
5,378
5,304
                        Table3.5-12b
Estimated National (50-State) PM2 5 and SO2 Emissions and Reductions
     from a 15 ppm Locomotive and Marine (LM) Fuel Program
Year
2000
2005
2010
2012
2015
2020
2030
Emissions (short tons)
Land-based full engine and fuel
program; LM fuel sulfur reduced
to 500 ppm in 2007
PM25
211,688
185,555
146,152
134,509
111,240
80,915
52,279
SO2
247,010
271,841
25,793
11,871
8,308
8,537
8,935
Land-based full engine and fuel
program; LM fuel sulfur further
reduced to 15 ppm in 2012
PM25
211,688
185,555
146,152
134,137
110,825
80,495
51,866
SO2
247,010
271,841
25,793
7,567
2,989
3,153
3,640
Reductions (short tons)
LM fuel sulfur reduced from
500 ppm to 1 5 ppm in 20 1 2
PM25
0
0
0
372
415
420
413
SO2
0
0
0
4,305
5,319
5,385
5,294
                            3-83

-------
Final Regulatory Impact Analysis
3.5.7 SO2 and Sulfate PM Reductions from Other Nonhighway Fuel

   The fuel sulfur requirements in this rule are also expected to indirectly affect diesel fuel for
other nonhighway end uses. This includes any application other than land-based nonroad
engines, locomotives, or marine vessels. Tables 3.5-13a and 3.5-13b provide the 48-state and 50-
state estimates of fuel volumes, fuel sulfur levels, and SO2 emissions and reductions for diesel
fuel for other nonhighway end uses. Tables 3.5-14a and 3.5-14b provide similar information for
sulfate PM emissions and reductions. Details regarding the estimated volumes and fuel sulfur
levels can be found in Chapter 7.

   The tables show the incremental reductions from controlling fuel sulfur: 1) to 500 ppm in
2007 for land-based, locomotive, and marine use (the  500 ppm NRLM fuel program), 2) further
control to 15 ppm in 2010 for land-based use only, and 3) further control to 15 ppm in 2010 for
locomotive and marine use (the 15 ppm LM fuel program).

   SO2 emissions are calculated similarly to the commercial marine and locomotive categories,
as described in Section 3.1.3. We estimate that 99 percent of the sulfur in other nonhighway fuel
is emitted in the form of SO2 and 1 percent in the form of sulfate PM.13

   For the incremental step of reducing LM fuel sulfur from 500 ppm to 15 ppm, heating oil
related benefits dominate those related to the LM fuel itself.  This occurs because the final rule
prohibits the use of downgraded distillate in NRLM fuel starting in mid-2010 in the
Northeast/Mid-Atlantic area, while this fuel would be able to be used in LM fuel in this area
under a 500 ppm cap.  When this downgraded distillate cannot be used in LM fuel, it will shift to
the heating oil market. The downgrade contains between 31 (highway-based) and 435 ppm (jet-
based) sulfur, well below that of heating oil. Thus, the sulfur content of heating oil decreases
significantly in the Northeast/Mid-Atlantic area with a 15 ppm cap on LM fuel.

   Chapter 8 provides details regarding the estimated number  of gallons of downgrade shifted to
the heating oil market and the corresponding sulfur content of this downgrade.  The resulting SO2
and sulfate PM emission reductions for the 15 ppm LM program given in Chapter 8 are
reproduced here. The 48-state and 50-state reductions for the 15 ppm LM program are the same,
since the benefits only occur in the Northeast/Mid-Atlantic area, which does not include Alaska
or Hawaii.

   Total SO2 reductions in 2030 for other nonhighway uses are estimated to be 19,000 tons with
the full fuel program. Of that, approximately 6,300 tons are due to the 500 ppm NRLM fuel
program and 12,000 tons are due to the 15 ppm LM fuel program.  Total  sulfate PM reductions in
2030 are  estimated to be 670 tons with the full fuel program. Of that, approximately 220 tons are
due to the 500 ppm NRLM fuel program and 420 tons are due to the 15 ppm LM fuel program.
These reductions are not included in Tables 3.5-la and 3.5-lb.
                                          3-84

-------
                                                              Table
                      Estimated National (48-State) SO7 Emissions
3.5-13a
and Reductions from Other Nonhighway Fuel
Year
2000
2005
2007
2008
2009
2010
2011
2012
2015
2020
2025
2030
Volume
(106 gals)
10,471
10,174
10,058
10,000
9,943
9,886
9,829
9,772
9,605
9,333
9,068
8,811
Sulfur (ppm)
Base
2,871
2,871
2,858
2,858
2,858
2,724
2,628
2,628
2,628
2,628
2,628
2,628
500 ppm
NRLM
Fuel
Program
(Control
to 500
ppm in
2007)
2,871
2,871
2,671
2,534
2,534
2,530
2,527
2,527
2,527
2,527
2,527
2,527
500 ppm
NRLM
Fuel
Program
andNR
only to 1 5
ppm in
2010
2,871
2,871
2,671
2,534
2,534
2,530
2,527
2,527
2,515
2,515
2,515
2,515
Full Fuel
Program
(MR
Control to
1 5 ppm in
2010; LM
in 20 12)
2,871
2,871
2,671
2,534
2,534
2,530
2,527
--
--
--
--
--
SO2 Emissions (tons/year)
Base
211,286
205,291
202,026
200,866
199,713
189,258
181,561
180,519
177,429
172,394
167,503
162,751
500 ppm
NRLM
Fuel
Program
(Control
to 500
ppm in
2007)
211,286
205,291
188,820
178,086
177,064
175,775
174,572
173,570
170,599
165,758
161,055
156,486
500 ppm
NRLM
Fuel
Program
andNR
only to 1 5
ppm in
2010
211,286
205,291
188,820
178,086
177,064
175,773
174,568
173,566
169,830
165,012
160,330
155,781
Full Fuel
Program
(MR
Control to
1 5 ppm in
2010; LM
in 20 12)
211,286
205,291
188,820
178,086
177,064
175,773
174,568
168,683
160,886
155,190
149,494
143,852
Incremental SO2 Reductions (tons/year)
500 ppm
NRLM
Fuel
Program
0
0
13,206
22,780
22,649
13,483
6,989
6,949
6,830
6,636
6,448
6,265
500 ppm
NRLM
Fuel
Program
andNR
only to 1 5
ppm in
2010
0
0
0
0
0
2
4
4
768
747
725
705
15 ppm
LM Fuel
Program
(LMtolS
ppm in
2012)
0
0
0
0
0
0
0
4,884
8,944
9,822
10,836
11,929
Full Fuel
Program
0
0
13,206
22,780
22,649
13,486
6,993
11,837
16,542
17,204
18,009
18,899
a NRLM refers to land-based diesel engines, locomotives, and recreational and commercial marine vessels.
NR refers to land-based diesel nonroad engines.
LM refers to locomotives, recreational and commercial marine vessels.

-------
                                                              Table
                      Estimated National (50-State) SO7 Emissions
3.5-13b
and Reductions from Other Nonhighway Fuel
Year
2000
2005
2007
2008
2009
2010
2011
2012
2015
2020
2025
2030
Volume
(106 gals)
10,819
10,512
10,392
10,332
10,273
10,214
10,155
10,097
9,924
9,643
9,369
9,103
Sulfur (ppm)
Base
2,859
2,859
2,846
2,846
2,846
2,717
2,624
2,624
2,624
2,624
2,624
2,624
500 ppm
NRLM
Fuel
Program
(Control
to 500
ppm in
2007)
2,859
2,859
2,666
2,533
2,533
2,529
2,526
2,526
2,526
2,526
2,526
2,526
500 ppm
NRLM
Fuel
Program
andNR
only to 1 5
ppm in
2010
2,859
2,859
2,666
2,533
2,533
2,529
2,526
2,526
2,515
2,515
2,515
2,515
Full Fuel
Program
(MR
Control to
1 5 ppm in
2010; LM
in 20 12)
2,859
2,859
2,666
2,533
2,533
2,529
2,526
--
--
--
--
--
SO2 Emissions (tons/year)
Base
217,431
211,262
207,911
206,717
205,531
195,041
187,310
186,235
183,047
177,853
172,807
167,904
500 ppm
NRLM
Fuel
Program
(Control
to 500
ppm in
2007)
217,431
211,262
194,712
183,944
182,889
181,561
180,321
179,286
176,217
171,217
166,359
161,639
500 ppm
NRLM
Fuel
Program
andNR
only to 1 5
ppm in
2010
217,431
211,262
194,712
183,944
182,889
181,559
180,317
179,282
175,448
170,471
165,634
160,934
Full Fuel
Program
(MR
Control to
1 5 ppm in
2010;LM
in 20 12)
217,431
211,262
194,712
183,944
182,889
181,559
180,317
174,399
166,504
160,649
154,798
149,006
Incremental SO2 Reductions (tons/year)
500 ppm
NRLM
Fuel
Program
0
0
13,199
22,773
22,642
13,481
6,989
6,949
6,830
6,636
6,448
6,265
500 ppm
NRLM
Fuel
Program
andNR
only to 1 5
ppm in
2010
0
0
0
0
0
2
4
4
768
747
725
705
15 ppm
LM Fuel
Program
(LMtolS
ppm in
2012)
0
0
0
0
0
0
0
4,884
8,944
9,822
10,836
11,929
Full Fuel
Program
0
0
13,199
22,773
22,642
13,483
6,993
11,837
16,542
17,204
18,009
18,899
a NRLM refers to land-based diesel engines, locomotives, and recreational and commercial marine vessels.
NR refers to land-based diesel nonroad engines.
LM refers to locomotives, recreational and commercial marine vessels.

-------
                                                              Table3.5-14a
                    Estimated National (48-State) Sulfate Emissions and Reductions from Other Nonhighway Fuel
Year
2000
2005
2007
2008
2009
2010
2011
2012
2015
2020
2025
2030
Volume
(106 gals)
10,471
10,174
10,058
10,000
9,943
9,886
9,829
9,772
9,605
9,333
9,068
8,811
Sulfur (ppm)
Base
2,871
2,871
2,858
2,858
2,858
2,724
2,628
2,628
2,628
2,628
2,628
2,628
500 ppm
NRLM
Fuel
Program
(Control
to 500
ppm in
2007)
2,871
2,871
2,671
2,534
2,534
2,530
2,527
2,527
2,527
2,527
2,527
2,527
500 ppm
NRLM
Fuel
Program
andNR
only to 1 5
ppm in
2010
2,871
2,871
2,671
2,534
2,534
2,530
2,527
2,527
2,515
2,515
2,515
2,515
Full Fuel
Program
(MR
Control to
1 5 ppm in
2010; LM
in 20 12)
2,871
2,871
2,671
2,534
2,534
2,530
2,527
--
--
--
--
--
Sulfate Emissions (tons/year)
Base
7,470
7,258
7,142
7,101
7,061
6,691
6,419
6,382
6,273
6,095
5,922
5,754
500 ppm
NRLM
Fuel
Program
(Control
to 500
ppm in
2007)
7,470
7,258
6,675
6,296
6,260
6,214
6,172
6,136
6,031
5,860
5,694
5,532
500 ppm
NRLM
Fuel
Program
andNR
only to 1 5
ppm in
2010
7,470
7,258
6,675
6,296
6,260
6,214
6,172
6,136
6,004
5,834
5,668
5,507
Full Fuel
Program
(MR
Control to
1 5 ppm in
2010;LM
in 20 12)
7,470
7,258
6,675
6,296
6,260
6,214
6,172
5,964
5,688
5,487
5,285
5,086
Incremental Sulfate Reductions (tons/year)
500 ppm
NRLM
Fuel
Program
0
0
467
805
801
477
247
246
241
235
228
221
500 ppm
NRLM
Fuel
Program
andNR
only to 1 5
ppm in
2010
0
0
0
0
0
0
0
0
27
26
26
25
15 ppm
LM Fuel
Program
(LMtolS
ppm in
2012)
0
0
0
0
0
0
0
173
316
347
383
422
Full Fuel
Program
0
0
467
805
801
477
247
418
585
608
637
668
a NRLM refers to land-based diesel engines, locomotives, and recreational and commercial marine vessels.
NR refers to land-based diesel nonroad engines.
LM refers to locomotives, recreational and commercial marine vessels.

-------
                                                              Table3.5-14b
                    Estimated National (50-State) Sulfate Emissions and Reductions from Other Nonhighway Fuel
Year
2000
2005
2007
2008
2009
2010
2011
2012
2015
2020
2025
2030
Volume
(106 gals)
10,819
10,512
10,392
10,332
10,273
10,214
10,155
10,097
9,924
9,643
9,369
9,103
Sulfur (ppm)
Base
2,859
2,859
2,846
2,846
2,846
2,717
2,624
2,624
2,624
2,624
2,624
2,624
500 ppm
NRLM
Fuel
Program
(Control
to 500
ppm in
2007)
2,859
2,859
2,666
2,533
2,533
2,529
2,526
2,526
2,526
2,526
2,526
2,526
500 ppm
NRLM
Fuel
Program
andNR
only to 1 5
ppm in
2010
2,859
2,859
2,666
2,533
2,533
2,529
2,526
2,526
2,515
2,515
2,515
2,515
Full Fuel
Program
(MR
Control to
1 5 ppm in
2010; LM
in 20 12)
2,859
2,859
2,666
2,533
2,533
2,529
2,526
--
--
--
--
--
Sulfate Emissions (tons/year)
Base
7,687
7,469
7,350
7,308
7,266
6,895
6,622
6,584
6,471
6,288
6,109
5,936
500 ppm
NRLM
Fuel
Program
(Control
to 500
ppm in
2007)
7,687
7,469
6,884
6,503
6,466
6,419
6,375
6,338
6,230
6,053
5,881
5,715
500 ppm
NRLM
Fuel
Program
andNR
only to 1 5
ppm in
2010
7,687
7,469
6,884
6,503
6,466
6,419
6,375
6,338
6,203
6,027
5,856
5,690
Full Fuel
Program
(MR
Control to
1 5 ppm in
2010;LM
in 20 12)
7,687
7,469
6,884
6,503
6,466
6,419
6,375
6,166
5,887
5,680
5,473
5,268
Incremental Sulfate Reductions (tons/year)
500 ppm
NRLM
Fuel
Program
0
0
467
805
800
477
247
246
241
235
228
221
500 ppm
NRLM
Fuel
Program
andNR
only to 1 5
ppm in
2010
0
0
0
0
0
0
0
0
27
26
26
25
15 ppm
LM Fuel
Program
(LMtolS
ppm in
2012)
0
0
0
0
0
0
0
173
316
347
383
422
Full Fuel
Program
0
0
467
805
800
477
247
418
585
608
637
668
a NRLM refers to land-based diesel engines, locomotives, and recreational and commercial marine vessels.
NR refers to land-based diesel nonroad engines.
LM refers to locomotives, recreational and commercial marine vessels.

-------
                                                                   Emission Inventory
3.6 Emission Inventories Used for Air Quality Modeling

   The emission inputs for the air quality modeling are required early in the analytical process to
conduct the air quality modeling and present the results.  The air quality modeling was based on a
preliminary control scenario.  Since the preliminary control scenario was developed, we have
gathered more information regarding the technical feasibility of the standards (see Section III of
the preamble for the final rule and Chapter 4 of the Final RIA). As a result, we have revised the
Tier 4 emission standards for land-based diesel  engines.  We have also made changes to the fuel
provisions of the rule for locomotives and diesel marine vessels. This section describes the
changes in the inputs and resulting emission inventories between the preliminary baseline and
control scenarios used for the  air quality modeling and the updated baseline and control scenarios
in this final rule. This section will focus on the four nonroad diesel categories that are affected by
the new emission standards and/or the fuel sulfur requirements: land-based diesel engines,
recreational marine diesel engines, commercial  marine diesel engines, and locomotives.

   The methodology used to develop the emission inventories for the air quality modeling is first
briefly described, followed by comparisons of the preliminary and final baseline  and control
inventories.

3.6.1 Methodology for Emission Inventory Preparation

   Air quality modeling was performed for calendar years 1996, 2020, and 2030. For these
years, county-level emission estimates were developed by Pechan under contract to EPA. These
inventories account for county-level differences in fuel characteristics and temperature. The
NONROAD model was used to generate the county-level emission estimates for all nonroad
sources, with the exception of commercial marine engines, locomotives, and aircraft.  The
methodology has been documented in detail.10

   For the nonroad diesel categories affected by the final rule, the only fuel characteristic that
affects emissions is the fuel sulfur level. The specific pollutants affected by fuel sulfur level are
PM and SO2. To develop the county-level emission estimates for each baseline and control
inventory, one diesel fuel sulfur level was used  to characterize all counties outside California. A
separate diesel fuel sulfur level was used to characterize all counties within California. Diesel
emissions as modeled are not affected by ambient temperature.

3.6.2 Baseline Inventories

   Table 3.6-1  presents the preliminary 48-state baseline inventories used for the air quality
modeling. These are an aggregation of the county-level results. Results expressed as short tons
are presented for 1996, 2020, and 2030 for the land-based diesel, recreational marine diesel,
commercial marine diesel, and locomotive categories.  The pollutants include PM2 5, NOX, SO2,
VOC, and CO.  VOC includes both exhaust and crankcase emissions.
                                           3-89

-------
Final Regulatory Impact Analysis
                                        Table 3.6-1
                            Modeled 48-State Baseline Emissions
                     Preliminary Baseline Used for Air Quality Modeling
Applications
Land-Based Diesel
Engines
Recreational Marine
Diesel Engines
Commercial Marine
Diesel Engines a
Locomotives
Year
1996
2020
2030
1996
2020
2030
1996
2020
2030
1996
2020
2030
NOX
[short tons]
1,583,641
1,144,686
1,231,981
19,438
34,814
41,246
960,153
819,544
815,162
921,556
612,722
534,520
PM25
[short tons]
178,500
127,755
143,185
511
876
1,021
37,203
42,054
46,185
22,396
17,683
16,988
S02
[short tons]
172,175
308,075
360,933
2,535
4,562
5,418
37,252
43,028
48,308
57,979
62,843
70,436
VOC
[short tons]
221,398
97,113
97,345
803
1,327
1,528
31,613
37,362
41,433
48,381
36,546
31,644
CO
[short tons]
1,010,501
702,145
793,899
3,215
5,537
6,464
126,523
160,061
176,708
112,171
119,302
119,302
a Includes emissions from vessels using both diesel and residual fuel, with the exception of SO2. For the pollutants other
than SO2, it was not possible to separate emissions from diesel-fueled and residual-fueled vessels.
   For the final baseline inventories, we have made minor changes to the diesel fuel sulfur levels.
The diesel fuel sulfur inputs used for the preliminary and final baseline inventories are provided
in Table 3.6-2. The diesel fuel sulfur level for land-based diesel engines is now reduced from
2500ppm to roughly 2200ppm, beginning in 2006. Both the preliminary and final sulfur levels
account for spillover of highway fuel, but the preliminary  sulfur levels did not properly account
for the 15ppm highway fuel sulfur content control phase-in beginning in 2006. The diesel fuel
sulfur levels for marine engines and locomotives are now higher prior to 2009 and lower
beginning in 2010.
                                            3-90

-------
                                                                   Emission Inventory
                                       Table 3.6-2
                    Modeled Baseline In-Use Diesel Fuel Sulfur Content
            Final Baseline vs. Preliminary Baseline Used for Air Quality Modeling
Applications
Land-Based Diesel Engines
Commercial and Recreational Marine
Engines and Locomotives
Final Baseline
Fuel Sulfur
ppm
2283
2249
2224
2167
2126
2637-2641
2616
2599
2444
2334-2350
Calendar Year
through 2005
2006
2007-2009
2010
2011+
through 2005
2006
2007-2009
2010
2011
Preliminary Baseline
Fuel Sulfur
ppm
2500"
2500a
Calendar Year
all years
all years
1 2500ppm is the 48-state average diesel fuel sulfur level, based on 2700ppm in 47 states and 120ppm in California.
   For the nonroad land-based diesel category, the preliminary inventories were generated with
the draft NONROAD2002 model. For the final inventory, the draft NONROAD2004 model was
used. The changes from draft NONROAD2002 to draft NONROAD2004 are described in
Section 3.1.1.8. The net difference in land-based diesel emissions with the two model versions is
generally within 3 percent, with the direction and variation of the change dependent on the
calendar year and pollutant of interest. Apart from the model changes, the lower fuel sulfur levels
will serve to reduce the PM and SO2 baseline inventories in 2020 and 2030.  Table 3.6-3
compares the preliminary and final 48-state baseline scenario inventories for land-based diesel
engines, as well as  recreational marine diesel engines, commercial marine diesel  engines,  and
locomotives.

   For recreational marine diesel engines, the  preliminary inventories were generated with the
draft NONROAD2002 model.  For the final inventory, the draft NONROAD2004 model was
used. The changes from draft NONROAD2002 to draft NONROAD2004 are more substantial for
this category. The  recreational marine populations, median life, and deterioration factors  for HC
and NOX were revised to match what was used in the 2002 final rulemaking that covers large
spark ignition engines (>25 hp), recreational equipment, and recreational marine diesel engines
(>50 hp).  The exhaust emission factors for HC, NOX, and PM were also revised in draft
NONROAD2004 to reflect the  final standards.
                                           3-91

-------
Final Regulatory Impact Analysis
   For locomotives, there have been reductions to the fuel volume estimates used to calculate
emissions for this category.  For the preliminary inventory development, railroad distillate values
were taken from the EIA Fuel and Kerosene Supply 2000 report. Fuel consumption specific to
locomotives was calculated by subtracting the rail maintenance fuel consumption as generated by
the draft NONROAD2002 model from the EIA railroad distillate estimates.

   For the final inventory, the EIA railroad distillate estimates were taken from the EIA Fuel and
Kerosene Supply 2001 report. The estimates were first adjusted to estimate the fraction of
distillate that is diesel fuel. The diesel fraction used was 0.95 for railroad distillate.  Fuel
consumption estimates from rail maintenance were  then subtracted. The estimate of rail
maintenance fuel consumption was also revised by  assuming these engines consume one percent
of the total railroad diesel fuel estimate, rather than using the estimate derived from draft
NONROAD2002. The revised estimate of rail maintenance fuel consumption is roughly half of
the NONROAD-derived estimate; however, the rail maintenance portion of the total railroad
diesel fuel consumption is small, so this change alone does not significantly affect the resulting
locomotive estimate.  The derivation of diesel fractions and the revised estimate of rail
maintenance fuel consumption is documented in Chapter 7.

   There have also been reductions to the fuel volumes assigned to commercial marine vessels.
For the preliminary inventory development, vessel bunkering distillate values were taken from the
EIA Fuel and Kerosene Supply 2000 report.  Fuel consumption specific to commercial marine
vessels was calculated by subtracting the recreational marine fuel consumption as generated by
the draft NONROAD2002 model from the EIA vessel bunkering estimates.

   For the final inventory, the EIA vessel bunkering distillate estimates were taken from the EIA
Fuel and Kerosene Supply 2001 report. The vessel  bunkering distillate estimates were first
adjusted to estimate the fraction of distillate that is diesel fuel. The diesel fraction used was 0.90
for vessel bunkering distillate. Fuel consumption estimates from recreational marine engines
were then subtracted. The estimate of recreational marine fuel consumption was that generated
by the draft NONROAD2004 model.  These revised fuel volumes were used to generate SO2 and
sulfate PM estimates for commercial marine diesel  engines in the final inventory. Emission
estimates for other pollutants emitted by commercial marine vessels were also revised in the final
inventory to reflect the January 2003 final rule for Category 3 commercial marine residual
engines.

   As a result, differences in total emissions between the final and preliminary baseline scenarios
are generally within 10 percent.  Exceptions include PM25 and SO2. Total PM25 emissions are
higher with the final baseline scenario, in part due to the upward revision of the PM2 5 fraction of
total PM.from 92 to 97 percent.  Total SO2 emissions are lower, due to reductions in fuel volumes
for some categories and reductions in fuel sulfur levels.
                                          3-92

-------
                                                              Table 3.6-3
                                    Modeled 48-State Emission Impact Due to Changes in Baseline
Applications
Land-Based
Diesel Engines
Recreational
Marine Diesel
Engines
Commercial
Marine Diesel
Engines2
Locomotives
Total
Year
1996
2020
2030
1996
2020
2030
1996
2020
2030
1996
2020
2030
1996
2020
2030
NOX [short tons]
Final
1,564,904
1,119,481
1,192,833
33,679
47,847
52,085
823,905
943,560
1,117,848
934,070
508,084
481,077
3,356,558
2,618,972
2,843,843
Preliminary
1,583,641
1,144,686
1,231,981
19,438
34,814
41,246
960,153
819,544
815,162
921,556
612,722
534,520
3,484,788
2,611,766
2,622,909
Difference
-18,737
(-1.2%)
-25,205
(-2.2%)
-39,148
(-3.2%)
14,241
(73.3%)
13,033
(37.4%)
10,839
(26.3%)
-136,248
(-14.2%)
124,016
(15.1%)
302,686
(37.1%)
12,514
(1.4%)
-104,638
(-17.1%)
-53,443
(-10.0%)
-128,230
(-3.7%)
7,206
(0.3%)
220,934
(8.4%)
VOC Emissions [short tons]
Final
220,971
97,513
96,374
1,297
1,604
1,669
28,986
41,588
52,880
38,035
30,125
28,580
289,289
170,830
179,503
Preliminary
221,398
97,113
97,345
803
1,327
1,528
31,613
37,362
41,433
48,381
36,546
31,644
302,195
172,348
171,950
Difference
-427
(0.0%)
400
(0.4%)
-971
(1.0%)
494
(61.5%)
277
(20.9%)
141
(9.2%)
-2,627
(-9.1%)
4,226
(11.3%)
11,447
(27.6%)
-10,346
(-21.4%)
-6,421
(-17.6%)
-3,064
(-9.7%)
-12,906
(-4.3%)
-1,518
(0.9%)
7,553
(4.4%)
CO [short tons]
Final
1,004,586
697,630
786,181
5,424
9,482
11,232
108,883
150,562
178,360
92,496
99,227
107,780
1,211,389
956,901
1,083,553
Preliminary
1,010,501
702,145
793,899
3,215
5,537
6,464
126,523
160,061
176,708
112,171
119,302
119,302
1,252,410
987,045
1,096,373
Difference
-5,915
(-0.6%)
-4,515
(-0.6%)
-7,718
(-1.0%)
2,209
(68.7%)
3,945
(71.2%)
4,768
(73.8%)
-17,640
(-13.9%)
-9,499
(-5.9%)
1,652
(0.9%)
-19,675
(-17.5%)
-20,075
(-16.8%)
-11,522
(-9.7%)
-41,021
(-3.3%)
-30,144
(-3.1%)
-12,820
(-1.2%)
a To provide direct comparisons, for pollutants other than SO2, emissions include vessels using both diesel and residual fuels.

-------
Final Regulatory Impact Analysis
                                      Table 3.6-3 (cont.)
                Modeled 48-State Emission Impact Due to Changes in Baseline
Applications
Land-Based
Diesel Engines
Recreational
Marine Diesel
Engines
Commercial
Marine Diesel
Engines a
Locomotives
Total
Year
1996
2020
2030
1996
2020
2030
1996
2020
2030
1996
2020
2030
1996
2020
2030
PM2 5 Emissions [short tons]
Final
186,507
129,058
142,484
923
1,261
1,371
33,908
52,197
70,319
22,266
17,213
16,025
243,604
199,729
230,199
Preliminary
178,500
127,755
143,185
511
876
1,021
37,203
42,054
46,185
22,396
17,683
16,988
238,610
188,368
207,379
Difference
8,007
(4.5%)
1,303
(1.0%)
-701
(-0.5%)
412
(80.6%)
385
(43.9%)
350
(34.3%)
-3,295
(-8.9%)
10,143
(24.1%)
24,134
(52.3%)
-130
(-0.6%)
-470
(-2.7%)
-963
(-5.7%)
4,994
(2.1%)
11,361
(6.0%)
22,820
(11.0%)
SO2 [short tons]
Final
143,572
237,044
279,511
4,286
6,850
8,158
30,136
29,268
33,020
56,193
53,352
58,103
234,187
326,514
378,792
Preliminary
172,175
308,075
360,933
2,535
4,562
5,418
37,252
43,028
48,308
57,979
62,843
70,436
269,941
418,508
485,095
Difference
-28,603
(-16.6%)
-71,031
(-23.1%)
-81,422
(-22.6%)
1,751
(69.1%)
2,288
(50.2%)
2,740
(50.6%)
-7,116
(-19.1%)
-13,760
(-32.0%)
-15,288
(-31.6%)
-1,786
(-3.1%)
-9,491
(-15.1%)
-12,333
(-17.5%)
-35,754
(-13.2%)
-91,994
(-22.0%)
-106,303
(-21.9%)
a To provide direct comparisons, for pollutants other than SO2, emissions include vessels using both diesel and residual
fuels.
                                            3-94

-------
                                                                     Emission Inventory
3.6.3 Control Inventories

    Table 3.6-4 presents the preliminary 48-state control inventories used for the air quality
modeling. These are an aggregation of the county-level results.  Results expressed as short tons
are presented for 2020 and 2030 for the land-based diesel, recreational marine diesel, commercial
marine diesel, and locomotive categories. Results are not presented for 1996, since controls will
affect only future-year emission estimates.

                                        Table 3.6-4
                           Modeled 48-State Controlled Emissions
                 Preliminary Control Scenario Used for Air Quality  Modeling
Applications
Land-Based Diesel
Engines
Recreational Marine
Diesel Engines
Commercial Marine
Diesel Engines
Locomotives
Year
2020
2030
2020
2030
2020
2030
2020
2030
NOX
[short tons]
481,068
222,237
34,814
41,246
819,544
815,162
612,722
534,520
PM25
[short tons]
36,477
14,112
552
636
38,882
42,625
13,051
11,798
SO2
[short tons]
1,040
1,159
20
24
184
206
272
305
voc
[short tons]
73,941
63,285
1,327
1,528
37,362
41,433
36,546
31,644
CO
[short tons]
249,734
133,604
5,537
6,464
160,061
176,708
119,302
119,302
    The certification standards used for the preliminary and final control scenarios are provided in
Tables 3.6-5 and 3.6-6, respectively.  In general, the preliminary control scenario is more
stringent in terms of levels and effective model years for PM and NOX than the final control
scenario for all horsepower categories.  The NMHC standard is 0.14 g/hp-hr with both scenarios
for <750 hp engines, although the phase-in of this standard is later in the final control scenario.
The final control scenario also has a transitional NMHC  standard of 0.30 g/hp-hr for engines over
750 hp. There are no Tier 4 CO standards in both control scenarios, although CO is assumed to
be reduced 90 percent in both scenarios with the advent of trap-equipped engines (corresponding
to the start of 0.02 or 0.01 g/hp-hr PM standards). As a result, the final standards will increase the
emissions of PM, NOX, NMHC, and CO in 2020 and 2030 relative to the preliminary standards.
                                           3-95

-------
Final Regulatory Impact Analysis
                                           Table 3.6-5
              Preliminary Tier 4 Emission Standards Used for Air Quality Modeling
Engine Power
hp<25
25 < hp < 50
50 < hp < 100
100 < hp < 175
175 < hp < 750
hp > 750
Emission Standards
g/hp-hr
transitional or
final
transitional
final
transitional
final
transitional
final
transitional
final
transitional
final
transitional
final
PM
0.01
0.01
0.01
0.01
0.01
0.01
0.01
0.01
0.01
0.01
0.01
0.01
NOX
NMHC
5.6 ^
0.30
0.14
5.6 ^
0.30
0.14
3.5^b
0.30
0.14
3.0^
0.30
0.14
3.0^
0.30
0.14
4.8 ^
0.30
0.14
CO
6.0/4.9 b
6.0/4.9 b
4.1 b
4.1 b
3.7 b
3.7 b
3.7 b
3.7 b
2.6 b
2.6 b
2.6 b
2.6 b
Model Year
2010
2012
2010
2012
2010
2012
2010
2012
2009
2011
2009
2011
a This is a combined NMHC + NOX standard.
b This emission standard is unchanged from the level that applies in the previous model year. For engines below 25 hp, the
    CO standard is 6.0 g/hp-hr for engines below 11 hp and 4.9 g/hp-hr for engines at or above 11 hp. There are no Tier 4
    CO standards.
                                               3-96

-------
                                                                                  Emission Inventory
                                                Table 3.6-6
                                        Tier 4 Emission Standards
Engine
Power
hp<25
25 < hp < 75
75 < hp < 175
175 < hp < 750
hp > 750
except Generator sets
Generator sets
750 < hp < 1200
Generator sets
hp > 1200
Emission Standard
(g/hp-hr)
transitional
or final
final
transitional
final
transitional
final
transitional
final
transitional
final
transitional
final
transitional
final
PM
0.30
0.22
0.02
0.01
0.01
0.01
0.01
0.075
0.03
0.075
0.02
0.075
0.02
N0xa
NMHCa
5.6 b'c
5.6/3. 5 b'c
3.5b
0.30
(50%)
0.30
0.30
(50%)
0.30
2.6
2.6
2.6
0.50
0.50
0.50
0.14
(50%)
0.14
0.14
(50%)
0.14
0.30
0.14
0.30
0.14
0.30
0.14
cod
6.0/4.9 c
4.1/3.7 c
4.1/3.7°
3.7 c
3.7 c
2.6 c
2.6 c
2.6 c
2.6 c
2.6 c
2.6 c
2.6 c
2.6 c
Model
Year(s)
2008
2008-2012
2013
2012-2013
2014
2011-2013
2014
2011-2014
2015
2011-2014
2015
2011-2014
2015
a Percentages are model year sales fractions required to comply with the indicated NOX and NMHC standards, for model
years where less than 100 percent is required. For a complete description of manufacturer options and alternative
standards, refer to Section II of the preamble.
b This is a combined NMHC + NOX standard.
0 This emission standard level is unchanged from the level that applies in the previous model year. For 25-75 hp engines,
    the transitional NMHC + NOX standard is 5.6 g/hp-hr for engines below 50 hp and 3.5 g/hp-hr for engines at or above
    50 hp.  For engines under 75 hp, the CO standard is 6.0 g/hp-hr for engines below 11 hp, 4.9 g/hp-hr for engines 11 to
    under 25 hp, 4.1 g/hp-hr for engines 25 to below 50 hp and 3.7 g/hp-hr for engines at or above 50 hp.
d There are no Tier 4 CO standards. The CO emission standard level is unchanged from the level that applies in the
previous model year.
                                                    3-97

-------
Final Regulatory Impact Analysis
   The diesel fuel sulfur inputs used for the preliminary and final control scenarios are provided
in Tables 3.6-7 and 3.6-8, respectively. For land-based diesel engines, the modeled in-use diesel
fuel sulfur content is 11 ppm in 2020 and 2030 for both scenarios. For recreational marine
engines, commercial marine engines and locomotives, the modeled in-use diesel fuel sulfur
content is 11 ppm in 2020 and 2030 for the preliminary control scenario, but 55 ppm in 2020 and
2030 for the final control scenario. As a result, the fuel sulfur levels required by the final rule
will serve to increase the PM and SO2 control inventories for the recreational marine, commercial
marine, and locomotive categories in 2020 and 2030.  This will be offset slightly by the reduced
fuel volumes assigned to the commercial marine and locomotive categories.

                                       Table 3.6-7
      Modeled 48-State In-Use Diesel Fuel Sulfur Content Used for Air Quality Modeling
Applications
All Diesel Categories
Standards
Baseline + hwy 500 ppm
"spillover"
Baseline +hwy 15 ppm
"spillover"
June intro of 1 5 ppm
Final 1 5 ppm standard
Modeled In-Use Fuel Sulfur
Content, ppm
2500
2400
1006
11
Calendar
Year
through 2005
2006-2007
2008
2009
                                          3-98

-------
                                                                    Emission Inventory
                                        Table 3.6-8
                     Modeled 48-State In-Use Diesel Fuel Sulfur Content
Applications
Land-based,
all power ranges
Recreational and
Commercial Marine Diesel
Engines and Locomotives
Calendar Year(s)
through 2005
2006
2007
2008-2009
2010
2011-2013
2014
2015+
through 2000
2001
2002-2003
2004-2005
2006
2007
2008-2009
2010
2011
2012
2013
2014
2015-2017
2018-2038
2039-2040
Modeled In-Use Fuel Sulfur
Content, ppm
2283
2249
1140
348
163
31
19
11
2641
2637
2638
2639
2616
1328
408
307
234
123
43
51
56
56
55
   To adjust PM emissions for these in-use fuel sulfur levels, the adjustment is made relative to
the certification diesel fuel sulfur levels in the model.  The modeled certification diesel fuel sulfur
inputs used for the preliminary and final control scenarios are provided in Tables 3.6-9 and 3.6-
                                           3-99

-------
Final Regulatory Impact Analysis
10, respectively. For 2020 and 2030, the certification diesel fuel sulfur levels are the same for
both the preliminary and final control scenarios.

   Table 3.6-11 compares the preliminary and final 48-state control scenario inventories for
land-based diesel engines, recreational marine diesel engines, commercial marine diesel engines,
and locomotives. Results are presented for PM2 5, NOX, SO2, VOC, and CO emissions.

   For land-based diesel engines, emissions of PM25, NOX, VOC, and CO emissions are higher
for the final control scenario.  This is due to the less stringent emission standards.  There were no
differences in either the in-use or certification diesel fuel sulfur levels in 2020 and 2030 for this
category.  The minor difference in SO2 emissions between the preliminary and final scenarios is
attributed to differences in the version of the NONROAD model used and aggregation of county-
level runs for the preliminary scenario compared with using one national level run for the final
control scenario.

   The recreational marine, commercial marine, and locomotive categories are controlled in both
scenarios; however, the in-use fuel sulfur level is 11 ppm for the preliminary  control scenario and
56 ppm for the final control scenario. This directly affects the SO2 emissions. Accordingly, the
SO2 emissions for these categories are higher for the final control scenario.

   For the recreational marine category, differences are also attributed to the version of the
NONROAD model used. For the commercial marine category, the final control scenario now
accounts for the latest rulemaking inventories, as well as updated fuel volumes.  For locomotives,
the final control scenario incorporates updated fuel volume estimates.
                                        Table 3.6-9
        Modeled Certification Diesel Fuel Sulfur Content Used for Air Quality Modeling
Engine Power
hp<50
50 < hp < 175
175 < hp < 750
hp > 750
Standards
Tier 2
Tier 4a
Tier 3
Tier 4a
Tier 3
Tier 4a
Tier 2
Tier 4a
Modeled Certification Fuel
Sulfur Content, PPM
2000
15
2000
15
2000
15
2000
15
Model
Year
through 2009
2010
through 2009
2010
through 2008
2009
through 2008
2009
          1 Tier 4 refers to both transitional and final standards.
                                           5-100

-------
                                                                         Emission Inventory
                                         Table 3.6-10
                     Modeled Certification Diesel Fuel Sulfur Content
Engine Power
hp<75
75 < hp < 100
100 < hp < 175
175 < hp < 750
hp > 750
Standards
Tier 2
transitional
final
Tier 3 transitional3
final
Tier 3
final
Tier 3
final
Tier 2
final
Modeled Certification Fuel
Sulfur Content, PPM
2000
500
15
500
15
2000
15
2000
15
2000
15
Model
Year
through 2007
2008
2013
2008-2011
2012
2007-2011
2012
2006-2010
2011
2006-2010
2011
a The emission standard here is still Tier 3 as in the Baseline case, but since the Tier 3 standard begins in 2008 for 50-
100 hp engines it is assumed that this new technology introduction will allow manufacturers to take advantage of the
availability of 500 ppm fuel that year.
                                             5-101

-------
                                                              Table 3.6-11
                                Modeled 48-State Emission Impact Due to Changes in Control Scenario
Applications
Land-Based Diesel
Engines
Recreational Marine
Diesel Engines
Commercial Marine
Diesel Engines a
Locomotives
Year
2020
2030
2020
2030
2020
2030
2020
2030
NO, [short tons]
Final
677,420
458,649
47,847
52,085
943,560
1,117,848
508,084
481,077
Preliminary
481,068
222,237
34,814
41,246
819,544
815,162
612,722
534,520
Difference
196,352
(40.8%)
236,412
(106%)
13,033
(37.4%)
10,839
(26.3%)
124,016
(15.1%)
302686
(37.1%)
-104,638
(-17.1%)
-53,443
(-10.0%)
PM2 , [short tons]
Final
50,065
21,698
738
749
49,968
67,804
13,149
11,599
Preliminary
36,477
14,112
552
636
38,882
42,625
13,051
11,798
Difference
13,588
(37.3%)
7,586
(53.8%)
186
(33.7%)
113
(17.8%)
11,086
(28.5%)
25,179
(59.1%)
98
(0.8%)
-199
(-1.7%)
SO2 [short tons]
Final
986
1,074
164
195
703
786
1,282
1,384
Preliminary
1,040
1,159
20
24
184
206
272
305
Difference
-54
(-5.2%)
-85
(-7.3%)
144
(720%)
171
(713%)
519
(282%)
580
(282%)
1,010
(371%)
1,079
(354%)
' To provide direct comparisons, for pollutants other than SO2, emissions include vessels using both diesel and residual fuels.

-------
                                                                           Emission Inventory
                                      Table 3.6-11, continued
Applications
Land-Based Diesel
Engines
Recreational Marine
Diesel Engines
Commercial Marine
Diesel Engines a
Locomotives
Year
2020
2030
2020
2030
2020
2030
2020
2030
VOC [short tons]
Final
79,372
66,344
1,604
1,669
41,589
52,880
30,125
28,580
Preliminary
73,941
63,285
1,327
1,528
37,362
41,433
36,546
31,644
Difference
5,431
(7.3%)
3,059
(4.8%)
277
(20.9%)
141
(9.2%)
4,227
(11.3%)
11,447
(27.6%)
-6,421
(-17.6%)
-3,064
(-9.7%)
CO [short tons]
Final
309,593
167,014
9,482
11,232
150,562
178,360
99,227
107,780
Preliminary
249,734
133,604
5,537
6,464
160,061
176,708
119,302
119,312
Difference
59,859
(24.0%)
33,410
(25.0%)
3,945
(71.2%)
4,768
(73.8%)
-9,499
(-5.9%)
1,652
(0.9%)
-20,075
(-16.8%)
-11,532
(-9.7%)
a To provide direct comparisons, for pollutants other than SO2, emissions include vessels using both diesel and residual
fuels.
                                                5-103

-------
Final Regulatory Impact Analysis
Chapter 3 References

1. U. S. Environmental Protection Agency.  Exhaust and Crankcase Emission Factors for
NonroadEngine Modeling: Compression Ignition. NR-009b. Assessment & Standards Division,
Office of Transportation & Air Quality. Ann Arbor, MI. November, 2002. (Docket A-2001-28,
Document II-A-29)

2. U.S. Environmental Protection Agency.  Median Life, Annual Activity, and Load Factor
Values for Nonroad Engine Emissions Modeling. NR-005b. Assessment & Standards Division,
Office of Transportation & Air Quality. Ann Arbor, MI. May, 2002. (Docket A-2001-28,
Document II-A-30)

3. U.S. Environmental Protection Agency.  Nonroad Engine Population Estimates. NR-006b.
Assessment & Standards Division, Office of Transportation & Air Quality. Ann Arbor, MI. July,
2002. (Docket A-2001-28, Document II-A-31)

4. U.S. Environmental Protection Agency.  Nonroad Engine Growth Estimates. NR-008b.
Assessment & Standards Division, Office of Transportation & Air Quality. Ann Arbor, MI. May,
2002. (Docket A-2001-28, Document II-A-32)

5. U.S. Environmental Protection Agency.  Calculation of Age Distributions in the Nonroad
Model: Growth and Scrappage. NR-007a Assessment & Standards Division, Office of
Transportation & Air Quality. Ann Arbor, MI. June, 2002.  (Docket A-2001-28, Document II-A-
33)

6. U.S. Environmental Protection Agency. Conversion Factors for Hydrocarbon Emission
Components. NR-002. Assessment and Standards Division, Office of Transportation & Air
Quality. November, 2002.  (Docket A-2001-28, Document II-A-34)

7. U.S. Environmental Protection Agency. Recommended revision of the fraction ofdiesel
paniculate emissions mass less than 2.5 microns in size. Memo to the docket from Bruce
Cantrell. October 17, 2003.  (Docket A-2001-28, Document IV-B-21)

8.U.S. Environmental Protection Agency. Control of Emissions From Nonroad Large Spark-
Ignition Engines, and Recreational Engines (Marine and Land-Based); Final Rule. 67 FR
68241-68290. November 8, 2002.  (Docket Number A-2001-01, Document V-B-05)

9. U.S. Environmental Protection Agency. Documentation For Aircraft, Commercial Marine
Vessel, Locomotive, and Other Nonroad Components of the National Emissions Inventory,
Volume I - Methodology. Emission Factor and Inventory Group, Emissions Monitoring and
Analysis Division. November 11, 2002.
(ftp: //ftp. epa.gov/Emi slnventory/draftnei 99ver3 /hap s/documentati on/nonroad/)

10. U.S. Environmental Protection Agency. Control of Emissions From Nonroad Large Spark-
Ignition Engines, and Recreational Engines (Marine and Land-Based); Final Rule. 67 FR
68241-68290. November 8, 2002.  (Docket Number A-2001-01, Document V-B-05)

                                         3-104

-------
                                                                Emission Inventory
11. E. H. Pechan & Associates, Inc. Procedures for Developing Base Year and Future Year
Mass Emission Inventories for the Nonroad Diesel Engine Rulemaking. Prepared for U.S.
Environmental Protection Agency, Office of Air Quality Planning and Standards. February,
2003.

12. U.S. Environmental Protection Agency. Control of Emissions of Air Pollution from 2004 and
Later Heavy-Duty Highway Engines and Vehicles. Office of Air and Radiation. July, 2000.
(Docket Number A-99-06, Document IV-A-01)

13. John E. Batey and Roger McDonald. Advantages of Low Sulfur Home Heating Oil. Interim
Report of Compiled Research, Studies, and Data Resources. National Oilheat Research Alliance
and U.S. Department of Energy. December, 2002. (E-Docket Number OAR-2003-0012-0933)
                                         5-105

-------
CHAPTER 4: Technologies and Test Procedures for Low-Emission Engines
    4.1 Feasibility of Emission Standards  	  4-1
        4.1.1 PM Control Technologies 	  4-2
             4.1.1.1 In-Cylmder PM Control 	  4-3
             4.1.1.2 Diesel Oxidation Catalyst (DOC) 	  4-5
             4.1.1.3 Catalyzed Diesel Particulate Filter (CDPF)  	  4-6
        4.1.2 NOx Control Technologies  	  4-19
             4.1.2.1  In-Cylinder NOx Control Technologies	  4-19
             4.1.2.2  Lean-NOx Catalyst Technology	  4-20
             4.1.2.3  NOx Adsorber Technology  	  4-21
             4.1.2.4 Selective Catalytic Reduction (SCR) Technology  	  4-69
        4.1.3 Can These Technologies Be Applied to Nonroad Engines and Equipment?	  4-70
             4.1.3.1 Nonroad Operating Conditions andExhaust Temperatures  	  4-71
             4.1.3.2 Durability andDesign	  4-80
        4.1.4 Are the Standards for Engines >25 hp and <75 hp Feasible?  	  4-82
             4.1.4.1 What makes the 25 - 75 hp category unique? 	  4-83
             4.1.4.2 What engine technology is used currently, and will be used for Tier 2 and Tier 3, in the 25-75hp
                 range?	  4-85
             4.1.4.3 Are the standards for 25 -75 hp engines technologically feasible?	  4-86
        4.1.5 Are the Standards for Engines <25 hp Feasible?	  4-94
             4.1.5.1 What makes the < 25 hp category unique? 	  4-94
             4.1.5.2 What engine technology is currently used in the <25 hp category? 	  4-95
             4.1.5.3 What data support the feasibility of the new standards?	  4-95
        4.1.6 Meeting the Crankcase Emission Requirements  	  4-101
        4.1.7 Why Do We Need 15 ppm Sulfur Diesel Fuel?	  4-101
             4.1.7.1 Catalyzed Diesel Particulate Filters and the Need for Low-Sulfur Fuel	  4-102
             4.1.7.2 Diesel NOx Catalysts and the Need for Low-Sulfur Fuel	  4-108
    4.2 Transient Emission Testing 	  4-110
        4.2.1 Background and Justification	  4-110
             4.2.1.1 Microtrip-Based Duty Cycles 	  4-112
             4.2.1.2 "Day-in-the-Life"-BasedDuty Cycles 	  4-112
        4.2.2 Data Collection and Cycle Generation  	  4-113
             4.2.2.1 Test Site Descriptions  	  4-113
             4.2.2.2 Engine and Equipment Description	  4-115
             4.2.2.3 Data Collection Process	  4-118
             4.2.2.4 Cycle Creation Process 	  4-119
        4.2.3 Composite Cycle Construction 	  4-126
        4.2.4 Cycle Characterization Statistics	  4-128
        4.2.5 Cycle Normalization/Denormalization Procedure	  4-129
        4.2.6 Cycle Performance Regression Statistics	  4-130
        4.2.7 Constant-Speed, Variable-Load Equipment Considerations	  4-130
             4.2.7.1. Background on Cycle Considered 	  4-131
             4.2.7.2. Follow-on Constant-Speed Engine Testing  and Analysis   	  4-132
        4.2.8 Cycle Harmonization	  4-134
             4.2.8.1 Technical Review  	  4-134
             4.2.8.2 Global Harmonization Strategy	  4-136
        4.2.9 Cold-Start Transient Test Procedure 	  4-144
        4.2.10 Applicability of Component Cycles to Nonroad Diesel Market	  4-146
             4.2.10.1 Market Representation of Component Cycles	  4-147
             4.2.10.2 Inventory Impact of Equipment Component Cycles	  4-147
             4.2.10.3 Power and Sales Analysis	  4-148
             4.2.10.4 Broad Application Control  	  4-148
        4.2.11 Final Certification Cycle Selection Process	  4-149
    4.3 Steady-State Testing	  4-150
        4.3.1 RampedModal Cycle  	  4-151
             4.3.1.1 Introduction and Background 	  4-151
             4.3.1.2 Comparison of Steady-State vs. RMC Testing	  4-154
        4.3.2 Transportation Refrigeration Unit Test Cycle	  4-164
    4.4 Not-to-Exceed Testing	  4-167

-------
 	Technologies and Test Procedures for Low-Emission Engines

  CHAPTER 4:  Technologies  and Test Procedures for Low-
                              Emission Engines
   The new emission standards will require both new engine technologies and new
measurement procedures.  Section 4.1 documents the technical analysis supporting the feasibility
of meeting the Tier 4 emission standards for nonroad diesel engines, including the not-to-exceed
standards. Section 4.2 describes the development and characteristics of the new transient duty
cycles and Section 4.3 describes issues related to steady-state duty cycles, including the
development of new ramped-modal duty cycles and new cycles for transportation refrigeration
units.

4.1 Feasibility of Emission Standards

   A description of the new emission standards and our reasons for setting those standards can
be found in Section II of the preamble to the final rule. This chapter documents the analysis we
completed to inform the decisions described in the preamble regarding new emission standards
for nonroad diesel engines. This analysis incorporates recent Agency analyses of emission-
control technologies for highway diesel engines and expands those analyses with more recent
data and additional analysis specific to the application of technology to nonroad diesel
engines.1'2'3

   This section is organized into subsections describing diesel emission-control technologies,
issues specific to the application of these technologies to new nonroad engines, specific analyses
for engines within distinct power categories (<25 hp and 25-75 hp) and an analysis of the need
for low-sulfur diesel fuel (15 ppm sulfur) to enable these emission-control technologies.

   For the past 30 or more years, emission-control development for gasoline vehicles and
engines has concentrated most aggressively on exhaust emission-control devices.  These devices
currently provide as much as or  more than 95 percent of the emission control on a gasoline
vehicle. In contrast, the emission-control development work for highway and nonroad diesel
engines has concentrated on improvements to the engine itself to limit the emissions leaving the
combustion chamber.

   During the past 15 years, however, more development effort has been put into  catalytic
exhaust emission-control devices for diesel engines, particularly in the area of particulate matter
(PM) control. Those developments, and recent developments in diesel NOx exhaust emission-
control devices, make the widespread commercial use of diesel exhaust emission controls
feasible. EPA has recently set new emission standards for diesel engines installed in highway
vehicles based on the emission-reduction potential of these devices.  We believe these devices
will make possible a level  of emission control for nonroad diesel engines that is similar to that
                                         4-1

-------
Regulatory Impact Analysis
attained by gasoline three-way-catalyst applications. However, without low-sulfur diesel fuel,
these technologies cannot be implemented.

   Although the primary focus of the Tier 4 emissions program and the majority of the analysis
contained in this RIA is directed at the application of catalytic emission control technologies
enabled by 15 ppm sulfur diesel fuel, there are also important elements of the program based
upon continuing improvements in engine-out emission controls. Like the advanced catalytic
based technologies, these engine-out emission solutions for nonroad diesel engines rely upon
technologies already applied to on-highway diesel engines. Additionally, these technologies
form the basis for the Tier 3 emission standards for some nonroad diesel engines in other size
categories.  Extensive analysis and discussion of these engine-out emission control technologies
can be found in the RIAs associated with the On-Highway Heavy-Duty 2004 emission standards
and the Nonroad Tier 2 and Tier 3 emission standards.4'5'6'7 Those detailed analyses are not
repeated here but are a fundamental underpinning of EPA's understanding of engine-out
emission controls for diesel engines and the feasibility of applying those controls to nonroad
diesel engines in the Tier 4 timeframe.

4.1.1 PM Control Technologies

   Particulate matter from diesel engines is made of four components;
       - solid carbon soot,
       - volatile and semi-volatile organic matter
       - inorganic solids (ash) , and
       - sulfate.

The formation of the solid carbon soot portion of PM is inherent in diesel engines due to the
heterogenous distribution of fuel and air in a diesel combustion system. Diesel combustion is
designed to allow for overall lean (excess oxygen) combustion giving good efficiencies and low
CO and HC emissions with a small region of rich (excess fuel) combustion within the fuel-
injection plume. It is within this excess fuel  region of the combustion that PM is  formed when
high temperatures and a lack of oxygen cause the fuel to pyrolize, forming soot. Much of the
soot  formed in the engine is burned during the combustion process as the soot is mixed with
oxygen in the cylinder at high temperatures.  Any soot that is not fully burned before the exhaust
valve is opened will be emitted from the engine as diesel PM.

   The volatile and semi-volatile organic material in diesel PM is often simply referred to as the
soluble organic  fraction (SOF) in reference to a test method used to measure its level.  SOF is
primarily composed of engine oil that passes through the engine with no oxidation or only partial
oxidation and condenses in the atmosphere to form PM. The SOF portion of diesel PM can be
reduced through reductions in engine oil consumption and through oxidation of the SOF
catalytically in the exhaust.

   The inorganic solids (ash) in diesel PM comes primarily from metals found in engine oil and
to certain extent from engine wear.  Ash makes up a very small portion of total PM such that it is
often not listed as a PM component and has no impact on compliance with PM emission

                                          4-2

-------
	Technologies and Test Procedures for Low-Emission Engines

standards. However, it does impact maintenance of PM filter technologies, as discussed later,
because in aggregate over a very long period of time ash accumulation in the PM filter can reach
a level such that it must be cleaned from the filter (see section 4.1.1.3.4 below).

   The sulfate portion of diesel PM is formed from sulfur present in diesel fuel and engine
lubricating oil that oxidizes to form sulfuric acid (H2SO4) and then condenses in the atmosphere
to form sulfate PM.  Approximately two percent of the sulfur that enters a diesel engine from the
fuel is emitted directly from the engine as sulfate PM.8 The balance of the sulfur content is
emitted from the engine as SO2.  Oxidation catalyst technologies applied to control the SOF and
soot portions of diesel PM  can inadvertently oxidize SO2 in the exhaust to form sulfate PM. The
oxidation of SO2 by oxidation catalysts to form sulfate PM is often called sulfate make. Without
low-sulfur diesel fuel, oxidation catalyst technology to control diesel PM is limited by the
formation of sulfate PM in the exhaust as discussed in more detail in the discussion below of the
need for low-sulfur fuel.

   4.1.1.1 In-Cylinder PM Control

   The soot portion of PM emissions can be reduced by increasing the availability of oxygen
within the cylinder for soot oxidation during combustion.  Oxygen can be made more available
by either increasing the oxygen content in-cylinder or by increasing  the mixing of the fuel and
oxygen in-cylinder.  Several current technologies can influence oxygen content and in-cylinder
mixing, including improved fuel-injection systems, air management systems, and combustion
system designs. Many of these PM-reducing technologies offer better control of combustion in
general, and better utilization of fuel allowing for improvements in fuel efficiency concurrent
with reductions in PM emissions. Improvements in combustion technologies and refinements of
these systems is an ongoing effort for highway engines and for some nonroad engines where
emission standards or high fuel use encourage their introduction. The application of better
combustion system technologies across the broad range of nonroad engines for meeting the new
emission standards offers an opportunity for significant reductions in engine-out PM emissions
and possibly for reductions in fuel consumption.

   In general, the application of these in-cylinder emission control solutions for PM are more
successful (reduce PM to a lower level) as engine size increases. This occurs for three reasons:
1) larger engines have a higher volume to surface area within the cylinder reducing the
proportion of the in-cylinder volume near a cooler cylinder wall and thus decreasing PM
formation in these cool regions; 2) larger engines operate over a narrow engine speed range
allowing for better matching of turbomachinery to the engine (i.e., higher boost and more
oxygen); and  3) larger engines operate at lower engine speeds reducing oil consumption which
contributes to SOF and providing longer residence time for combustion to complete (i.e., at
slower speeds the combustion event measured in time is longer). In  the Tier  4 program, we are
setting an emission standard of 0.075 g/bhp-hr for some nonroad diesel engines >750 hp
beginning in 2011.  This emission level is approximately 25 percent lower than the level for

-------
Regulatory Impact Analysis
most current on-highway diesel engines (using 500 ppm sulfur fuel).A We are projecting that in-
cylinder PM emission control technologies along with 15 ppm sulfur diesel fuel will allow these
very large nonroad diesel engines to meet this emission standard.  Given the inherent PM control
advantage that these larger diesel engines enjoy when compared to the smaller on-highway
counterparts and the use of lower sulfur diesel fuel which lowers sulfate PM, we can conclude
that the 0.075 g/bhp-hr emission standard is clearly feasible for these engines in 2011.

   Another means to reduce the soot portion of engine-out PM emissions from diesel
(compression-ignited) engines is to operate the engine with a homogenous method of operation,
rather than the typical heterogeneous operation. In homogenous diesel combustion, also called
premixed diesel combustion, the fuel is dispersed evenly with the air throughout the combustion
system. This means there are no fuel-rich/oxygen-deprived regions of the system where fuel can
be pyrolized rather than burned.  Rather, combustion occurs globally initiating at an
indeterminate number of locations. Because there are no fuel-rich/oxygen deprived regions in
homogenous combustion, the carbon (soot) PM emissions are eliminated. The resulting PM
emissions are very low, consisting primarily of SOF and sulfate.

   Homogenous diesel combustion has been under development for more than twenty years, yet
it is still unable to overcome a number of developmental issues.9'10 Fundamental among these
issues is the ability to control the start of combustion.11 Conventional diesel engines control the
start of combustion by controlling the start of fuel injection: injection-timing control.
Homogenous diesel combustion  systems cannot readily use fuel-injection timing to control the
start of combustion because it is  difficult to inject fuel into the engine without initiating
combustion. If combustion is initiated while the fuel is being injected, the engine will operate
under heterogenous combustion resulting in high PM emissions. Techniques used to delay the
start of combustion such as decreasing intake air temperatures or reducing the engines
compression ratio can lead to misfire, a failure to ignited the fuel at all. Engine misfire results in
no engine power and high hydrocarbon (raw fuel) emissions. Conversely, techniques to advance
the start of combustion such as increasing intake air temperatures or increasing the engine
compression ratio can lead to premature uncontrolled  combustion called engine knock.  Engine
knock causes exceedingly high in-cylinder pressure spikes that can irreversibly damage  a diesel
engine at all but low-load conditions.

   Controlled homogenous combustion is possible with a diesel engine under certain
circumstances, and is used in limited portions of engine operation by some engine
manufacturers. Nissan, a passenger car manufacturer, has developed a modified version of
premixed combustion that they call modulated-kinetics, or MK, combustion.12'13 When operated
under MK combustion the PM and NOx emissions of the engine are dramatically decreased.
Unfortunately, the range of engine operation for which the MK combustion process can  function
is limited to low-load conditions. At higher engine loads the combustion process is not  stable
and the engine reverts to operation with conventional  diesel combustion. This dual mode
operation allows the engine to benefit from the homogenous combustion approach when
   A On-highway diesel engines used in urban buses must meet an even lower PM standard of 0.05 g/bhp-hr.

                                          4-4

-------
	Technologies and Test Procedures for Low-Emission Engines

possible, while still providing the full range of engine operation. Other approaches that are
similarly limited to low-load engine operation have been proposed to produce a dual combustion
mode engine.14'15'16

   4.1.1.2 Diesel Oxidation Catalyst (DOC)

   Diesel oxidation catalyst (DOCs) are the most common form of diesel aftertreatment
technology today and have been used for compliance with the PM standards for some highway
engines since the early 1990s. DOCs reduce diesel PM by oxidizing a small fraction of the soot
emissions and a significant portion of the SOF emissions. Total DOC effectiveness to reduce
PM emissions is normally limited to approximately 30 percent because the SOF portion of diesel
PM for modern diesel engines is typically less than 30 percent and because the DOC increases
sulfate emissions, reducing the overall effectiveness of the catalyst.  Limiting fuel sulfur levels to
15ppm allows DOCs to be designed for maximum effectiveness (nearly 100% control of SOF
with highly active catalyst technologies) since their control effectiveness is not reduced by
sulfate make (i.e., their sulfate make rate is high but because the sulfur level in the fuel is low the
resulting PM  emissions are well controlled).

   DOC effectiveness to control HC and CO emissions are directly related to the "activity" of
the catalyst material used in DOC washcoating.  Highly active (hence effective) DOCs can
reduce HC emissions by 97 percent while low activity catalysts realize approximately 50 percent
HC control.17 Today, highly active DOC formulations cannot be used for NMHC and CO
control because the sulfur in  current diesel fuel leads to unacceptable sulfate PM  emissions, as
discussed later in this section. However, with the low sulfur diesel fuel that will be available
under this program, DOCs will be able to provide substantial control of these pollutants. We
have projected the use of DOCs  as part of an overall compliance strategy for engines meeting the
interim PM standards in 2008. For those engines, DOC would also provide significant
reductions in  CO and HC including over the various emission test cycles for these engines.
Oxidation catalyst technologies generally (i.e., DOCs and CDPFs) will be an effective tool to
ensuring compliance over the NTE provisions of the Tier 4 program and to ensuring compliance
with the CO standards under the new test cycles.

   Data presented by one engine manufacturer regarding the existing Tier 2 PM  standard  show
that while a DOC can be used to reduce PM emissions when tested on 2,000 ppm sulfur fuel,
lowering the fuel sulfur level to 380 ppm enabled the DOC to reduce PM by 50 percent from the
2,000 ppm sulfur fuel.18  Without the availability of 500 ppm sulfur fuel in 2008,  DOCs would be
of limited use for nonroad engine manufacturers and would not provide the emission-control
necessary for most engine manufacturers to meet the 2008 interim Tier 4 standards. With  the
availability of 500 ppm sulfur fuel, DOCs can be designed to provide PM reductions on the order
of 20 to 50%, while suppressing particulate sulfate reduction.19 These levels of reductions have
been seen on transient duty cycles as well as on highway and nonroad steady-state duty cycles.

   DOCs are also very effective at reducing the air toxic emissions from diesel engines.  Test
data show that emissions of toxics such as poly cyclic aromatic hydrocarbons (PAHs) can be
reduced by more than 80 percent with a DOC.20

                                          4-5

-------
Regulatory Impact Analysis
   DOCs are less effective at controlling the solid carbon soot portion of PM. The solid (soot)
typically constitutes 60 to 90 percent of the total diesel PM. Even with 15 ppm sulfur fuel,
DOCs would therefore not be able to achieve the level of PM control needed to meet the PM
filter based PM emission standards (i.e., PM standards at or below 0.03 g/bhp-hr).  As noted
above however, DOCs can be effective tools to accomplish emission reductions on the order of
20 to  50 percent even when operated on 500 ppm sulfur diesel fuel and thus may be used by
some  manufacturers as a means to reduce emissions in order to comply with the 2008 interim
Tier 4 standards for engines <75 hp.

   4.1.1.3 Catalyzed Diesel Particulate Filter (CDPF)

   4.1.1.3.1 CDPF PM and HC Control Effectiveness

   Emission levels from a catalyzed diesel particulate filter (CDPF) are determined by several
factors.  Filtering efficiencies for solid particle emissions like soot are determined by the
characteristics of the PM filter, including  wall thickness and pore size.  Some of these
characteristics represent a tradeoff between mechanical strength, weight, size and filtering
efficiency. Filtering  efficiencies for ceramic based diesel soot filters can be as high as 99
percent with the appropriate filter design.21  Given an appropriate PM filter design, the
contribution of the soot portion of PM to the total PM emissions can be negligible (less than
0.001 g/hp-hr). For some wire mesh or ceramic fiber filter technologies the filtering efficiency is
lower, around 70 percent, but the mechanical strength (resistance to thermal and mechanical
stress) especially for very large filter sizes is improved.8'22'23 The level of soot emission control
is much less dependent on engine  test cycle or operating conditions due to the mechanical
filtration characteristics  of the particulate filter.

   Control of the SOF portion of diesel soot is accomplished on a CDPF through catalytic
oxidation. At the elevated temperature of diesel exhaust, the SOF portion of diesel PM consists
primarily of gas-phase hydrocarbons which later form particulate matter in the environment
when the SOF condenses. Catalytic materials applied to  CDPFs can oxidize a substantial
fraction of the SOF in diesel PM just as the SOF portion is oxidized by a DOC. However, we
believe that for engines with very  high SOF emissions the emission rate may be higher than can
be handled by a conventionally sized catalyst resulting in higher than zero SOF emissions.  If a
manufacturer's base engine technology has high oil consumption rates, and therefore high
engine-out SOF emissions (i.e., higher than 0.04 g/hp-hr), compliance with the 0.01 g/hp-hr
   B There are a number of different ways to measure mechanical strength and toughness. One metric for
comparison is tensile strength. Comparing the tensile strength of fiber based filter technologies (approximately
1,000 MPa) to a ceramic filter technology such as Silicon Nitride (5.1 MPa) is illustrative of the higher strength of
the fiber based technology.

                                           4-6

-------
	Technologies and Test Procedures for Low-Emission Engines

emission standard may require additional technology beyond the application of a CDPF system
alone.c

   Modern highway diesel engines have controlled SOF emission rates to comply with the
existing 0.1 g/hp-hr emission standards. Typically the SOF portion of PM from a modern
highway diesel engine contributes less than 0.02 g/hp-hr to the total PM emissions. This level of
SOF control is accomplished by controlling oil consumption through the use of engine
modifications (e.g., piston ring design, the use of 4-valve heads, the use of valve stem seals,
etc.).24 Nonroad diesel engines may similarly need to control engine-out SOF emissions to
comply with the new emission standards.  The means to control engine-out SOF emissions are
well known and have additional benefits, as they decrease oil consumption reducing operating
costs. With good control of engine-out SOF emissions (i.e.,  engine-out SOF < 0.02 g/hp-hr) and
the application of catalytic material to the DPF, SOF emissions from CDPF equipped nonroad
engines will contribute only a very small fraction of the total tailpipe PM emissions (less than
0.004 g/hp-hr). Alternatively, it may be less expensive or more practical for some applications
to ensure that the SOF control realized by the CDPF is in excess of 90 percent, thereby allowing
for higher engine-out SOF emission levels.

   The catalytic materials used on a CDPF to promote soot regeneration and to control SOF
emissions are also effective to control NMHC emissions including toxic hydrocarbon emissions.
CDPFs designed for operation on low-sulfur diesel fuel  (i.e., with highly active catalyst
technologies) can reduce total hydrocarbon emissions by more than 90 percent.25 Toxic
hydrocarbon emissions are typically reduced in proportion to total hydrocarbon emissions.
Table 4.1-1 shows hydrocarbon compound reduction data for two different CDPF technologies.26
   c SOF oxidation efficiency is typically better than 80 percent and can be better than 90 percent. Given a base
engine SOF rate of 0.04 g/hp-hr and an 80 percent SOF reduction a tailpipe emission of 0.008 can be estimated from
SOF alone. This level may be too high to comply with a 0.01 g/hp-hr standard once the other constituents of diesel
PM (soot and sulfate) are added.  In this case, engine-out SOF emissions will need to be reduced or the CDPF will
need to reduce SOF emissions by more than 90 percent.

                                           4-7

-------
Regulatory Impact Analysis
              Table 4.1-1 Polyaromatic Hydrocarbon Reductions with a CDPF
Polyaromatic Hydrocarbon Reductions with Catalyzed Diesel Particulate Filters
Compound
Napthalene
2-Methylnapthalene
Acenapthalene
Acenapthene
Fluorene
Phenanthrene
Anthracene
Fluoranthene
Pyrene
Benzo(a)anthracene
Chrysene
Benzo(b)fluoranthene
Benzo(k)fluoranthene
Benzo(e)pyrene
Perylene
lndeno(123-cd)pyrene
Dibenz(ah)anthracene
Benzo(ghi)perylene
Baseline
295
635
40
46
72
169
10
7.7
14
0.22
0.51
0.26
0.15
0.26
0.01
0.13
0.01
0.32
DPF-A
50
108
0.8
6.7
29
33
1
0
0
0
0
0
0
0
0
0
0
0
DPF-B
0
68
1
11
12
26
1
2
2
0.01
0
0
0
0
0
0
0
0
%Red DPF-A
83%
83%
98%
85%
60%
81%
90%
100%
100%
100%
100%
100%
100%
100%
100%
100%
100%
100%
%Red DPF-B
100%
89%
98%
76%
83%
85%
90%
74%
86%
95%
100%
100%
100%
100%
100%
100%
100%
100%
   The best means to reduce sulfate emissions from diesel engines is by reducing the sulfur
content of diesel fuel and lubricating oils.  This is one of the reasons that we are limiting sulfur
levels in nonroad diesel fuel to 15ppm or less.  The catalytic material on the CDPF is crucial to
ensuring robust regeneration and high SOF oxidation; however, it can also oxidize the sulfate in
the exhaust with high efficiency.  The result is that the predominant form of PM emissions from
CDPF equipped diesel engines is  sulfate PM. Even with 15ppm sulfur diesel fuel a CDPF
equipped diesel engine can have total PM emissions including sulfate emissions as high as 0.009
g/hp-hr over some representative  operating cycles using conventional diesel engine oils. This
level of emissions will meet the new PM emission standard of 0.01 g/hp-hr for engines between
75 hp and 750 hp. We further believe there is room for reductions from this level to provide
engine manufacturers with additional compliance margin. Our recently released Highway Diesel
Progress Review Report 2 documents progress  by a consortium of engine manufacturers, oil
companies and other stakeholders to develop a  new engine oil formulation with reduced Sulfur,
Ash, and Phosphorous (SAP) content for diesel engines.  The new engine oil formulation is
expected to be ready in 2006.  Any reduction in the sulfur level of engine lubricating oils will be
beneficial.  Similarly, as discussed above, we expect engine manufacturers to reduce engine oil
consumption to reduce SOF emissions and secondarily to reduce sulfate PM emissions. While
we believe sulfate PM emissions will be the single largest source of the total PM from diesel
engines, we believe that with the combination of technology, and the appropriate control of
engine-out PM emissions, sulfate and total PM emissions will be low enough to allow
                                          4-8

-------
	Technologies and Test Procedures for Low-Emission Engines

compliance with a 0.01 g/hp-hr standard, except in the case of small engines with higher fuel
consumption rates, as described later in this section.0

   CDPFs have been shown to be very effective at reducing PM mass by reducing dramatically
the soot and SOF portions of diesel PM.  In addition, recent data show that they are also very
effective at reducing the overall number of emitted particles when operated on low-sulfur fuel.
Hawker, et al, found that a CDPF reduced particle count by over 95 percent, including some of
the smallest  measurable particles (< 50 nm), at most of the tested conditions. The lowest
observed efficiency in reducing particle number was 86 percent. No generation of particles by
the CDPF was observed under any tested conditions.27 Kittelson, et al, confirmed that ultrafine
particles can be reduced by a factor often by oxidizing volatile  organics, and by an additional
factor often by reducing sulfur in the fuel. Catalyzed PM traps efficiently oxidize nearly all of
the volatile organic PM precursors (SOF), and the reduction of diesel fuel sulfur levels to 15ppm
or less will substantially reduce the number of ultrafine PM emitted from diesel engines.  The
combination of CDPFs with low-sulfur fuel is expected to result in very large reductions in both
PM mass and the number of ultrafine particles.

   Engine operating conditions have little impact on the particulate trapping efficiency of
carbon particles by CDPFs, so the greater than 90 percent efficiency for elemental carbon
particulate matter will apply to engine operation within the NTE zone and over the regulated
transient cycles, as well as to the test modes that comprise the steady-state test procedures such
as the ISO Cl.  However, engine operation will affect the CDPF regeneration and oxidation of
SO2 to sulfate PM (i.e., "sulfate-make").  Sulfate-make will reduce the measured PM removal
efficiency at some NTE operating conditions and some steady-state modes, even at the 15 ppm
fuel sulfur cap.  This increased sensitivity to fuel sulfur is caused by the higher temperatures that
are found at  some of the steady-state modes. High exhaust temperatures promote the oxidation
of SO2 to SO3 (which then combines with water in the exhaust, forming a hydrated sulfate)
across the precious metals found in CDPFs. The sulfate emissions condense in the atmosphere
(as well as in the CFR mandated dilution tunnel used for PM testing) forming PM.

   Under contract from the California Air Resources Board, two nonroad diesel engines were
recently tested for control  of PM emissions with the application of a CDPF over several transient
and steady-state test cycles.28  The first engine was a 1999 Caterpillar 3408 (480 hp, 18 liter
displacement) nonroad diesel  engine certified to the Tier 1 standards. The engine was tested
with and without a CDPF on 12 ppm sulfur diesel fuel. The transient emission results for this
engine are summarized in Table 4.1-2. The steady-state emission results are summarized in
Table 4.1-3.  The test results confirm the excellent PM control performance realized by a CDPF
with low-sulfur diesel fuel across a wide range of nonroad operating cycles in spite of the
relatively high engine-out PM emissions from this Tier 1 engine. We expect engine-out PM
emissions to be lower for production engines meeting Tier 3 standards, which will form the
   D We have also set slightly higher PM standards for >750 hp engines predicated on the use of alternative PM
filter technologies. These higher levels (standards of 0.02 g/bhp-hr for gensets, and 0.03 g/bhp-hr for mobile
machines) are not based on higher sulfate emission rates, as for the <75 hp engines, but instead on slightly less
effective PM filtration efficiencies and differing engine out emission rates.

                                           4-9

-------
Regulatory Impact Analysis
technology baseline for the Tier 4 engines. The engine demonstrated PM emissions of 0.009
g/hp-hr on the Nonroad Transient Cycle (NRTC) from an engine-out emission level of 0.256
g/hp-hr, a reduction of 0.247 g/hp-hr (a greater than 96% reduction). The engine also
demonstrated excellent PM performance on the existing steady-state ISO Cl cycle with PM
emissions of 0.010 g/hp-hr from an engine-out emission level of 0.127, a reduction of 0.107
g/hp-hr. Thus, this engine would meet the new emission standards for 75-750 hp variable-speed
nonroad engines.

       Table 4.1-2 Transient PM Emissions for a Tier 1 NR Diesel Engine with a CDPF
                     1999 (Tier 1) Caterpillar 3408 (480hp, 181)
Test Cycle
Proposed Nonroad TransientCycle (NRTC)
Proposed Constant Speed Variable Load Cycle (CSVL)
On-Highway U.S. FTP Transient Cycle (FTP)
Agricultural Tractor Cycle (ACT)
Backhoe Loader Cycle (BHL)
Crawler Tractor Dozer Cycle (CRT)
Composite Excavator Duty Cycle (CEX)
Skid Steer Loader Typical No. 1 (SST)
Skid Steer Loader Typical No. 2 (SS2)
Skid Steer Loader Highly Transient Speed (SSS)
Skid Steer Loader Highly Transient Torque (SSQ)
Arc Welder Typical No.1 (AWT)
Arc Welder Typical No. 2 (AW2)
Arc Welder Highly Transient Speed (AWS)
Rubber-Tired Loader Typical No.1 (RTL)
Rubber-Tired Loader Typical No. 2 (RT2)
Rubber-Tired Loader Highly Transient Speed (RTS)
Rubber-Tired Loader Highly Transient Torque (RTQ)
PM [g/bhp-hr]
Engine Out
0.256
0.407
0.239
0.181
0.372
0.160
0.079
0.307
0.242
0.242
0.351
0.510
0.589
0.424
0.233
0.236
0.255
0.294
w/ CDPF
0.009
0.016
0.019
0.009
0.022
0.014
0.009
0.016
0.013
0.008
0.004
0.018
0.031
0.019
0.010
0.011
0.008
0.009
Reduction
%
96%
96%
92%
95%
94%
91%
88%
95%
95%
97%
99%
96%
95%
96%
96%
96%
97%
97%
   Table 4.1-2 also shows results over a large number of additional test cycles developed from
real-world in-use test data to represent typical operating cycles for different nonroad equipment
applications (see Section 4.2 for information on these test cycles). The results show that the
CDPF technology is highly effective to control in-use PM emissions over any number of
disparate operating conditions.  Remembering that the base Tier 1 engine was not designed to
meet a transient PM standard, the CDPF emissions demonstrated here show that very low
emission levels are possible even when engine-out emissions are exceedingly high (e.g., a
reduction of 0.558 g/hp-hr is demonstrated on the AW2 cycle).

   The results summarized in the two tables support the feasibility of the NTE provisions in this
rulemaking. In spite of the Tier 1 baseline of this engine, there are only three test results with
emissions higher than the permissible limit for the NTE standards. The first, in Table 4.1-2,
shows PM emissions  of 0.031 over the AW2 cycle, but from a very high baseline level of nearly
0.6 g/hp-hr. We believe that simple improvements to the engine-out PM  emissions as needed to
comply with the Tier  2 emission standard would reduce these emission below the 0.02 level
required by the NTE standard.  There are two other test points in Table 4.1-3 that are  above the
                                         4-10

-------
                      Technologies and Test Procedures for Low-Emission Engines
NTE standard, both at 10 percent engine load.  However, both test points are outside the NTE
zone, which excludes emissions for engine loads below 30 percent. It is important to note that,
although the engine would not be constrained to meet NTE standards under these conditions, the
resulting reductions at both points are still substantially greater than 96 percent.

       Table 4.1-3 Steady-State PM Emissions from a Tier 1 NR Diesel Engine w/ CDPF
1999 (Tier 1) Caterpillar 3408 (480hp, 181)
Engine Speed
%
100
100
100
100
100
60
60
60
60
60
91
80
63
0

Engine Load
%
100
75
50
25
10
100
75
50
25
10
82
63
40
0
ISO C1 Composite
PM ([g/bhp-hr]
Engine Out
0.059
0.103
0.247
0.247
0.925
0.028
0.138
0.180
0.370
0.801
0.091
0.195
0.240
-
0.127
w/ CDPF
0.010
0.009
0.012
0.000
0.031
0.011
0.009
0.010
0.007
0.018
0.006
0.008
0.008
-
0.011
Reduction
%
83%
91%
95%
100%
97%
61%
93%
95%
98%
98%
93%
96%
97%
-
91%
    The second engine tested was a prototype engine developed at Southwest Research Institute
(SwRI) under contract to EPA.29 The engine, dubbed Deere Development Engine 4045 (DDE-
4045) because the prototype engine was based on a John Deere 4045 production engine, was also
tested with a CDPF from a different manufacturer on the same 12 ppm diesel fuel.  The engine is
very much a prototype and experienced a number of part failures during testing, including to the
turbocharger actuator. Nevertheless, the transient emission results summarized in Table 4.1-4
and the steady-state results summarized in Table 4.1-5 show that substantial PM reductions are
realized on this engine as well. The emission levels on the NRTC and the ISO Cl duty cycles
would meet the PM standard of 0.01 g/hp-hr once the appropriate rounding convention is
applied.E Note also that measured emissions over the transient highway FTP cycle are higher
than for either of the new nonroad transient duty cycles.  This suggests that developing PM-
compliant engines on the new nonroad transient cycles may not be substantially different from
developing compliant technologies for highway engines.
     The rounding procedures in ASTM E29-90 are applied to the emission standard. The emission results are
therefore rounded to the same number of significant digits as the specified standard, i.e., 0.014 g/hp-hr is rounded to
0.01 g/hp-hr, while 0.015 g/hp-hr would be rounded to 0.02 g/hp-hr.
                                          4-11

-------
Regulatory Impact Analysis
      Table 4.1-4 Transient PM Emissions for a Prototype NR Diesel Engine with a CDPF
                   EPA Prototype Tier 3 DDE-4045 (108hp, 4.5I)
Test Cycle
Proposed Nonroad TransientCycle (NRTC)
Proposed Constant Speed Variable Load Cycle (CSVL)
On-Highway U.S. FTP Transient Cycle (FTP)
Agricultural Tractor Cycle (ACT)
Backhoe Loader Cycle (BHL)
Crawler Tractor Dozer Cycle (CRT)
Composite Excavator Duty Cycle (CEX)
Skid Steer Loader Typical No. 1 (SST)
Skid Steer Loader Typical No. 2 (SS2)
Skid Steer Loader Highly Transient Speed (SSS)
Skid Steer Loader Highly Transient Torque (SSQ)
Arc Welder Typical No.1 (AWT)
Arc Welder Typical No. 2 (AW2)
Arc Welder Highly Transient Speed (AWS)
Rubber-Tired Loader Typical No.1 (RTL)
Rubber-Tired Loader Typical No. 2 (RT2)
Rubber-Tired Loader Highly Transient Speed (RTS)
Rubber-Tired Loader Highly Transient Torque (RTQ)
PM [g/bhp-hr]
Engine Out
0.143
0.218
0.185
0.134
0.396
0.314
0.176
0.288
0.641
0.298
0.536
0.290
0.349
0.274
0.761
0.603
0.721
0.725
w/ CDPF
0.013
0.018
0.023
0.008
0.021
0.008
0.009
0.012
0.013
0.011
0.014
0.018
0.019
0.019
0.014
0.012
0.010
0.009
Reduction
%
91%
92%
88%
94%
95%
97%
95%
96%
98%
96%
97%
94%
95%
93%
98%
98%
99%
99%
   As with the results from the Caterpillar engine, the two low-load (10 percent load) steady-
state emission points (see Table 4.1-5) have some of the highest brake specific emission rates.
However, these rates are not high enough to preclude compliance with the steady-state emission
cycle. The test points are also not within the NTE zone and still show substantial levels of PM
reduction.

      Table 4.1-5 Steady-State PM Emissions for a Prototype NR Diesel Engine w/CDPF
EPA Prototype Tier 3 DDE-4045 (108hp, 4.5I)
Engine Speed
%
100
100
100
100
100
60
60
60
60
60
91
80
63
0

Engine Load
%
100
75
50
25
10
100
75
50
25
10
82
63
40
0
ISO C1 Composite
PM [g/bhp-hr]
Engine Out
0.178
0.116
0.126
0.218
0.470
0.045
0.062
0.090
0.146
0.258
0.094
0.099
0.136
-
0.129
w/ CDPF
0.012
0.006
0.006
0.013
0.029
0.007
0.014
0.009
0.019
0.046
0.004
0.006
0.011
-
0.010
Reduction
%
93%
95%
96%
94%
94%
84%
78%
90%
87%
82%
95%
94%
92%
-
92%
                                        4-12

-------
	Technologies and Test Procedures for Low-Emission Engines

   The new NTE requirement, unlike the nonroad transient cycle (NRTC) or the existing ISO
Cl cycle, is not a composite test.  In fact, several of the individual modes within the Cl cycle
test fall within the NTE zone.  As discussed above, CDPFs are very efficient at capturing
elemental carbon PM (up to 99 percent), but sulfate-make under certain operating conditions
may exceed the standard of 0.01 g/hp-hr over the NRTC or Cl duty cycles, which is part of the
reason the NTE standard for PM is greater than the PM standards that apply for testing over the
NRTC and Cl duty cycles.

   In this rulemaking, we are making changes to the test procedures for nonroad Cl engines.
The switch to the test procedures specified in part 1065 and part 86 (from those specified in part
89) will generally improve the repeatability of emission measurements. These changes do not
change our analysis of the feasibility to comply with the Tier 4 standards as they are designed to
improve accuracy and repeatability and as  such do not adversely impact stringency. Also,  as
described in section III.G.3 of the preamble, we are considering in a separate proceeding
additional changes to the part  1065 regulations to further improve the test procedures. Like the
changes finalized in this rulemaking, these planned changes will not impact stringency only
accuracy and repeatability, and thus, will not impact feasibility.

   The new NTE requirements apply  not only during standard laboratory conditions, but also
during the expanded ambient temperature,  humidity, and altitude limits defined in the
regulations. We believe the new NTE PM standard is technologically feasible across this range
of ambient conditions. As discussed above, CDPFs are mechanical filtration devices, and
ambient temperature changes will have minimal effect on CDPF performance.  Ambient altitude
will also have minimal, if any, effects  on CDPF filtration efficiencies, and ambient humidity
should have no effect on CDPF performance.  As discussed above, particulate sulfate make is
sensitive to high exhaust gas temperatures; however, at sea-level conditions, the NTE
requirements apply up to ambient temperatures that are only 14°F greater than standard test cell
conditions (100°F under the NTE standards, versus 86°F for Cl laboratory conditions). At an
altitude of 5,500 feet above sea level, the NTE standards  apply only up to an ambient
temperature within the range of standard laboratory conditions (i.e., 86°F).  These small  or non-
existent differences in ambient temperature should have little effect on the sulfate make  of
CDPFs, and as can be seen in Tables 4.1-3 and 4.1-5 above, even when tested at an engine
operating test mode representative of the highest particulate sulfate generating conditions (peak-
torque operation) with 12 ppm sulfur diesel fuel, the results show the engine would easily meet
the NTE PM standard. Based on the available test data and the expected impact of the expanded,
but constrained, ambient conditions under which engines must comply with the NTE standards,
we conclude that the NTE PM standard for engines > 75 hp is technologically feasible (including
engines >750 hp), provided low-sulfur diesel fuel (15 ppm or lower) is available. Although we
do not have data available specific to the application of wire or fiber mesh filter technologies on
diesel engines >750 hp, the same filtration principles and control mechanisms apply to this
technology as to the ceramic technology described here.  A discussion of the technical feasibility
for engines with rated power lower than 75 hp is given in Sections 4.1.4 and 4.1.5.
                                          4-13

-------
Regulatory Impact Analysis
    4.1.1.3.2 CDPF Regeneration

    Diesel particulate filters (DPFs) control diesel PM by capturing the soot portion of PM in a
filter media, typically a ceramic wall flow substrate, and then by oxidizing (burning) it in the
oxygen-rich atmosphere of diesel exhaust. The SOF portion of diesel PM can be controlled
through the addition  of catalytic materials to the DPF to form a catalyzed diesel particulate filter
(CDPF).F The catalytic material  is also very effective to promote soot burning. This burning off
of collected PM is referred to as "regeneration."  In aggregate over an extended period of
operation, the PM must be regenerated at a rate equal to or greater that its accumulation rate, or
the DPF will clog.

    For a non-catalyzed DPF the soot can regenerate only at very high temperatures, in excess of
600°C, a temperature range that occurs infrequently  in normal diesel engine operation (exhaust
temperatures for many  engines might never reach 600°C).  With the addition of a catalytic
coating to make a CDPF, the temperature necessary  to ensure regeneration is decreased
significantly to approximately 250°C, a temperature within the normal operating range for most
diesel engines.30

    The catalytic materials that most effectively promote soot and SOF oxidation, however, are
significantly impacted by sulfur in  diesel fuel.  Sulfur both degrades catalyst oxidation efficiency
(i.e., poisons the catalyst) and forms sulfate PM.  Both catalyst poisoning by sulfur and increases
in PM emissions due to sulfate make influence our decision to limit the sulfur level  of diesel fuel
to 15 ppm as discussed in greater detail in the discussion below of the need for low-sulfur diesel
fuel.

    Filter regeneration is affected by catalytic materials used to promote oxidation, sulfur in
diesel fuel, engine-out soot rates, and exhaust temperatures. At higher exhaust temperatures,
soot oxidation occurs at a higher rate. Catalytic materials accelerate soot oxidation  at a single
exhaust temperature  compared with non-catalyst  DPFs, but even with catalytic materials
increasing the exhaust temperature further accelerates soot oxidation.

    Having applied 15 ppm sulfur diesel fuel and the best catalyst technology to promote low-
temperature oxidation (regeneration), the regeneration balance of soot oxidation equal to or
greater than soot accumulation over aggregate operation simplifies to the following  question: are
the exhaust temperatures high enough on aggregate to oxidize the engine-out PM emission rate?0
The answer is yes, for most highway applications and many nonroad applications, as
demonstrated by the  widespread  success of retrofit CDPF systems for nonroad equipment and
the use of both retrofit and original equipment CDPF systems for highway vehicles.3132'33
   F With regard to gaseous emissions such as NMHCs and CO, the CDPF works in the same manner with similar
effectiveness as the DOC (i.e., NMHC and CO emissions are reduced by more than 80 percent).

   G If the question was asked, "without 15 ppm sulfur fuel and the best catalyst technology, are the exhaust
temperatures high enough on aggregate to oxidize the engine-out PM emission rate?" the answer would be no, for all
but a very few highway or nonroad diesel engines.

                                           4-14

-------
	Technologies and Test Procedures for Low-Emission Engines

However, it is possible that for some nonroad applications the engine-out PM emission rate may
exceed the soot oxidation rate even with low-sulfur diesel fuel and the best catalyst technologies.
Should this occur, successful regeneration requires that either engine-out PM emission rates be
decreased or exhaust temperatures be increased, both feasible strategies. In fact, we expect both
to occur as highway-based technologies are transferred to nonroad engines.  As discussed earlier,
engine technologies to lower PM emissions while improving fuel consumption are continuously
being developed and refined. As these technologies are applied to nonroad engines driven by
both new emission standards and market pressures for better products, engine-out PM emissions
will decrease.  Similarly, techniques to raise exhaust temperatures periodically for initiating soot
oxidation in a PM filter have been developed for highway diesel vehicles as typified by the PSA
system used on more than 400,000 vehicles in Europe.34

   During our 2002 Highway Diesel Progress Review, we investigated the plans of highway
engine manufacturers to use CDPF systems to comply with the HD2007 emission standards for
PM.  We learned that all diesel engine manufacturers intend to comply through the application of
CDPF system technology.  We also learned that the manufacturers are developing  means to raise
the exhaust temperature, if necessary, to ensure that CDPF regeneration occurs.35 These
technologies include modifications to fuel-injection strategies, modifications to EGR strategies,
and modifications to turbocharger control strategies. These systems are based upon the
technologies used by the engine  manufacturers to comply with the 2004 highway emission
standards. In general, the systems anticipated to be used by highway manufacturers to meet the
2004 emission standards are the  same technologies that engine manufacturers have indicated to
EPA that they will use to comply with the Tier 3 nonroad regulations (e.g., electronic fuel
systems).36  In a manner similar to highway engine manufacturers, we  expect nonroad engine
manufacturers to adapt their Tier 3 emission-control technologies to provide back-up
regeneration systems for CDPF technologies to comply with the new emission standards. We
have estimated costs for such systems in our cost analysis.

   4.1.1.3.3 Current Status of CDPF Technology

   More than one emission control manufacturer is developing CDPFs.  In field trials, they have
demonstrated highly efficient PM control and promising durability.  A recent publication
documents results from a sample of these field test engines after years of use in real-world
applications.37 The sampled CDPFs had on average four years of use covering more than
225,000 miles in applications ranging from city  buses to garbage trucks to intercity trains, with
some units accumulating more than 360,000 miles. When tested on the highway FTP cycle, they
continued to demonstrate PM reductions in excess of 90 percent.

   Another program evaluating CDPFs in the field is the ARCO Emission Control Diesel  (EC-
D) program.11 In that program, a technology validation is being run to evaluate EC-D and
CDPFs using diesel vehicles  operating  in southern California. The fuel's performance, impact
   H EC-D is a diesel fuel developed recently by ARCO (Atlantic Richfield Company) from typical crude oil using
a conventional refining process and having a fuel sulfur content less than 15 ppm.

                                          4-15

-------
Regulatory Impact Analysis
on engine durability and vehicle performance, and emission characteristics are being evaluated
in several fleets in various applications.  The program is still ongoing, but interim results have
been made available.38 These interim results have shown that vehicles retrofitted with CDPFs
and fueled with EC-D (7.4 ppm sulfur) emitted 91 percent to 99 percent less PM compared with
the vehicles fueled with California diesel fuel (121 ppm sulfur) having no exhaust filter
equipment.  Further, the test vehicles equipped with the CDPFs and fueled with EC-D have
operated reliably during the program start-up period and no significant maintenance issues have
been reported for the school bus, tanker truck and grocery truck fleets that have been operating
for over six months (approximately 50,000 miles).39 These results from on-highway diesel
engines are  significant because in form and function the engines are virtually the same as those
used for nonroad diesel applications. In fact, in many cases on-highway diesel engines have
directed nonroad counterparts that are virtually identical. Further, even for nonroad engines
which may differ in physical size  or horsepower range, the underlying chemistry and filtration
efficiency of CDPFs is the same.

   Even with the relatively mature state of the CDPF technology, progress is still being made to
improve catalytic-based soot regeneration technologies and to develop system solutions to
ensure that even under the most extreme conditions soot regeneration can be ensured.
Improvements in catalytic soot oxidation are important because more active soot oxidation can
help to improve fuel economy  and to ensure robust soot regeneration. A PM filter with a more
effective soot oxidation catalyst would be expected to have a lower average soot loading and
therefore would be less restrictive to exhaust flow, thus decreasing the pressure drop across the
PM filter and leading to better fuel economy.  Improved effectiveness in oxidizing soot will also
further ensure that excessive soot loading that might lead to PM filter failure will not occur.

   A paper presented at a recent conference of the Society of Automotive Engineers (SAE)
documented design improvements in catalyzed diesel particulate filters with improved soot
oxidation effectiveness.  The paper showed that changes in where catalytic materials were coated
within a PM filter system (on an upfront flow-through catalyst, on the surface of the PM filter or
a combination of both) influenced the effectiveness of the catalyst material to promote soot
oxidation.40 This kind of system analysis suggests that there remain opportunities to further
improve how diesel particulate filters are designed to promote soot oxidation and that different
solutions may be chosen dependent upon expected nonroad equipment operation (expected
exhaust temperature history), packaging constraints and cost.

   Alhough highly effective catalytic soot oxidation, enabled by clean diesel fuel (15 ppm S),
suggests that PM filters will regenerate passively for most vehicle and many nonroad  equipment
applications, there remains the possibility that for some conditions active regeneration systems
(backup systems) may be desirable.1 For this reason, some vehicle manufacturers have
   1 We are defining backup regeneration to include any number of methods for raising exhaust temperatures in
order to promote PM filter regeneration. These could include changes to engine management to change engine
operation and raise exhaust temperature, any external mechanism to add heat into the exhaust, or a combination of
engine management to increase hydrocarbon (fuel) emissions from the engine in order to oxidize those emissions
across a diesel oxidation catalyst (DOC) and thus raise exhaust temperatures.

                                           4-16

-------
	Technologies and Test Procedures for Low-Emission Engines

developed systems to help ensure that PM soot regeneration can occur under all conditions.  One
example of this is a current production product sold in Europe by PSA/Peugeot. On diesel
powered Peugeot 607 passenger cars (a Ford Taurus-sized passenger car) a PM filter system is
installed that includes mechanisms for engine-promoted soot oxidation.  The vehicle estimates
soot loading from several parameters, including exhaust backpressure and can periodically
promote more rapid soot oxidation by injecting additional fuel late in the combustion cycle.  This
fuel is injected so late in the cycle that it does not contribute to engine power but instead is
combusted (oxidized) across an oxidation catalyst in front of the PM filter. The combustion of
the fuel across the catalyst increases the exhaust temperature substantially, encouraging rapid
soot oxidation. Peugeot has sold more than 400,000 passenger cars with this technology and
expects to expand the use of the system  across all of its diesel vehicle lines.41 Other European
vehicle manufacturers indicated to EPA during our progress review, that they intend to introduce
similar technologies in the near future. They noted that this was not driven by regulation but by
customer demand for clean diesel technologies.  The fact that manufacturers  are introducing PM
filter technologies in advance of mandatory regulations suggests that the technology is well
developed and mature.

   The potential for synergistic benefits to the application of both PM filters and NOx adsorbers
was highlighted in the HD2007 Regulatory Impact Analysis, but at that time  little was known as
to the extent of these synergistic benefits.42 Toyota has developed a combined  diesel particulate
filter and NOx adsorber technology dubbed DPNR (Diesel Particulate NOx Reduction). The
mechanism for synergistic PM soot regeneration with programmed NOx regeneration was
recently documented by Toyota in a SAE publication. The paper showed that active oxygen
molecules created both  under lean conditions as part of the NOx storage function and under rich
conditions created by the NOx regeneration function were effective at promoting soot oxidation
at low temperatures.43  This suggests that the combination of a NOx adsorber catalyst function
with a diesel particulate filter can provide a more robust soot regeneration system than a PM
filter-only technology.

   4.1.1.3.4 CDPF'Maintenance

   Inorganic solid particles present in diesel exhaust can be captured by diesel particulate filters.
Typically these inorganic materials are metals derived from engine oil, diesel fuel or even engine
wear.  Without a PM filter these materials are normally exhausted from the engine as diesel PM.
While the PM filter is effective at capturing inorganic materials it is not  typically effective at
removing them, since they do not tend to be oxidized into a gaseous state (carbon soot is
oxidized to CO2 which can easily pass through the PM filter walls). Because these inorganic
materials are not typically combusted and remain after the bulk of the PM is  oxidized from the
filter they are typically referred to as ash. While filtering metallic ash from the exhaust is an
environmental benefit of the PM filter technology it also creates a maintenance need for the PM
filter to remove the ash from the filter periodically.

   The maintenance function for the removal of ash is relatively straightforward, and itself does
not present a technical challenge for the industry.  We have estimated cost for ash removal as
one of the costs of this rule (see RIA Chapter 6).  However, both the industry and EPA would

                                          4-17

-------
Regulatory Impact Analysis
like to see ash-related PM filter maintenance reduced as much as possible. EPA has specific
guidelines for acceptable maintenance intervals for nonroad diesel engines with CDPFs intended
to ensure robust emission-control technologies (3,000hrs for engines <175 hp and 4,500hrs for
engines >175hp). Nonroad engine manufacturers are similarly motivated to improve reliability
to minimize end-user maintenance costs. The issue of ash accumulation was raised consistently
during our progress review visits with the industry. The industry is investigating several ways to
address this issue including means to improve ash tolerance and to reduce the amount of ash
present in diesel exhaust.

   For most current PM filter designs ash accumulates at the end of the inlet passages of the PM
filter. As more ash is accumulated, the effective filter size is reduced because the ash fills the
end of the passage  shortening the effective filter length.  Increasing PM filter size to tolerate
higher levels of ash accumulation  is one simple approach to address ash.  This approach, though
effective, is undesirable due to the added cost  and size of the resulting PM filter.  Several
companies are investigating means to develop PM filter mechanisms that are more ash-tolerant.
These approaches include concepts to increase storage area within the filter itself and concepts
that promote self-cleaning of the filter, perhaps driven by engine and vehicle vibrations during
normal vehicle operation. Our recent Highway Diesel Progress Review Report 2 described two
such systems recently introduced for on-highway applications. For light-duty vehicle
applications the technologies are described as  fit for life, meaning that ash cleaning maintenance
will not be necessary over the life of a light-duty diesel vehicle. For heavy-duty diesel engines
(and for nonroad diesel engines >250 hp) the technologies are expected to increase the interval
between ash cleaning by 50 percent.

   In addition to concepts to improve ash handling, possibilities exist to decrease the amount of
ash present in diesel exhaust. The predominant source of ash in diesel exhaust is inorganic
materials contained in engine oil (oil ash). A significant fraction of the ash in engine oil is from
additives necessary to control acidification of engine oil due in part to sulfuric acid  derived from
sulfur in diesel fuel. As the sulfur content of diesel fuel is decreased, the need for additives to
neutralize the acids in engine oil should also decrease.  The concept of an engine oil with less
ash content is often referred to as "low-ash oil." Several technical programs are ongoing to
determine the impact of changes in oil ash content and other characteristics of engine oil on
exhaust emission-control technologies and engine wear and performance. Historically, as engine
technologies have changed (often  due to changes in emission regulations) engine oil
formulations have also changed.  These changes have been accomplished through industry
consensus on oil specifications based on defined test protocols.  This process of consensus
definition has begun to develop engine oils specifications for highway diesel engines for the
2007 model year. This engine oil  will also be appropriate for application to nonroad diesel
engine designed with the same technologies (i.e., an engine oil specification designed for
highway HD2007 engines would also be appropriate for use with Tier 4 engines).

   It may also be possible to reduce the ash level in diesel exhaust by reducing oil consumption
from diesel engines. Diesel engine manufacturers over the years have reduced engine oil
consumption to reduce PM emissions and to reduce operating costs for engine owners.  Further
improvements in oil consumption  may be possible to reduce ash accumulation rates in PM

                                           4-18

-------
	Technologies and Test Procedures for Low-Emission Engines

filters. If oil accumulation rates could be halved and engine oil ash content similarly decreased,
the PM filter maintenance interval would be increased four-fold.  Current retrofit PM filter ash
maintenance intervals can range from 50k miles to more than 200k miles.44

4.1.2 NOx Control Technologies

   Oxides of nitrogen (NO and NO2, collectively called NOx) are formed at high temperatures
during the diesel combustion process from nitrogen and oxygen present in the intake air.  The
NOx formation rate is exponentially related to peak cylinder temperatures and is also strongly
related to nitrogen and oxygen content (partial pressures).  NOx control technologies for diesel
engines have focused on reducing emissions by lowering the peak cylinder temperatures and by
decreasing the oxygen content of the intake air.

   4.1.2.1  In-Cylinder NOx Control Technologies

   Several technologies have been developed to accomplish these objectives, including fuel-
injection timing retard, fuel-injection rate control, charge air cooling, exhaust gas recirculation
(EGR) and cooled EGR. The use of these technologies can result in significant reductions in
NOx emissions, but are limited due to practical and physical constraints of heterogeneous diesel
combustion.45

   Our recent Highway Diesel Progress Review Report 2, investigated the extent to which in-
cylinder NOx control technologies had advanced. The report noted that a number of diesel
engine manufacturers introduced cooled EGR systems on their heavy-duty diesel engines in
2002 compliant with the 2004 emission standards for NOx and NMHC of 2.5 g/bhp-hr.  The
engines circulate a portion of the exhaust gases through a heat exchanger cooling the exhaust
before reintroducing the gases into the engine intake manifold. The systems control NOx
emissions by providing a diluent (spent exhaust gases) reducing the oxygen content of the intake
air and recirculated exhaust mixture. Engine manufacturers have now demonstrated that these
systems can be further refined to allow NOx emissions compliant with the 2007 NOx averaging
level of approximately 1.2 g/bhp-hr. To reduce NOx emissions below 1.2 g/bhp-hr engine
manufacturers will likely need to increase EGR rates (use higher levels of EGR), thus we are
referring to such refinements for on-highway 2007 diesel engines as high flow EGR. Although
there are nonroad specific challenges to applying similar technologies to nonroad diesel engines
(most notably the lack  of ram-air for cooling), the fundamental NOx control technologies are
applicable to all diesel  engines. We are confident based on the continuing development of on-
highway  technologies for in-cylinder NOx control using cooled EGR or ACERT™ that nonroad
diesel engines between 25 and 75 hp and mobile machine nonroad engines >750 hp will be able
to comply with their respective Tier 4 standards (i.e., 3.5 g/bhp-hr NOx+NMHC for 25-50 hp
engines, the same standard certified using the NRTC and NTE for 50-75 hp engines, and 2.6
g/bhp-hr NOx for >750 hp engines), including the NRTC (with cold-start) and the NTE
standards all of which are similar in difficulty to the heavy-duty FTP (with cold-start) and the
NTE standards for on-highway engines. For additional discussion of these emission control
technologies and the impact of the NTE and cold-start, see the RIA for the on-highway FID 2004
emission standards and the RIA for the Tier 2/3 emission standards.46'47

                                         4-19

-------
Regulatory Impact Analysis
   A new form of diesel engine combustion, commonly referred to as homogenous diesel
combustion or premixed diesel combustion, can give very low NOx emissions over a limited
range of diesel engine operation.  In the regions of diesel engine operation over which this
combustion technology is feasible (light-load conditions), NOx emissions can be reduced enough
to comply with the 0.3 g/hp-hr NOx emission standard.48  Some engine manufacturers are
already producing engines that utilize this technology over a narrow range of engine operation.49
Unfortunately, it is not possible today to apply this technology over the full range of diesel
engine operation. We believe that more engine manufacturers will utilize this alternative
combustion approach in the limited range over which it is effective, but will have to rely on
conventional heterogenous diesel combustion for the bulk of engine operation.  See Section
4.1.1.1 for additional discussion of homogenous diesel combustion and PM emission control.

   4.1.2.2  Lean-NOx Catalyst Technology

   Lean-NOx catalysts have been under  development for some time, and two methods have
been developed for using a lean NOx catalyst depending on the level of NOx reduction desired
though neither method can produce more than a 30 percent NOx reduction. The "active" lean-
NOx catalyst injects a reductant that serves to reduce NOx to N2 and O2 (typically diesel fuel is
used as the reductant). The reductant is introduced upstream of, or into, the catalyst. The
presence of the reductant provides locally oxygen-poor  conditions that allow the NOx emissions
to be reduced by the catalyst.

   The lean-NOx catalyst washcoat incorporates a zeolite catalyst that acts to adsorb
hydrocarbons from the exhaust stream. Once adsorbed  on the zeolite,  the hydrocarbons will
oxidize and create a locally oxygen-poor  region that is more conducive to reducing NOx. To
promote hydrocarbon oxidation at lower temperatures, the washcoat can incorporate platinum or
other precious  metals. The platinum also helps to eliminate the emission of unburned
hydrocarbons that can occur if too much reductant is injected, referred to as "hydrocarbon slip."
With platinum, the NOx conversion can take place at the low exhaust temperatures that are
typical of diesel engines. However, the presence of the  precious metals can lead to production of
sulfate PM, as  already discussed for PM control technologies.

   Active lean-NOx catalysts have been  shown to provide up to 30 percent NOx reduction
under limited steady-state conditions.  However, this NOx control is achieved with a fuel
economy penalty upwards of 7 percent due to the need to inject fuel into the exhaust stream.50
NOx reductions over the transient highway FTP cycle are only on the order of 12 percent due to
excursions outside the optimum NOx reduction efficiency temperature range for these devices.51
Consequently,  the active lean-NOx catalyst does not appear to be capable of enabling the
significantly lower NOx emissions required by the Tier  4 NOx standards.

     The "passive" lean-NOx catalyst uses no reductant injection.  The passive lean-NOx
catalyst is therefore even more limited in its ability to reduce NOx because the exhaust gases
normally contain very few hydrocarbons. For that reason, today's passive lean-NOx catalyst is
capable of best steady-state NOx reductions of less than 10 percent. Neither approach to lean-
                                          4-20

-------
	Technologies and Test Procedures for Low-Emission Engines

NOx catalysis listed here can provide the significant NOx reductions necessary to meet the Tier
4 standards.

    4.1.2.3 NOx Adsorber Technology

    NOx emissions from gasoline-powered vehicles are controlled to extremely low levels
through the use of the three-way catalyst technology first introduced in the 1970s. Three-way-
catalyst technology is very efficient in the stochiometric conditions found in the exhaust of
properly controlled gasoline-powered vehicles.  Today, an advancement upon this well-
developed three-way catalyst technology, the NOx adsorber, has shown that it too can make
possible extremely low NOx emissions from lean-burn engines such as diesel engines/ The
potential of the NOx adsorber catalyst is limited only by its need for careful integration with the
engine and engine control system (as was done for three-way catalyst equipped passenger cars in
the 1980s  and 1990s) and by poisoning of the catalyst from sulfur in the fuel. The Agency  set
stringent new NOx standards for highway diesel engines beginning in 2007 predicated upon the
use of the  NOx adsorber catalyst enabled by significant reductions in fuel sulfur levels (15 ppm
sulfur or less). The final rule includes similarly stringent NOx emission standards for nonroad
engines from 75-750 hp and for certain engines >750 hp, again based on using technology
enabled by a reduction in fuel sulfur levels.

    NOx adsorbers work to control NOx emissions by storing NOx on the surface of the catalyst
during the lean engine operation typical of diesel engines. The adsorber then undergoes
subsequent brief rich regeneration events where the NOx is released and reduced across
precious-metal catalysts. The NOx storage period can be as short as 15 seconds and as along as
10 minutes depending upon engine-out NOx emission rates and exhaust temperature. Several
methods have been developed to accomplish the necessary brief rich exhaust conditions
necessary  to regenerate the NOx adsorber technology including late-cycle fuel injection, also
called post injection, in exhaust fuel injection, and  dual bed technologies with off-line
regeneration.52'53'54 This method for NOx control has been shown to be highly effective when
applied to diesel engines but has some technical challenges associated with it. Primary among
these is sulfur poisoning of the catalyst, as described in Section 4.1.2.3.4.2 below.

    4.1.2.3.1 How do NOx Adsorbers Work?

    As noted, the NOx adsorber catalyst is a further development of the three-way catalyst
technology developed for gasoline powered vehicles more than twenty years ago.  The NOx
adsorber enhances the three-way catalyst function through the addition of storage materials on
the catalyst surface that can adsorb NOx under oxygen-rich conditions. This enhancement
means that a NOx adsorber can allow for control of NOx emissions under lean-burn (oxygen-
rich) operating conditions typical of diesel engines.
   J NOx adsorber catalysts are also called, NOx storage catalysts (NSCs), NOx storage and reduction catalysts
(NSRs), and NOx traps.

                                          4-21

-------
Regulatory Impact Analysis
   Three-way catalysts reduce NOx emissions as well as HC and CO emissions (hence the name
three-way) by promoting oxidation of HC and CO to water and CO2 using the oxidation potential
of the NOx pollutant, and, in the process, reducing the NOx emissions to atomic nitrogen, N2.
Said another way, three-way catalysts work with exhaust conditions where the net oxidizing and
reducing chemistry of the exhaust is approximately equal, allowing the catalyst to promote
complete oxidation/reduction reactions to the desired exhaust components, carbon dioxide
(CO2), water (H2O) and nitrogen (N2).  The oxidizing potential in the exhaust comes from NOx
emissions and from oxygen (O2) that is not consumed during combustion. The reducing
potential in the exhaust comes from HC and CO emissions, which are products of incomplete
combustion. Operation of the engine to ensure that the oxidizing and reducing potential of the
combustion and exhaust conditions is precisely balanced is referred to as stoichiometric engine
operation.

   If the exhaust chemistry varies from stoichiometric conditions emission control is decreased.
If the exhaust chemistry is net "fuel-rich," meaning there is an excess of HC and CO emissions
in comparison to the oxidation potential of the NOx and O2 present in the exhaust, the excess HC
and CO pollutants are emitted from the engine.  Conversely, if the exhaust chemistry is net
"oxygen-rich" (lean-burn), meaning there is an excess of NOx and O2 in comparison to the
reducing potential of the HC and CO present in the exhaust, the excess NOx pollutants are
emitted from the engine. It is this oxygen-rich operating condition that typifies diesel engine
operation. Because of this, diesel engines equipped with three-way catalysts (or simpler
oxidation catalysts) have very low HC and CO emissions while NOx (and O2) emissions remain
almost unchanged from the high engine-out emission levels. For this reason, when diesel
engines are equipped with catalysts (diesel oxidation catalysts (DOCs)) they have HC and CO
emissions that are typically lower, but have NOx emissions that are an order of magnitude
higher, than for gasoline engines equipped with three-way catalysts.

   The NOx adsorber catalyst works to overcome this situation by storing NOx emissions when
the exhaust conditions are oxygen-rich.  Unfortunately the storage capacity of the NOx adsorber
is limited, requiring that the stored NOx be periodically purged from the storage component. If
the exhaust chemistry is controlled such that when the stored NOx emissions are released the net
exhaust chemistry is at stoichiometric or net fuel-rich conditions, then the three-way catalyst
portion of the catalyst can reduce the NOx emissions in the same way as for a gasoline three-way
catalyst equipped engine.  Simply put, the NOx adsorber works to control NOx emissions by
storing NOx on the catalyst surface under lean-burn conditions typical of diesel engines and then
by reducing the NOx emissions with a three-way catalyst function by periodically operating
under stoichiometric or fuel-rich conditions.

   The NOx storage process can be further broken down into two steps. First the NO in the
exhaust is oxidized to NO2 across an oxidation promoting catalyst, typically platinum.  Then the
NO2 is further oxidized and stored on the surface of the catalyst as a metallic nitrate (MNO3).
The storage components are typically alkali or alkaline earth metals that can form stable metallic
nitrates.  The most common storage component is barium carbonate (BaCO3), which can store
NO2 as barium nitrate (Ba(NO3)2) while releasing CO2. For the NOx storage function to work,
the NOx must be oxidized to NO2 prior to storage and a storage site must be available (the

                                         4-22

-------
	Technologies and Test Procedures for Low-Emission Engines

device cannot be "full"). During this oxygen-rich portion of operation, NOx is stored while HC
and CO emissions are oxidized across the three-way catalyst components by oxygen in the
exhaust.  This can result in near zero emissions of NOx, HCs, and CO under the net oxygen-rich
operating conditions typical of diesel engines.

   The NOx adsorber releases and reduces NOx emissions under fuel-rich operating conditions
through a similar two step process, referred to here as NOx adsorber regeneration. The metallic
nitrate becomes unstable under net fuel-rich operating conditions, decomposing and releasing the
stored NOx. Then the NOx is reduced by reducing agents in the exhaust (CO and HCs) across a
three-way catalyst system, typically containing platinum and rhodium. Typically, this NOx
regeneration step occurs at a significantly faster rate than the period of lean-NOx storage such
that the fuel-rich operation constitutes only a small fraction of the total operating time. Since
this release and reduction step, NOx adsorber regeneration, occurs under net fuel-rich operating
conditions, NOx emissions can be almost completely eliminated. But for some of the HC and
CO emissions, "slip"(failure to remove all of the HC and CO) may occur during this process.
The HC and CO slip can be controlled with a downstream "clean-up" catalyst that promotes their
oxidation or potentially by controlling the  exhaust constituents such that the excess amount of
the HC and CO pollutants at the fuel-rich operating condition is as low as possible, that is, as
close to stoichiometric conditions as possible.

   The difference between stoichiometric three-way catalyst function and the newly developed
NOx adsorber technology can be summarized as follows.  Stoichiometric three-way catalysts
work to reduce NOx, HCs and CO by  maintaining a careful balance between oxidizing (NOx and
O2) and reducing (HCs and CO) constituents and then promoting their mutual destruction across
the catalyst on a continuous basis.  The newly developed NOx adsorber technology works to
reduce the pollutants by balancing the oxidation and reduction chemistry on a discontinuous
basis, alternating between net  oxygen-rich and net fuel-rich operation to control the pollutants.
This approach allows lean-burn engines (oxygen-rich operating), like diesel engines, to operate
under their normal operating mode most of the time, provided that they can  periodically switch
and operate such that the exhaust conditions are net fuel-rich for brief periods.  If the
engine/emission-control system can be made to operate in this manner, NOx adsorbers offer the
potential to employ the highly effective  three-way catalyst chemistry to lean-burn engines.

   4.1.2.3.2 NOx Adsorber Regeneration Mechanisms

   NOx adsorbers work to control NOx emissions by storing the NOx pollutants on the catalyst
surface during oxygen-rich engine operation (lean-burn engine operation) and then by
periodically releasing and reducing the NOx emissions under fuel-rich exhaust conditions.  This
approach to controlling NOx emissions can work for a diesel engine provided that the engine and
emission-control system can be designed to work in concert, with relatively long periods of
oxygen-rich operation (typical diesel engine operation) followed by brief periods of fuel-rich
exhaust operation. The ability to control the NOx emissions in this manner  is the production
basis  for lean-burn NOx emission control in stationary power systems and for lean-burn gasoline
engines.  As outlined below, we believe there are several approaches to accomplish the required
periodic operation on a diesel  engine.

                                          4-23

-------
Regulatory Impact Analysis
   The most frequently mentioned approach for controlling the exhaust chemistry of a diesel
engine is through in-cylinder changes to the combustion process.  This approach roughly mimics
the way in which lean-burn gasoline engines function with NOx adsorbers.  That is, the engine
itself changes in operation periodically between "normal" lean-burn (oxygen-rich) combustion
and stoichiometric or even fuel-rich combustion to promote NOx control with the NOx adsorber
catalyst.  For diesel engines this approach typically requires  the use of common rail fuel systems,
which allow for multiple fuel-injection events, along with an air handling system that includes
exhaust gas recirculation (EGR).

   The normal lean-burn engine operation can last from as little time as 15  seconds to more than
three minutes as the exhaust NOx emissions are stored on the surface of the NOx adsorber
catalyst.  The period of fuel-lean, oxygen-rich, operation is determined by the NOx emission rate
from the  engine and the storage capacity of the NOx adsorber. Once the NOx adsorber catalyst
is full (once  an unacceptable amount of NOx is slipping through the catalyst without storage) the
engine must switch to fuel-rich operation to regenerate the NOx adsorber.

   The engine typically changes to fuel-rich operation by increasing the EGR rate, by throttling
the fresh  air  intake, and by introducing an additional fuel-injection event late in the combustion
cycle.  The increased EGR rate works to decrease the oxygen content of the intake air by
displacing fresh air that has a high oxygen content with exhaust gases that have a much lower
oxygen content.  Intake air throttling further decreases the amount of fresh air in the intake gases
again lowering the amount of oxygen entering the combustion chamber. The combination of
these first two steps serves to lower the oxygen concentration in the combustion chamber,
decreasing the amount of fuel required to reach a fuel-rich condition. The fuel is metered then
into the combustion chamber in two steps under this mode of operation.  The first, or primary,
injection  event meters a precise amount of fuel to deliver the amount of torque (energy) required
by the operator demand (accelerator pedal input). The second injection event is designed to
meter the amount of fuel necessary to achieve a net fuel-rich operating condition. That is, the
primary plus secondary injection events introduce an excess of fuel when compared with the
amount of oxygen in the combustion chamber. The secondary injection event occurs very late in
the combustion cycle, so it does not generate additional torque. This is necessary so the
switching between the normal lean-burn operation and this periodic fuel-rich operation is
transparent to the user.

   Additional ECM capability will be necessary to monitor the NOx adsorber and determine
when the NOx regeneration events are necessary. This can be done in a variety of ways, though
they fall into two general categories:  predictive and reactive. First, the predictive method
estimates or  measures the NOx flow into the adsorber in conjunction with the predicted adsorber
performance to determine when the adsorber is near capacity.  Then, upon entering optimal
engine operating conditions, the system performs a NOx regeneration. This particular step is
similar to an on-board diagnostic (OBD) algorithm waiting for proper conditions to perform a
functionality check. During the NOx regeneration, sensors determine how accurately the
predictive algorithm performed, and adjust it accordingly. Second, the reactive method is
envisioned to monitor NOx downstream of the NOx adsorber and detection  of NOx slippage
triggers a regeneration event. This method is dependent on good NOx-sensor technology.  This

                                          4-24

-------
	Technologies and Test Procedures for Low-Emission Engines

method also depends on the ability to regenerate under any given engine operating condition,
since the algorithm reacts to indications that the adsorber had reached its NOx storage capacity.
In either case, we believe these algorithms are not far removed from the systems that will be
used by nonroad manufacturers to meet Tier 3 emission standards and will be virtually identical
to the systems used by highway engine manufacturers to comply with the HD2007 emission
regulations.  When used in combination with the sophisticated control systems that will be
available, we expect that NOx regeneration events can be seamlessly integrated into engine
operation such that the operator may not be aware that the events are taking place.

   Using this approach of periodic switching between normal lean-burn operation and brief
periods of fuel-rich operation all accomplished within the combustion chamber of a diesel engine
is one way in which an emission-control system for a diesel engine can be optimized to work
with the NOx adsorber catalyst.  This approach requires no new engine hardware beyond the air
handling and advanced common rail fuel systems that many advanced diesel engines will have
already applied to meet the Tier 3 NOx standard.  For this reason an in-cylinder approach is
likely to appeal to engine manufacturers for product lines where initial purchase cost or package
size is the most important factor in determining engine purchases.

   Another approach to accomplish the NOx adsorber regeneration is through the use of a so-
called "dual-bed" or "multiple-bed" NOx adsorber catalyst system. Such a system is designed so
the exhaust flow can be partitioned and routed through two or more catalyst "beds" operating in
parallel. Multiple-bed NOx adsorber catalysts restrict exhaust flow to part of the catalyst during
its regeneration. By doing so, only a portion of the exhaust flow need be made rich, reducing
dramatically the amount of oxygen needing to be depleted and thus the fuel required to be
injected to generate a rich exhaust stream. One simple example of a multiple bed NOx adsorber
is the dual-bed system in Figure 4.1-1. In this example, the top half of the adsorption catalyst
system is regenerating under a low exhaust flow condition (exhaust control valve nearly closed),
while the remainder of the exhaust flow is bypassed to a lower half of the  system.  A system of
this type has the following characteristics:

   •   Half of the system operates with a major flow in an "adsorption mode," where most of
       the exhaust is well lean of stoichiometric (A > 1 or »1, typical diesel exhaust), NO is
       converted to NO2 over a Pt-catalyst, and stored as a metallic nitrate within the NOx
       adsorbent material.K

       The other half of the system has its exhaust flow restricted to just a small fraction (~5
       percent) of the total flow and operates in a regeneration mode.
          - While the flow is restricted for regeneration, a small quantity of fuel is sprayed into
          the regenerating exhaust flow at the beginning of the regeneration event.
          - The fuel is oxidized by the  oxygen in the exhaust until sufficient oxygen is depleted
          for the stored NOx to be released. This occurs at exhaust conditions of A  < 1.
   K A condition of A = 1 means that there are precisely the needed quantity of reactants for complete reaction at
   ilibrium. A < 1 means that there is insufficient oxygen, A > 1 means that there is excess oxygen.
equilibrium.

                                          4-25

-------
Regulatory Impact Analysis
          - At these conditions, NOx can also be very efficiently reduced to N2 and O2 over a
          precious-metal catalyst.

       At the completion of regeneration, the majority of the flow can then be reintroduced into
       the regenerated half of the system by opening the flow control valve.

       Simultaneously, flow is restricted to the other half of the system to allow it to regenerate.

                                      Figure 4.1-1
      Schematic Representation of the Operation of a Dual-Bed NOx Adsorption Catalyst
                                                Secondary
                                               Fuel Injector
                                                   (on)
  NOx
Adsorber
^

Flow-

                        Partially Closed
                        Exhaust-flow
                                                 Exhaust Flow
Fully Open /
Exhaust-flow U
Control Valve
/
^. '
NUX H 1 Flow
Adsorber 1 -•


r-i-S
/ Secondary
J~cT Fuel Injector
1 Diesel Engine

   Although the schematic shows two separate systems, the diversion of exhaust flow can occur
within a single catalyst housing, and with a single catalyst monolith.  There may also be
advantages to using more than one partition for the NOx adsorber system such as the use of
multiple beds allows desulfation of one bed while normal NOx adsorption and regeneration
events occur in other beds.

   The NOx adsorber performance  can be enhanced by incorporating a catalyzed diesel
particulate filter (CDPF) into the system.  A number of synergies exist between NOx adsorber
systems and CDPFs. Both systems rely on conversion of NO to NO2 over a Pt catalyst for part
of their functioning. Partial oxidation reforming of diesel fuel to hydrogen and CO over a Pt-
catalyst has been demonstrated for fuel-cell applications. A similar reaction to reform the fuel
upstream of the NOx adsorber during regeneration provides a more reactive reductant for
desorption and reduction of NOx. Heavier fuel hydrocarbons are known to inhibit NOx
reduction on the NOx adsorption catalyst since competitive adsorption by hydrocarbons on the
precious-metal sites inhibits NOx reduction during adsorber regeneration.55 Partial oxidation of
the secondary fuel injected into the exhaust during regeneration could lead to sooting of the fuel.
Using a CDPF upstream of the NOx adsorber, but downstream of the secondary fuel injection,
allows partial oxidation of the fuel hydrocarbons to occur over the Pt catalyst on the surface of
                                          4-26

-------
	Technologies and Test Procedures for Low-Emission Engines

the CDPF.  The wall-flow design of the CDPF efficiently captures any soot formed during partial
oxidation of the fuel injected into the exhaust, preventing any increase in soot emissions.  The
partial oxidation reaction over the CDPF is exothermic, which can be used increase the rate of
temperature rise for the NOx adsorber catalyst after cold starts, similar to the use of light-off
catalysts with cascade three-way catalyst systems.56

   4.1.2.3.3 How Efficient are Diesel NOx Adsorbers ?

   Research into applying the NOx adsorber catalyst to diesel exhaust is only a few years old
but benefits from the larger body of experience with stationary power sources and with lean-burn
gasoline systems. In simplest terms the question is how well does the NOx adsorber store NOx
under normal lean-burn diesel engine operation, and then how well does the control system
perform the NOx regeneration function. Both of these functions are affected by the temperature
of the exhaust and of the catalyst surface.  For this reason efficiency is often discussed as a
function of exhaust temperature under steady-state conditions.  This is the approach used in this
section and is extended in Section 4.1.3.1.2 below to predict the effectiveness of the NOx
adsorber technology when engines operate over the new transient duty cycles.  The potential for
both NOx storage and reduction to operate at very high efficiencies can be realized through
careful emission-control system design, as described below.

   The NOx storage function consists of oxidation of NO to NO2 and then storage of the NOx as
a metallic nitrate on the catalyst surface.  The effectiveness of the catalyst at accomplishing these
tasks is dependent upon exhaust temperature, catalyst temperature, precious-metal dispersion,
NO storage volume, and transport time (mass flow rates through the catalyst).  Taken as a whole,
these factors determine how effectively a NOx adsorber-based control system can store NOx
under lean-burn diesel engine operation.

   Catalyst and exhaust temperature are important because the rate at which the desirable
chemical reactions occur is a function of the local temperature where the reaction occurs. The
reaction rate for NO to NO2 oxidation and for NOx storage increases with increasing
temperature.  Beginning at temperatures as low as 100°C NO oxidation to NO2 can be promoted
across a platinum catalyst at a rate high enough to allow for NOx storage to occur. Below 100°C
the reaction can still occur (as it does in the atmosphere);  however, the reaction rate is so  slow as
to make NOx storage ineffective below this temperature in a mobile source application. At
higher exhaust temperatures, above 400°C, two additional mechanisms affect the ability of the
NOx adsorber to store NOx. First the NO to NO2 reaction products are determined by an
equilibrium reaction that favors NO rather than NO2.  That is across the oxidation catalyst, NO is
oxidizing to form NO2 and NO2 is decaying to form NO at a rate that favors a larger fraction of
the gas being NO rather than NO2.  As this is an equilibrium reaction when the NO2 is removed
from the gas stream by storage on the catalyst surface, the NOx gases quickly "re-equilibrate"
forming more NO2.  This removal of NO2 from the gas stream and the rapid oxidation of NO to
NO2 means that in spite of the NO2 fraction of the NOx gases in the catalyst being low at
elevated conditions (30 percent at 400°C) the storage of NOx can continue to occur with high
efficiencies, near 100 percent.
                                          4-27

-------
Regulatory Impact Analysis
   Unfortunately, the other limitation of high-temperature operation is not so easily overcome.
The metallic nitrates that are formed on the catalyst surface and that serve to store the NOx
emissions under fuel-lean operating conditions can become unstable at elevated temperatures.
That is, the metallic nitrates thermally decompose releasing the stored NOx under lean operating
conditions allowing the NOx to exit the exhaust system "untreated."  The temperature at which
the storage metals begin to thermally release the stored NOx emissions varies dependent upon
the storage metal or metals used, the relative ratio of the storage metals, and the washcoat
design. Changes to catalyst formulations can change the upper temperature threshold for thermal
NOx desorption by as much as 100°C.57 Thermal stability is the primary factor determining the
NOx control efficiency of the NOx adsorber at temperatures higher than 400-50CTC. NOx
adsorber  catalyst developers are continuing to work to improve this aspect of NOx adsorber
performance, and as documented in EPA's 2002 Highway Progress Review improving
temperature performance is being realized.

   The NOx adsorber catalyst releases stored NOx emissions under fuel-rich operating
conditions and then reduces the NOx over a three-way catalyst function. While the NOx storage
function determines the NOx control efficiency during lean operation, it is the NOx release and
reduction function that determines the NOx control efficiency during NOx regeneration.  Since
NOx storage can approach near 100 percent effectiveness for much of the temperature range of
the diesel engine, the NOx reduction function often determines the overall NOx control
efficiency.

   NOx  release can occur under relatively cool exhaust temperatures even below 200°C for
current NOx adsorber formulations. Unfortunately, the three-way NOx reduction function is not
operative at such cool exhaust temperatures. The lowest temperature at which a chemical
reaction is promoted at a defined efficiency (often 50 percent) is referred to as the "light-off
temperature.  The 80 percent light-off temperature for the three-way catalytic NOx reduction
function of current NOx adsorbers is between 200°C and 250°C. Even though NOx storage and
release can occur at cooler temperatures, NOx control is therefore limited under steady-state
conditions to temperatures greater than this light-off temperature.

   Under transient operation, however, NOx control can be accomplished at temperatures below
this NOx reduction light-off temperature provided that the period of operation at the lower
temperature is preceded by operation at higher temperatures and provided that the low-
temperature operation does not continue for an extended period.  This NOx  control is possible
due to two characteristics of the system specific to transient operation. First, NOx control can be
continued below the light-off temperature because storage can continue below that temperature.
If the exhaust temperature again rises above the NOx reduction light-off temperature before the
NOx adsorber storage function is full, the NOx reduction can then precede at high efficiency.
Said another way, if the excursions to very low temperatures are brief enough, NOx storage can
proceed under this mode of operation, followed by NOx reduction when the exhaust
temperatures are above the light-off temperature. Although this sounds like a limited benefit
because NOx storage volume is limited, in fact it can be significant, because the NOx emission
rate from the engine is low at low temperatures. While the NOx storage rate may be limited such
that at high-load conditions the lean-NOx storage period is as short as 30 seconds, at the very

                                          4-28

-------
	Technologies and Test Procedures for Low-Emission Engines

low NOx rates typical of low-temperature operation (operation below the NOx reduction light-
off temperature) this storage period can increase dramatically. This is due to the NOx mass flow
rate from the engine changing dramatically between idle conditions and full load conditions.
The period of lean-NOx storage is expected to increase in inverse proportion to the NOx
emission rate from the engine.  The period of NOx storage under light load conditions therefore
can likewise be expected to increase dramatically.

   Transient operation can further allow for NOx control below the NOx reduction light-off
temperature  due to the thermal inertia of the emission-control system itself.  The thermal inertia
of the emission-control system can work to warm the exhaust gases to a local temperature high
enough to promote the NOx reduction reaction even though the inlet exhaust temperatures are
below the light-off temperature for the catalyst.

   The combination of these two effects was observed during testing of NOx adsorbers at the
National Vehicle and Fuel Emissions Laboratory (NVFEL), especially regarding NOx control
under idle conditions.  It was observed that when idle conditions followed loaded operation, for
example when cooling the engine down after a completing an emission test, that the NOx
emissions were effectively zero (below background levels) for extended periods of idle operation
(for more than 10 minutes). It was also discovered that the stored NOx can be released and
reduced in this operating mode, even though the exhaust temperatures were well below 250°C,
provided that the regeneration event was triggered within the  first 10 minutes of idle operation
(before the catalyst temperature decreased significantly). However, if the idle mode was
continued for extended periods (longer than 15 minutes) NOx control eventually diminished.
The loss of NOx control at extended idle conditions appeared to be due to the inability to reduce
the stored NOx leading to high NOx emissions during NOx regeneration cycles.

   NOx control efficiency with the NOx adsorber technology under steady-state operating
conditions can be seen to be limited by the light-off temperature threshold of the three-way
catalyst NOx reduction function.  Further, a mechanism for extending control below this
temperature  is described for transient operation and is observed in testing of NOx adsorber-based
catalyst systems.  In addition,  as described later in this section, new combustion strategies such
as Toyota's low-temperature combustion technology can raise exhaust temperatures at low loads
to promote improved NOx performance with a NOx adsorber catalyst.

   Overall, NOx adsorber efficiency reflects the composite effectiveness of the NOx adsorber in
storing, releasing and reducing NOx over repeated lean/rich cycles. As detailed above, exhaust
temperatures play a critical role in determining the relative effectiveness of each of these catalyst
functions. These limits on the individual catalyst functions can explain the observed overall
NOx control efficiency of the  NOx adsorber, and can be used to guide future research to improve
overall NOx adsorber efficiency and the design of an integrated NOx emission-control system.

   At low exhaust temperatures overall NOx control is  limited by the light-off temperature
threshold of the three-way NOx reduction function in the range from 200°C to 250°C.  At high
temperatures (above 400° to 500°C) overall NOx control is limited by the thermal stability of the
                                          4-29

-------
Regulatory Impact Analysis
NOx storage function. For exhaust temperatures between these two extremes NOx control can
occur at virtually 100 percent effectiveness.

   The ability of the complete system, including the engine and the emission-control system, to
control NOx emissions consistently (well in excess of 90 percent) is therefore dependent upon
the careful management of temperatures within the system.  Figure 4.1-2 provides a pictoral
representation of these constraints and indicates how well a diesel engine can match the
capabilities of a NOx adsorber-based NOx control system.  The figure shows accumulated NOx
emissions (grams) over the highway FTP cycle for both a light heavy-duty and a heavy heavy-
duty engine. The engine-out NOx emissions are shown as the dark bars on the graphs. The
accumulated NOx emissions shown here, divided by the integrated work over the test cycle gives
a NOx emission rate of 4 g/hp-hr (the 1998 emission standard for highway heavy-duty diesel
engines) for each of these engines.  Also shown on the graph as a solid line is the steady-state
NOx conversion efficiency for a NOx adsorber, MECA "B", used in testing at NVFEL (see
Section 4.1.2.3.5.2 below for more details on testing at NVFEL). The line has been annotated to
show the constraint under low-temperature operation (three-way catalyst light-off). The white
bars on the graph represent an estimate of the tailpipe NOx  emissions that can be realized from
the application of the NOx adsorber based upon the steady-state efficiency curve for adsorber
MECA "B". These estimated tailpipe emissions are highest in the temperature range below
250°C even though the engine-out NOx emissions are the lowest in this region.  This is due to
the light-off temperature threshold for the NOx three-way reduction function.
                                         4-30

-------
 	Technologies and Test Procedures for Low-Emission Engines

                             Figure 4.1-2
  NOx Adsorber Efficiency Characteristics versus Exhaust Temperature
      LHD Diesel Estimated NOx Adsorber Effectiveness over HD FTP
      - MECA "B" NOx Adsorber (%)
    • Engine Out NOx FTP (4g NOx Engine)
    I   I Projected FTP Tailpipe NOx
                                                                        E
                                                                        £
                                                                        D)
                                                                            O
                                                                            CL
                                                                            Q
                                                                            I
                                                                            o

                                                                            o

                                                                            I
                                                                            LLJ
                                                                            X
                                                                            O
                                                                            z
                                                                            T3
                                                                            "(0
                                                                    500
                      Catalyst Inlet Temperature (  C)
      HMD Diesel Estimated NOx Adsorber Effectiveness over HD FTP
100
 90
 - MECA "B" NOx Adsorber (%)
^| Engine Out NOx FTP (4g NOx Engine)
I   I Projected FTP Tailpipe NOx
                                                    ~r
                                                                       10
                                           350
                                                   400
                                                            450
                                                                    500
                      Catalyst Inlet Temperature (°C)
E
(0
D)

"o

O
                                                                            Q
                                                                            I
                                                                            E
                                                                            LU
                                                                            X
                                                                            O
                                   4-31

-------
Regulatory Impact Analysis
    Since the conversion efficiencies are based upon steady-state operation, it is likely that the
low-temperature performance can be better than estimated here due to a catalyst's ability to store
the NOx emissions at these low temperatures and then to reduce them when transient operation
raises the exhaust temperatures above the three-way light-off temperature.  This assertion
provides one explanation for differences noted between this approximation of the NOx
efficiency over the highway FTP cycle for the light heavy-duty engine shown in Figure 4.1-2 and
actual NOx adsorber efficiency demonstrated with this engine in the NVFEL test program.
Based upon the figure above (using the steady-state conversion estimate) the  NOx adsorber
catalyst should have provided less than an 84 percent reduction in NOx emissions over the
highway FTP cycle. However, testing at NVFEL (detailed in Section 4.1.2.3.5) has
demonstrated a greater than 90 percent reduction in NOx emissions with this  same engine and
catalyst pair without significant optimization of the system. Clearly then, steady-state NOx
adsorber performance estimates can underestimate the real NOx reductions realized in transient
vehicle operation. Nevertheless, we have used this approach as a screening analysis to predict
performance for nonroad engines equipped with NOx adsorber catalysts in Section 4.1.3.1.2
below.

    The tailpipe NOx emissions are the lowest in the range from 250°C to 450°C, even though
this is where the majority of the engine-out NOx emissions are created, because of the high
overall NOx reduction efficiency of the NOx adsorber system under these conditions.  At
temperatures above 500°C the NOx conversion efficiency of the NOx adsorber can be seen to
decrease.

   Figure 4.1-2 shows that the temperature window of a current technology NOx adsorber
catalyst is well matched to the exhaust temperature profiles of a light heavy-duty and a heavy
heavy-duty diesel engine operated over the highway FTP cycle.  The discussion in Section
4.1.3.1.2 below shows similarly that the nonroad transient cycle (NRTC) is also well matched to
the performance of the NOx adsorber catalyst.  Testing at NVFEL on the same engine operated
over a wide range of steady-state points, shows that even for extended high-load operation, as
typified by the 100 percent load test points in the test, NOx conversion efficiencies remained
near or above 90 percent (see discussion of the NVFEL test program in Section 4.1.2.3.5 below).

    The discussion above makes it clear that when the engine and NOx adsorber-based emission-
control system are well matched, NOx reductions can be far in excess of 90 percent. Conversely,
it can be inferred that if exhaust temperatures are well in excess of 500°C or well below 200°C
for significant periods of engine operation then NOx control efficiency may be reduced.
Researchers are developing and testing new NOx adsorber formulations designed to increase the
high temperature stability of the NOx adsorber and to therefore widen this window of
      •   S8
operation.

       How effective are NOx adsorbers for cold-start emissions?

   In addition to broadening the catalyst temperature window, the exhaust temperature from the
diesel engine can be managed to align with the temperature window of the catalyst.
                                          4-32

-------
	Technologies and Test Procedures for Low-Emission Engines

    The steady-state analysis discussed above is based on steady-state emission results (i.e., after
exhaust temperatures have stabilized), but the NRTC also includes a cold-start test where the
catalyst initial temperature will be at ambient conditions (see Section 4.2).  The NRTC emission
level for the engine is determined by weighting the cold-start emissions by 1/20 (5 percent), and
weighting the hot-start emission results by 19/20 (95 percent).  Historically, for highway heavy-
duty diesel engines that are similar to current technology nonroad diesel engines not equipped
with an exhaust emission-control device, the cold-start and hot-start emissions have been nearly
identical.  However, with the application of emission-control devices which have optimal
temperature operating windows, such as a NOx adsorber, the cold-start test will become a design
challenge for highway diesel engine manufacturers and for nonroad diesel engine manufacturers,
just as it has been a design challenge for light-duty gasoline vehicle manufacturers for more than
20 years.

    Manufacturers have several available tools to overcome this challenge:

    •   The volume, shape,  and substrate material have a significant effect on the warm-up time
       of a NOx adsorber (just as they do for light-duty three-way catalysts). Manufactures will
       optimize the make-up of the adsorber for best light-off characteristics, such as the thin-
       walled ceramic monolith catalysts typical of modern low-emission light-duty gasoline
       applications.

    •   The packaging of the exhaust emission-control devices, including the use of insulating
       material and air-gap exhaust systems, will also decrease light-off time, and we expect
       manufacturers to explore those opportunities.

       The location of the adsorber, with respect to it's proximity to the exhaust manifold, will
       have a significant impact on the light-off characteristics.

    •   As discussed above, NOx adsorbers have the ability to store NOx at temperatures  much
       less than the three-way catalyst function temperature operating window, on the order of
       100°C. This is unlike the performance of catalysts for light-duty gasoline engines, and  it
       allows the NOx adsorber to store NOx for some period of time before the light-off time
       of the three-way function of its catalyst, resulting in an overall lower effective
       temperature for the device.

    These first four tools available to manufacturers all deal with system design opportunities to
improve the cold-start performance of the NOx adsorber system. In addition, manufacturers
have several active tools that can be used to enhance the cold-start performance of the system,  all
based on technologies that may be used to comply with the Tier 3 emission standards (i.e.,
technologies that will  form the baseline for most engines meeting the Tier 4 standards). These
include the use of engine start-up routines that have a primary purpose of adding heat to the
exhaust to enhance NOx adsorber light-off.  For example:

    •   retarded injection timing;
                                          4-33

-------
Regulatory Impact Analysis
   •   intake air throttling;

   •   post-injection addition of fuel; or

       or increasing back-pressure with an exhaust brake or a VGT system.

   We anticipate manufacturers will explore all these tools to choose the best combination
necessary to minimize light-off time and improve the cold-start NRTC performance. Highway
manufacturers must overcome this same challenge to comply with the HD2007 emission
standards some number of years before these nonroad emission standards go into effect.
Additionally, highway manufacturers must do this with a higher cold-start weighting of 1/7,
rather than 1/20 we are adopting for nonroad diesel engines.  This means that highway engine
manufacturers must have lower cold-start emissions relative to their hot-start emissions than will
nonroad engine manufacturers meeting the Tier 4 standards.  We therefore believe that
manufacturers of nonroad diesel engines will be able to use the technologies described above to
comply with the Tier 4 standards over the NRTC, including the cold-start test.

   One light-duty passenger car manufacturer,  Toyota, has already demonstrated such an
approach to comply with light-duty cold-start requirements. Toyota has shown with its low-
temperature combustion technology one mechanism for raising exhaust temperatures even at
extremely low-load conditions.  The approach, called Low Temperature Combustion (LTC),
increases exhaust temperatures at low-load conditions by more than 50°C while decreasing
engine-out NOx emissions.59 As a result, exhaust temperature are increased into the region for
effective NOx adsorber operation even  at light loads.  The technologies that Toyota uses to
accomplish LTC, cooled EGR and advanced common rail fuel systems, are similar to the
systems that we expect many nonroad engine manufacturers will use to comply with the Tier 3
standards.

   Another example of system integration approaches for diesel engines designed to allow
compliance with transient emission control standards including hot and cold emissions can be
seen in recent work by the Department of Energy and contractors under the Advanced Petroleum
Based Fuels Program - Diesel Emission Control (APBF-DEC).  This work documented in a
recent  SAE paper and in EPA's Highway Diesel Progress Review Report 2, shows that NOx
emission can be reduced adequately on a combined hot and cold start FTP test procedure to
demonstrate emissions below 0.3 g/bhp-hr.60'61 The work illustrates both the ability to control
NOx emissions under cold-start conditions using rapid warm-up procedures and the ability to
reduce NOx emissions below the regulated standards under hot-start conditions to compensate
for the slightly elevated emissions levels experienced under cold-start conditions.

       How  effective are NOx adsorbers over the NTE zone?

   We are adopting an NTE standard for nonroad Tier 4 engines that replicates the provisions
for highway  diesel trucks. A complete discussion of the NTE provisions can be found in Section
IIIJ  of the preamble to the final rule. In short, we are setting an NTE emission limit, over a
broad range of engine operating conditions, that is 1.5 times the limit that applies for testing over

                                         4-34

-------
	Technologies and Test Procedures for Low-Emission Engines

the NRTC and over the steady-state tests. As discussed below, a 90 percent NOx reduction is
technologically feasible across the range of engine operating conditions and ambient conditions
subject to the NTE standards. Also, as discussed below, some modifications to the NTE
provisions to address technical issues that result from the application of advanced NOx catalyst
systems were included in the HD2007 standards and are carried over into this final rule.

   Section 4.1.2.3.5.2 contains a description of the ongoing NOx adsorber evaluation test
program run by our EPA laboratory. Included in that section are test data on four different NOx
adsorbers for which extensive steady-state mapping was performed to calculate various steady-
state emission levels (See Figures 4.1-10 through 4.1-13). Several of the test modes presented in
these figure are not within the NTE zone for NOx, and so would not be subject to the NTE
standard. The following modes listed  in these four figures are within the NTE zone for NOx:
EPA modes 6 - 13, 15, 17, 19, 20. For all of the adsorbers, efficiencies of 90 percent or greater
were achieved across the majority of the NTE zone.  The region of the NTE zone for which
efficiencies less than 90 percent were achieved were concentrated on or near the  torque curve
(EPA modes 8, 9, 15 and 17), with the exception of Adsorber D, for which  EPA modes 6 and 7
achieved 87 percent and 89 percent NOx reduction, respectively. However, Adsorber D was
able to achieve NOx reductions greater than 90 percent along the torque curve. The test modes
along the torque curve represent the highest exhaust gas temperature conditions for this test
engine, on the order of 500°C. Exhaust temperatures of 500°C are near the current upper
temperature limit of the peak NOx reduction efficiency range for NOx adsorbers. It is therefore
not unexpected that the NOx reductions along the torque curve for the test engine are not as high
as in other regions of the NTE zone. We expect manufacturers to choose a NOx  adsorber
formulation that matches the operating range of exhaust gas temperatures for the engine. In
addition, the steady-state mode  data in Figures 4.1-10 through 4.1-13 were  collected under
stabilized conditions. In reality, actual in-use operation of a heavy-duty diesel vehicle likely
does not experience periods of sustained operation along the torque curve, which diminishes the
likelihood that the NOx adsorber bed itself will achieve temperatures in excess of 500°C.
Regardless, as observed in our ongoing diesel progress review and documented in the 2002
diesel progress report,  catalyst developers are realizing incremental improvements  in the high-
temperature NOx reduction capabilities of NOx adsorbers through improvements in NOx
adsorber formulations.62'63'64 As discussed above, only  small improvements in  the current
characteristics are necessary to achieve 90 percent NOx reductions or greater across the NTE
zone.

   As discussed above, the use of advanced NOx adsorber-based catalyst systems  will present
cold-start challenges for highway heavy-duty diesel engines, and for nonroad diesel engines,
under our Tier 4 program, similar to what light-duty gasoline manufacturers have faced in the
past, due to the light-off characteristics of the NOx adsorber.  We have previously discussed the
tools available to engine manufacturers to overcome these challenges to achieve the NOx
standard. The majority of engine operation within the NTE zone will occur at  exhaust gas
temperatures well above the light-off requirement of the NOx adsorbers. Figures 4.1-10 through
4.1-13 below show that all test modes  within the NTE zone have exhaust gas temperatures
greater than 300°C, which is well within the peak NOx reduction efficiency range of current
generation NOx adsorbers. However,  although NTE testing does not include engine start-up

                                         4-35

-------
Regulatory Impact Analysis
conditions, a diesel engine that has not been warmed up could conceivably be started and very
quickly be operated under conditions that are subject to NTE testing; for example, within a
minute or less of vehicle operation after the vehicle has left an idle state.  The final rule specifies
a minimum emission sampling period of 30 seconds for NTE testing. Conceivably, vehicle
emissions could be measured against the NTE standards during that first minute of operation,
and in all likelihood it would not meet the NTE NOx standard.  Given that the NRTC standards
will require control of cold-start emissions, manufacturers will be required to pay close attention
to cold start to comply with the NRTC.  As discussed above, engine operation during NTE
testing will be at exhaust gas temperatures within the optimum NOx reduction operating window
of the NOx adsorbers. In addition, the NOx adsorber is capable of adsorbing NOx at
temperatures on the order of 100°C. Figures 4.1-10 through 4.1-13 all show NOx emission
reductions on the order of 70 - 80 percent are achieved at temperatures as low as 250°C. We are
therefore setting a threshold for exhaust gas temperatures of 250°C, below which the specified
NTE requirements do not apply; we also adopted this provision for the same reason for highway
engines in our HD2007 program.

   The NTE requirements apply not only during laboratory conditions applicable to the transient
test, but also under the wider range of ambient conditions for altitude, temperature and humidity
specified in the regulations.  These expanded conditions will have minimal impact on the
emission-control systems expected to be used to meet the NTE NOx standard. In general, it can
be said that the performance of the NOx adsorbers are only affected by the exhaust gas stream to
which the adsorbers are exposed.  The impact of ambient  humidity, temperature, and altitude will
therefore affect the performance of the adsorber only to the extent these ambient conditions
change the exhaust gas conditions (i.e., exhaust gas temperature and gas constituents).  The
ambient humidity  conditions subject to the NTE requirement will have minimal, if any, impact
on the performance of the NOx adsorbers.  The exhaust gas itself, independent of the ambient
humidity, contains a very high concentration of water vapor, and the impact of the ambient
humidity on top of the products of dry air and fuel  combustion are minimal.  The effect of
altitude on NOx adsorber performance should also be minimal or negligible. NTE testing is
limited to altitudes below 5,500 feet above sea level.  The decrease in atmospheric pressure at
5,500 feet should have minimal impact on the NOx adsorber performance. Increasing altitude
can decrease the air-fuel ratio for diesel engines, which can in turn increase exhaust gas
temperatures.  However, as discussed in the final rule for  the highway 2004 standards (Phase 1),
highway engines with Phase 1 technology (and thus the similar Tier 3 nonroad diesel engines)
can be  designed to target air-fuel ratios at altitude that will maintain appropriate exhaust gas
temperatures within the ambient conditions specified by the highway NTE test procedure and
thus the similar NTE procedure for Tier 4 engines. This approach also allows manufacturers to
maintain engine-out PM emission levels near the 0.1 g/hp-hr level. Finally, the NTE regulations
specify ambient temperatures that are broader than the NRTC temperature range of 68-86°F.
The NTE test procedure specifies no lower ambient temperature bounds.  However, as discussed
above,  we limit NTE requirements on NOx (and NMHC) for engines equipped with NOx (and/or
NMHC) catalysts to include only engine operation with exhaust gas temperatures greater than
250°C. Low ambient temperatures will therefore not present any difficulties for NTE NOx
compliance. NTE standards also apply under ambient temperatures that are higher than the
laboratory conditions. The NTE standards apply up to a temperature of 100°F at sea level, and

                                         4-36

-------
	Technologies and Test Procedures for Low-Emission Engines

up to 86°F at 5,500 feet above sea level. At altitudes in between, the upper NTE ambient
temperature requirement is a linear fit between these two conditions.  At 5,500 feet, the NTE
ambient temperature requirement is the same as the upper end of the temperature range (86°F)
for testing with prescribed duty cycles,  and will therefore have no impact on the performance of
the NOx adsorbers, considering that majority of the test data described throughout this chapter
were collected under laboratory conditions.  The NTE upper temperature limits at sea level is
100°F, which is 14°F (7.7°C) greater than the NRTC range.  This increase is relatively minor, and
while it will increase the exhaust gas temperature; in practice the increase should be passed
through the engine to the exhaust gas, and the exhaust gas would be on the order of 8°C higher.
Within the exhaust gas temperature range for a diesel engine during NTE operation, an 8°C
increase is very small.  As discussed above, we expect manufacturers to choose an adsorber
formulation matched to a particular engine design and we expect the small increase in exhaust
gas temperature that can  occur from the expanded ambient temperature requirements for the
NTE to be taken into account by manufacturers when designing the complete emission-control
system.

   To summarize, based on the information presented in this chapter, and the analysis and
discussion presented in this section, we conclude the NTE NOx requirement (1.5 x NRTC/C1
standard) contained in this final rule will be feasible.

   Further discussion of feasibility of the NOx requirement under transient testing conditions
can be found in Section 4.1.3.1.2.

   4.1.2.3.4 Are Diesel NOx Adsorbers Durable ?

   The considerable success in demonstrating NOx adsorbers makes us confident that the
technology is capable of providing the level of conversion efficiency needed to meet the Tier 4
NOx standard. However, there are several engineering challenges that will need to be addressed
in going from this level of demonstration to implementation of durable and effective emission-
control systems on nonroad equipment. In addition to the generic need to optimize engine
operation to match the NOx adsorber performance, engine and catalyst manufacturers will
further need to address issues of system and catalyst durability.  The nature of these issues are
well understood. The hurdles  that must be overcome have direct analogues in technology issues
that have been addressed previously in automotive applications and are expected to be overcome
with many of the same solutions. With the transfer of highway technologies to nonroad engines
anticipated in this rulemaking, we believe we have already addressed the issues highlighted in
this section for highway engines well before the start of this nonroad program.

    In this section, we will describe the major technical hurdles that must be addressed to ensure
that the significant emission reductions from NOx adsorbers occur throughout the life of nonroad
diesel engines. This section is organized into separate durability discussions for the system
components (hardware) and various near-term and long-term durability issues for the NOx
adsorber catalyst itself.

       4.1.2.3.4.1 NOx Adsorber Regeneration Hardware Durability

                                          4-37

-------
Regulatory Impact Analysis
    The system we have described in Figure 4.1-1 represents but one possible approach for
generating the necessary exhaust conditions to allow for NOx adsorber regeneration and
desulfation.  The system consists of three catalyst substrates (for a CDPF/Low Temperature NOx
Adsorber, a High Temperature NOx Adsorber and an Oxidation Catalyst), a support can that
partitions the exhaust flow through the first two catalyst elements, three fuel injectors, and a
means to divert exhaust flow through one or more of the catalyst partitions. Though not shown
in the figure,  a NOx /O2 sensor is also likely to be needed for control feedback and on-board
diagnostics (OBD). All of these elements have already been applied in one form or another to
either diesel or gasoline engines in high volume long life applications.

    The NOx adsorber system we described earlier borrows several components from the
gasoline three-way catalyst systems and benefits from the years of development on three way
catalysts.  The catalyst substrates (the ceramic support elements on which a catalyst coating is
applied) have developed through the years to  address concerns  with cracking due to thermal
cycling  and abrasive damage from vehicle vibration.  The substrates applied for diesel NOx
adsorbers will be virtually identical to the ones used for today's passenger cars in every way but
size. They are expected to be equally durable when applied to diesel applications as has already
been shown in the successful application of diesel oxidation catalysts (DOCs) on some diesel
engines over the last 15 years. Retrofit catalyst-based systems have similarly been applied to
nonroad diesel engines with good durability, as described in Section 4.1.3.2 below.

    The NOx/O2 sensor needed for regeneration control and OBD is another component
originally designed and developed for gasoline powered vehicles (in this case lean-burn gasoline
vehicles) that are already well developed and  can be applied with confidence in long life for
NOx adsorber-based diesel emission control.  The NOx/O2 sensor is an evolutionary technology
based largely on the current Oxygen (O2) sensor technology developed for gasoline three-way
catalyst-based systems. Oxygen sensors have proven to be extremely reliable and long lived in
passenger car applications, which see significantly higher temperatures than are normally
encountered on a diesel engine.65'66 Diesel engines do have one characteristic that makes the
application of NOx/O2 sensors more difficult.  Soot in diesel exhaust can cause fouling of the
NOx/O2 sensor damaging its performance.  However this issue  can be addressed through the
application of a catalyzed diesel particulate filter (CDPF)  in front of the sensor.  (See Section
4.1.2.3.2 above, noting  synergies that can result from use in tandem of NOx adsorbers and
CDPFs.) The CDPF then provides a protection for the sensor from PM while not hindering its
operation. Since the NOx adsorber will likely be located downstream of a CDPF in each of the
potential technology scenarios we  have considered this solution to the issue of PM sooting is
readily addressed.

    Fuel is metered into a modern gasoline engine with relatively low pressure pulse-width-
modulated fuel-injection valves. These valves are designed to cycle well over a million times
over the life of a vehicle while continuing to accurately meter fuel. Applying this technology to
provide diesel fuel as a  reductant for a NOx adsorber system is  a relatively straightforward
extension of the technology.  A NOx adsorber system cycles far fewer times over its life when
compared with the current long life of gasoline injectors.  However, these gasoline fuel injectors

                                          4-38

-------
	Technologies and Test Procedures for Low-Emission Engines

designed to meter fuel into the relatively cool intake of a car cannot be directly applied to the
exhaust of a diesel engine. In the testing done at NVFEL, a similar valve design was used that
had been modified in material properties to allow application in the exhaust of an engine. While
benefitting from the extensive experience with gasoline-based injectors a designer can therefore,
in a relatively straightforward manner, improve the characteristics of the injector to allow
application for exhaust reductant regeneration.  Toyota has shown with its Avensis DPNR diesel
passenger car how to use a gasoline direct injection (GDI)-based fuel injector to inject diesel fuel
in the exhaust manifold of a diesel engine to allow for NOx adsorber regeneration and
desulfation.67

   The NOx adsorber system we describe in Figure 4.1-1 requires a means to partition the
exhaust during regeneration and to control the relative amounts of exhaust flow between two or
more regions of the exhaust system.  Modern diesel engines already  employ a valve designed to
carry out this very task. Most modern turbochargers employ a wastegate valve that allows some
amount of the exhaust flow to bypass the exhaust turbine to control maximum engine boost and
limit turbocharger speed. These valves can be designed to be proportional, bypassing a specific
fraction of the exhaust flow to track a specified boost pressure for the system.  Turbocharger
wastegate valves applied to heavy-duty diesel engines typically last the life of the engine in spite
of the extremely harsh environment within the turbocharger. This same valve approach can be
applied to accomplish the flow diversion required for diesel NOx adsorber regeneration and
desulfation.  Since temperatures will typically be cooler at the NOx adsorber compared with the
inlet to the exhaust turbine on a turbocharger, the control valve should be equally reliable in this
application.

       4.1.2.3.4.2 NOx Adsorber Catalyst Durability

   In many  ways a NOx adsorber, like other engine catalysts, acts like a small chemical process
plant.  It has specific chemical processes that it promotes under specific conditions with different
elements of the  catalyst materials.  There is often an important sequence to the needed reactions
and a need to match process rates to keep this sequence of reactions going.  Because of this need
to promote specific reactions under the right conditions early catalysts were often easily
damaged. This  damage prevents or slows  one or more the reactions  causing a loss in emission
control. For example, contaminants from engine oil, like phosphorous or zinc, can attach to
catalysts sites partially  blocking the site from the exhaust constituents and slowing reactions.
Similarly, lead added to gasoline to increase octane levels bonds to the catalyst sites, causing
poisoning as well. Likewise, sulfur, which occurs naturally in petroleum products like gasoline
and diesel fuel,  can poison many catalyst functions preventing or slowing the desired reactions.
High exhaust temperatures experienced under some conditions can cause the catalyst materials to
sinter (thermally degrade) decreasing the surface area  available for reactions to decrease.

   All of these problems have been addressed over time for the gasoline three-way catalysts,
resulting in the high efficiency and long life durability now typical of modern vehicles. To
accomplish this, changes were made to fuels and oils used in vehicles (e.g., lead additives
banned from gasoline, sulfur levels reduced in gasoline distillates, specific  oil formulations for
                                          4-39

-------
Regulatory Impact Analysis
aftertreatment equipped cars), and advances in catalysts designs were needed to promote
sintering-resistant catalyst formulations with high precious-metal dispersion.

    The wealth of experience gained and technological advancements made over the last 30 years
of gasoline catalyst development can now be applied to the development of the NOx adsorber
catalyst. The NOx adsorber is itself an incremental advancement from current three-way catalyst
technology. It adds one important additional component not currently used on three-way
catalysts, NOx storage catalyst sites. The NOx storage sites (normally alkali or alkaline earth
metals) allow the catalyst to store NOx emissions with extremely high efficiency under the lean-
burn conditions typical of the diesel exhaust. It also adds a new durability concern due to sulfur
storage on the catalyst.

    This section will explore the durability issues of the NOx adsorber catalyst applied to diesel
engines. It describes the effect of sulfur in diesel fuel on catalyst performance, the methods to
remove the sulfur from the catalyst through active control processes, and the implications for
durability of these methods.  It then discusses these durability issues relative to similar issues for
existing gasoline three-way catalysts and the engineering paths to solve these issues. This
discussion shows that the NOx adsorber is an incremental improvement upon the existing three-
way catalyst, with many of the same solutions for the expected durability issues.

       Sulfur Poisoning of the NOx Storage Sites

    The NOx adsorber technology is extremely efficient at storing NOx as a nitrate on the
surface of the catalyst, or adsorber (storage) bed, during lean operation. Because of the
similarities in chemical properties of SOx and NOx, the SO2 present in the exhaust is also stored
on the catalyst surface as a sulfate.  The sulfate compound that is formed is significantly more
stable than the nitrate compound and is typically not released during the NOx release and
reduction step (NOx regeneration step) (i.e., it is stored preferentially to NOx). Since the NOx
adsorber is virtually 100 percent effective at capturing SO2 in the adsorber bed, sulfate
compounds quickly occupy the NOx storage sites on the catalyst thereby reducing and
eventually rendering the catalyst ineffective for NOx reduction (poisoning the catalyst).

    Figure 4.1-3 shows the effect of sulfur poisoning of a NOx adsorber catalyst as reported by
the DOE DECSE program. The graph shows the NOx adsorber efficiency versus exhaust inlet
temperature under steady-state conditions for a diesel engine-based  system. The three dashed
lines that overlap each other show the NOx conversion efficiency of the catalyst when sulfur has
been removed from the catalyst.  The three solid lines show the effect of sulfur poisoning on the
catalyst at three different fuel sulfur levels over different periods of extended aging (up to 250
hours). From the figure, it can be seen that even with three ppm  sulfur fuel a significant loss in
NOx efficiency can occur in as little as 250 hours. Further, it can be seen that quite severe sulfur
poisoning can occur with elevated fuel sulfur levels. Catalyst performance was degraded by
more than 70 percent over only 150 hours of operation when 30 ppm sulfur fuel was used.68
                                          4-40

-------
                   Technologies and Test Procedures for Low-Emission Engines
                                    Figure 4.1-3
        Comparison of NOx Conversion Efficiency before and after Desulfation
    100
     80
 c
 •|   60

 LU
 C
 o
 '(/)
 c
 o
 O
     40
     20
             250
300         350         400
      Catalyst Inlet Temperature [C]
450
500
• -X- - DECSE II after desulfation (3-ppm)
• -*— - DECSE II after desulfation (16-ppm)
• ••- - DECSE II after desulfation (30-ppm)
                 —X—DECSE II before desulfation (3-ppm, 250 hrs aging)
                 —*—DECSE II before desulfation (16-ppm, 200 hrs aging)
                 —•—DECSE II before desulfation (30-ppm, 150 hrs aging)
The DECSE researchers drew three important conclusions from Figure 4.1-3:

•  Fuel sulfur, even at very low levels like three ppm, can limit the performance of the NOx
   adsorber catalyst significantly.

•  Higher fuel sulfur levels, like 30 ppm, dramatically increase the poisoning rate, further
   limiting NOx adsorber performance.

•  Most importantly though, the figure shows that if the sulfur can be removed from the
   catalyst through a desulfation (or desulfurization) event, the NOx adsorber can provide
   high NOx control even after exposure to sulfur in diesel fuel. This is evidenced by the
   sequence of the data presented in the figure. The three high conversion efficiency lines
   show the NOx conversion efficiencies after a desulfation event that was preceded by the
   sulfur poisoning and degradation shown in the solid lines.
                                        4-41

-------
Regulatory Impact Analysis
   It is clear from this data that higher fuel sulfur levels dramatically reduce the efficiency of
NOx adsorber catalysts.  Sulfur accumulates in the NOx storage sites preventing their use for
NOx storage. In other words, they decrease the storage volume of the catalyst.  The rate at
which sulfur fills NOx storage sites is expected to be directly proportional to the amount of
sulfur that enters the catalyst. A doubling in fuel-sulfur levels should therefore correspond to a
doubling in the  SOx poisoning rate.

   The design of a NOx adsorber will need to address accommodating an expected volume of
sulfur before experiencing unacceptable penalties in either lost NOx control efficiency or
increased fuel consumption due to more frequent NOx regenerations. The amount of operation
allowed before that limit is realized for a specific adsorber design will be inversely proportional
to fuel sulfur quantity. In the theoretical case of zero sulfur, the period of time before the sulfur
poisoning degraded  performance excessively would be infinite. For a more practical fuel sulfur
level like the 10 ppm average expected with a  15 ppm fuel sulfur cap, the period of operation
before unacceptable poisoning levels have been reached is expected to be less than 40 hours
(with today's NOx adsorber formulations).69

   Future improvements in the NOx adsorber technology are expected due to its relatively early
state of development.  Some of these improvements are likely to include improvements in the
kinds of materials used in NOx adsorbers to increase the means and ease of removing stored
sulfur from the catalyst bed.  However, because the stored sulfate species are inherently more
stable than the stored nitrate compounds (from stored NOx emissions), we expect that future
NOx adsorbers will  continue to be poisoned by sulfur in the exhaust.  A separate sulfur release
and reduction cycle  (desulfation cycle) will therefore always be needed to remove the stored
sulfur.

       NOx Adsorber Desulfation

   Numerous test programs have shown that sulfur can be removed from the catalyst surface
through a sulfur regeneration step (desulfation step) not dissimilar from the NOx regeneration
function.70'71'72'73'74^75 The stored sulfur compounds are removed by exposing the catalyst to hot
and rich (air-fuel ratio below the stoichiometric ratio of 14.5 to 1) conditions for a brief period.
Under these conditions, the stored sulfate is released and reduced in the catalyst.  This sulfur
removal process, called desulfation or desulfurization in this document, can restore the
performance of the NOx adsorber to near new  operation.

   Most of the  information in the public domain on NOx adsorber desulfation is based upon
research done either in controlled bench reactors using synthetic gas compositions or on
advanced lean-burn  gasoline engine vehicles.  As  outlined above, these programs have shown
that desulfation of NOx adsorber catalysts can be accomplished under certain conditions but the
work does not directly answer whether NOx adsorber desulfation is practical for diesel  engine
exhaust conditions.  The DECSE Phase II program answers that question.

   Phase II of the DECSE program developed and demonstrated a desulfurization (desulfation)
process to restore NOx conversion efficiency lost to sulfur contamination. The engine used in

                                          4-42

-------
	Technologies and Test Procedures for Low-Emission Engines

the testing was a high-speed direct-injection diesel selected to provide a representative source of
diesel exhaust and various exhaust temperature profiles to challenge the emission-control
devices.  The desulfation process developed in the DECSE Phase II program controlled the air-
fuel ratio and catalyst inlet temperatures to achieve the high temperatures required to release the
sulfur from the device.  Air-fuel ratio control was accomplished in the program with exhaust gas
recirculation (EGR) and a  post-injection of fuel to provide additional reductants. Using this
approach the researchers showed that a desulfation procedure can be developed for a diesel
engine with the potential to meet in-service engine operating conditions and acceptable levels of
torque fluctuation. The NOx efficiency recovery accomplished in DECSE Phase II using this
approach is shown in Figure 4.1-3, above.

   The effectiveness of NOx adsorber desulfation appears to be closely related to the
temperature of the exhaust gases during desulfation, the exhaust chemistry (relative air-fuel
ratio), and to the NOx adsorber catalyst formulation.76'77  Lower air-fuel ratios (more available
reductant) works to promote the release of sulfur from the surface, promoting faster and more
effective desulfation. Figure 4.1-4 shows results from Ford testing on NOx adsorber conversion
efficiency with periodic aging and desulfation events in a control flow reactor test.78  The control
flow reactor test uses controlled gas constituents that are meant to represent the potential exhaust
gas constituents from a lean-burn engine.  The solid line with the open triangles labeled "w/o
regen" shows the loss of NOx control over thirteen hours of testing without a desulfation event
and with eight ppm sulfur  in the test gas (this is roughly equivalent to 240 ppm fuel sulfur,
assuming an air-fuel ratio for diesel engines of 30:1).79 From the figure it can be seen that
without a desulfation event, sulfur rapidly degrades the performance of the NOx adsorber
catalyst.  The remaining two lines show the NOx adsorber performance with periodic sulfur
regeneration events timed  at one-hour intervals and lasting for  10 minutes (a one-hour increment
on 240 ppm fuel sulfur is approximately equivalent to 34 hours of operation on 7 ppm fuel). The
desulfation events were identical to the NOx regeneration events, except that the desulfation
events occurred at elevated temperatures.  The base NOx regeneration temperature for the testing
was 350°C. The sulfur regeneration, or desulfation, event was  conducted at two different gas
temperatures of 550°C and 600°C to show the effect of exhaust gas temperature on desulfation
effectiveness, and thus NOx adsorber efficiency.  From Figure  4.1-4 it can be seen that, for this
NOx adsorber formulation, the NOx recovery after desulfation is higher for the desulfation event
at 600°C than at 550°C.
                                          4-43

-------
Regulatory Impact Analysis
                                       Figure 4.1-4
             Flow Reactor Testing of a NOx Adsorber with Periodic Desulfations
         95%
         60%
                          # of SOx and DeSOx events (1hr periods)
   As suggested by Figure 4.1-4, it is well known that the rate of sulfur release (also called
sulfur decomposition) in a NOx adsorber increases with temperature.80'81  However, while
elevated temperatures directionally promote more rapid sulfur release, they also can directionally
promote sintering of the precious metals in the NOx adsorber washcoat. The loss of conversion
efficiency due to exposure of the catalyst to elevated temperatures is referred to as thermal
degradation in this document.

           Thermal Degradation

   The catalytic metals that make up most exhaust emission-control technologies, including
NOx adsorbers, are designed to be dispersed throughout the catalyst into as many small catalyst
"sites" as possible. By spreading the catalytic metals into many small catalyst sites, rather than
into a fewer number large sites, catalyst efficiency is improved. This is because smaller catalyst
sites have more surface area per mass, or volume, of catalyst when compared with larger catalyst
sites.  Since most of the reactions being promoted by the catalyst occur on the surface, increasing
surface area increases catalyst availability and thus conversion efficiency. While high dispersion
(many small catalyst sites) is in general good for most catalysts, it is even more beneficial to the
NOx adsorber catalyst because of the need for the catalytic metal sites to perform multiple tasks.
NOx adsorber catalysts typically rely on platinum to oxidize NO to NO2 prior to adsorption of
                                          4-44

-------
	Technologies and Test Procedures for Low-Emission Engines

the NO2 on an adjacent NOx storage site.  Under rich operating conditions, the NOx is released
from the adsorption site, and the adjacent platinum (or platinum + rhodium) catalyst site can
serve to reduce the NOx emissions into N2 and O2. High dispersion, combined with NO
oxidation, NOx storage and NOx reduction catalyst sites being located in close proximity,
provide the ideal catalyst design for a NOx adsorber catalyst.  But high temperatures, especially
under oxidizing conditions, can promote sintering of the platinum and other PGM catalyst sites,
permanently decreasing NOx adsorber performance.

   Catalyst sintering is a process by which adjacent catalyst sites can "melt" and regrow into a
single larger catalyst site (crystal growth). The single larger catalyst site has less surface area
available to promote catalytic activity than the original two or more catalyst sites that were
sintered to form it. This loss in surface area decreases the efficiency of the catalyst.82 High
temperatures, promote sintering of platinum catalysts especially under oxidizing conditions.83 It
is therefore important to limit the exposure of platinum-based catalysts to high exhaust
temperatures especially during periods of lean operation. Consequently, the desire to promote
rapid desulfation of the NOx adsorber catalyst technology by maximizing the desulfation
temperature and the need to limit the exposure of the catalyst to  the high temperatures that
promote catalyst sintering must be carefully balanced. An example of this tradeoff can be seen
in Figure 4.1-5, which shows the NOx conversion efficiency of three NOx adsorber catalysts
evaluated after extended periods of sulfur poisoning followed by sulfur regeneration periods.84
The three catalysts (labeled A,  B, and C) are identical in formulation and size but were  located at
three different positions in the exhaust system of the gasoline direct injection engine used for this
testing.  Catalyst A was located 1.2 meters from the exhaust manifold, catalyst B 1.8 meters from
the exhaust manifold and catalyst C was located 2.5 meters from the exhaust manifold.  Locating
the catalysts further from the engine lowered the maximum exhaust temperature and thus catalyst
bed temperature experienced during the programmed sulfur regeneration cycle.  Catalyst A
experienced the highest catalyst bed temperature of 800°C, while catalyst C experienced the
lowest catalyst bed temperature of 650°C. Catalyst B experienced a maximum catalyst bed
temperature of 730°C. Figure 4.1-5 shows that there is an optimum desulfation temperature that
balances the tradeoffs between rapid sulfur regeneration and thermal degradation (thermal
sintering) at high temperatures.
                                          4-45

-------
Regulatory Impact Analysis
                                   Figure 4.1-5
         Influence of Maximum Catalyst Bed Temperature During Desulfation
               Influence of Maximum Temperature in Durability Cycle on Engine Bench
     100%
  o
  o
  O
  IO
  fO
  +J
  re
   c
   o
   0)
   c
   o
   o
   X
   O
80%
      60% --
40%
      20% -
       0%
             Lower Temperatures Decrease
             Sulfur Deconposition Rates and
             the Effective Period of Desulfation
                                        B
             Aging Time:
             100 hours
                                            High Temperatures Promote
                                            Sulfate Decomposition but
                                            Increase Precious Metal Sintering
          600          650           700          750           800
                Maximum Catalyst Bed Temperature (°C) in Durability Cycle
                              From SAE 1999-01-3501 Figure 7
                                                                       850
    The DECSE Phase II program, in addition to investigating the ability of a diesel engine /
NOx adsorber-based emission-control system to desulfate, provides a preliminary assessment of
catalyst durability when exposed to repeated aging and desulfurization cycles. Two sets of tests
were completed using two different fuel sulfur levels (three ppm and 78 ppm) to investigate
these durability aspects. The first involved a series of aging, performance mapping,
desulfurization and performance mapping cycles. An example of this testing is shown in Figure
4.1-6.  The graph shows a characteristic "sawtooth" pattern of gradual sulfur poisoning followed
by an abrupt improvement in performance after desulfation.  The results shown in Figure 4.1-6
are for two identical catalysts one operated on 3 ppm sulfur fuel (catalyst S5) and the other
operated on 78 ppm sulfur fuel (catalyst S7). For the catalyst operated on 3 ppm sulfur fuel the
loss in performance over the ten hours of poisoning is noted to be very gradual. There appears to
be little need to desulfate that catalyst at the ten-hour interval set in the experiment. In fact it can
be seen that in several cases the performance after desulfation is worse than prior to desulfation.
This suggests, as discussed above, that the desulfation cycle can itself be damaging to the
catalyst. In actual use, we would expect an engine operating on 3 ppm sulfur fuel not to
desulfate until well beyond a ten-hour interval and be engineered to better withstand the damage
caused by desulfation, as discussed later in this section.  For the catalyst operated on 78 ppm
sulfur fuel the loss in performance over the ten-hour poisoning period is dramatic.  To ensure
continued high performance when operating on 78 ppm sulfur fuel, the catalyst requires frequent
                                          4-46

-------
	Technologies and Test Procedures for Low-Emission Engines

desulfations. From the figure it can be inferred that the desulfation events need to be spaced at
intervals as short as one to two hours to maintain acceptable performance.

                                      Figure 4.1-6
           Integrated NOx Conversion Efficiency following Aging and Desulfation
    100%
  O
 "•5
  O
  x
 O
           Aging 10 hrs
           of 3 ppm fuel
          Aging 10 hrs
          of 78 ppm fuel  =
                                                                DECSE Catalyst S7
                                                                - Aged on 78 ppm S
                                                                DECSE Catalyst S5
                                                                - Aged on 3 ppm S
     30% -
20% -
10% -
     0%
                       10
              Time (hours) cycle of 10 hrs Sulfur Aging / 6 min Desulfation
   As a follow on to the work shown in Figure 4.1-6, the desulfation events were repeated an
additional 60 times without sulfur aging between desulfation events.  This was done to
investigate the possibility of deleterious affects from the desulfation event itself even without
additional sulfur poisoning.  As can be seen in Figure 4.1-7, the investigation did reveal that
repeated desulfation events even without additional sulfur aging can cause catalyst deterioration.
As described previously, high temperatures can lead to a loss in catalyst efficiency due to
thermal degradation (sintering of the catalytic metals). This appears to be the  most likely
explanation for the loss in catalyst efficiency shown here. For this testing, the catalyst inlet
temperature was controlled to approximately 700°C; however, the catalyst bed temperatures may
have been higher.85

   Based on the work in DECSE Phase II, the researchers concluded that:
   • The desulfurization procedure developed has the potential to meet in-service engine
   operating conditions and to provide acceptable driveability conditions.
                                         4-47

-------
Regulatory Impact Analysis
   • Although aging with 78 ppm sulfur fuel reduced NOx conversion efficiency more than
   aging with three ppm sulfur fuel as a result of sulfur contamination, the desulfurization
   events restored the conversion efficiency to nearly the same level of performance. However,
   repeatedly exposing the catalyst to the desulfurization procedure developed in the program
   caused a continued decline in the catalyst's desulfated performance.

   • The rate of sulfur contamination during aging with 78 ppm sulfur fuel increased with
   repeated aging / desulfurization cycles (from 10 percent per ten hours to 18 percent per ten
   hours). This was not observed with the three ppm sulfur fuel, where the rate of decline
   during aging was fairly constant at approximately two percent per ten hours.

                                      Figure 4.1-7
              Integrated NOx Conversion Efficiency after Repeated Desulfation
        100%
  C
  O
  ?
  £
  3 ^
  V) C
  0) O
  O +3
  »- O
  0) 3
  O X
  .
  fc
  LU
  X
  O
         90% -
80% -
70% -
60% -
         50% -
         40% -
         30% -
20% -
         10% -
          0%
                                                      •DECSE Catalysts?
                                                               •DECSE Catalyst S8
                      10          20         30         40         50         60         70
                      Desulfation Events (# of desulfation cycles)
   Currently available data on NOx adsorber formulations show clearly that sulfur can be
removed from the surface of the NOx adsorber catalyst. The initial high performance after a
desulfation event is then degraded over time by the presence of sulfur until the next desulfation
event.  The resulting characteristic NOx adsorber performance level over time exhibits a saw-
tooth pattern with declining performance followed by rapid recovery of performance following
desulfation. The rate of this decline increases substantially with higher fuel sulfur levels. Tto
ensure a gradual and controllable decline in performance, fuel sulfur levels must be minimized.
                                          4-48

-------
	Technologies and Test Procedures for Low-Emission Engines

However, even given very low fuel sulfur levels, gradual decline in performance must be
periodically overcome.  The development experience so far shows that diesel engines can
accomplish the required desulfation event. The circumstances that effectively promote rapid
desulfation also promote thermal degradation.  It will therefore be important to limit thermal
degradation.

          Limiting Thermal Degradation

   The issue of thermal degradation of NOx adsorber catalyst components is similar to the
thermal sintering issues faced by light-duty three-way catalysts for vehicles developed to meet
current California LEV and future Federal Tier 2 standards using platinum+rhodium (Pt+Rh)
catalysts. Initial designs were marked by unacceptable levels of platinum sintering that limited
the effectiveness of Pt+Rh catalysts.  This problem has been overcome through modifications to
the catalyst supports and surface structures that stabilize the precious metals at high temperatures
(>900 °C).  Stabilization of ceria components using Zirconium (Zr) has pushed the upper
temperature limits of ceria migration to well over 1000 °C.86'87 Stabilization components can
function in different ways. Some are used to "fill" structural vacancies, for example "open"
locations within a crystalline lattice, thus strengthening the lattice structure. Such strengthening
of crystalline lattice structures is particularly important at high temperatures.  Other types of
stabilizing components can act as obstructions within a matrix to prevent migration of
components, or can enhance the mobility of other molecules or atoms, such as oxygen.  An
approach stabilizing NOx adsorber catalyst components similar to the approaches taken with
LEV three-way catalyst designs should help to minimize thermal sintering of components during
desulfation.

   In many ways, limiting the thermal degradation of the NOx adsorber catalyst should be easier
than for the gasoline three-way catalyst. Typical exhaust gas temperatures for a heavy light-duty
gasoline truck (e.g., a Ford Expedition) commonly range from 450°C to more than 800°C during
normal operation.88 A heavy-duty diesel engine in contrast rarely has exhaust gas temperatures
in excess of 500°C. Further, even during the desulfation  event, exhaust temperatures are
expected to be controlled below 700°C.  The NOx adsorber applied to diesel engines is therefore
expected to see both lower average temperatures and lower peak temperatures when compared
with an equivalent gasoline engine.  Once thermal degradation improvements are made to NOx
adsorber  catalysts, thermal degradation will reasonably be expected to be less than the level
predicted for future Tier 2 gasoline applications.

   In addition to the  means to improve the thermal stability of the NOx adsorber by applying
many of the same techniques being perfected for the  Tier 2 gasoline three-way catalyst
applications, an additional possibility exists that the desulfation process itself can be improved to
give both high sulfur  removal and to limit thermal degradation.  The means to do this might
include careful control of the maximum temperature  during desulfation to limit the exposure to
high temperatures. Also, improvements in how the regeneration process occurs may provide
avenues for improvement. Low air-fuel ratios (high  levels of reductant) are known to improve
the desulfation process. The high level of reductant may also help to suppress oxygen content in
the exhaust to further limit thermal degradation.

                                         4-49

-------
Regulatory Impact Analysis
   Researchers at Ford Scientific Research Labs have investigated NOx adsorber catalyst
desulfation (called DeSOx in their work) to answer the question: "if a regeneration process
(sulfur regeneration) is required periodically, will the high temperatures required for the
regeneration have deleterious, irreversible effects on NOx efficiency?" To explore the issue of
NOx adsorber durability after repeated desulfation events, Ford conducted repeated sequential
sulfur poisoning and desulfation cycles with a NOx  adsorber catalyst. The results of their
experiment are shown in Figure 4.1-8.89 As shown in Figure 4.1-8, the NOx adsorber sample
underwent more than 90 poisoning and desulfation cycles with 12 hours occurring between the
end of one desulfation to the end of the next desulfation without a measurable loss in post-
desulfation performance. This testing was done using a laboratory tool called a pulsator, used to
study ceramic monolith catalyst samples. The ceramic test samples were heated to between
700°C  and 750°C. These results indicate that for some combinations of temperatures and
reductant chemistries the NOx  adsorber can be repeatedly desulfated without a significant loss in
NOx reduction efficiency.  This work indicates that it is possible to optimize the desulfation
process to allow for adequate sulfur removal without a significant decrease in NOx reduction
efficiency.

                                      Figure 4.1-8
               Repeated Sulfur Poisoning and Desulfation on a Bench Pulsator
         100
                   10
                           20
                                   30       40      50
                                         DeSOx event
                                                          60
                                                                  70
    These results indicate that, with further improvements to the NOx adsorber catalyst design
incorporating the experience gained on gasoline three-way catalysts and continuing
improvements in the control of the desulfation,  degradation of the NOx adsorber catalyst with
each desulfation event can be limited. However, the expectation remains that there will be some
                                          4-50

-------
	Technologies and Test Procedures for Low-Emission Engines

level of deterioration with desulfation that must be managed to ensure long-term high efficiency
of the NOx adsorber. This means that the number and frequency of desulfation events must be
kept to a minimum.  The key to this is to limit the amount of sulfur to which the catalyst is
exposed over its life. In this way, the deterioration in performance between desulfation events is
controlled at a gradual rate and the period between desulfations can be maximized to limit
thermal degradation.

       Overall System Durability

   NOx emission control with a NOx adsorber catalyst-based systems is an extension of the
very successful three-way catalyst technology. NOx adsorber technology is most accurately
described as incremental and evolutionary with system components that are straightforward
extensions of existing technologies.  The technology therefore benefits substantially from the
considerable experience gained over the past 30 years with the today's highly reliable and
durable three-way catalyst systems.

   The following observations can be made from the data provided in the preceding sections on
NOx adsorber durability:

   •   NOx adsorber catalysts are poisoned by sulfur in diesel fuel, even at fuel sulfur levels as
       low as three  ppm.

   •   A sulfur regeneration event (desulfation) can restore NOx adsorber performance.

   •   A diesel engine can produce exhaust conditions that are conducive to desulfation.

   •   Desulfation events, which require high catalyst temperatures, can cause sintering of the
       catalytic metals in the NOx adsorber, thereby reducing NOx-control efficiency.

       The means exist from the development of gasoline three-way catalysts to improve the
       NOx adsorber's thermal durability.

   •   In carefully controlled experiments, NOx adsorbers can be desulfated repeatedly without
       an unacceptable loss in performance.

       The number  and frequency of desulfation events must be limited to ensure any gradual
       thermal degradation over time does not excessively deteriorate the catalyst.

   Based on these observations, we  are confident that NOx adsorber technology for HD2007
and later engines will be durable over the life of heavy-duty diesel vehicles, provided that the
engines use fuel with a 15 ppm sulfur cap and that the technology will prove to be similarly
durable when applied some years later to nonroad diesel engines to comply with the Tier 4
emission standards.  Without the use of this low-sulfur fuel, we can no longer be confident that
the increased number of desulfation cycles that will be required to address the impact of sulfur
on efficiency can be accomplished without unrecoverable thermal degradation and thus loss of

                                          4-51

-------
Regulatory Impact Analysis
NOx adsorber efficiency. Limiting the number and frequency of these deleterious desulfation
events through the use of diesel fuel with sulfur content less than 15 ppm allows us to conclude
with confidence that NOx adsorber catalysts will be developed that are durable throughout the
life of a nonroad diesel engine.

    4.1.2.3.5 Current Status of NOx Adsorber Development

    NOx adsorber catalysts were first introduced in the power generation market less than five
years ago.  Since then, NOx adsorber systems in stationary source applications have enjoyed
considerable success.  In 1997, the South Coast Air Quality Management District of California
determined that a NOx adsorber system provided the "Best Available Control Technology" NOx
limit for gas turbine power systems.90  Average NOx control for these power generation facilities
is in excess of 92 percent.91 A NOx adsorber catalyst applied to a natural gas fired powerplant
has demonstrated better than 99 percent reliability for more than 21,000  hours of operation while
controlling NOx by more than 90 percent.92 The experience with NOx adsorbers in these
stationary power applications shows that NOx adsorbers can be highly effective for controlling
NOx emissions for extended periods of operation with high reliability.

       4.1.2.3.5.1 Lean-Burn Gasoline Engines

    The NOx adsorber's ability to control NOx under oxygen-rich (fuel-lean) operating
conditions has led industry to begin applying NOx adsorber technology to lean-burn engines in
mobile source applications. NOx adsorber catalysts have been developed and are now in
production for lean-burn gasoline vehicles in Japan, including several vehicle models sold by
Toyota Motor Corporation.L The 2000 model year saw the first application of this technology in
the United States with the introduction of the Honda Insight, certified to the California LEV-I
ULEV category standard.  Table 4.1-6 lists some of the 2002  European lean-burn direct-injection
gasoline vehicles that use NOx adsorber catalyst technology.93 These lean-burn gasoline
applications are of particular interest because they are similar to diesel vehicle applications in
terms of lean-NOx storage and the need for periodic NOx regeneration under transient driving
conditions. The fact that they have been successfully applied to these mobile source applications
shows clearly that NOx adsorbers can work under transient conditions provided that engineering
solutions can be found to periodically cause normally lean-burn exhaust conditions to operate in
a rich regeneration mode.
   L Toyota requires that their lean-burn gasoline engines equipped with NOx adsorbers are fueled on premium
gasoline in Japan, which has an average sulfur content of six ppm.

                                           4-52

-------
          	Technologies and Test Procedures for Low-Emission Engines

          Table 4.1-6 2002 European Lean-Burn Gasoline Direct-Injection Engines
Model
Audi A2 FSI
Audi A4 FSI
BMW 760 il_
Citroen C5H PI
Mercedes CLK 200 CGI
Mercedes C 200 CGI
Mitsubishi Carisma GDI
Mitsubishi Space Star GDI
Mitsubishi Space Wagon 2.4 GDI
Mitsubishi Space Runner 2.4 GDI
Mitsubishi Galant2.4GDI
Mitsubishi Pajero Pinin 2.0 GDI
Mitsubishi Pajero 3.2 V6 GDI
Peugeot 406 HPI
VW Lupo FSI
VW Polo FSI
VW Golf FSI
VW Bora FSI
Volvo S40 1.8
Displacement(liter)
1.6
2
6
2
1.8
1.8
1.8
1.8
2.4
2.4
2.4
2
3.5
2
1.4
1.4
1.6
1.6
1.6
Power(KVWPS)
81/110
110/150
ca. 300/408
103/140
125/170
125/170
90/122
90/122
108/147
110/150
106/144
90/122
149/202
103/140
77/105
63/85
81/110
81/110
90/122
       4.1.2.3.5.2 EPA National Vehicle and Fuel Emissions Laboratory

   As part of an ongoing effort to evaluate the rapidly developing state of this technology, the
Manufacturers of Emission Control Association (MECA) have provided numerous NOx
adsorber catalyst formulations to EPA for evaluation. Testing of some of these catalysts at
NVFEL revealed that formulations were capable of reducing NOx emissions by more than 90
percent over the broad range of operation in the highway steady-state SET procedure (sometimes
called the EURO 4 test).  At operating conditions representative of "road-load" operation for a
highway trucks, the catalysts showed NOx reductions as high as 99 percent resulting in NOx
emissions well below 0.1 g/hp-hr from an engine-out level of nearly 5 g/hp-hr.  Figure 4.1-9
shows  an engine torque vs. speed map with the various steady-state test modes used in this
testing as well  as the 8 modes of the ISO-C1 cycle used for nonroad certification. Though not
included in the test results shown in Figures 4.1-10 through 4.1-12, the ISO-C1 modes are
closely approximated by  some other test modes, as can be seen in Figure 4.1-9. We therefore
expect similarly good performance on the ISO-C1 test modes.  Testing on the highway transient
test procedure has shown similarly good results, with hot-start NOx  emissions over the highway
FTP cycle reduced by more than 90 percent. These results demonstrate that significant NOx
reductions are possible over a broad range of operating conditions with current NOx adsorber
technology, as typified by the highway FTP cycle and the SET procedure.

   The test program at NVFEL can be divided into phases.  The first phase began with an
adsorber screening process using a single leg of the planned dual-leg system. The goals of this
screening process, a description of the test approach, and the results  are described below.  The
                                         4-53

-------
Regulatory Impact Analysis
next phase of the test program consisted of testing the dual-leg system on a more advanced Tier
3 like diesel engine (i.e., with common rail fuel system and cooled EGR) using a NOx adsorber
chosen during the first phase in each of two legs.  The current ongoing phase is working on
improved systems approaches including a demonstration of an improved package four "leg"
system.

    Testing Goals—Single-Leg NOx Adsorber System

    The goal of the NOx adsorber screening process was to evaluate available NOx adsorber
formulations from different manufacturers with the objective of choosing an adsorber with 90
percent or better NOx reduction for continued evaluation.  To this end, four different adsorber
formulations were provided from three different suppliers.  Since this was a screening process
and since a large number of each adsorber formulation would be required for a full dual-leg
system, it was decided to run half of a dual-leg system (a single-leg system) and mathematically
correct the emissions and fuel economy impact to reflect a full dual-leg system. The trade-off
was that the single-leg system would be able to run only steady-state modes, as the emissions
could not be corrected over a transient cycle. The configuration used for this test was similar to
that shown  in Figure 4.1-1, but with a  catalyst installed only on one side of the system.

    Test Approach—Single-Leg NOx  Adsorber System

    The single-leg system consisted of an exhaust brake, a fuel injector, CDPF, and a NOx
adsorber in one test leg.  The other leg, the "bypass leg," consisted of an exhaust brake that
opened when the test-leg brake was closed; this vented the remainder of the exhaust out of the
test cell. Under this setup,  the test leg, i.e., the leg with  the adsorber, was directed into the
dilution tunnel where the emissions were measured and  then compensated to account for
emissions from the bypass  leg.  The restriction in the bypass leg was set to duplicate the
backpressure of the test leg so that, while bypassing the  test leg to conduct a NOx regeneration,
the backpressure of the bypass leg simulated the presence of a NOx adsorber system. A clean-up
diesel oxidation catalyst (DOC) downstream of the NOx adsorber was not used for this testing.

    The measured emissions had to be adjusted to account for the lack of any NOx adsorber in
the bypass leg. For this correction, it was assumed that the bypass leg's missing (virtual)
adsorber would adsorb only while the  actual leg was regenerating. It was also assumed the
virtual adsorber would have regeneration fuel requirements in proportion to its adsorbing time.
The emission-control performance of the virtual adsorber was assumed to be the same as the
performance of the actual adsorber. With these assumptions, the gaseous emissions could be
adjusted.94

    Test Results—Single-Leg NOx Adsorber System
                                          4-54

-------
	Technologies and Test Procedures for Low-Emission Engines

   Two sets of steady-state modes were run with each adsorber formulation. These modes
consisted of the SET modes and the AVL 8 mode composite FTP predict!on.M The modes are
illustrated in Figure 4.1-9 and are numbered sequentially one through 20 to include both the
eight AVL modes and the 13 SET modes (the idle mode is repeated in both tests).  The mode
numbers shown in the figure are denoted as "EPA" modes in the subsequent tables to
differentiate between the AVL and SET modes that have duplicate mode numbers.  The highway
NTE zone (which is the same as the nonroad NTE zone) is also shown in Figure 4.1-9 to show
that these two sets of modes give  comprehensive coverage of the NTE zone.  The ISO Cl
steady-state modes used for nonroad engines are closely represented by the test modes shown
here. The only Cl mode not well represented is the 10 percent load point (ISO Mode  5), which
is outside of the nonroad NTE zone.  The modes were run with varying levels of automation,
with the general strategy being to inject sufficient fuel during regeneration to obtain a lambda at
or slightly fuel-rich of stoichiometric (A < 1). The NOx regenerations were then  timed to achieve
the desired NOx reduction performance. The adsorber formulations were identified as A, B, D,
and E. Prior to testing,  each set of adsorbers were aged at 2500 rpm,  150 Ib-ft for 40 minutes,
then 2500 rpm full load for 20 minutes, repeated for a total of 10 hours.
       Figure 4.1-9 Steady-State Test Modes from NVFEL Testing and ISO C-l Modes
    700
    600 --
    500 --
   C- 400 - -
   Torque
- - NOx NTE
   PM NTE
»  AVL Modes
A  SET modes
•  ISOC-1 Modes
    300
    200 -
     100
                             1400
                                     1600     1800     2000
                                          Speed (rpm)
                                                            2200
                                                                   2400
                                                                           2600
                                                                                   2800
   M
      The AVL 8 mode test procedure is a steady-state test procedure developed by Anstalt fur
Verbrennungskraftmaschinen, Prof. Dr. Hans List (or Institute for Internal Combustion Engines) to approximate
emission levels that would occur while operating the engine over the transient highway FTP cycle.
                                          4-55

-------
Regulatory Impact Analysis
   The SET and AVL Composite emission results, along with the NOx reduction performance
vs. adsorber inlet temperature, are shown in Figures 4.1-10 through 4.1-13 for each of the tested
NOx adsorber formulations. The SET composites for all four adsorber formulations had NOx
reductions in excess of 90 percent with under a three percent impact on fuel economy. The HC
emissions varied most widely, most likely due to differences in regeneration strategies, and to
some extent, adsorber formulation. The HC emissions with the exception of adsorber "A" were
very good, less than 0.1 g/hp-hr over the SET and less than 0.2 g/hp-hr over the AVL composite.
Note that no DOC was used to clean up the HC emissions.

   Another point to note is that the EPA mode 1 (ISO-C1 Mode 11) data for each composite is
the same. This is because EPA mode 1, low idle, is too cold for effective steady-state
regeneration, but efficient NOx adsorption can occur for extended periods of time. (Note that the
exhaust temperature at idle is well below the NTE threshold of 250°C discussed earlier.) For
either of these composite tests, a regeneration would not be needed under such conditions. EPA
mode 1 has very little impact on either composite in any case because of the low power and
emission  rate. EPA mode 2 also had very  low steady-state temperatures, and the difficulty
regenerating at this mode can be seen in the impacts on HC emissions and on fuel economy.
But, like EPA mode 1, the engine would adsorb during EPA mode 2 for extended periods
without needing regeneration. None of the ISO-C1 modes, other than the idle mode, are similar
to EPA mode 2. Further, no attempt was made to apply new combustion approaches such as the
Toyota low-temperature combustion technology to raise exhaust temperatures at these operating
modes.

   The AVL composite showed greater differences between the adsorber formulations than the
SET.  Three of the adsorbers achieved greater than 90 percent NOx reduction over the AVL
composites with the other adsorber at 84 percent NOx reduction. The greater spread in NOx
reduction performance was, in part, due to this composite's emphasis on EPA mode 8, which
was at the upper end of the NOx reduction efficiency temperature window. Adsorber E had an
EPA mode  8 NOx reduction of 66 percent, and the NOx reduction efficiency vs. inlet
temperature graph clearly shows that this formulation's performance falls off quickly above
450°C. In contrast, the other formulations do not show  such an early, steep loss in performance.
The fuel economy impacts vary  more widely also, partly due to the test engineers' regeneration
strategies, particularly with the low-temperature modes, and to the general inability to regenerate
at very low-temperature modes at steady-state. Note also that none of the regeneration strategies
here can be considered fully optimized, as they reflect the product of trial and error
experimentation by the test engineers.  With further testing and understanding of the technology
a more systematic means for optimization  should be possible. In spite of the trial and error
approach the results shown here are quite promising.

   The AVL composite was developed as a steady-state test that would predict engine-out
emission  levels  over the transient highway FTP cycle.  As discussed in  4.1.3.1.2 below, NOx
adsorber control effectiveness is projected to be more effective over the NRTC than over the
highway FTP cycle. The AVL cycle loses some accuracy when testing engines with NOx
adsorbers, since regeneration does not occur at the low-temperature modes (EPA modes 1, 2, 5).
In real-world conditions, diesel engines do not come to  steady-state temperatures at any of these

                                         4-56

-------
	Technologies and Test Procedures for Low-Emission Engines

modes, and the adsorber temperatures will be higher at EPA modes 1, 2, and 5 than the stabilized
steady-state values used for this modal testing.  Consequently, the actual performance over a
transient duty cycle should be much better than the composites would suggest (see the discussion
of transient testing below).

   Based on the composite data and the temperature performance charts, amongst other factors,
adsorber formulation B was chosen for further dual-leg performance work. Both composites for
this formulation were well above 90 percent. The NOx vs. temperature graph, Figure 4.1-11,
also shows that this formulation was a very good match for this engine.
                                        4-57

-------
Regulatory Impact Analysis
Base
EPA
Mode
1
9
10
11
12
13
14
15
16
17
18
19
20
SET
Mode
1
2
3
4
5
6
7
8
9
10
11
12
13
SET
Weighting
15%
8%
10%
10%
5%
5%
5%
9%
10%
8%
5%
5%
5%
Speed
(rpm)
Idle
1619
1947
1947
1619
1619
1619
1947
1947
2275
2275
2275
2275
Torque
(Ib-ft)
0
630
328
493
332
498
166
630
164
599
150
450
300
BSNOx
(g/hp-hr)
13.0
4.6
4.7
5.0
5.0
5.0
5.5
4.0
5.0
4.0
4.8
5.0
4.8
Composite Results 4.6
Base
EPA
Mode
1
2
3
4
5
6
7
8
AVL
Mode
1
2
3
4
5
6
7
8
AVL
Weighting
42%
8%
3%
4%
10%
12%
12%
9%
Speed
(rpm)
Idle
987
1157
1344
2500
2415
2415
2313
Torque
(Ib-ft)
0
86
261
435
94
228
394
567
BSNOx
(g/hp-hr)
13.00
8.80
8.40
5.90
5.50
4.60
4.90
4.10
Composite Results 4.9
Adsorber
Inlet T
(C)
144
461
357
411
384
427
287
498
293
515
282
404
357
BSNOx
(g/hp-hr)
0.16
0.11
0.07
0.06
0.13
0.24
0.25
0.89
0.14
0.48
0.42
0.08
0.14
NOx Red

100%
98%
98%
99%
97%
95%
95%
78%
97%
88%
91%
98%
97%
HC*
(g/hp-hr)
0.00
0.92
1.02
1.35
0.11
0.81
1.39
0.36
1.88
1.12
0.68
0.62
0.70
FE Impact
*
0.0%
2.4%
2.0%
2.6%
1 .3%
1 .6%
3.3%
1 .9%
4.1%
3.8%
3.5%
3.0%
2.8%
0.31 93% 0.91 * 2.6% *
Adsorber
Inlet T
(C)
144
172
346
430
286
325
386
505
BSNOx
(g/hp-hr)
0.16
0.83
0.36
0.20
0.37
0.08
0.10
1.06
NOx Red

100%
91%
96%
97%
93%
98%
98%
74%
HC*
(g/hp-hr)
0.00
0.75
1.10
2.16
4.93
2.30
2.38
0.03
FE Impact
*
0.0%
7.7%
3.1%
3.0%
3.6%
3.6%
3.1%
1 .9%
0.44 91% 1.69* 2.9%*
             * HC results & FE Impacts do not reflect future potential as they are derived using a 5 g NOx engine which requires more frequent NOx regens

                   than would result using a 2.5 g engine and the tested system was not a fully optimized engine & emission control system.
                         100%
                          80%
                       o

                       
-------
            Technologies and Test Procedures for Low-Emission Engines
Base
EPA
Mode
1
9
10
11
12
13
14
15
16
17
18
19
20
SET
Mode
1
2
3
4
5
6
7
8
9
10
11
12
13
SET
Weighting
15%
8%
10%
10%
5%
5%
5%
9%
10%
8%
5%
5%
5%
Speed
(rpm)
Idle
1619
1947
1947
1619
1619
1619
1947
1947
2275
2275
2275
2275
Torque
(Ib-ft)
0
630
328
493
332
498
166
630
164
599
150
450
300
BSNOx
(g/hp-hr)
13.0
4.6
4.7
5.0
5.0
5.0
5.5
4.0
5.0
4.0
4.8
5.0
4.8
Composite Results 4.6
Base
EPA
Mode
1
2
3
4
5
6
7
8
AVL
Mode
1
2
3
4
5
6
7
8
AVL
Weighting
42%
8%
3%
4%
10%
12%
12%
9%
Speed
(rpm)
Idle
987
1157
1344
2500
2415
2415
2313
Torque
(Ib-ft)
0
86
261
435
94
228
394
567
BSNOx
(g/hp-hr)
13.00
8.80
8.40
5.90
5.50
4.60
4.90
4.10
Composite Results 4.9
Adsorber
Inlet T
(C)
144
498
366
446
375
420
296
524
293
537
280
426
357
BSNOx
(g/hp-hr)
0.16
0.18
0.07
0.14
0.06
0.07
0.18
0.46
0.36
0.56
0.29
0.24
0.11
NOx Red

100%
96%
98%
97%
99%
98%
97%
89%
93%
86%
94%
95%
98%
HC*
(g/hp-hr)
0.00
0.01
0.04
0.01
0.08
0.10
0.10
0.01
0.05
0.04
0.03
0.04
0.02
FE Impact
*
0.0%
1 .2%
0.5%
1 .5%
0.7%
2.3%
0.3%
3.2%
0.4%
4.3%
0.4%
4.3%
0.9%
0.27 94% 0.03* 2.2%*
Adsorber
Inlet T
(C)
144
162
355
446
263
346
403
544
BSNOx
(g/hp-hr)
0.16
0.56
0.30
0.09
0.66
0.11
0.05
0.73
NOx Red

100%
94%
96%
98%
88%
98%
99%
82%
HC*
(g/hp-hr)
0.00
2.11
0.16
0.23
0.25
0.03
0.02
0.35
FE Impact
*
0.0%
1 .8%
0.3%
0.9%
1 .6%
0.4%
1 .4%
4.0%
0.33 93% 0.19* 2%*
* HC results & FE Impacts do not reflect future potential as they are derived using a 5 g NOx engine which requires more frequent NOx regens

      than would result using a 2.5 g engine and the tested system was not a fully optimized engine & emission control system.
      100%
       80%
.3!

£
UJ
c
o

'u


I
OL
       60%
       40%
       20%
       0%
         200      250      300      350     400      450

                            Adsorber Inlet Temperature (C)
                                                             500
                                                                     550
     Figure 4.1-11.  SET & AVL Composites, and Temperature vs.

                       NOx Chart for Adsorber B
                                    4-59

-------
Regulatory Impact Analysis
Base
EPA
Mode
1
9
10
11
12
13
14
15
16
17
18
19
20
SET
Mode
1
2
3
4
5
6
7
8
9
10
11
12
13
SET
Weighting
15%
8%
10%
10%
5%
5%
5%
9%
10%
8%
5%
5%
5%
Speed
(rpm)
Idle
1619
1947
1947
1619
1619
1619
1947
1947
2275
2275
2275
2275
Torque
(Ib-ft)
0
630
328
493
332
498
166
630
164
599
150
450
300
BSNOx
(g/hp-hr)
13.00
4.60
4.70
5.00
5.00
5.00
5.50
4.00
5.00
4.00
4.80
5.00
4.80
Composite Results 4.6
Adsorber
Inlet T
(C)
144
451
356
400
377
431
305
501
303
489
278
391
330
BSNOx
(g/hp-hr)
0.16
0.18
0.14
0.09
0.07
0.11
0.23
0.16
0.15
0.93
0.57
0.12
0.21
NOx Red

100%
96%
97%
98%
99%
98%
96%
96%
97%
93%
88%
98%
96%
HC*
(g/hp-hr)
0.00
0.07
0.15
0.05
0.01
0.02
0.14
0.04
0.14
0.09
0.18
0.10
0.09
FE Impact
*
0.0%
1 .3%
1 .7%
1 .6%
1 .2%
1 .6%
2.3%
2.1%
3.1%
1 .7%
3.5%
1 .8%
2.9%
0.28 94% 0.08* 1.9%*
Base
EPA
Mode
1
2
3
4
5
6
7
8
AVL
Mode
1
2
3
4
5
6
7
8
AVL
Weighting
42%
8%
3%
4%
10%
12%
12%
9%
Speed
(rpm)
Idle
987
1157
1344
2500
2415
2415
2313
Torque
(Ib-ft)
0
86
261
435
94
228
394
567
BSNOx
(g/hp-hr)
13.00
8.80
8.40
5.90
5.50
4.60
4.90
4.10
Composite Results 4.9
Adsorber
Inlet T
(C)
144
162
359
427
273
301
363
493
BSNOx
(g/hp-hr)
0.16
0.56
0.08
0.14
1.25
0.52
0.66
0.31
NOx Red

100%
94%
99%
98%
77%
89%
87%
92%
HC*
(g/hp-hr)
0.00
2.11
0.30
0.19
0.26
0.13
0.04
0.08
FE Impact
*
0.0%
1 .8%
3.1%
1 .7%
6.4%
1 .9%
1 .4%
1 .6%
0.51 90% 0.14* 1.9%*
                * HC results & FE Impacts do not reflect future potential as they are derived using a 5 g NOx engine which requires more frequent NOx regens

                      than would result using a 2.5 g engine and the tested system was not a fully optimized engine & emission control system.
                      100%
                       80%
                    u
                    C
                    0)
                    'o

                    it
                    UJ

                    C
                    o
60%
                       40%
                       20%
                        0%
                          200      250       300       350       400      450

                                               Adsorber Inlet Temperature (C)
                                                                                  500
                                                                                           550
                     Figure 4.1-12.  SET & AVL Composites, and Temperature vs.

                                         NOx Chart for Adsorber D
                                                    4-60

-------
         Technologies and Test Procedures for Low-Emission Engines
Base
EPA
Mode
1
9
10
11
12
13
14
15
16
17
18
19
20
SET
Mode
1
2
3
4
5
6
7
8
9
10
11
12
13
SET
Weighting
15%
8%
10%
10%
5%
5%
5%
9%
10%
8%
5%
5%
5%
Speed
(rpm)
Idle
1619
1947
1947
1619
1619
1619
1947
1947
2275
2275
2275
2275
Torque
(Ib-ft)
0
630
328
493
332
498
166
630
164
599
150
450
300
BSNOx
(g/hp-hr)
13.00
4.60
4.70
5.00
5.00
5.00
5.50
4.00
5.00
4.00
4.80
5.00
4.80
Composite Results 4.6
Adsorber
Inlet T
(C)
144
455
343
442
377
419
412
392
294
492
388
391
327
BSNOx
(g/hp-hr)
0.16
0.47
0.07
0.36
0.08
0.29
0.14
0.05
0.09
0.95
0.11
0.12
0.22
NOxRed

100%
89%
98%
93%
98%
94%
98%
99%
98%
76%
98%
98%
95%
HC*
(g/hp-hr)
0.00
0.02
0.05
0.07
0.01
0.03
0.05
0.02
0.26
0.03
0.03
0.10
0.02
FE Impact
*
0.0%
2.1%
0.9%
9.0%
1 .5%
1 .6%
1 .7%
2.1%
4.4%
2.0%
2.4%
1 .8%**
1 .4%
** Md 19 data from Adsorber D
0.33 93% 0.05* 2.9%*
Base
EPA
Mode
1
2
3
4
5
6
7
8
AVL
Mode
1
2
3
4
5
6
7
8
AVL
Weighting
42%
8%
3%
4%
10%
12%
12%
9%
Speed
(rpm)
Idle
987
1157
1344
2500
2415
2415
2313
Torque
(Ib-ft)
0
86
261
435
94
228
394
567
BSNOx
(g/hp-hr)
13.00
8.80
8.40
5.90
5.50
4.60
4.90
4.10
Composite Results 4.9
Adsorber
Inlet T
(C)
144
166
339
449
256
313
372
508
BSNOx
(g/hp-hr)
0.16
7.39
0.09
0.65
1.36
0.35
0.12
1.39
NOx Red

100%
16%
99%
89%
75%
92%
97%
66%
HC*
(g/hp-hr)
0.00
1.02
0.05
0.01
0.91
0.21
0.10
0.04
FE Impact
*
0.0%
71 .9%
2.3%
2.1%
15.8%
5.6%
2.6%
3.3%
0.80 84% 0.16* 5.4%*
* HC results & FE Impacts do not reflect future potential as they are derived using a 5 g NOx engine which requires more frequent NOx regens
      than would result using a 2.5 g engine and the tested system was not a fully optimized engine & emission control system.
     100%
      80%
      60%
u
C
0)
'o
it
UJ
C
o
   =  40%


   1
      20%
       0%
         200      250      300       350      400       450

                            Adsorber Inlet Temperature (C)
                                                            500
                                                                     550
    Figure 4.1-13. SET & AVL Composites, and Temperature vs.

                       NOx Chart for Adsorber E
                                 4-61

-------
Regulatory Impact Analysis
   Testing Goals—Dual-Leg NOx Adsorber System

   After completing the screening process and selecting NOx adsorber "B," the dual-leg system
was developed. The dual-leg system was first tested on the same ISB engine as was used for the
single-leg testing. The results from that portion of the testing were similar to the single-leg
results (i.e., >90 percent NOx reductions for most test modes) and were reported in the HD2007
Regulatory Impact Analysis.95 Subsequent testing of the NOx adsorber system was made at
NVFEL but with a new ISB engine that had been upgraded to include nonroad Tier 3 type
technologies, such as common rail fuel injection and cooled EGR.  The change in engine
technology led to significantly lower engine-out emissions (similar to the levels expected for
2004 highway engines Tier 3 nonroad engines) and to different exhaust gas temperature
characteristics.  As a result of the engine changes, the overall system performance was improved
on both the steady-state test points and on the transient highway FTP cycle.96  As discussed
further in Section 4.1.3.1.2 below, performance  over the NRTC is projected to be better than for
the highway FTP cycle. Also, as can be seen in Figure 4.1-9 above, the SET steady-state test
points are not significantly different from the ISO Cl test points (to which  nonroad engines
would be subject). Emission reductions are therefore expected to be similar.

   Testing Approach—Dual-Leg NOx Adsorber System

   The steady-state SET testing was conducted in a manner similar to that used in the screening
process described above.  The modes were run with varying levels of automation, with the
general strategy being to inject sufficient fuel during regeneration to obtain a lambda at or
slightly fuel-rich of stoichiometric (A < 1).  The NOx regenerations were then timed to achieve
the targeted 90 percent NOx reduction. The regeneration control and optimization strategies are
described in more detail in an SAE paper included in the docket for this rule.97

   Transient regeneration control over the highway FTP cycle was accomplished using a time-
based regeneration schedule. This control regenerated on a prescribed schedule of time and fuel
quantities, so regenerations occurred at predetermined engine conditions during the transient
cycle.

   The emission results presented here are only for hot-start portions of the highway FTP cycle.
The adsorber system was not optimized for cold-start performance and does not provide a
meaningful assessment of adsorber warmup performance.  To better simulate the "cold-soak-
hot" procedure called for in highway FTP cycle, a preconditioning mode was chosen to provide
adsorber temperatures at the start of the "hot" cycle similar to those found  following the "cold-
soak" portion of the test. The mode chosen was EPA mode 10 (1947 rpm,  328 Ib-ft), which
resulted in adsorber inlet temperatures (i.e., at the outlet of the CDPF) at the start of the hot cycle
of about 280°C.  Another purpose for the preconditioning was to ensure the adsorbers were in the
same condition at the start of each test. Given that our regeneration control system did not
automatically take into account the starting condition of the NOx adsorbers, this preconditioning
was necessary to provide repeatable transient test results.
                                          4-62

-------
                     Technologies and Test Procedures for Low-Emission Engines
   Test Results—Dual-Leg NOx Adsorber System

   The highway SET is made up of the 13 Euro III modes. Several modes were run twice by
different engineers, and the best calibration was chosen for the SET composite. Table 4.1-7
shows the SET composite test results. These data show that 90 percent NOx reductions were
possible over the SET composite, with a modal NOx reduction range from 89 percent to nearly
100 percent. The adsorber NOx and HC reduction performance varied primarily as a function of
exhaust temperature.

                  Table 4.1-7 SET Results for Dual-Leg System at NVFEL
       Modal and composite SET NOx and HC emissions results for the Modified Cummins ISB engine.
Modified Cummins ISB
(HPCR, cooled EGR)
SET
Mode
1
2
3
4
5
6
7
8
9
10
11
12
13
SET
Weighting
15%
8%
10%
10%
5%
5%
5%
9%
10%
8%
5%
5%
5%
Speed
(rpm)
Idle
1649
1951
1953
1631
1626
1623
1979
1951
2348
2279
2275
2274
Torque
(Ib-ft)
0
633
324
490
328
496
161
609
159
560
145
447
296
SET Weighted Composite Results:
BSNOx
(g/hp-hr)
6.95
3.10
1.79
1.98
1.90
2.35
2.05
2.09
1.68
1.95
1.66
1.84
1.76
2.10
BSHC
(g/hp-hr)
6.77
0.08
0.21
0.12
0.22
0.09
0.56
0.08
0.49
0.11
0.57
0.14
0.25
0.17
Modified Cummins ISB
(Baseline + CDPF and NOx adsorber catalysts)
Outlet T
(°C)
144
529
403
486
403
504
313
524
323
524
306
465
400

BSNOx
(g/hp-hr)
0.16
0.33
0.06
0.07
0.10
0.07
0.02
0.19
0.01
0.10
0.01
0.10
0.03
0.12
NOx (%-
Reduction)
100%
89%
96%
96%
95%
97%
99%
91%
100%
95%
99%
95%
98%
94%
BSHC
(g/hp-hr)
0.00
0.03
0.01
0.02
0.01
0.02
0.03
0.03
0.02
0.04
0.02
0.01
0.01
0.03
Reductant FE
Impact (%)*
0.0%
1.6%
1.0%
1.3%
0.9%
1.6%
0.9%
1.7%
0.8%
2.3%
0.7%
0.9%
0.9%
1.4%**
Notes:
* Fuel economy impact of fuel-reductant addition for NOx adsorber regeneration.
** Increased exhaust restriction from the wall-flow and flow through monoliths results in a further FE impact of approximately 1-2% over
the SET composite.
   The fuel economy impact was defined as the percent increase in fuel consumption caused by
the adsorber regeneration fuel, or the mass of fuel used for regeneration, divided by the mass of
fuel consumed by the engine during one regeneration and adsorption cycle. The fuel economy
impact varied from virtually zero to 2.3 percent depending on the mode with a composite fuel
economy impact of 1.4 percent. We anticipate significant improvements in regeneration
strategies are possible with different system configurations.  Also, changes in engine operation
designed to increase exhaust temperatures, not attempted in this work, can provide substantial
improvements in catalyst performance and potentially a lower fuel economy impact.

   Test Results over the Highway FTP Cycle

   As with the steady-state test results, the test results over the hot-start portion of the highway
FTP cycle showed NOx and PM emission reductions greater than 90 percent. The baseline
(without the catalyst system) NOx emissions of 2.7 g/hp-hr were reduced to 0.1 g/hp-hr with the
addition of the catalyst system, a better than 95 percent reduction in NOx emissions.  Similarly,
                                         4-63

-------
Regulatory Impact Analysis
the PM emissions were reduced to below 0.003 g/hp-hr from a baseline level of approximately
0.1 g/hp-hr, a reduction of more than 95 percent. The fuel economy impact associated with
regeneration of the NOx adsorber system was measured as 1.5 percent over the highway FTP
cycle. The fuel economy impact associated with increased exhaust restriction from the CDPF
was less than the measurement variability for the test cycle (i.e., less than 0.5 percent).98

   Durability Baseline NOx Adsorber Catalyst Testing

   Additional testing was conducted at NVFEL to provide baseline performance data to gauge
improvements in NOx adsorber durability performance in support of the FID2007 technology
reviews.  The data provide a look at the state of adsorber technology in 2001, with  a glimpse of
improvements that will be made in the future and is documented in a SAE paper."  It is clear
from the analysis that there were vast differences in the durability performance of the
formulations over these short tests. Adsorber suppliers were early on in their development and
rapid improvements were being made.  Two adsorbers representing one company's progress over
two years showed significantly better aging performance (i.e., less degradation over time). This
performance was evidenced by its NOX adsorbing and regeneration performance after 100
hours.100  In support of the U.S. EPA's continuing effort to monitor NOX adsorber progress, new
formulations are continuing to be evaluated.

   Development of a Four "Leg" System Design

   At NVFEL, developments have continued on methods and system designs for NOx adsorber
catalyst technologies.  A novel four-leg NOx adsorber/PM trap system was developed as an
evolution of the proof-of-concept two-leg system that was used for previous testing at NVFEL
(the system used  in the test results reported here). The four-leg system has a catalyst volume that
is less than half of the volume of the two-leg system.  This allows the four-leg system to be
packaged in a volume not much larger than a muffler for a medium heavy duty truck application
as can be seen in Figure 4.1-14. Efforts have also been made to reduce the cost of  the system by
using simpler injectors and valve actuators.
                                         4-64

-------
           	Technologies and Test Procedures for Low-Emission Engines

            Figure 4.1 -14 Prototype 4-leg System Compared with a Truck Muffler
   Initial testing indicates that the four-leg system at least matches the previous two-leg systems
NOx reduction efficiency with similar fuel consumption as can be seen in Figure 4.1-15. Note
that the results shown in the figure are based upon the NOx sensor data used in the control
system. Work is underway to confirm these steady-state results and to demonstrate the
performance over transient cycles.

              Figure 4.1-15 Preliminary Results for Prototype Four-Leg System
             95%
                     Mode FE Penalty
                                                                     3.6
                                         4-65

-------
Regulatory Impact Analysis
       4.1.2.3.5.3 Department of Energy (DOE) Test Programs

   The U.S. Department of Energy (DOE) has funded several test programs at national
laboratories and in partnership with industry to investigate the NOx adsorber technology.  Most
of these test programs are part of the Advanced Petroleum Based Fuel (APBF) program of
DOE's Office of Transportation Technology (OTT). The initial phases of the programs are often
referred to as the Diesel Emission Control Sulfur Effects (DECSE) program, which are part of
the APBF programs. Five reports documenting  the DECSE program are available from the DOE
OTT website (www.ott.doe.gov/decse) and were used extensively throughout our
analysis.101'102'103104'105

   In the DECSE program, an advanced diesel engine equipped with common rail fuel injection
and exhaust gas recirculation (EGR) was combined with a NOx adsorber catalyst to control NOx
emissions. The system used an in-cylinder control approach. Rich regeneration conditions are
created for the NOx adsorber catalyst regeneration through increased EGR rates and a secondary
injection event designed to occur late enough in the engine cycle so as not to change engine
torque output. Using this approach, the DECSE program has shown NOx conversion
efficiencies exceeding 90 percent over a catalyst inlet operating temperature window of 300°C to
450°C. This performance level was achieved while staying within the four percent fuel economy
penalty target defined for regeneration calibration.106

   Subsequent work organized under the APBF program is commonly referred to as the APBF-
Diesel Emission Control program, or APBF-DEC. The ongoing APBF-DEC work includes
additional phases to develop prototype CDPF/NOx adsorber systems for a heavy-duty truck, a
large sport utility vehicle and a passenger car. The program is looking at all important issues
related to the technology including, packaging systems, effective regeneration, emission
performance and durability.107

       4.1.2.3.5.4 Heavy-Duty Engine Manufacturers

   Heavy-duty diesel engine manufacturers (highway manufacturers) are currently developing
systems to comply with the FID2007 emission standards including the NOx adsorber technology.
As noted in EPA's Highway Diesel Progress Review Report 2, which documents in more  detail
progress by the highway diesel engine industry to  develop CDPF and NOx adsorber technology,
the progress to develop these emission-control systems is progressing rapidly.  Although much
of the work being done is protected as confidential business information, a recent public
presentation by Daimler Chrysler Powersystems is illustrative some of the work that has been
done prior to 2003.108 The presentation reviews three possible system configurations for a
combined CDPF / NOx adsorber system and compares the trade-offs among the approaches.
Similar to the results shown in Section 4.1.2.3.5.3 by EPA, a dual-leg system demonstrated 90
percent or higher NOx emission control over a wide range of operation.

   Two Japanese truck manufacturers, Toyota and Hino have recently introduced light heavy-
duty diesel trucks in Japan using the Toyota developed Diesel Particulate NOx Reduction

                                         4-66

-------
	Technologies and Test Procedures for Low-Emission Engines

(DPNR) catalyst system. The DPNR system described in a light-duty application in our 2002
Highway Diesel Progress Review, consists of a diesel paniculate filter with NOx storage catalyst
coated onto the PM filter substrate.  In some applications, the system can be further enhanced
with the addition of an oxidation catalyst and an additional NOx adsorber catalyst applied to a
conventional flow through catalyst substrate. The new trucks introduced in Japan, the Toyota
Dyna and the Hino Dutro are commonly used as urban delivery vehicles and as refuse hauling
vehicles.

   In July 2003, EPA engineers visited Toyota's Higashifuji Technical Center in Japan to
participate in testing of the engine and DPNR catalyst system being introduced later in the year
as the Toyota Dyna product.  EPA participated in several days of testing and reviewed detailed
technical information regarding the emission control system and its potential for further
development. The information shared with EPA in that test program was  designated as
confidential business information by Toyota. However, Toyota has published a relatively
detailed SAE paper in Japan describing the system and its performance.109 The paper
summarizes the demonstrated emission reduction of the vehicle as aged to an estimated 250,000
kilometers with NOx emissions controlled below 0.5 g/bhp-hr and PM emissions controlled
below 0.01 g/bhp-hr.

       4.1.2.3.5.5 Light-Duty Diesel Vehicle Manufacturers

   Diesel passenger car manufacturers are developing emission-control systems using NOx
adsorbers and PM filters in a combined control strategy to meet upcoming Euro IV emission
standards for larger passenger cars and sedans in Europe and the light-duty Tier 2 emission
standards in the United States. EPA has tested five prototype diesel passenger cars with these
technologies over the last year and a half. The results shown in Figure 4.1-16 demonstrate the
potential for substantial reductions with NOx adsorber and PM filter technologies when tested
with low-sulfur diesel fuel. All five vehicles demonstrated substantial reductions in NOx and
PM emissions when compared with a current relatively clean (compared with only a few years
ago) diesel passenger cars as represented by the  solid black diamond and solid black square in
Figure 4.1-16.110
                                          4-67

-------
Regulatory Impact Analysis
     Figure 4.1-16 Tier 2 Passenger Car Prototypes Tested at NVFEL on the FTP75 Cycle
       0.90
       0.85 --
       0.80 j
       0.75 :
       0.70 :
    | 0.65 --
    £ o.so -
    M
    5 0.55 .
0.50 -
0.45
0.40
0.35
0.30
0.25
0.20
0.15
0.10
0.05
0.00
   0.0
 Toyota Avensis D-CAT Station Wagon
 I VW Golf TDI Station Wagon
 ', Mercedes E320 Sedan
 "' Mercedes E320 Sedan
 APBF-DEC Audi A4 Station Wagon
 Vehicle "E"
* Tier 1 VW Beetle TDI
• Tier 2 Bin 10 VW Jetta Wagon TDI
195% Confidence Interval
                                       Ok miles NOx standard
                                           120k mies PM standard
                 5.0    10.0    15.0   20.0   25.0    30.0   35.0
                                       PM Emissions (mg/rni)
                                              40.0
45.0    50.0
55.0
    One vehicle in the test program, the Mercedes E320, was tested with both new catalyst
hardware and aged catalyst hardware. The aged catalyst had experienced the equivalent of the
100,000 km of aging. The aged test results show that the aged catalyst system has lost some
amount of NOx storage volume, causing the NOx emissions to breakthrough as the catalyst fills
with NOx prior to the periodic NOx regenerations. In this testing, the NOx regeneration period
was fixed for the new and aged catalyst at the same interval. It appears from the data that the
regeneration interval for the fresh catalyst was too infrequent for the aged catalyst, which  had a
reduced NOx-storage volume. At the very low NOx emission levels shown in the figure, it takes
only a very small breakthrough in NOx emissions to significantly increase the emissions over the
lowest control levels. Manufacturers are currently working to keep the number of regeneration
episodes to the minimum number to minimize  stress on catalyst materials (i.e., limit thermal
degradation as discussed in Section 4.2 above). We believe manufacturers are continuing to
develop more heat-resistant materials that will reduce overall aging of the catalyst.  If such
materials had been available at this time, we believe the NOx results for the aged vehicle would
have been better.  Note however, that the PM emissions show no deterioration for the aged
system compared with the new system.

    The most recently tested vehicle, vehicle "E" was tested after aging of the catalyst system to
the equivalent of 50,000 miles of vehicle operation.  The emissions results even after this
extended aging where very good demonstrating NOx emission levels below 0.07 g/mile and PM
                                           4-68

-------
	Technologies and Test Procedures for Low-Emission Engines

emissions below 0.01 g/mile. Relative to vehicle "D" this demonstrates substantial progress by
manufacturers to improve the overall durability of NOx adsorber catalysts.

    4.1.2.4 Selective Catalytic Reduction (SCR) Technology

    Another NOx catalyst-based emission-control technology is selective catalytic reduction
(SCR).  SCR catalysts require a reductant, ammonia, to reduce NOx emissions. Because of the
significant safety concerns with handling and storing ammonia, most SCR systems make
ammonia within the catalyst system from urea. Such systems are commonly called urea SCR
systems. Throughout this document, the term SCR and urea SCR may be used interchangeably
and should be considered as referring to the same urea-based catalyst system. With the
appropriate control system to meter urea in proportion to engine-out NOx emissions, urea SCR
catalysts can reduce NOx emissions by over 90 percent for a significant fraction of the diesel
engine operating range.111  Although EPA has not done  an extensive analysis to evaluate its
effectiveness, we believe it may be possible to  reduce NOx emissions with a urea SCR catalyst
to levels consistent with compliance with Tier 4 NOx standards.

    We have significant concerns regarding a technology that requires extensive user
intervention to function properly and the lack of the urea delivery infrastructure necessary to
support this technology. Urea SCR systems consume urea in proportion to the engine-out NOx
rate. The urea consumption rate can be on the  order of five percent of the engine fuel
consumption rate. Unless the urea tank is prohibitively large, the urea must therefore be
replenished frequently. Most urea systems are designed to be replenished every time fuel is
added or at most every few times that fuel is added. There is not a system in place today to
deliver or dispense automotive-grade urea to diesel fueling stations. One study conducted for the
National Renewable Energy Laboratory (NREL), estimated that if urea were to be distributed to
every diesel fuel station in the United States, the cost would be more than $30 per  gallon.112

    We are not aware of a proven mechanism that ensures that the user will replenish the urea
supply as necessary to maintain emission-control performance.  Further, we believe that,  given
the additional cost for urea, there will be significant disincentives for the end-user to replenish
the urea because the cost of urea can be avoided without equipment performance loss. See
NRDC v. EPA, 655 F.  2d 318, 332 (D.C. Cir. 1981) (referring to "behavioral barriers to periodic
restoration of a filter by a [vehicle] owner" as a valid basis for EPA considering  a technology
unavailable).  Due to the lack of an infrastructure to deliver the needed urea, and the lack of a
track record of successful ways to ensure urea use, we have concluded that the urea SCR
technology is not likely to be available for general use in the time frame of the Tier 4 standards.
We have therefore not based the feasibility or cost analysis of this emission-control program on
the use or availability of the urea SCR technology.  However, we do not preclude its use for
compliance with the emission standards, provided that a manufacturer can demonstrate
satisfactorily that the engine will use urea under all conditions.  We believe that  consistent use of
urea can only be ensured only for a few unique installations and therefore believe it is
inappropriate to base a national emission-control program on a technology that can effectively
serve only in a few niche applications.
                                          4-69

-------
Regulatory Impact Analysis
   This section has described several technologies that can reduce emissions from diesel
engines.  The following section describes the challenges to applying these diesel engine
technologies to engines and equipment designed for nonroad applications.

4.1.3 Can These Technologies Be Applied to Nonroad Engines and Equipment?

   The emission standards and the introduction dates for those standards, as described earlier in
Section III of the preamble,  are premised on the transfer of diesel engine technologies being, or
already developed, to meet light-duty and heavy-duty vehicle standards that begin in 2007.  The
Tier 4 aftertreatment based standards for engines from 75-750 hp will begin to go into effect four
years later. This time lag between equivalent highway and nonroad diesel engine standards is
necessary to allow time for engine and equipment manufacturers to further develop these
highway  engine technologies for nonroad engines and to align this program with nonroad Tier 3
emission standards that begin to go into effect in 2006.

   The test procedures and regulations for the HD2007 highway engines include a transient test
procedure, a broad steady-state procedure and NTE provisions that require compliant engines to
emit at or below 1.5 times the regulated emission levels under virtually all conditions.  An
engine designed to comply with the HD2007 emission standards will meet the Tier 4 standards if
it is tested over the transient and steady-state duty cycles specified in the final rule, which cover
the same regions and types of engine operation. Said in another way, a highway diesel engine
produced in 2007 may be certified in compliance with the transient and steady-state standards in
this final rule for nonroad diesel engines several years in advance of the date when these
standards are scheduled to go into effect. However, that engine, while compliant with certain of
the nonroad emission standards, would not necessarily be  designed to address the various
durability and performance requirements of many nonroad equipment manufacturers. We expect
that the engine manufacturers will need additional time to further develop the necessary
emission-control systems to address  some of the nonroad issues described below as well as to
develop the appropriate calibrations for engine rated speed and torque characteristics required by
the diverse range of nonroad equipment. Furthermore, not all nonroad engine manufacturers
produce highway diesel engines or produce nonroad engines that are  developed from highway
products.  There is therefore a need for lead time between  the Tier 3 emission standards, which
go into effect in 2006-2008, and the Tier 4 emission  standards.  We believe the technologies
developed to comply with the Tier 3  emission standards such as improved air handling systems
and electronic fuel systems will form an essential technology baseline that manufacturers will
need to initiate and control the various regeneration functions required of the catalyst-based
technologies for Tier 4. The Agency has given consideration to all these issues in setting the
levels and timing of the Tier 4 emission standards.

   This section presents some of the challenges of applying advanced emission-control
technologies to nonroad engines and equipment and describes why we believe technologies
developed for highway diesel engines can be further refined to address these issues in a timely
manner for nonroad engines consistent with the Tier 4 emission standards.
                                          4-70

-------
	Technologies and Test Procedures for Low-Emission Engines

   4.1.3.1 Nonroad Operating Conditions and Exhaust Temperatures

   Nonroad equipment is highly diverse in design, application, and typical operating conditions.
This variety of operating conditions affects emission-control systems through the resulting
variation in the torque and speed demands (i.e., power demands). This wide range in what
constitutes typical nonroad operation makes the design and implementation of advanced
emission-control technologies more difficult. The primary concern for catalyst-based emission-
control technologies is exhaust temperature. In general, exhaust temperature increases with
engine power and can vary dramatically as engine power demands vary.

   For most catalytic emission-control technologies there is a minimum temperature below
which the chemical reactions necessary for emission control  do not occur. The temperature
above which substantial catalytic activities is realized is often called the light-off temperature.
For gasoline engines, the light-off temperature is typically important only in determining cold-
start emissions.  Once gasoline vehicle exhaust temperatures exceed the light-off temperature,
the catalyst is "lit-off" and remains fully functional under all operating conditions.  Diesel
exhaust is significantly cooler than gasoline exhaust due to the  diesel engine's higher thermal
efficiency and its operation under predominantly lean conditions. Absent control action taken by
an electronic engine control system, diesel exhaust may fall below the light-off temperature of
catalyst technology even when the engine is fully warmed up.

   The relationship between the exhaust temperature of a nonroad diesel engine and light-off
temperature is an important factor for both CDPF and NOx adsorber technologies.  For the
CDPF technology, exhaust temperature determines the rate of filter regeneration and if too low
causes a need for supplemental means to ensure proper filter regeneration. In the case of the
CDPF, it is the aggregate soot regeneration rate that is important, not the regeneration rate at any
particular moment in time. A CDPF controls PM emissions under all conditions and can
function properly (i.e., not plug) even when exhaust temperatures are low for an extended time
and the regeneration rate is lower than the soot accumulation rate, provided that occasionally
exhaust temperatures and thus the soot regeneration rate are increased enough to regenerate the
CDPF. A CDPF can passively (without supplemental heat addition) regenerate if exhaust
temperatures remain above 250°C for more than 40 percent of engine operation.113  Similarly
(and as discussed in more detail earlier), there is a minimum temperature (e.g., 200°C) for NOx
adsorbers below which NOx regeneration is not readily possible and a maximum temperature
(e.g., 500°C) above which NOx adsorbers are unable to effectively store NOx.  These minimum
and maximum temperatures define a characteristic temperature window of the NOx adsorber
catalyst.  When the exhaust temperature is within the temperature window (above the minimum
and below the maximum) the catalyst is highly effective. When exhaust temperatures fall
outside this window of operation, NOx adsorber effectiveness is diminished. There is therefore a
need to match diesel exhaust temperatures to conditions for effective catalyst operation under the
various operating conditions of nonroad engines.

   Although the range of products for highway vehicles is not as diverse as for nonroad
equipment, the need to match exhaust temperatures to catalyst characteristics is still present.
This is a significant concern for highway engine manufacturers and has been a focus of our

                                         4-71

-------
Regulatory Impact Analysis
ongoing diesel engine progress review.  There we have learned that substantial progress is being
made to broaden the operating temperature window of catalyst technologies, while at the same
time, engine systems are being  designed to better control exhaust temperatures. Highway diesel
engine manufacturers are working to address this need through modifications to engine design,
modifications to engine control strategies and modifications to exhaust system designs. Engine
design changes including the ability for multiple late fuel injections and the ability to control
total air flow into the engine give controls engineers additional flexibility to change exhaust
temperature characteristics. Modifications to the exhaust system, including the use of insulated
exhaust manifolds and exhaust  tubing, can help to preserve the temperature of the exhaust gases.
New engine control strategies designed to take advantage of engine and exhaust system
modifications can then be used to manage exhaust temperatures across a broad range of engine
operation. The technology solutions being developed for highway engines to better manage
exhaust temperature are built upon the same emission-control technologies (i.e., advanced air
handling systems and electronic fuel-injection systems) that we expect nonroad engine
manufacturers to use for meeting the Tier 3 emission standards.

    4.1.3.1.1 CDPFS and Nonroad Operating Temperatures

    EPA has conducted a screening analysis to better understand the effect of engine operating
cycles  and engine power density on exhaust temperatures, specifically to see if passive CDPF
regeneration can be expected under all conditions for nonroad engine applications. Our
approach for assessing the likelihood of passive regeneration by a CDPF is based on what we
learned from the literature as well  as information submitted by various catalyst manufacturers for
product verification to our voluntary diesel retrofit program.

    For this analysis three representative nonroad engines were tested.  The engines are
described in Table 4.1-8.  In the case of the Cummins engine, the testing was done at three
different engine ratings (250hp, 169hp, and 124hp) to evaluate the effect of engine power density
on expected exhaust temperatures  and therefore the likelihood of passive PM filter regeneration.

                                       Table 4.1-8
                    Engines Tested to Evaluate PM Filter Regeneration
Engine Model
Lombardini
LDW1003-FOCS
Kubota V2203-E
Cummins I SB
Model
Year
2001
1999
2000
Displacement
(L)
1.0
2.2
5.9
Cylinder
Number
3
4
6
Rated
Power (hp)
26
50
260
Air Induction
naturally
aspirated
naturally
aspirated
turbocharged
intercooled
Engine
Type
IDI
IDI
DI
   As described in 4.1.1.3 above, passive filter regeneration occurs when the exhaust
temperatures are high enough that on aggregate the PM accumulation rate on the filter is less
                                          4-72

-------
	Technologies and Test Procedures for Low-Emission Engines

than the PM oxidation rate on the filter over an extended time period. During that time period
there can be periods of low-temperature operation where the PM accumulation rate is higher than
the oxidation rates, provided that there are other periods of higher temperature operation where
the PM oxidation rate is significantly higher than the accumulation rate.  CDPF manufacturers
provide guidelines for CDPF applications where passive regeneration is necessary (i.e., no
provision for occasional active regeneration is provided).  These guidelines are based on the
cumulative amount of typical engine operation above and below a particular exhaust
temperature. One CDPF manufacturer has stated that passive regeneration will occur if
temperatures exceed 250°C for more than 30 percent of engine operation.114 Another CDPF
manufacturer has stated that catalyzed diesel particulate filters will work properly in the field if
the engine exhaust temperature is at least 250-275°C for about 40-50 percent of the duty cycle.115
   EPA used the more restrictive of these guidelines to evaluate the likelihood that passive
regeneration will during typical nonroad operating cycles. To do this, the exhaust temperatures
collected from testing each engine on various nonroad transient duty cycles were sorted in an
ascending order. Upon sorting, we identified the 50th and 60th percentile mark of the temperature
obtained for a transient cycle run, which lasted anywhere between 8 to 20 minutes for an entire
cycle duration.  The temperatures associated with the 50th and 60th percentile mark correspond to
the minimum temperatures for 50 and 40 percent of the duty cycle, respectively. In addition, we
also calculated the average temperature obtained throughout a given cycle.

    Tables 4.1-9, 4.1-10, and 4.1-11 show the 50th and 60th percentile temperatures representing
the minimum temperatures for 50% and 40% of the duty cycle, respectively.  The tables show
that the 60th percentile temperature exceeded 250 C for most of the engine tests on all three
engines. The runs that did not result in at least 250°C for 40% of the duty cycle were from the
highway FTP cycle for the two small engines, and from the backhoe cycle for the lowest power
rating,  i.e., 124 hp, on the Cummins ISB engine.
                                          4-73

-------
Regulatory Impact Analysis
                                     Table 4.1-9
         Engine-out Exhaust Gas Temperature Data - 124, 163, 260 hp Cummins ISB
Cycle

Agricultural Tractor 260 hp (test #1454)
124 hp (test #15 18)
Wheel Loader 260 hp (test #1449)
169 hp (test #1530)
124 hp (test #1526)
Backhoe 260 hp (test #1455)
169 hp (test #1528)
124 hp (test #1523)
JRC Composite 260 hp (test # 1 660)
260 hp (test #1661)
1 69 hp (test #1529)
1 24 hp (test #1525)
Average
T(°C)
418
319
295
264
221
261
236
185
311
317
289
252
50th %tile
T(°C)
444
336
323
277
222
280
238
194
323
326
290
243
60th %tile
T(°C)
452
339
295
311
258
303
254
201
337
339
304
265
Operation at
T m 275°C
92%
89%
57%
50%
29%
52%
24%
0%
75%
78%
61%
37%
                                     Table 4.1-10
             Engine-out Exhaust Gas Temperature Data - 50 hp Kubota V2203E
Cycle
Agricultural Tractor
Nonroad Composite
Skid Steer Loader
Federal Test Procedure
Average
T(°C)
518
289
259
232
50th %tile
T(°C)
544
286
257
210
60th %tile
T(°C)
561
310
268
238
Operation at
T m 275°C
96%
56%
34%
30%
                                     Table 4.1-11
          Engine-out Exhaust Gas Temperature Data - 26 hp Lombardini LDW1003
Cycle
Arc Welder
Nonroad Composite
Skid Steer Loader
Federal Test Procedure
Agricultural Tractor
Average
T(°C)
262
274
243
177
516
50th %tile
T(°C)
257
271
239
148
548
60th %tile
T(°C)
263
290
252
175
554
Operation at
T m 275°C
26%
48%
24%
15%
97%
   The results shown here lead us to conclude that, for a significant fraction of nonroad diesel
engine operation, exhaust temperatures are likely to be high enough to ensure passive
regeneration of CDPFs. However, the results also indicate that for some operating conditions it
                                        4-74

-------
	Technologies and Test Procedures for Low-Emission Engines

may be that passive filter regeneration is not realized. In the case of those operating conditions,
we believe that active backup regeneration systems (systems designed to increase exhaust
temperature periodically to initiate filter regeneration) can be used to ensure CDPF regeneration.
Additional data regarding in-use temperature operation are contained in a recent report from the
Engine Manufacturers Association (EMA) and the European Association of Internal Combustion
Engine Manufacturers (Euromot).116  This report contains data from a range of applications and
power categories. Similar to the data presented above, the EMA/Euromot data indicate that,
while several nonroad applications generate temperatures high enough to passively regenerate a
filter, there are also some  applications that require active regeneration.

   We have assumed in our cost analysis that all nonroad engines complying with a PM
standard of 0.03 g/hp-hr or lower (those engines that we are projecting will use a CDPF) will
have an active  means to control temperature (i.e., we have costed a backup active regeneration
system,  though some applications may not need one). We have made this assumption believing
that manufacturers will not be able to predict,  accurately, in-use conditions for every piece of
equipment and will thus choose to provide the technologies on a back-up basis. As explained
earlier, the technologies necessary to accomplish this temperature management are
enhancements  of the Tier  3 emission-control technologies that will form the baseline for Tier 4
engines, and the control strategies being developed for highway diesel engines. We believe there
are no nonroad engine applications above 25 hp for which these highway engine approaches will
not work.  However, given the diversity in nonroad equipment design and application, we
believe that additional time will be needed to match the engine performance characteristics to the
full range  of nonroad equipment.

   Matching the operating temperature window of the broad range of nonroad equipment may
be somewhat more challenging for nonroad engines than for many highway diesel engines
simply because of the diversity in equipment design and equipment use. Nonetheless, the
problem has been successfully solved in highway applications facing low-temperature
performance situations as  difficult to address as any encountered faced by nonroad applications.
The most challenging temperature regime for highway engines are encountered at very light-
loads as typified by congested urban driving. Under congested urban driving conditions exhaust
temperatures may be too low for effective NOx reduction with a NOx adsorber catalyst.
Similarly,  exhaust temperatures may be too low to ensure passive CDPF regeneration.  To
address  these concerns, light-duty diesel engine manufacturers have developed active
temperature management  strategies that provide effective emissioncontrol even under these
difficult light-load conditions. Toyota has shown with their prototype DPNR vehicles that
changes to EGR and fuel-injection strategies can realize an increase in exhaust temperatures of
more than 50°C under even very light-load conditions allowing the NOx adsorber catalyst to
function under these normally cold exhaust conditions.117 Similarly, PSA has demonstrated
effective CDPF regeneration under demanding light-load taxi cab conditions with current
production technologies.118 Both of these are examples of technology paths available to nonroad
engine manufacturers to increase temperatures under light-load conditions.

   We are not aware of any in-use operating cycles for nonroad equipment that are more
demanding of low-temperature performance than highway urban driving. Both the Toyota and

                                          4-75

-------
Regulatory Impact Analysis
PSA systems are designed to function even with extended-idle operation typical of a taxi waiting
to pick up a fare.N By actively managing exhaust temperatures engine manufacturers can ensure
highly effective catalyst-based emission-control performance (i.e., compliance with the emission
standards) and reliable filter regeneration (failsafe operation) across a wide range of engine
operation typical of the broad range of nonroad engine operation in use and the new nonroad
transient duty cycle.

    The systems described here from  Toyota and PSA are examples of highly integrated engine
and exhaust emission-control  systems based upon active engine management designed to
facilitate catalyst function. Because these systems are based upon the same engine control
technologies likely to be used to comply with the Tier 3 standards and because they allow great
flexibility to trade-off engine control  and catalyst control approaches depending on operating
mode and need, we believe most nonroad engine manufacturers will use similar approaches to
comply with the Tier 4 emission standards.  However, there are other technologies available that
are designed to be added to existing engines without the need for extensive integration and
engine management strategies. One example of such a system is an active DPF system
developed by Deutz for use  on a wide range on nonroad equipment. The Deutz system has been
sold as an OEM retrofit technology that does not require changes to the base engine technology.
The system is electronically controlled and uses supplemental in-exhaust fuel injection to raise
exhaust temperatures periodically to regenerate the DPF.  Deutz has sold over 2,000 of these
units and reports that the systems have been reliable and effective.  Some manufacturers may
choose to use this approach  for compliance with the Tier 4 PM standard, especially in the case of
engines that may be able to meet the NOx standards with engine-out emission-control
technologies (i.e., engines rated between 25 and 75 hp and mobile machines  >750 hp).

    We believe that, given the timing  of the Tier 4 emission standards and the availability and
continuing development of technologies  to address temperature management for highway
engines (whose technologies are transferrable to all nonroad engines with greater than 25 hp
power rating), nonroad engines can be designed to meet the emission standards adopted in this
final rule in a timely manner.
    N There is one important distinction between the current PSA system and the kind of system that we project
industry will use to comply with the Tier 4 standards: the PSA system incorporates a cerium fuel additive to help
promote soot oxidation.  The additive serves a similar function to a catalyst to promote soot oxidation at lower
temperatures. Even with the use of the fuel additive, passive regeneration is not realized on the PSA system and an
active regeneration is conducted periodically involving late cycle fuel injection and oxidation of the fuel on an
up-front diesel oxidation catalyst to raise exhaust temperatures.  This form of supplemental heating to ensure
infrequent but periodic PM filter regeneration has proven to be robust and reliable for more than 500,000 PSA
vehicles. Our 2002 progress review found that other manufacturers will be introducing similar systems in the next
few years without the use of a fuel additive. One vehicle manufacturer, Renault has recently announced that they
will introduce this year a CDPF system on a diesel passenger car that does not rely on an additive to help ensure that
regeneration occurs.

                                            4-76

-------
	Technologies and Test Procedures for Low-Emission Engines

   4.1.3.1.2 NOx Adsorbers andNonroad Operating Temperatures

   Section 4.1.2.3.3 above describes a method to directionally evaluate the match between the
operating temperature characteristics of a diesel engine in typical use and the range of
temperatures over which a NOx adsorber catalyst is highly effective, the operating window of
the NOx adsorber catalyst technology.  The analysis is not effective to accurately predict exact
emission results as it does not account for the thermal inertia of the catalyst technologies nor the
ability of the NOx adsorber to store NOx at lower temperatures as discussed in more fully in
Section  4.1.2.3.3. Nevertheless, this analysis approach can be used to compare predicted
performance of an engine with a NOx adsorber catalyst on various test cycles and with various
engine configurations.

   In this case, we have used this analysis approach to better understand the characteristics of
the NRTC and the Cl composite cycle relative to the highway FTP cycle. We have extensive
experience testing NOx adsorber catalyst systems on the highway FTP cycle (see discussion
above in Section 4.2) showing that NOx reductions in excess of 90% can be expected. Here, we
are trying to understand if the NOx performance on the NRTC  and the Cl composite cycle
should be expected to be better or worse than the highway FTP cycle. To accomplish that, we
tested a  Cummins ISB (see Table 4.1-8) engine at three different power ratings representative of
the range of engine power density currently seen for nonroad diesel engines (250hp, 169hp,  and
124hp).  Following the technique described in Section 4.1.2.3.3, we estimated a notional NOx
adsorber efficiency for the various test cycles and engine power ratings described here. Further,
we performed this analysis for several different NOx adsorber mounting locations (i.e., we
measured exhaust temperatures at several locations in the exhaust system, a catalyst is not
actually installed for this testing). By measuring temperature at several  locations, we could
further understand the impact of heat loss in the exhaust system on NOx adsorber performance.
The results of this testing and analysis are presented in Tables 4.1-12, 4.1-13 and 4.1-14.

                                      Table 4.1-12
          Estimated NOx Adsorber Efficiency on Cummins ISB  ISO-C1 Composite"
Engine Power
(hp)
124
169
250
6" from turbo
outlet (%)
90.5
86.2
79.5
25 "from turbo
outlet (%)
90.7
87.1
84.2
4' from turbo
outlet (%)
90.6
88.7
85.2
6' 7" from turbo
outlet (%)
89.8
90.8
87.9
       ' The estimates are based on the absorber B curve shown in Figure 4.1-11.
                                         4-77

-------
Regulatory Impact Analysis
                                      Table 4.1-13
            Estimated NOx Adsorber Efficiency on Cummins ISB - NRTC Cycle"
Engine Power
(hp)
124
169
250
6" from turbo
outlet (%)
85.6
93.0
91.6
25" from turbo
outlet (%)
83.9
92.2
92.9
4' from turbo
outlet (%)
81.7
91.1
93.6
6' 7" from turbo
outlet (%)
77.4
88.6
93.5
       1 The estimates are based on the absorber B curve shown in Figure 4.1-11.

                                      Table 4.1-14
         Estimated NOx Adsorber Efficiency on Cummins ISB - Highway FTP Cycle"
Engine Power (hp)
124
169
250
6" from turbo outlet (%)
60.3
72.4
83.0
          1 The estimates are based on the absorber B curve shown in Figure 4.1-11.
   Results of the analysis show that for many nonroad engines, the expected exhaust
temperatures are well matched for NOx adsorber control giving high NOx conversion
efficiencies with today's NOx adsorber technology. The NOx-reduction potential by these
devices was higher over nonroad cycles when compared with that achieved from the highway
FTP cycle. This higher efficiency obtained from the engine testing results was due to
comparatively higher engine-out exhaust temperatures obtained from running on various
nonroad transient cycles compared with the highway FTP cycle, thus indicating that the transfer
of highway engine technologies developed for the FID2007 emission standards will be able to
provide similar or better control for nonroad diesel engines designed to comply with the Tier 4
standards.

   4.1.3.1.3 Power Density Trends in Nonroad

   An analysis  of power density trends in nonroad diesel engines was undertaken to understand
what levels of power density to expect in the future for nonroad diesel engines.  This analysis
included  consideration of data from the Power Systems Research 2002 database (PSR). The
PSR data includes estimates of nonroad diesel engine  model  specifications and sales going back
at least 20 years. This data set represents the most comprehensive nonroad engine database of
this nature available.

   This analysis specifically examined trends in power density within various power categories
from 1985 to 2000.  The PSR database reports both rated power and  engine displacement, from
                                         4-78

-------
                     Technologies and Test Procedures for Low-Emission Engines
which power was calculated.0 The data were divided into 5 power categories: 70-100 hp; 100 -
175hp; 175 - 300hp, 300 - 600hp, and >600hp.  For each power category, a sales-weighted
average of power density was calculated for each year.  Table 4.1-15 shows the resulting data, as
well as the percent change from 1985 to 2000. Figure 4.1-17 is a graphical representation of the
data in Table 4.1-15.

                                      Table 4.1-15
           Sales-Weighted Power Density by Power Category (hp/liter), 1985 - 2000
Year
1985
1986
1987
1988
1989
1990
1991
1992
1993
1994
1995
1996
1997
1998
1999
2000
% Change
1985-2000
50-100hn
20.5
20.5
20.9
21.1
20.7
21.2
21.5
21.9
22.3
22.3
22.0
22.2
22.1
22.6
23.1
22.9
11%
100-175hn
24.0
23.4
23.3
23.6
24.2
24.8
25.2
25.6
25.5
25.6
25.8
25.7
25.9
26.3
26.4
26.4
9%
175-300hn
25.2
25.9
25.9
26.3
27.8
28.3
28.7
29.1
29.6
30.2
30.1
30.1
30.0
30.0
30.1
30.4
17%
300-600hn
30.2
30.1
30.6
29.8
31.8
30.5
30.6
30.2
30.0
30.7
32.7
35.1
35.4
35.1
35.5
35.6
15%
600hn+
27.5
27.6
27.9
28.1
31.9
32.7
33.4
35.0
33.9
34.7
35.2
35.5
35.4
35.3
34.9
34.9
21%
   Figure 4.1-7 shows reasonably steady increase in power density for engines all power
categories from 1985 until approximately 1994/1995, though the rate of increase varies between
the power categories. From 1994/95 until 2000 most power categories saw either no change or a
slight increase in power density, with the exception of the >600hp category, which saw a small
decrease.  Power density increases by engine rated power, with the 70-100hp category showing
the lowest values, with year 2000 being 22.9 hp/liter, and the 300-600hp and 600+hp categories
have sales-weighted power densities on the order of 35 hp/liter.
     Power density is equal to the engine's rated power divided by the engines total displacement.  The data in this
memorandum is presented in terms of horsepower per liter.
                                         4-79

-------
Regulatory Impact Analysis
               Figure 4.1-17 Power Density Trends for Nonroad Diesel Engines

                               1985 - 2000, >50 horsepower engines
       15
        1985
                  1987
                           1989
                                     1991
                                               1993
                                                         1995
                                                                   1997
                                                                            1999
                                            Year
                          -50-100hp'
                                   -100-175hp-
                                            -175-300hp -*-300-600hp •
                                                              -600hp+
   4.1.3.2 Durability and Design

   Nonroad equipment is designed to be used in a wide range of tasks in some of the harshest
operating environments imaginable, from mining equipment to crop cultivation and harvesting to
excavation and loading. In the normal course of equipment operation the engine and its
associated hardware will experience levels of vibration, impacts, and dust that may exceed
conditions typical of highway diesel vehicles. Failing to consider differences in operating
conditions in engine and equipment design would be expected to lead to eventual failure of the
equipment.

   Specific efforts to design for the nonroad operating conditions will be required to ensure that
the benefits of these new emission-control technologies are realized for the life of nonroad
equipment. Much of the engineering knowledge and experience to address these issues already
exists with the nonroad equipment manufacturers. Vibration and impact issues are
fundamentally mechanical durability concerns (rather than issues of technical feasibility of
achieving emission reductions) for any component mounted on a piece of equipment (e.g., an
engine coolant overflow tank). Equipment manufacturers must design mounting hardware such
as flanges, brackets, and bolts to support the new component without failure. Further, the
catalyst substrate material itself must be able to withstand the conditions encountered on nonroad
equipment without itself cracking  or failing.  There is a large body of real-world testing with
retrofit emission-control technologies that demonstrates the durability of the catalyst components
themselves even in the harshest of nonroad equipment applications.
                                          4-80

-------
	Technologies and Test Procedures for Low-Emission Engines

   Deutz, a nonroad engine manufacturer, sold approximately 2,000 diesel particulate filter
systems for nonroad equipment in the period from 1994 through 2000. The very largest of these
systems were limited to engine sizes below 850 hp.  The majority of these systems were sold into
significantly smaller applications. Many of these systems were sold for use in mining
equipment. No other applications are likely to be more demanding than this.  Mining equipment
is exposed to extraordinarily high levels of vibration, experiences impacts with the mine walls
and face, and high levels of dust.  Yet in meetings with the Agency, Deutz shared their
experience that no system had failed due to mechanical failure of the catalyst or catalyst
housing.119 The Deutz system utilized a conventional cordierite PM filter substrate as is
commonly used for heavy-duty highway truck CDPF systems. The canning and mounting of the
system was a Deutz design. Deutz was able to design the catalyst housing and mounting in such
a way as to protect the catalyst from the harsh environment as evidenced by its excellent record
of reliable function.

        Other nonroad equipment manufacturers have also offered OEM diesel  particulate filter
systems to comply with requirements of some mining and tunneling worksite standards.
Liebherr, a nonroad engine and equipment manufacturer, offers diesel particulater filter systems
as an OEM option on 340 different nonroad equipment models.120  We believe this experience
shows that appropriate design considerations, as are necessary with any component on a piece of
nonroad equipment, will be adequate to address concerns with the vibration and impact
conditions that can occur in some nonroad applications. This experience applies equally well to
the NOx adsorber catalyst technologies, as the mechanical properties of DOCs, CDPFs, and NOx
adsorbers are all similar. We believe that no new or fundamentally different solutions are
needed to address the vibration and impact constraints for nonroad equipment below 750 hp.
Engines above 750 hp are fundamentally similar to smaller engines with the most obvious
difference being their larger size.  Their larger size does create some additional issues regarding
the size and physical strength of emission control technologies. While we believe that it may be
possible to address these concerns using the same technologies as for engines <750 hp, we
recognize that today we have limited evidence to draw that conclusion definitively. As
described in Preamble Section II, we have therefore made some revisions to the proposed
emission standards for engines >750 hp reflecting technologies (e.g., wire or fiber mesh PM
filters) that we can say with confidence will be appropriate and available in the timeframe of this
rulemaking.

   Certain nonroad applications, including some forms of harvesting equipment and mining
equipment, may have specific limits on maximum surface temperature for equipment
components to ensure that the components do not serve as ignition sources for flammable dust
particles (e.g. coal dust or fine crop dust).  Some have suggested that these design constraints
might limit the equipment manufacturers ability to install advanced diesel catalyst technologies
such as NOx adsorbers and CDPFs.  This concern seems to be largely based upon anecdotal
experience with gasoline catalyst technologies where, under certain circumstances, catalyst
temperatures can exceed 1,000°C and, without appropriate design considerations, could
conceivably serve as an ignition source. We do not believe these concerns are justified in the
case of either the NOx adsorber catalyst or the CDPF technology.  Catalyst temperatures for
NOx adsorbers and CDPFs should not exceed the maximum exhaust manifold temperatures

                                         4-81

-------
Regulatory Impact Analysis
already commonly experienced by diesel engines (i.e., catalyst temperatures are expected to be
below 800°C).P CDPF temperatures are not expected to exceed approximately 700°C in normal
use and are expected to reach the 650°C temperature only during periods of active regeneration.
Similarly, NOx adsorber catalyst temperatures are not expected to exceed 700°C and again only
during periods of active sulfur regeneration, as described in Section 4.1.7 below. Under
conditions where diesel  exhaust temperatures are naturally as high as 650°C, no supplemental
heat addition from the emission-control system will be necessary and therefore exhaust
temperatures will not exceed their natural level.  When natural exhaust temperatures are too low
for effective functioning of the emission-control system, then supplemental heating  (as described
earlier) may be necessary, but this is not expected to produce temperatures higher than the
maximum levels normally encountered in diesel exhaust. Furthermore, even if it were necessary
to raise exhaust temperatures  to a higher level to promote effective emission control, there are
technologies available to isolate the higher exhaust temperatures from flammable materials such
as dust. One approach is the use of air-gapped exhaust systems (i.e., an exhaust pipe inside
another concentric exhaust pipe separated by an air-gap) that serve to insulate the inner high-
temperature surface from the  outer surface, which could come into contact with the  dust.  The
use of such a system may be additionally desirable to maintain higher exhaust temperatures
inside the catalyst to promote better catalyst function. Another technology  to control surface
temperature already used by some nonroad equipment manufacturers is water cooled exhaust
systems.121 This approach is similar to the air-gapped system but uses engine coolant water to
actively cool the exhaust system. Flammable dust concerns should not prevent the use of either
a NOx adsorber or a CDPF, because catalyst temperatures are  not expected to be unacceptably
high and because remediation technologies exist to address these concerns.  In fact,  exhaust
emission-control technologies (i.e., aftertreatment) have already been applied on both an OEM
basis  and for retrofit to nonroad equipment for use in potentially explosive environments. Many
of these applications must undergo Underwriters Laboratory (UL) approval before they can be
used.122

   We agree that nonroad equipment must be designed to address durable performance for a
wide range of operating conditions and applications that are not commonly  experienced by
highway vehicles.  We believe further, as demonstrated by retrofit experiences around the world,
that there are technical solutions that allow catalyst-based emission-control technologies to be
applied to nonroad equipment.

4.1.4 Are the Standards for  Engines >25 hp and <75 hp Feasible?

   As discussed in Section II of the preamble, the emission standards for engines between 25
and 75 hp consist of a 2008 transitional standard and long-term 2013 standards. The transitional
standard is a 0.22 g/hp-hr PM standard. The 2013 standards consist of a 0.02 g/hp-hr PM
   p The hottest surface on a diesel engine is typically the exhaust manifold, which connects the engines exhaust
ports to the inlet of the turbocharger. The hot exhaust gases leave the engine at a very high temperature (800° C at
high power conditions) and then pass through the turbo where the gases expand driving the turbocharger providing
work and are cooled in the process. The exhaust leaving the turbocharger and entering the catalyst and the
remaining pieces of the exhaust system is normally at least 100°C cooler than in the exhaust manifold.

                                           4-82

-------
	Technologies and Test Procedures for Low-Emission Engines

standard and a 3.5 g/hp-hr NMHC+NOx standard. The transitional standard is optional for 50-
75 hp engines, as the 2008 implementation date is the same as the effective date of the Tier 3
standards. Manufactures may decide, at their option, not to undertake the 2008 transitional PM
standard, in which case their implementation date for the 0.02 g/hp-hr PM standard begins in
2012.

    The remainder of this section discusses (1) what makes the 25-75 hp category unique, (2)
which engine technology is used currently, (3) which engine technology will be used for
applicable Tier 2 and Tier 3 standards, and (4) why the Tier 4 standards are technologically
feasible.

    4.1.4.1 What makes the 25 - 75 hp category unique?

    Many of the nonroad diesel engines >75 hp are either a direct derivative of highway heavy-
duty diesel engines, or share some common traits with highway diesel engines.  These include
similarities in displacement, aspiration, fuel systems, and electronic controls.  At the time of the
proposal, we summarized some of the key engine parameter using data from the 2001 engines
certified for  sale in the United States.  For this final rule, we have also added to this data set by
including the 2004 engines certified for sale in the U.S.  A comparison of these two data sets
show a number of important trends, as discussed below.

    Table 4.1-16 contains a summary of some key engine parameters from the 2001 engines
certified for  sale in the United States, and Table 4.1-17 is a summary of the 2004 engines.Q

                                      Table 4.1-16
          Summary of Model Year 2001 Key Engine Parameters by Power Category
Engine Parameter
IDI Fuel System
DI Fuel System
Turbocharged
1 or 2 Cylinder Engines
Electronic fuel systems
Percent of 2001 U.S. Production"
0-25 hp
83%
17%
0%
47%
0%
25-75 hp
47%
53%
7%
3%
0%
75-100 hp
4%
96%
62%
0%
0%
>100 hp
<0.1%
>99%
91%
0%
14%
       ' Based on sales weighting of 2001 engine certification data.
   Q Data in Table 4.1-16 are derived from a combination of the publically available certification data for model
year 2001 engines, as well as the manufacturers reported estimates of 2001 production targets, which is not public
information.
                                          4-83

-------
Regulatory Impact Analysis
                                      Table 4.1-17
          Summary of Model Year 2004 Key Engine Parameters by Power Category
Engine Parameter
IDI Fuel System
DI Fuel System
Turbocharged
1 or 2 Cylinder Engines
Electronic fuel systems
Percent of 2004 U.S. Production3
0-25 hp
85%
15%
0%
18%
0%
25-75 hp
54%
46%
22%
<1%
0%
75-100 hp
6%
94%
78%
0%
18%
>100 hp
0.0%
100.0%
99.9%
0%
61%
       a Based on sales weighting of 2004 engine certification data.

   As can be seen in Table 4.1-16 & 4.1-17, the engines in the 25-75 hp category have some
important technology differences from the larger engines.  These include a higher percentage of
indirect-injection fuel systems, and a lower fraction of turbocharged engines. (The distinction in
the <25 hp category is even more different, with no turbocharged engines, a large number of the
engines have two cylinders or less,  and a significant majority of the engines have indirect-
injection fuel systems.)

   The distinction is particularly marked with respect to electronically controlled fuel systems.
These are commonly available in the > 75 hp power categories (see Table 4.1.17 above showing
that the technology is already migrating in significant amounts even into the 75-100 hp power
band), but, based on the available certification data as well as our discussions with engine
manufacturers, we believe there are very limited, if any in the 25-75 hp category (and no
electronic fuel systems in the less than 25 hp category) at this time.  The research and
development work currently being performed for the  heavy-duty highway market is targeted at
turbocharged and electronically  controlled, direct-injection engines with at least four cylinders
and per-cylinder displacements greater than 0.5 liters. As discussed in more detail below (and in
the preamble), as well as in Section 4.1.5.1 (regarding the <25 hp category),  these engine
distinctions are important from a technology perspective and warrant a different set  of standards
and implementation time-frame  for the 25-75 hp category (as well as for the  <25 hp category).

   At the same time, the data in Tables 4.1-16 and 4.1-17  shows that engine technology is
steadily progressing in the nonroad diesel engine market, and the penetration of that technology
has increased in the past few years, i.e., from 2001 to 2004. In 2001, only engines in the 300-
600hp range were required to comply with Tier 2. Today,  in 2004,  all engines in the 25-750hp
range must comply with the Tier 2 emission standards.  As a result of the inherent benefits of
electronically controlled fuel systems and turbocharging, and as a response to the Tier 2
emission standards, the penetration of these engine technologies in the past few years has been
dramatic.  Nearly all engines >100 hp are turbocharged and have direct injection fuel systems.
In addition, more than 60  percent of the engines > 100 hp now have electronically controlled fuel
systems (up from 14% in 2001). In the 75-100hp range, turbocharging has increased from 62 to
                                          4-84

-------
	Technologies and Test Procedures for Low-Emission Engines

78 percent, and electronically controlled fuel systems have increased from 0 percent in 2001 to
18 percent in 2004.  The certification data shows that these electronically controlled fuel systems
are available across the full 75-100 hp range, with some engines which use these fuel  systems
having a rated power of 75 hp and others having a rated power of 99 hp. The data also indicate
that the engines in the 75-100 hp range with electronically controlled fuel systems are designed
for use in nonroad equipment such as agricultural tractors, mobile cranes, dozers, loaders, fork
lifts, and a range of other nonroad equipment. Even in the 25-75 hp range, turbocharging has
increased by a factor of 3, from 7 to 21 percent. We expect all of these trends to continue as Tier
2 is fully implemented by 2006, and as the Tier 3 standards are phased-in from 2006 to 2008.
Another reason we expect that the trend will continue is because of the inherent benefits for the
end-user which result from the use of electronically controlled fuel systems and turbocharging.

   4.1.4.2 What engine technology is used currently, and will be used for Tier 2 and Tier 3,
   in the  25-75hp range?

   In the  1998 nonroad diesel rulemaking, we established Tier 1 and Tier 2 standards for
engines in the 25-50 hp category. Tier 1  standards were implemented in 1999, and the Tier 2
standards take effect in 2004. The 1998 rule also established Tier 2 and Tier 3 standards for
engines between 50 and 75 hp. The Tier 2 standards take effect  in 2004, and the Tier 3 standards
take effect in 2008.  The Tier 1 standards for engines between 50 and 75 hp took effect in 1998.
All engines in the 25-75 hp range were first required to meet Tier 2 standards in the 2004 model
year, and the MY 2004 data presented in Table 4.1-17 represent  Tier 2 technology for this power
range.

   Engines in the 25-75 hp category use either indirect injection (IDI)  or direct injection (DI)
fuel  systems. The IDI system injects fuel into a pre-chamber rather than directly into  the
combustion chamber as in the DI system.123  This difference in fuel systems results in
substantially different emission characteristics, as well as  several important operating
parameters.  In general, the IDI engine has lower engine-out PM and NOx emissions,  while the
DI engine  has better fuel efficiency and lower heat rejection.124

   We expect a significant shift in the engine technology that will be used in this power
category as a result of the upcoming Tier 2 and Tier 3 standards, in particular for the 50-75 hp
engines. In the 50-75 hp category, the 2008 Tier 3 standards will likely result in the significant
use of turbocharging and electronic fuel systems, as well as the introduction of both cooled and
uncooled exhaust gas recirculation by some engine manufacturers and possibly the use of
charge-air-cooling.125 To some extent this has already begun to occur as a result of the Tier 2
standards,  as discussed above in relation to Tables 4.1-16  and 4.1-17. In addition, we have heard
from some engine manufacturers that the engine technology used to meet Tier 3 for engines in
the 50-75 hp range will also be made available on those engines  in the 25-50 hp range that are
built on the same engine platform. For the Tier 2 standards for the 25-50 hp products, a large
number of engines already meet these standards; we therefore expect to see more moderate
changes in these engines, including an increased penetration of turbocharging.126
                                          4-85

-------
Regulatory Impact Analysis
   4.1.4.3 Are the standards for 25 -75 hp engines technologically feasible?

   This section discusses the feasibility of both the interim 2008 PM standard and the long-term
2013 standards.

   4.1.4.3.1 2008 PM Standards

   As just discussed in Section 4.1.4.2, engines in the 25-50 hp category must already meet Tier
1 NMHC+NOx and PM standards. We have examined the model year 2002 engine certification
data for engines in the 25-50 hp category.127  We have also examined the model year 2004
certification data for engines in the 25-50hp category. For the model year 2002 data, there is no
Tier 1 PM standard for engines in the 50-75 hp range, and engine manufacturers are therefore
not required to report PM emission levels until  Tier 2 starts in 2004, so there is no 2002 data to
summarize for the 50-75hp range.

   Summary of 2002 Model Year Certification Data for 25-50 hp

   A summary of the 2002 model year certification data for the 25-50 hp engines is presented in
Table 4.1-18, and Figure 4.1-18 is a graph of the HC+NOx and PM results from these same
engines.   These data indicate that over 10 percent of the engine families already meet the 2008
0.22 g/hp-hr PM standard and 5.6 g/hp-hr NMHC+NOx standard (unchanged from Tier 2 in
2008). These include a variety of engine families using a mix of engine technologies (IDI and
DI, turbocharged and naturally aspirated) tested on a variety of certification test cycles.R Five
engine families are more than 20  percent below the 0.22 g/hp-hr PM standard; an additional 24
engine families that already meet the 2008 NMHC+NOx standards will require no more than a
30 percent PM reduction to meet  the 2008 PM standards. Unfortunately, similar data do not
exist for engines between 50 and  75 hp for the 2002 model year.
      The Tier 1 standards for this power category must be demonstrated on one of a variety of different engine
test cycles. The appropriate test cycle is selected by the engine manufacturer based on the intended in-use
application of the engine.

                                          4-86

-------
                     Technologies and Test Procedures for Low-Emission Engines

                                     Table 4.1-18
          2002 Model Year Certification Data for 25-50 hp Nonroad Diesel Engines
PM Emissions Relative to the 0.22
g/hp-hr Standard
0 - 5 % below T4a
5 - 20 % below T4a
>20 % below T4a
require <30% PM
reduction to meet T4a
requires >30%PM reduction
and/or
Total # of Engine Families
IDI Engines
5 -mode/
NA
0
1
2
3
2
8
8-mode/
NA
0
5
1
15
17
38
5 -mode/
TO
0
1
0
0
1
2
8-mode/
TO
0
2
1
4
3
10
DI Engines
5 -mode/
NA
0
0
0
0
8
8
8-mode/
NA
1
0
1
2
40
44
8-mode/
TO
0
0
0
0
8
8
Totals
1
9
5
24
79
118
  Engine also meets 2008 NMHC+NOx
   The model year 2002 engines in this power range use well known engine-out emission-
control technologies, such as optimized combustion chamber design and fuel-injection timing
control strategies, to comply with the existing standards. These data have a two-fold
significance. First, they indicate that some engines in this power range can already achieve the
2008 standard for PM using only engine-out technology, and that other engines should be able to
achieve the standard making improvements just to engine-out performance.
                                        4-87

-------
Regulatory Impact Analysis
       Figure 4.1-18 Emission Certification Data for 25-50 HP Model Year 2002 Engines
                                                      n  g    n
                                                     xx   n  n
                                                    x

                            x
                             x
                        x
                                                        x
                    x   x
 y A
~*XT  x
     X
     X
               x
x
   x
  o
                                X
                                      X  X
                                                                         Xx
                                                x
                                                x
                                                  x
                                                                   X
      0.0
0.1
0.2           0.3           0.4
         PM (g/bhp-hr)
                                  0.5
                     0.6
                x All IDI & other Dl engines   D Naturally Aspirated DI/8-mode cycle engines
    Summary of 2004 Model Year Certification Data for 25-75 hp

    Table 4.1-19 contains a summary of the model year 2004 certification data for PM and
NMHC+NOx as it relates to the 2008 Tier 4 emission standards for engines in the 25-75hp
range.  The data represented in Table 4.1-19 is also shown graphically in Figure 4.1-19. As can
be seen, the 2004 data shows 35 percent of the engine families in the 25-50hp range already meet
the 2008 0.22 g/hp-hr PM standard and a 5.6 g/hp-hr NMHC+NOx standard (which standard is
unchanged from Tier 2 in 2008). In the 50-75 hp range, the data shows 7 percent of the families
can meet the 2008 Tier 4 standards (0.22 g/bhp-hr PM and 3.5 g/bhp-hr NMHC+NOx).  The
relatively low percentage of engines in the 50-75 hp category  which meet the Tier 4 standards
today is largely a result of the stringency of the Tier 3 NMHC+NOx emission standards, which
are required for this power category in 2008. As discussed in our Tier 3 Staff Technical Paper
which reviewed the feasibility of the Tier 3 standards, we believe in-cylinder technologies such
as cooled EGR will be necessary to comply with the Tier 3 emission standards (and we  included
the cost of such systems in our assessment of costs for the Tier 3 rule). Technologies such as
cooled EGR and advanced fuel systems have been shown to be capable of reducing NOx

-------
                     Technologies and Test Procedures for Low-Emission Engines
emissions by 50 percent or more without increasing PM emissions.  As can be seen by the data in
Figure 4.1-19, more than 70 percent of the engines in the 50-75 hp range are below the 0.22
g/bhp-hr PM level, and a NOx reduction of 50 percent would easily bring these engines into
compliance with the Tier 3 NMHC+NOx standards.  Finally, when considered as a whole, nearly
one-quarter of the model year 2004 engine families in the 25-75 hp range could comply with the
Tier 4 2008 PM and NMHC+NOx requirements today.

                                    Table 4.1-19
          2004 Model Year Certification Data for 25-75hp Nonroad Diesel Engines
Power Catergory
25-50 hp,
# of Engine Families

50-75 hp,
# of Engine Families

Power Catergory
25-50 hp,
% of Engine Families

50-75 hp,
% of Engine Families

25-75hp,
% of Engine Families
PM Emissions < 0.22 g/bhp-hr and
2008 NMHC+NOx standards?
No
Yes
Total

No
Yes
Total

PM Emissions < 0.22 g/bhp-hr and
2008 NMHC+NOx standards?
No
Yes

No
Yes

No
Yes
Naturally Aspirated
DI
23
12
35

22
0
22

IDI
35
17
52

11
4
15

NA
58
29
87

33
4
37

Naturally Aspirated
DI
66%
34%

100%
0%

79%
21%
IDI
67%
33%

73%
27%

69%
31%
NA
Total
67%
33%

89%
11%

73%
27%
Turbocharged
DI
6
2
8

15
0
15

IDI
4
5
9

9
0
9

TC
10
7
17

24
0
24

Turbocharged
DI
75%
25%

100%
0%

91%
9%
IDI
44%
56%

100%
0%

72%
28%
TC
Total
59%
41%

100%
0%

83%
17%
Grand
W
36
104

57
4
61

Grand
Total
65%
35%

93%
72%

76%
24%
                                        4-89

-------
Regulatory Impact Analysis
      Figure 4.1-19 Emissions Certification Data for 25-75 HP Model Year 2004 Engines
         2008 NMHC+NOx Standard, 25-50hp
        2008 NMHC+NOx Standard, 50-75hp
                        2008 Tier 4 PM Standard
     0.0
0.1
0.2           0.3
      PM Emissions (g/bhp-hr)
0.4
0.5
0.6
                                A 25-50 hp Engines o 50-75 hp Engines
   Discussion of Certification Data and 2008 Feasibility

   Despite the fact that the certification data from recent model years indicates that engine-out
techniques are capable of meeting the Tier 4 2008 PM standards for some engines, we are not
basing the feasibility of the 2008 PM standard on engine-out techniques alone, as discussed
below.

   As can be seen from the 2002 model year data in Figure 4.1-18, while the engines are all
certified to the same emission standard (Tier  1) with similar engine technology, the emission
levels from these engines vary widely. The same can be seen for the 2004 model year data
shown in Figure 4.1-19, in particular for the 25-50hp engines. Figure 4.1-18 highlights a
specific example of this wide range: engines using naturally aspirated DI technology and tested
on the 8-mode test cycle.  Even for this  subset of DI engines achieving approximately the same
HC+NOx level of-6.5 g/hp-hr, the PM rates vary from approximately 0.2 to more than 0.5 g/hp-
hr. There is limited information available to indicate why for these small diesel engines with
similar technology operating at approximately the same HC+NOx level the PM emission rates
cover such a broad range.  We are therefore not predicating the 2008 PM standard on the  lowest
engine-out emissions being achieved today, because it is uncertain whether or not additional
                                          4-90

-------
	Technologies and Test Procedures for Low-Emission Engines

engine-out improvements will lower all engines to the 2008 PM standard. Instead, we believe
there are two likely means by which companies can comply with the 2008 PM standard. First,
some engine manufacturers can comply with this standard using known engine-out techniques
(e.g., optimizing combustion chamber designs, fuel-injection strategies). However, based on the
available data as shown in Figure 4.1-18 and 4.1-19, it is unclear whether engine-out techniques
will work in all cases.  We therefore believe some engine companies will choose to use a
combination of engine-out techniques and diesel oxidation catalysts, as discussed below.

   Emission Reductions from Engine-out Techniques

   For some of the engines not already meeting the 2008 Tier 4 PM standard, engine-out
techniques may bring the  engines into compliance with the 2008 standards.  In our recent Staff
Technical Paper on the feasibility of the Tier 2 and Tier 3 standards, we projected that engines
greater than 50 hp will rely on some combination of technologies—including electronic fuel
systems such as electronic rotary pumps or common-rail fuel systems—to comply with the Tier
3 NMHC+NOx standards.128  In addition to enabling the Tier 3 NMHC+NOx standards,
electronic fuel systems with high injection pressure and the capability to perform pilot-injection
and rate-shaping, have the potential to substantially reduce PM emissions.129 Even for
mechanical fuel systems, increased injection pressures can reduce PM emissions substantially.130
As discussed above, we are projecting that the Tier 3 engine technologies used in engines
between 50 and 75 hp, such as turbocharging and electronic fuel systems, will make their way
into engines in the 25-50 hp range.  However, we do not believe this technology will be required
to achieve the Tier 4 2008 PM standard. As demonstrated by the 2002 and 2004 certification
data, engine-out techniques such as optimized combustion chamber design, fuel-injection
pressure increases and fuel-injection timing can be used to achieve the 2008 Tier 4 standards for
many of the engines in the 25-75 hp category without the need to add turbocharging or electronic
fuel systems.

   Emission Reductions from Diesel Oxidation Catalysts

   For those engines not  able to achieve the Tier 4 standards with known engine-out techniques,
we project that these engines can meet the standards with diesel oxidation catalysts.  DOCs are
passive flow-through emission-control devices that are typically coated with a precious metal or
a base-metal washcoat. DOCs have been proven to be durable in use on both light-duty and
heavy-duty diesel applications. In addition, DOCs have already been used to control PM or
carbon monoxide on some nonroad applications.131

   Certain DOC formulations can be sensitive to diesel fuel sulfur levels, and depending on the
level of emission reduction necessary, sulfur in diesel fuel can be an impediment to PM
reductions. Precious-metal oxidation catalysts can oxidize the sulfur in the fuel and form
particulate sulfates.  However, even with current high-sulfur nonroad fuel, some manufacturers
have demonstrated that a properly formulated DOC can be used in combination with other
technologies to achieve the existing Tier 2 PM standards for larger engines (i.e., the 0.15 g/hp-hr
standard).132 However, given the high level of sulfur in  current nonroad fuel, the use of DOCs as
a PM-reduction technology is  severely limited.  Data presented by one engine manufacturer

                                          4-91

-------
Regulatory Impact Analysis
regarding the existing Tier 2 PM standard show that, while a DOC can be used to meet the
current standard even when tested on 2,000 ppm sulfur fuel, lowering the fuel sulfur level to 380
ppm enabled the DOC to reduce PM by 50 percent from the 2,000 ppm sulfur fuel.133 Without
the availability of 500 ppm sulfur fuel in 2008, DOCs would be of limited use for nonroad
engine manufacturers and would not provide the emission-control necessary for most
manufacturers to meet the Tier 4 standards. With the availability of 500 ppm sulfur fuel, DOCs
can be designed to provide PM reductions on the order of 20 to 50%, while suppressing
particulate sulfate reduction.134 These levels of reductions have been seen on transient duty
cycles as well as highway and nonroad steady-state duty cycles.  As discussed above, 24 engine
families in the 25-50 hp range are within 30 percent of the 2008 PM standard and are at or below
the 2008 NMHC+NOx standard for this power range, indicating that use of DOCs should
achieve the incremental improvement necessary to meet the 2008 PM standard.  However, we
also do not believe that an  emission level lower than 0.22 g/bhp-hr will be generally feasible in
2008 due to  the diesel fuel sulfur level of 500 ppm and consequent potential for sulfate PM
formation.

    4.1.4.3.2    2013 Standards

    For engines in the 25-50 range, we are adopting  standards starting in 2013 of 3.5 g/hp-hr for
NMHC+NOx and 0.02 g/hp-hr for PM.  Additionally,  compliance with the existing CO emission
standards will need to be demonstrated over new test cycles including the NRTC with cold-start,
and NTE. For the 50-75 hp engines, we are adopting a 0.02 g/hp-hr PM standard that will be
implemented in 2013, and for those manufacturers who choose to pull-ahead the standard one-
year, 2012 (manufacturers  who choose to pull-ahead the 2013 standard for engine in the 50-75
range do not need to comply with the transitional 2008 PM standard).

       4.1.4.3.2.1PMStandard

    Sections 4.1.1 through  4.1.3 have already discussed catalyzed diesel particulate filters,
including explanations  of how CDPFs reduce PM emissions, and how to apply CDPFs to
nonroad engines.  We concluded there that CDPFs can be used to achieve the Tier 4 PM
standard for engines >75 hp. Specifically we discussed the ability of ceramic based filter
technologies to meet the 0.01 g/bhp-hr standard for engines from 75-750 hp and the ability of
alternate depth filter technologies to meet a slightly less stringent standard of 0.02 g/bhp-hr
standard (0.03 for mobile machines) for engines > 750 hp.  As also discussed in Section 4.1.3,
PM filters may require  active back-up regeneration systems for many nonroad applications.
Secondary technologies will likely be needed in addition to  enable proper regeneration,  possibly
including electronic fuel systems such as common rail, which makes possible multiple post-
injections for raising exhaust gas temperatures to aid in filter regeneration.

    Particulate filter technology, with the requisite trap regeneration technology, can also be
applied to engines in the 25 to 75 hp range. The fundamentals of how a filter is able to reduce
PM emissions, as described in Section 4.1.1, are not a function of engine power, and CDPFs are
just as effective at capturing soot emissions and oxidizing SOF on smaller engines as on larger
engines.  As discussed in more detail below, particulate sulfate generation rates are slightly

                                          4-92

-------
	Technologies and Test Procedures for Low-Emission Engines

higher for the smaller engines; however, we have addressed this issue in the final rule.  The PM
filter regeneration systems described in Sections 4.1.1 and 4.1.3 are also applicable to engines in
this size range and are therefore likewise feasible. Engine manufacturers may prefer some
specific trap-regeneration technologies over others in the 25-75 hp category. Specifically, an
electronically controlled secondary fuel-injection system (i.e., a system that injects fuel into the
exhaust upstream of a PM filter).  Such a system has been commercially used successfully by at
least one nonroad engine manufacturer, and other systems have been tested by technology
companies.135

   We are, however, adopting a slightly higher PM standard (0.02 g/hp-hr rather than 0.01) for
these engines.  As discussed in  Section 4.1.1, with the use of a CDPF, the PM emissions emitted
by the filter are primarily derived from the fuel sulfur.  The smaller power category engines tend
to have higher fuel consumption than larger engines. This occurs for a number of reasons.  First,
the lower power categories include a high fraction of IDI engines, which by their nature consume
approximately 15 percent more fuel than a DI engine. Second, as engine displacements get
smaller, the engine's combustion  chamber surface-to-volume ratio  increases. This leads to
higher heat-transfer losses and therefor lower efficiency and higher fuel consumption. In
addition, frictional losses are a higher percentage of total power for the smaller displacement
engines, which also contributes to higher fuel consumption.  Because of the higher fuel
consumption rate, we expect a higher paniculate sulfate level, and are therefore adopting a 0.02
g/hp-hr standard.

   Test data confirm that this standard, as well as the NTE of 1.5 times the standard, is
achievable. In 2001, EPA completed a test program on two small nonroad diesel engines (a 25
hp IDI engine and a  50 hp IDI engine) that demonstrated the 0.02 g/hp-hr standard can be
achieved with the use of a CDPF.136  This test program included testing on the existing 8-mode
steady-state duty cycle as well as  the new nonroad transient cycle.  The 0.02g/hp-hr level was
achieved on each engine over both test cycles.  In addition, the 0.02 g/hp-hr level was achieved
on a variety of nonroad test cycles that are intended to represent several specific applications,
such as skid-steer loaders, arc-welders, and agricultural tractors.  We believe these data indicate
the robust emission-reduction capability of particulate filters and demonstrate that  the NTE
standard  of 1.5 x 0.02  g/hp-hr standard (0.03 g/hp-hr) can be achieved under the NTE test
requirements, because the data was generated over a number of test cycles which are intended to
represent real in-use operation,  such as we would expect the NTE to represent. This test
program  also demonstrates why we  have adopted a slightly higher PM standard for the 25 - 75
hp category (0.02 g/hp-hr vs. 0.01).  The data from the test program described above showed fuel
consumption rates over the 8-mode test procedure between 0.4 and 0.5 Ibs/bhp-hr,  while typical
values for a modern  turbocharged DI engine with 4-valves per cylinder in the >75  hp categories
are on the order of 0.3 to 0.35 Ibs/hp-hr.

   The CDPF technology applied to meet the PM standard will also serve to ensure compliance
with the existing CO emission standards over the new test procedures.  CDPFs can reduce CO
emissions by more than 80 percent,  a level of control that will more than offset any increase in
CO emission due to the new test cycles.
                                          4-93

-------
Regulatory Impact Analysis
       4.1.4.3.2.2 NMHC+NOx Standard

   We are adopting a 3.5 g/hp-hr NMHC+NOx standard for engines in the 25 - 50 hp range
starting in 2013. This will align the NMHC+NOx standard for engines in this power range with
the Tier 3 standard for engines in the 50 - 75 hp range, which starts in 2008.  EPA's recent Staff
Technical paper, which reviewed the technological feasibility of the Tier 3 standards, contains a
detailed discussion of a variety of technologies capable of achieving a 3.5 g/hp-hr standard.
These include cooled EGR, uncooled EGR, as well as advanced in-cylinder technologies relying
on electronic fuel systems and turbocharging.137 These technologies are  capable of reducing
NOx emission by more than 50 percent including when measured over transient test cycles
including cold-start.138  Given the Tier 2 NMHC+NOx standard of 5.6 g/hp-hr, a 50 percent
reduction will allow a Tier 2 engine to comply with the 3.5 g/hp-hr NMHC+NOx standard. In
addition, because this NMHC+NOx standard is concurrent with the 0.02 g/hp-hr PM standards,
which we project will be achievable with particulate filters, engine designers will have
significant additional flexibility in reducing NOx because the PM filter will eliminate the
traditional concerns with the engine-out NOx vs. PM trade-off. Further, the CDPF technology
will substantially reduce NMHC emissions (by more than 80 percent) providing additional
control effective to help meet the NOx+NMHC emission standards.

4.1.5 Are the Standards  for Engines <25 hp Feasible?

   As discussed in Section III of the preamble, there is a new PM standard of 0.30 g/hp-hr for
engines less than 25 hp beginning in 2008. As discussed below, the NMHC+NOx and CO levels
for this power category is unchanged from Tier 2 levels although compliance will need to be
demonstrated over additional test cycles beginning in 2013.  This section describes (1) what
makes the <25 hp category unique, (2) which engine technologies are currently used in the <25
hp category, and (3) data  showing that the new emission standards are technologically feasible.

   4.1.5.1 What makes the < 25 hp category unique?

   Nonroad engines less  than 25 hp are the least sophisticated nonroad diesel engines from a
technological perspective. All of the engines currently sold in this power category lack
electronic fuel systems  and turbochargers (see Table 4.1-17). Nearly 20 percent of the products
have two-cylinders or less, and 14 percent of the engines sold in this category are single-cylinder
products, several of these have no batteries and are crank-start machines, much like a simple
walk-behind lawnmower. In addition, given the available data and taking into account the Tier 2
standards that have not yet been implemented, we are not projecting any significant penetration
of advanced engine technology, such as electronically controlled fuel systems, into this category
in the next five to ten years.

   4.1.5.2 What engine technology is currently used in the <25 hp category?

   In the 1998 nonroad diesel rulemaking we established Tier 1 and Tier 2 standards for these
products.  Tier 1 was implemented in model year 2000, and Tier 2 will be implemented in model
year 2005.  As discussed in EPA's recent Staff Technical Paper, we project the Tier 2 standards

                                          4-94

-------
	Technologies and Test Procedures for Low-Emission Engines

will be met by basic engine-out emission-optimization strategies.139 We are not predicting that
Tier 2 will require electronic fuel systems, EGR, or turbocharging.  As discussed in Section
4.1.5.3 of this RIA and in the Staff Technical Paper, a large number of engines in this power
category already meet the Tier 2 standards by a wide margin.140

   Two basic types of engine fuel-injection technologies are currently present in the less than 25
hp category, mechanical indirect injection (IDI) and mechanical direct injection (DI).  The IDI
system injects fuel into a pre-chamber rather than directly into the combustion chamber as in the
DI system.  This difference in fuel systems results in substantially different emission
characteristics, as well as several important operating parameters.  In general, as noted earlier,
the IDI engine has lower engine-out PM and NOx emissions, while the DI engine has better fuel
efficiency and lower heat rejection.

   4.1.5.3 What data support the feasibility of the new standards?

   We project that the Tier 4 PM standard can  be met by 2008 based on:
   —the existence of a large number of engine families already meeting the standards,
   —the use of engine-out reduction techniques and
   —the use of diesel oxidation catalysts.

   We have examined model year 2002 and 2004 engine certification data for nonroad diesel
engines less than 25 hp category.141 Tier 2  does not begin for these engines until model year
2005, and thus all of the data we examined are certified to the Tier 1 emission standards.  As
described below, there is little difference between these data sets, and it is likely that many of the
2004 model year engine families are carry overs from the model year 2002.

   Summary of 2002 Model Year Certification Data for Engines <25 hp

   A summary of the model year 2002 certification data for engines <25hp  is presented in Table
4.1-20.  The data is also shown in graphical form in Figure 4.1-20. These data indicate that
some engine families already meet the Tier 4 PM standard (and the 2008 NMHC+NOx standard,
unchanged  from Tier 2). The current data indicate that approximately 28% of the engine
families are already at or below the Tier 4 PM standard, while meeting the 2008 NMHC+NOx
standard.  These data reflect a range of certification test cycles, and include both IDI and DI
engines.8 Many of the engine families are certified well below the Tier 4 standard while meeting
the 2008 NMHC+NOx level. Specifically, 15 percent of the engine families are more than 20
percent below the Tier 4 PM standard. An additional 15 percent of the engine families already
meeting 2008 NMHC+NOx standards will  require no more than a 30 percent PM reduction to
meet the 2008 PM standards.  The public certification data indicate that these engines do not use
turbocharging, electronic fuel systems, exhaust  gas recirculation, or aftertreatment technologies.
   s The Tier 1 and Tier 2 standards for this power category must be demonstrated on one of a variety of different
engine test cycles. The appropriate test cycle is selected by the engine manufacturer based on the intended in-use
applications(s) of the engine.

                                          4-95

-------
Regulatory Impact Analysis
                                      Table 4.1-20
           2002 Model Year Certification Data for <25 hp Nonroad Diesel Engines
PM Emissions Relative
to the 0.30 g/hp-hr Standard
0-5% below T4a
5-20% below T4a
>20% below T4a
requires 3 0% PM
reduction to meet T4a
requires >30%PM reduction and/or
> 2008 NMHC+NOx std.
Total # of Engine Families
IDI Engines
5-mode
1
4
1
5
7
18
6-mode
0
6
9
4
8
27
8-mode
1
1
5
4
4
15
DI Engines
5-mode
0
0
0
0
18
18
6-mode
0
0
1
2
18
21
8-mode
0
0
0
0
3
3
Totals
2
11
ie
15
58
102
       ' Engine also meets the 2008 NMHC+NOx standard.
     Figure 4.1-20 Emission Certification Data for <25 HP Model Year 2002 Engines
      0.0       0.1       0.2       0.3       0.4       0.5
                                   PM Emissions (g/bhp-hr)
0.6
0.7
0.8
                              X Other IDI & DI D IDI 6-mode data
                                          4-96

-------
   	Technologies and Test Procedures for Low-Emission Engines

   Summary of 2004 Model Year Certification Data for 25-50 hp
   The certification data for model year 2004 engines is summarized in Table 4.1 .-21. In
general, this data is similar to the 2002 data shown in Table 4.1-20. The data shows that 31% of
the certified engines are below the 2008 Tier 4 standards, as compared to 28% in the 2002 data.
However, one of the differences is a higher number of 2004 direct-injection engines are below
the Tier 4 levels in 2004 (5 out of 48) as compared to 2002 (1 out of 42).  This data is also shown
in Figure 4.1-21.

                                     Table 4.1-21
           2004 Model Year Certification Data for <25hp Nonroad Diesel Engines

Engine Family
Count

% of Engine
Families
PM Emissions
Below 0.30 g/bhp-
hr?
No
Yesa
Total

No
Yesa
Direct Injection
Fuel System
43
5
48

90%
10%
Indirect Injection
Fuel System
38
32
70

54%
46%
Totals
81
37
118

69%
31%
   a Engine also meets the 2008 NMHC+NOx standard.
                                        4-97

-------
Regulatory Impact Analysis
      Figure 4.1-21 Emissions Certification Data for <25 HP Model Year 2004 Engines
        2008 NMHC+NOx Standard
               0.1
                        0.2
                                  0.3       0.4       0.5
                                    PM Emissions (g/bhp-hr)
                                                              0.6
                                                                       0.7
                                                                                0.8
   Discussion of Certification Data and 2008 Feasibility

   These model year 2002 and 2004 engines use well known  engine-out emission-control
technologies, such as combustion chamber design and fuel-injection timing control strategies, to
comply with the existing standards (Tier 1 in both cases).  As with 25-75 hp engines, these data
have a two-fold significance. First, they indicate that some engines in this power category can
already achieve the 2008 PM standard using only engine-out technology, and that other engines
should be able to achieve the standard making improvements just to engine-out performance.
However, the data does not indicate that all  engines could comply with the 2008 PM standard
using engine-out techniques alone. Despite being certified to the same emission standards with
similar engine technology, the emission levels from these engines vary widely. As can be seen
in the Figure 4.1-20, the emission levels cover a wide range. Figure 4.1-20 highlights a specific
example of this wide range:  engines using naturally aspirated IDI technology and tested on the
6-mode test cycle. Even for this subset of IDI engines achieving approximately the same
HC+NOx level of~4.5 g/hp-hr, the PM rates vary from approximately 0.15 to 0.5 g/hp-hr. There
is limited information available to indicate why for these small diesel engines with similar
technology operating at approximately the same HC+NOx level the PM emission rates cover
such a broad range.  The model year 2004 data in Figure 4.1-21 shows a similarly large spread in
PM emissions.  We are therefore not predicating the 2008 PM standard on the combination of

                                         4-98

-------
	Technologies and Test Procedures for Low-Emission Engines

diesel oxidation catalysts and the lowest engine-out emissions in the final rule, because it is
uncertain whether or not additional engine-out improvements would lower all engines to the
2008 PM standard. Instead, we believe there are two likely means by which companies can
comply with the 2008 PM standard. First, some engine manufacturers can comply with this
standard using known engine-out techniques (e.g., optimizing combustion chamber designs, fuel-
injection strategies). However, based on the available data, it is unclear whether engine-out
techniques will work in all cases. We therefore believe some engine companies will choose to
use a combination of engine-out techniques and  diesel oxidation catalysts, as discussed below.

   Emission Reductions from Engine-out Techniques

   PM emissions can be reduced through in-cylinder techniques for small nonroad diesel
engines using similar techniques as used in larger nonroad and highway engines.  As discussed
in Section 4.1.1 there several technologies that can influence oxygen content and in-cylinder
mixing (and thus lower PM emissions) including improved fuel-injection systems and
combustion system designs. For example, increased injection pressure can reduce PM emissions
substantially.142 The wide-range of emission characteristics present in the existing engine
certification data likely result from differences in fuel systems and combustion chamber designs.
For many of the engines with higher emission levels, further optimization of the fuel system and
combustion chamber can provide additional PM reductions.

   Emission Reductions from Diesel Oxidation Catalysts

   Diesel oxidation catalysts (DOCs) also offer the opportunity to reduce PM emissions from
the engines in this power category. As explained earlier, DOCs are passive flow-through
emission-control devices that are typically coated with a precious metal or a base-metal wash-
coat. DOCs have been proven to be durable in-use on both light-duty and heavy-duty  diesel
applications. In addition, DOCs have already  been used to control either PM or in some cases
carbon monoxide on some nonroad applications.143 However, as discussed in Section 4.1.1,
certain DOC formulations can be sensitive to diesel fuel sulfur level.  Specifically, precious-
metal oxidation catalysts (which have the greatest potential for reducing PM) can oxidize the
sulfur in the fuel  and form particulate sulfates. Given the high level of sulfur in current nonroad
fuel, the use of DOCs as a PM reduction technology is severely limited.  Data presented by one
engine manufacturer regarding the existing Tier  2 PM standard show that while a DOC can be
used to meet the current standard when tested  on 2,000 ppm sulfur fuel, lowering  the fuel sulfur
level to 380 ppm enabled the DOC to reduce PM by  50 percent from the 2,000 ppm sulfur
fuel.144 Without the availability of 500 ppm sulfur fuel in 2008, DOCs would be of limited use
for nonroad engine manufacturers and would not provide the  emission-control necessary for
most engine manufacturers to meet the Tier 4  standards.  With the availability of 500 ppm sulfur
fuel, DOCs can be designed to provide PM reductions on the  order  of 20 to 50%, while
suppressing particulate sulfate reduction.145 These levels of reductions have been seen on
transient duty cycles as well as on highway and nonroad steady-state duty cycles.

   DOCs are also effective to control HC and CO emissions. The application of DOC as a
means  to comply with the PM standard in 2008 will also provide an effective means to meet the

                                          4-99

-------
Regulatory Impact Analysis
existing standards for NOx+NMHC and CO over the new test cycles in 2013. The increase in
NOx emissions over transient test conditions with typical in-cylinder controls are very small as
indicated by the transient adjustment factors estimated in the NONROAD model. HC emissions
may increase during transient testing conditions, however the ability of a DOC to reduce HC
emissions in excess of 80 percent would more than offset any increase in NOx+NMHC
emissions observed over the new test cycles.  Similarly for CO, the additional CO control
allowed by the use of the DOC will more than offset any increase in CO emissions as measured
over the new test cycles. For purposes of our cost analysis contained in Chapter 6, we have
assumed that all engines certifying to the 2008 interim PM standards will use DOCs for
compliance.
                                         4-100

-------
	Technologies and Test Procedures for Low-Emission Engines

4.1.6 Meeting the Crankcase Emission Requirements

   The most common way to eliminate crankcase emissions has been to vent the blow-by gases
into the engine air intake system, so the gases can be recombusted. Prior to the HD2007
rulemaking, we have required that crankcase emissions be controlled only on naturally aspirated
diesel engines.  We had made an exception for turbocharged diesel engines (both highway and
nonroad) because of concerns in the past about fouling that could occur by routing the diesel
particulates (including engine oil) into the turbocharger and aftercooler.  However, this is an
environmentally significant exception since most nonroad equipment over 75hp use
turbocharged engines,  and a single engine can emit over 100 pounds of NOx, NMHC, and PM
from the crankcase over its lifetime.

   Given the available means to control crankcase emissions, we are eliminating this exception
for nonroad diesel engines, as we did for highway engines in 2007. We anticipate that the diesel
engine manufacturers will be able to control crankcase emissions through the use of  closed
crankcase filtration systems or by routing unfiltered blow-by gases directly into the exhaust
system upstream of the emission-control equipment.  However, the crankcase provision has been
written such that if adequate control can be had without "closing" the crankcase, then the
crankcase vent to the atmosphere. Manufacturers show that they meet this requirement by
adding the emissions from the crankcase ventilation system to the emissions from the engine's
exhaust system, either by measuring them separately and adding  together mathematically or by
routing crankcase emissions into the exhaust stream before sampling for emission measurement.

   We expect that manufacturers will have to utilize closed crankcase approaches, as described
here to meet the stringent tailpipe emission standards in this final rule. Closed crankcase
filtration  systems work by separating oil and particulate matter from the blow-by gases through
single or dual stage filtration approaches, routing the blow-by gases into the engine's intake
manifold and returning the filtered oil to the oil sump. Oil separation efficiencies in excess  of 90
percent have been demonstrated with production ready prototypes of two stage filtration
systems.146  By eliminating 90 percent of the oil that would otherwise be vented to the
atmosphere, the system works to reduce oil consumption and to eliminate concerns over fouling
of the intake system when the gases are routed through the turbocharger. Hatz, a nonroad engine
manufacturer, currently has closed crankcase systems on many of its turbocharged engines.

4.1.7 Why Do We Need 15 ppm Sulfur Diesel Fuel?

   As stated earlier, we strongly believe that fuel sulfur control is critical to ensuring the
success of NOx and PM aftertreatment technologies.  To evaluate the effect of sulfur on diesel
exhaust control  technologies, we used three key factors for categorizing the impact of sulfur in
fuel on emission-control function.  These factors were efficiency, reliability, and fuel economy.
Taken together, these three factors support the position that the Tier 4 standards are feasible only
with diesel fuel  sulfur levels of 15 ppm or lower.  Brief summaries of these factors are provided
below.
                                         4-101

-------
Regulatory Impact Analysis
    The efficiency of emission-control technologies to reduce harmful pollutants is directly
affected by sulfur in diesel fuel.  Initial and long-term conversion efficiencies for NOx, NMHC,
CO and diesel PM emissions are significantly reduced by catalyst poisoning and catalyst
inhibition due to sulfur. NOx conversion efficiencies with the NOx adsorber technology in
particular are dramatically reduced in a very  short time due to sulfur poisoning of the NOx
storage bed. In addition, total PM control efficiency is negatively impacted by the formation of
sulfate PM.  As explained in the following sections, the CDPF, NOx adsorber, and urea SCR
catalyst technologies described here have the potential to make significant amounts of sulfate
PM under operating conditions typical of many nonroad engines. We believe that the formation
of sulfate PM will be in excess of the total PM standard, unless diesel fuel sulfur levels are at or
below 15 ppm. Based on the strong negative impact of sulfur on emission-control efficiencies
for all of the technologies evaluated, we believe that 15 ppm represents an upper threshold of
acceptable diesel fuel sulfur levels.

    Reliability refers to the expectation that emission-control technologies must continue to
function as required under all operating conditions for the life of the engine. As discussed in the
following sections,  sulfur in diesel fuel can prevent proper operation of both NOx and PM
control technologies. This can lead to permanent loss in emission-control effectiveness and even
catastrophic failure of the systems.  Sulfur in diesel fuel impacts reliability by  decreasing catalyst
efficiency (poisoning of the catalyst), increasing diesel particulate filter loading, and negatively
impacting system regeneration functions. Among the most serious reliability concerns  with
sulfur levels greater than 15  ppm are those associated with failure to properly regenerate. In the
case of the NOx adsorber, failure to regenerate the stored sulfur (desulfate) will lead to rapid loss
of NOx emission control as a result of sulfur poisoning of the NOx adsorber bed.  In the case of
the diesel parti culate filter, sulfur in the fuel reduces the reliability of the regeneration function.
If regeneration does not occur, catastrophic failure of the filter could occur.  It is only by the
availability of low-sulfur diesel fuels that these technologies become feasible.

    Fuel economy impacts due to sulfur in diesel fuel affect both NOx and PM control
technologies.  The NOx adsorber sulfur regeneration cycle (desulfation cycle) can consume
significant amounts of fuel unless fuel sulfur levels are very low. The larger the amount of
sulfur in diesel fuel, the greater the adverse effect  on fuel economy. As sulfur levels increase
above 15 ppm, the adverse effect on fuel economy becomes more significant, increasing above
one percent and doubling with each doubling of fuel sulfur level. Likewise, PM trap
regeneration is inhibited by sulfur in diesel fuel. This leads to increased PM loading in the diesel
particulate filter and increased work to pump exhaust across this restriction. With low-sulfur
diesel fuel, diesel particulate filter regeneration can be optimized to give  a lower (on average)
exhaust backpressure and thus better fuel economy.  As a result, for both NOx and PM
technologies, reducing the fuel sulfur level decreases the operating costs  of the vehicle.

    4.1.7.1 Catalyzed Diesel Particulate Filters and the Need for Low-Sulfur Fuel

    CDPFs function to  control diesel PM through mechanical filtration of the solid PM (soot)
from the diesel exhaust stream and then oxidation  of the stored soot (trap regeneration) and
oxidation of the SOF. Through oxidation in the catalyzed diesel particulate filter the stored PM

                                          4-102

-------
	Technologies and Test Procedures for Low-Emission Engines

is converted to CO2 and released into the atmosphere.  Failure to oxidize the stored PM leads to
accumulation in the trap, eventually causing the trap to become so full that it severely restricts
exhaust flow through the device, leading to trap or vehicle failure.

   Uncatalyzed diesel paniculate filters require exhaust temperatures in excess of 650°C in
order for the collected PM to be oxidized by the oxygen available in diesel exhaust. That
temperature threshold for oxidation of PM by exhaust oxygen can be decreased to 450°C through
the use of base metal catalytic technologies. For a broad range of operating conditions typical of
in-use diesel engine operation, diesel exhaust can be significantly cooler than 400°C. If
oxidation of the trapped PM would occur only at exhaust temperatures lower than 300°C, then
diesel paniculate filters would be more robust for most applications and operating regimes.
Oxidation of PM (regeneration of the trap) at such low exhaust temperatures can occur by using
oxidants that are more readily reduced than oxygen. One such oxidant is NO2.

   NO2 can be produced in diesel exhaust through the oxidation of the nitrogen monoxide (NO),
created in the engine combustion process, across a catalyst. The resulting NO2-rich exhaust is
highly oxidizing in nature and can oxidize trapped diesel PM at temperatures as cool as 250°C.147
Some platinum group metals are known to be good catalysts to promote the oxidation of NO to
NO2.  To promote more effective passive regeneration of the diesel paniculate filters, significant
amounts of platinum group metals (primarily platinum) are therefore being used in the wash-coat
formulations of advanced CDPFs. The use of platinum to promote the oxidation of NO to NO2
introduces several system vulnerabilities affecting both the durability and the effectiveness of the
CDPF when sulfur is present in  diesel exhaust. (In essence, diesel engine exhaust temperatures
are in a  range necessitating use of precious-metal catalysts to adequately regenerate the PM
filter, but precious-metal catalysts are in turn highly sensitive to sulfur in diesel fuel.) The two
primary mechanisms by which sulfur in diesel fuel limits the robustness and effectiveness of
CDPFs are inhibition of trap regeneration, through inhibition of the oxidation of NO to NO2, and
a dramatic loss in total PM control effectiveness  due to the formation of sulfate PM.
Unfortunately, these two mechanisms trade-off against one another in the design of CDPFs.
Changes to improve the reliability of regeneration by increasing catalyst loadings lead to
increased sulfate emissions and, thus, loss of PM control effectiveness.  Conversely, changes to
improve PM control by reducing the use of platinum group metals and, therefore, limiting
"sulfate make" leads to less reliable regeneration. We believe the best means of achieving good
PM emission control  and reliable operation is to reduce sulfur in diesel fuel, as shown in the
following subsections.

   4.1.7.1.1 Inhibition of Trap Regeneration Due to Sulfur

   The CDPF technology relies on the generation of a very strong oxidant, NO2, to ensure that
the carbon captured by the PM trap's filtering media is oxidized under the exhaust temperature
range of normal operating conditions. This prevents plugging and failure of the PM trap. NO2 is
produced through the oxidation  of NO in the exhaust across a platinum catalyst.  This oxidation
is inhibited by sulfur poisoning of the catalyst surface.148  This inhibition limits the total  amount
of NO2 available for oxidation of the trapped diesel PM, thereby raising the minimum exhaust
temperature required to ensure trap regeneration.  Without sufficient NO2, the amount of PM

                                          4-103

-------
Regulatory Impact Analysis
trapped in the diesel particulate filter will continue to increase and can lead to excessive exhaust
back pressure and low engine power.

    The failure mechanisms experienced by diesel particulate filters due to lowNO2 availability
vary significantly in severity and long-term consequences. In the most fundamental sense, the
failure is defined as an inability to oxidize the stored particulate at a rate fast enough to prevent
net particulate accumulation over time. The excessive accumulation of PM over time blocks the
passages through the filtering media, making it more restrictive to exhaust flow. To continue to
force the exhaust through the now more restrictive filter, the exhaust pressure upstream of the
filter must increase. This increase in exhaust pressure is commonly referred to as increasing
"exhaust backpressure" on the engine.

    The increase in exhaust backpressure represents increased work being done by the engine to
force the exhaust gas through the increasingly restrictive particulate filter. Unless the filter is
frequently cleansed of the trapped PM, this  increased work can lead to reductions in engine
performance and increases in fuel consumption. This  loss in performance may be noted by the
equipment operator in terms of sluggish engine response.

    Full field test evaluations and retrofit applications of these catalytic trap technologies are
occurring in parts of the United States and Europe where low-sulfur diesel fuel is already
available.1 The experience gained in these field tests  helps to clarify the need for low-sulfur
diesel fuel. In Sweden and some European  city centers where 10 ppm diesel fuel sulfur is
readily available, more than 3,000 catalyzed diesel particulate filters have been  introduced into
retrofit applications without a single failure. Given the large number of vehicles participating in
these test programs, the diversity of the vehicle applications, and the extended time periods of
operation, there is a strong indication of the robustness of this technology  on 10 ppm low-sulfur
diesel fuel.149 Vehicle applications included intercity trains, airport buses, mail  trucks, city buses
and garbage trucks. Some vehicles have been operating with traps for more than 5 years and in
excess of 300,000 miles.  The field experience in areas where sulfur is capped at 50 ppm has
been less definitive. In regions without extended periods of cold ambient  conditions, such as the
United Kingdom, field tests on 50 ppm cap  low-sulfur fuel have also been positive, matching the
durability at 10 ppm, though sulfate PM emissions are much higher.  However,  field tests on 50
ppm fuel in Finland, where colder winter conditions are sometimes encountered (similar to many
parts of the United States), showed a significant number of failures (-10 percent) due to trap
plugging. This 10 percent failure rate has been attributed to insufficient trap regeneration due to
fuel sulfur in combination with low ambient temperatures.150  Other possible reasons for the high
failure rate in Finland when contrasted with the Swedish experience appear to be unlikely. The
Finnish and  Swedish fleets were substantially similar, with both fleets consisting of transit buses
powered by Volvo and Scania engines in the 10 to 11 liter range. Further, the buses were
operated in city areas  and none of the vehicles were operated in northern extremes such as north
of the Arctic Circle.151 Given that the fleets  in Sweden and Finland were substantially similar,
   T Through tax incentives 50 ppm cap sulfur fuel is widely available in the United Kingdom and 10 ppm sulfur
fuel is available in Sweden and in certain European city centers.

                                          4-104

-------
	Technologies and Test Procedures for Low-Emission Engines

and given that ambient conditions in Sweden are expected to be similar to those in Finland, we
believe that the increased failure rates noted here are due to the higher fuel sulfur level in a 50
ppm cap fuel versus a 10 ppm cap fuel.u

   Testing on an even higher fuel sulfur level of 200 ppm was conducted in Denmark on a fleet
of 9 vehicles.  In less than six months all of the vehicles in the Danish fleet had failed due to trap
plugging.152 The failure of some fraction of the traps to regenerate when operated on fuel with
sulfur caps of 50 ppm and 200 ppm is believed to be primarily due to inhibition of the  NO to
NO2 conversion, as described here.  Similarly the increasing frequency of failure with  higher fuel
sulfur levels is believed to be due to the further suppression of NO2 formation when higher sulfur
level diesel fuel is used.  Since this loss in regeneration effectiveness is due to sulfur poisoning
of the catalyst this real-world experience is expected to apply equally well to nonroad  engines
(i.e., operation on lower-sulfur diesel fuel, 15 ppm versus 50 ppm, will increase regeneration
robustness).

   As shown above, sulfur in diesel fuel inhibits NO oxidation leading to increased exhaust
backpressure and reduced fuel economy. We therefore believe that sulfur levels in nonroad
diesel fuel must be at or below 15 ppm to ensure reliable and economical operation over the wide
range of expected operating conditions.

   4.1.7.1.2 Loss ofPM Control Effectiveness

   In addition to inhibiting the oxidation of NO  to NO2, the sulfur dioxide (SO2)  in the exhaust
stream is itself oxidized to sulfur trioxide (SO3) at very high conversion efficiencies by the
precious metals in the catalyzed particulate filters.  The SO3 serves as a precursor to the
formation of hydrated sulfuric acid (H2SO4+H2O), or sulfate PM, as the exhaust leaves the
vehicle tailpipe. Virtually all of the SO3 is converted to sulfate under dilute exhaust conditions
in the atmosphere as well in the dilution tunnel used in heavy-duty engine testing.  Since
virtually all sulfur present in diesel fuel is converted to SO2, the precursor to SO3,  as part of the
combustion process, the total sulfate PM is directly proportional to the amount of  sulfur present
in diesel fuel.  Therefore, even though diesel particulate filters are very effective at trapping the
carbon and the  SOF portions of the total PM, the overall PM reduction efficiency of catalyzed
diesel particulate filters drops off rapidly with increasing sulfur levels due to the formation of
sulfate PM downstream of the CDPF.

   SO2 oxidation is promoted across a catalyst in a manner very similar to the oxidation of NO,
except it is converted at higher rates, with peak conversion rates in excess of 50 percent.  The
SO2 oxidation rate for a platinum-based oxidation catalyst typical of the type that might be used
    u The average temperature in Helsinki, Finland, for the month of January is 21 °F. The average temperature in
Stockholm, Sweden, for the month of January is 26°F.  The average temperature at the University of Michigan in
Ann Arbor, Michigan, for the month of January is 24°F. The temperatures reported here are from
www.worldclimate.com based upon the Global Historical Climatology Network (GHCN) produced jointly by the
National Climatic Data Center and Carbon Dioxide Information Analysis Center at Oak Ridge National Laboratory
(ORNL).

                                           4-105

-------
Regulatory Impact Analysis
in conjunction with, or as a washcoat on, a CDPF can vary significantly with exhaust
temperature. At the low temperatures the oxidation rate is relatively low, perhaps no higher than
ten percent. However at the higher temperatures that might be more typical of agricultural
tractor use pulling a plow and the highway Supplemental Emission Test (also called the EURO 4
or 13 mode test), the oxidation rate may increase to 50 percent or more. These high levels of
sulfate make across the catalyst are in contrast to the very low SO2 oxidation rate typical of
diesel exhaust (typically less than 2 percent). This variation in expected diesel exhaust
temperatures means that there will be a corresponding range of sulfate production expected
across a CDPF.

    The U.S. Department of Energy in cooperation with industry conducted a study entitled
DECSE to provide insight into the relationship between advanced emission-control technologies
and diesel fuel sulfur levels. Interim report number four of this program gives the total
particulate matter emissions from a heavy-duty diesel engine operated with a diesel particulate
filter on several  different fuel sulfur levels. A straight line fit through this data is presented in
Table 4.1-19 showing the expected total direct PM emissions from a diesel engine on the
supplemental emission test cycle.v The SET test cycle, a  13 mode steady-state cycle that these
data were developed on, is similar to the Cl  eight mode steady-state nonroad test cycle. Both
cycles include operation at full and intermediate load points at approximately rated-speed
conditions and torque peak-speed conditions. As a result, the sulfate make rate for the Cl cycle
and the SET cycle are expected to be similar. The data can be used to estimate the PM emissions
from diesel engines operated on fuels with average fuel sulfur levels in this range.

                                         Table 4.1-19
       Estimated PM Emissions from a Diesel Engine at the Indicated Fuel Sulfur Levels
Fuel Sulfur
[ppm]
3
T
15a
30
150
Steady-State Emission-Control Performance3
Tailpipe PMb
[g/hp-hr]
0.003
0.006
0.009
0.017
0.071
PM Increase
Relative to 3 ppm Sulfur
-
100%
200%
470%
2300%
       1 The PM emissions at these sulfur levels are based on a straight-line fit to the DECSE data; PM emissions
           at other sulfur levels are actual DECSE data. (Diesel Emission Control Sulfur Effects (DECSE)
           Program - Phase II Interim Data Report No. 4, Diesel Particulate Filters-Final Report, January 2000.
           Table Cl.) Although DECSE tested diesel particulate filters at these fuel sulfur levels, they do not
           conclude that the technology is feasible at all levels, but they do note that testing at 150 ppm is a moot
           point as the emission levels exceed the engine's baseline emission level.
       ' Total exhaust PM (soot, SOF, sulfate).
      Note that direct emissions are those pollutants emitted directly from the engine or from the tailpipe depending
on the context in which the term is used, and indirect emissions are those pollutants formed in the atmosphere
through chemical reactions between direct emissions and other atmospheric constituents.
                                            4-106

-------
                      Technologies and Test Procedures for Low-Emission Engines
    Table 4.1-19 makes it clear that there are significant PM emission reductions possible with
the application of catalyzed diesel paniculate filters and low-sulfur diesel fuel.  At the observed
sulfate PM conversion rates, the DECSE program results show that the 0.01 g/hp-hr total PM
standard is feasible for CDPF equipped engines operated on fuel with a sulfur level at or below
15 ppm. The results also show that diesel particulate filter control effectiveness is rapidly
degraded at higher diesel fuel sulfur levels due to the high sulfate PM make observed with this
technology. It is clear that PM reduction efficiencies are limited by sulfur in diesel fuel and that,
to realize the PM emission benefits sought in this rule, diesel fuel sulfur levels must be at or
below 15 ppm.

    4.1.7.1.3 Increased Maintenance Cost for Diesel P articulate Filters Due to Sulfur

    In addition to the direct performance and durability concerns caused by sulfur in  diesel fuel,
it is also known that sulfur can lead to increased maintenance costs, shortened maintenance
intervals, and poorer fuel economy for CDPFs.  CDPFs are  highly effective at capturing the
inorganic ash produced from metallic additives in engine oil.  This ash is accumulated in the
filter and is not removed through oxidation,  unlike the trapped soot PM. Periodically the ash
must be removed by mechanical cleaning of the filter with compressed air or water.  This
maintenance step is anticipated to occur on intervals of well over 1,500 hours (depending on
engine size).  However, sulfur in diesel fuel  increases this ash accumulation rate through the
formation of metallic sulfates in the filter, which increases both the size and mass of the trapped
ash. By increasing the ash accumulation rate, the sulfur shortens the time interval between the
required maintenance of the filter and negatively impacts fuel economy.

    4.1.7.2 Diesel NOx Catalysts and the Need for Low-Sulfur Fuel

    NOx adsorbers are damaged by sulfur in diesel fuel because the adsorption function itself is
poisoned by the presence of sulfur. The resulting need to remove the stored sulfur (desulfate)
leads to a need for extended high temperature operation that can deteriorate the NOx adsorber.
These limitations due to sulfur in the fuel affect the overall  performance and feasibility of the
NOx adsorber technology.

    4.1.7.2.1 Sulfur Poisoning (Sulfate Storage) on NOx Adsorbers

    The NOx adsorber technology relies on the ability of the catalyst to store NOx as a metallic
nitrate (MNO3) on the surface of the catalyst, or adsorber (storage) bed, during lean operation.
Because of the similarities in chemical properties of SOx and NOx, the SO2 present in the
exhaust is also stored by the  catalyst surface as a sulfate (MSO4). The sulfate compound that is
formed is significantly more stable than the  nitrate compound and is not released and reduced
during the NOx release and reduction step (NOx regeneration step). Since the NOx adsorber is
essentially 100 percent effective at capturing SO2 in the adsorber bed, the sulfur build up on the
adsorber bed occurs rapidly.  As a result, sulfate compounds quickly occupy all of the NOx
storage sites on the catalyst thereby rendering the catalyst ineffective for NOx storage and
subsequent NOx reduction (poisoning the catalyst).

                                          4-107

-------
Regulatory Impact Analysis
   The stored sulfur compounds can be removed by exposing the catalyst to hot (over 650°C)
and rich (air-fuel ratio below the stoichiometric ratio of 14.5 to 1) conditions for a brief period.153
Under these conditions, the stored sulfate is released and reduced in the catalyst.154  While
research to date on this procedure has been very favorable regarding sulfur removal from the
catalyst, it has revealed a related vulnerability of the NOx adsorber catalyst.  Under the high
temperatures used for desulfation, the metals that make up the storage bed can change in
physical structure. This leads to lower precious-metal dispersion, or "metal sintering," (a less
even distribution of the catalyst sites) reducing the effectiveness of the catalyst.155 This
degradation of catalyst efficiency due to high temperatures is often referred to as thermal
degradation. Thermal degradation is known to be a cumulative effect. That is, with each
excursion to high temperature operation, some additional degradation of the catalyst occurs.

   One of the best ways to limit thermal degradation is by limiting the accumulated number of
desulfation events over the life of the engine.  Since the period of time between desulfation
events will likely be determined by the amount of sulfur accumulated on the catalyst (the higher
the sulfur accumulation rate,  the shorter the period between desulfation events), the desulfation
frequency should be proportional to the fuel sulfur level. In other words, for each doubling in
the average fuel sulfur level,  the frequency and accumulated number of desulfation events are
expected to double.  We concluded in the HD2007 rulemaking, that this thermal degradation
would be unacceptable high for fuel sulfur levels greater than  15 ppm. Some commenters to the
HD2007 rule suggested that the NOx adsorber technology can meet the HD2007 NOx standard
using diesel fuel with a 30 ppm average  sulfur level. This implies that NOx adsorbers can
tolerate as much as a four-fold increase in  desulfation frequency (when compared with an
expected seven to 10 ppm average) without any increase in thermal  degradation.  That
conclusion was inconsistent with our understanding of the technology at the time of the HD2007
rulemaking and remains inconsistent with  our understanding of progress made by industry since
that time.  Diesel fuel sulfur levels must be at or below 15 ppm to limit the number and
frequency of desulfation events. Limiting the number and frequency of desulfation events will
limit thermal degradation and thus enable the NOx adsorber technology to meet the NOx
standard.

        This conclusion remains true for the highway NOx adsorber catalyst technology and
will be equally true for nonroad engines applying the NOx adsorber technology to comply with
the Tier 4 standards.

   Nonroad and highway diesel engines are similarly  durable, so they consume a similar amount
of diesel fuel their lifetimes.  This means that both nonroad and highway diesel engines will have
the same exposure to sulfur in diesel fuel and will therefore require the same number of
desulfation cycles over their lifetimes. This is true independent of the test cycle or in-use
operation of the nonroad engine.

   Sulfur in diesel fuel for NOx adsorber  equipped engines will also have an adverse effect on
fuel economy. The desulfation event requires controlled operation under hot and net fuel-rich
exhaust conditions.  These conditions, which are not part of a normal diesel engine operating
                                         4-108

-------
	Technologies and Test Procedures for Low-Emission Engines

cycle, can be created through the addition of excess fuel to the exhaust. This addition of excess
fuel causes an increase in fuel consumption.

   Future improvements in the NOx adsorber technology, as we have observed in our ongoing
diesel progress reviews, are expected and needed to meet the Tier 4 NOx standards. Some of
these improvements are likely to include improvements in the means and ease of removing
stored sulfur from the catalyst bed. However because the  stored sulfate species are inherently
more stable than the stored nitrate compounds (from stored NOx emissions) and so will always
be stored preferentially to NOx on the adsorber storage sites, we expect that a separate release
and reduction cycle (desulfation cycle) will always be needed to remove the stored sulfur.  We
therefore believe that fuel with a sulfur level at or below 15 ppm sulfur will be necessary to
control thermal  degradation of the NOx adsorber catalyst and to limit the fuel economy impact of
sulfur in diesel fuel.

   4.1.7.2.2 Sulfate P articulate Production and Sulfur Impacts on Effectiveness of NOx Control
   Technologies

    The NOx adsorber technology relies on a platinum-based oxidation function to ensure high
NOx-control efficiencies. As discussed more fully in Section 4.F.I, platinum-based oxidation
catalysts form sulfate PM from sulfur in the exhaust gases significantly increasing PM emissions
when sulfur is present in the exhaust stream.  The NOx adsorber technology relies on the
oxidation function to convert NO to NO2 over the catalyst bed.  For the NOx adsorber this is a
fundamental step prior to the storage of NO2 in the catalyst bed as a nitrate.  Without this
oxidation function the catalyst will trap only that small portion of NOx emissions from a diesel
engine that is NO2. This would reduce the NOx adsorber effectiveness for NOx reduction from
in excess of 90 percent to something well below 20 percent.  The NOx adsorber relies on
platinum to provide this oxidation function due to the need for high NO oxidation rates under the
relatively cool exhaust temperatures typical of diesel engines.  Because of this fundamental need
for a precious-metal catalytic oxidation function, the NOx adsorber inherently forms sulfate PM
when sulfur is present in diesel fuel, since sulfur in fuel invariably leads to sulfur in the exhaust
stream.

   The Compact-SCR technology, like the NOx adsorber technology, uses an oxidation catalyst
to promote the oxidation of NO to NO2 at the low temperatures typical of much of diesel engine
operation.  By converting a portion of the NOx emissions to NO2 upstream of the  ammonia SCR
reduction catalyst, the overall NOx reductions are improved significantly at low temperatures.
Without this oxidation function, low-temperature SCR NOx effectiveness is dramatically
reduced, making compliance with the NOx standard impossible. Future Compact-SCR systems
therefore need to rely on a platinum oxidation catalyst to provide the required control of NOx
emissions.  This use of an oxidation catalyst to enable good NOx control means that Compact-
SCR systems will produce  significant amounts of sulfate PM when operated on anything but the
lowest fuel sulfur levels due to the oxidation of SO2 to sulfate PM promoted by the oxidation
catalyst.
                                         4-109

-------
   Without conversion of NO to NO2 promoted by oxidation catalysts, neither of these control
technologies can meet the Tier 4 NOx standard. Each of these technologies will therefore
require low-sulfur diesel fuel to  control the sulfate PM emissions inherent in the use of highly
active oxidation catalysts.  The NOx adsorber technology may be able to limit its impact on
sulfate PM emissions by releasing stored sulfur as SO2 under rich operating conditions. The
Compact-SCR technology, on the other hand, has no means to limit sulfate emissions other than
through lower catalytic function or lowering sulfur in diesel fuel. The degree to which the NOx
emission-control technologies increase the production of sulfate PM through oxidation of SO2 to
SO3 varies somewhat from technology to technology, but it is expected to be similar in
magnitude and environmental impact to that for the PM control technologies discussed
previously, since both the NOx and the PM control catalysts rely on precious metals to achieve
the required NO to NO2 oxidation reaction.

   At fuel sulfur levels below 15 ppm this sulfate PM concern is greatly diminished. Without
this low-sulfur fuel, the NOx control technologies are expected to create PM emissions well in
excess of the PM standard regardless of the engine-out PM levels. We therefore believe that
diesel fuel sulfur levels will need to be at or below 15 ppm to apply the NOx control technology.

4.2 Transient Emission  Testing

4.2.1 Background and Justification

   In the 1998 Rulemaking for Nonroad Compression Ignition Engines, we acknowledged that
effective in-use control of emissions from nonroad sources would be positively impacted by
having a duty cycle that more accurately characterized the transient nature of nonroad activity.
While no certification cycle may guarantee complete in-use emission control, a cycle that
appropriately characterizes the activity of the subject equipment achieves a greater level of
control.  The basics of any nonroad transient duty cycle should fulfill the following goals:

   •   Represent nonroad activity broadly, with a basis in real-world activities through diverse
       data segments;
   •   Exercise the engine over its operating range; cycle not limited to a specific speed or load,
       but traverses the operating range over the engine's full power range;
   •   Measure particulate matter (PM) on a transient basis;
   •   Capture the basic characteristics of PM, as currently defined, including:
          - organic and inorganic carbon fractions
          - volatile fraction
          - sulfate fraction
          - ash, etc., and
   •   Ensure that measures developed to control emissions over the cycle encourage and afford
       greater assurance of adequate control  of emissions in-use.

   Since that rulemaking, we have embarked on a strategy for cataloging operational data,
generating a duty cycle from those data sets, and compiling a transient composite duty cycle that
represents a broad range of activity for nonroad diesel equipment.  Working cooperatively with
the Engine Manufacturers Association (EMA) and through contract with the Southwest Research
Institute (SwRI), we created a set of duty cycles based on the following nonroad applications:

-------
	Technologies and Test Procedures for Low-Emission Engines

   - Agricultural Tractor
   - Backhoe Loader
   - Crawler Tractor
   - Arc Welder
   - Skid Steer Loader
   - Wheel Loader
   - Excavator

   These application duty cycles were created from actual speed and load data recorded in-use
on each of these pieces of equipment.  The strategy for generating the duty cycles and the base
data  sets differed slightly. However, combining these two strategies has ensured that the
strengths of both approaches are integrated into the resultant composite  duty cycle. Each of the
pieces of equipment represented the top tier of nonroad equipment as defined by their
contribution to nonroad diesel inventory as defined by the 1991 Nonroad Engine and Vehicles
Emissions Study (NEVES).  The pieces of equipment selected have retained their historical
significance even today as they appear to match fairly well with EPA modeling data for the
impacts of those applications.

   The existing steady-state duty cycle affords good coverage of the range of activity seen by
nonroad diesel applications; however, it is incomplete.  The range of nonroad activity is much
broader and much more varied than can be captured by a set of steady-state points (see Figure
4.2-1). No single transient cycle, of reasonable length, could capture the full body of nonroad
diesel activity from the various equipment applications. However, it is possible to capture
typical operation of nonroad equipment and to extrapolate the applicability of available data to
the remainder of nonroad equipment for purposes of certification and modeling.  This can't
replace an in-use characterization, but it does drive development of engine design strategies to
focus emission-control and performance parameters on a broader set of  activity that is much
more likely to be seen in use.

   A much broader set of data from the nonroad duty cycle generation may be found in
Memorandum from Cleophas Jackson to EPA Air Docket A-2001-28. This operational and
cycle data demonstrate the amount of nonroad activity that can occur outside the modes of the
ISO  Cl  duty cycle.

   4.2.1.1 Microtrip-Based Duty Cycles

   The microtrip-based cycles were created based on a range of activity the equipment is likely
to see in use. The weighting of each microtrip impacted the duration of each segment within the
resulting duty cycle. Each microtrip was extracted from a full set of data with the equipment
being operated within the targeted implement application. The data from the extracted segment
were compared with the full body of data for the targeted implement application based on a chi
square analysis, with a 95% confidence level,  of the nature of the operation. This included a
characterization of the speeds, loads, velocities,  and accelerations over the full operating map,
for the given piece of equipment. Experienced operators conducting actual work operated each
unit.  The projects ranged from an actual farmer plowing to a backhoe digging a trench for a

                                         4-111

-------
Regulatory Impact Analysis
municipal works project to a wheel loader in a rock quarry loading a truck to a skid steer loader
preparing plots in a subdivision under construction. The microtrip-based application duty cycles
were the Agricultural Tractor cycle, the Backhoe Loader cycle and the Crawler/Dozer cycle.

   4.2.1.2 "Day-in-the-Life"-Based Duty Cycles

   In attempting to address real-world activity, another strategy was employed for the second
set of nonroad duty cycles.  This approach was termed the "day-in-the-life" strategy.  It could be
said that this approach yielded only a single or perhaps two microtrips per piece of equipment.
This approach was employed to capture data for work that would have otherwise have been done
regardless of EPA data collection needs.  With these pieces, the data recorded was simply data
generated as selected pieces of equipment were used by contractors or construction personnel
during their normal operation versus being asked to do certain types of operation. The day-in-
the-life duty cycles consisted of the Skidsteer Loader cycle, the Arc Welder cycle, the Rubber
Tire Loader cycle, and the Excavator cycle.  The Excavator Cycle is in fact a composite duty
cycle assembled from three equal time segments of operating data from two different excavators.
                                         4-112

-------
                      Technologies and Test Procedures for Low-Emission Engines
                                       Figure 4.2-1
       Backhoe Loader and Crawler Tractor Cycle Data
     Backhoe\Loader Cycle Data (with ISOC-1 V\feightings)

     100
      versus the ISO 8178-4 Cl Cycle

 Crawler Tractor Cycle Data(with ISOC-1 V\feightings)
                	            ^	s.

100 T
       -20
                20   40   60   80   100
                  Speed(%)
           20   40   60   80
              Speed(%)
4.2.2 Data Collection and Cycle Generation

   4.2.2.1 Test Site Descriptions

   Operators were instructed to complete a job commensurate with the functionality of the
vehicle and at their customary pace. Experienced operators conducted their normal work with a
given piece of nonroad equipment.  The work conducted by the equipment during the data
collection was actual work and not  artificial scenarios, which ensured the accuracy of the data.

   4.2.2.1.1 Agricultural Tractor Cycle Operation

   The John Deere agricultural tractor was operated by an experienced farmer on his farm. The
farmer was asked to conduct the following activities as he normally would on any given work
day.  This activity formed the basis for the microtrips for the agricultural tractor duty cycle. The
microtrip activity segments included: planter, tandem offset discing (35 foot), bedder, cultivator,
ripper (10 row), folding chisel plow, and turnaround. The work was conducted during spring
planting season in Hamlin, Texas, using an actual in-use field being prepared for cultivation.
The tractor was used to make passes with each selected implement. The normal load operation
retained for inclusion in the cycle generation was the "normal" operation with each implement.
The data from the intentionally, highly loaded pass were not included in the eventual
Agricultural Tractor cycle.
                                          4-113

-------
Regulatory Impact Analysis
   4.2.2.1.2 Backhoe Loader Cycle Operation

   The Caterpillar backhoe loader was utilized on a site by the City of Houston, Utility
Maintenance Division, Fleet Management Department to conduct the following activities:
reading, trenching, loading and grade and level.  The operation was conducted by a municipal
employee experienced in the operation of the backhoe conducting that activity.  Engine data
were collected during the repair of a collapsed city sewage line in a residential neighborhood.
The activity included demolishing the road over the sewage line, trenching to reach the pipe,
craning to remove the old pipe and install the new pipe, backfilling, loading, spreading gravel,
and finish-grading the site.

   4.2.2.1.3 Crawler Tractor Cycle Operation

   The Caterpillar D4 Tractor was used to conduct the following activity on the grounds of
Southwest Research Institute by an experienced  operator.  The microtrips included road bed
preparation, clearing activity, and pit activity.  The operation was examined at three independent
sites.  Site 1 included clearing trees and brush for a construction site.  At Site 2 the equipment
dug and prepared a road bed. At Site 3 V-trench and pit operations were examined. This
activity was similar to preparing a site for a small building foundation.

   4.2.2.1.4 Wheel Loader Operation

   The Caterpillar 988F Wheel Loader was operated at Redland Stone Products Company
(quarry) in  San Antonio, Texas. Data were  collected between June 8 and June 10, 1998. The
equipment was operated from morning until midnight, working to fill construction and mining
trucks, open-topped trailers of Class-8 highway trucks, and rail cars.156 The material being
moved was typical  for a quarry application, including aggregate of various densities, such as
crushed stone, gravel, and sand. Twenty-six hours of data were gathered at the quarry for the
wheel loader.

   4.2.2.1.5 Skid Steer Loader Operation

   The Daewoo skid steer loader was operated at a construction site for a new complex of
townhouses in the San Antonio, Texas, area by a commercial site preparation company.  The
equipment was used to create drives for individual homes. Specifically, the skid steer loader was
used to haul and position aggregate foundation material to prepare the driveway and sidewalk
areas prior to laying asphalt. Over twelve hours of data were gathered over three work days for
the skid steer loader. The implement used by the skid steer loader during this operation was its
bucket.

   4.2.2.1.6 Arc Welder Operation

   The Lincoln Electric 250-amp arc welder was operated at Redland Stone Products Company
(quarry) in  San Antonio, Texas. Data were  collected over a single work day. The equipment
                                         4-114

-------
	Technologies and Test Procedures for Low-Emission Engines

was used to perform repairs on a large, mobile steel crusher tower by a private contract firm,
Holt.  Eight hours of data were gathered at the quarry for the arc welder.

   4.2.2.1.7 Excavator Operations

   The Hitachi EX300LC excavator was operated at 3 different sites over 7 days in the greater
San Antonio metropolitan area. Data were collected during Winter 1998 and Spring 1999. The
equipment was used to level ground at a building site, to load aggregate materials into trucks at a
quarry and to dig trenches and transport pipes for a sewer project.  Almost thirty-nine hours of
data were gathered for this excavator.

   The Caterpillar 320BL excavator was operated at 4 different sites over 6 days in the greater
San Antonio metropolitan area. Data were collected during Winter 1998 and Spring 1999. The
equipment was used to perform digging, trenching, pipe transport and placement and backfilling
associated with an on-going sewer project. More than thirty-eight hours of data were gathered
for this excavator.

   4.2.2.2 Engine and Equipment Description

   In generating the microtrip-based and the day-in-the-life duty cycles, the equipment selected
were based on the highest sales volume applications  and the contribution of those applications to
the ambient inventories for NOx and PM. Those cycles were created based on a John Deere
4960 Agricultural Tractor, Caterpillar 446B Backhoe Loader, and a Caterpillar D4H Crawler
Tractor.  The detailed description of the engines may be seen in Table 4.2-1 through Table 4.2-
o 157
                                      Table 4.2-1
                          Agricultural Tractor—John Deere 4960
Engine Characteristic
Rated Speed (rpm)
Peak Torque (Nm)
Peak Power (kW)
Low Idle Speed (rpm)
Operating Range (rpm)
Other Engine Descriptors
Value
2200
970
189.2
850
850-2400
7.6L displacement, electronic controls
                                         4-115

-------
Regulatory Impact Analysis
                                      Table 4.2-2
                           Backhoe Loader—Caterpillar 446B
Engine Characteristic
Rated Speed (rpm)
Peak Torque (Nm)
Peak Power (kW)
Low Idle Speed (rpm)
Operating Range (rpm)
Other Engine Descriptors
Value
2200
405
76.8
800
800-2300
CAT 3 114-D17 engine
                                      Table 4.2-3
                            Crawler Tractor—Caterpillar D4H
Engine Characteristic
Rated Speed (rpm)
Peak Torque (Nm)
Peak Power (kW)
Low Idle Speed (rpm)
Other Engine Descriptors
Value
2200
442
85
800
3204-D17 engine
   The engines used for data generation for the day-in-the-life approach were from a skid steer
loader, an arc welder, and a wheel loader. The engine parameters of the Caterpillar 988F Series
II rubber tire loader, the Lincoln arc welder and the Daewoo skidsteer loader are listed in Table
4.2-4 through Table 4.2-6.
                                         4-116

-------
  Technologies and Test Procedures for Low-Emission Engines
                  Table 4.2-4
Rubber Tired Loader—1997 Caterpillar 988F Series II
Engine Characteristic
Rated Speed (rpm)
Peak Torque (Nm)
Peak Power (kW)
Low Idle Speed (rpm)
Operating Range (rpm)
Other Engine Descriptors
Value
2080
2908
321
850
850-2250
CAT 3408E-TA engine,
Caterpillar HEUI Fuel System, electronic
                  Table 4.2-5
 Arc Welder—1997 Lincoln Electric Shield-Arc 250
Engine Characteristic
Rated Speed (rpm)
Peak Torque (Nm)
Peak Power (kW)
Low Idle Speed (rpm)
Operating Range (rpm)
Other Engine Descriptors
Value
1,725
162
28.3
1375
800-1900
Perkins D3 .152 engine
                  Table 4.2-6
    Skid Steer Loader—1997 Daewoo DSL-601
Engine Characteristic
Rated Speed (rpm)
Peak Torque (Nm)
Peak Power (kW)
Low Idle Speed (rpm)
Peak Torque Speed (rpm)
Other Engine Descriptors
Value
2,800
121 Nm
30.6 kW
800
1,700
Yanmar 4TNE84 engine, 2.0 L Displacement,
in-line 4 cyl, naturally aspirated
                    4-117

-------
Regulatory Impact Analysis
   Two pieces of equipment were selected for generating the excavator duty cycle based on
estimates of equipment population and power distribution among excavators in the nonroad
equipment inventory in the United States at that time.158 With the highest excavator sales
volumes being in the 60-130 kW and 130-225 kW ranges, the Agency created its excavator duty
cycle based on both a Hitachi EX300LC excavator at 155  kW (208 hp) and a Mitsubishi/CAT
320 BL excavator at 95 kW (128 hp). The detailed description of the engines may be seen in
Table 4.2-7 and Table 4.2-8.

                                      Table 4.2-7
                 Excavator (higher power output)—1997 Hitachi EX300LC
Engine Characteristic
Rated Speed (rpm)
Peak Torque (Nm)
Peak Power (kW)
Low Idle Speed (rpm)
Peak Torque Speed (rpm)
Other Engine Descriptors
Value
2,200
Nm (636 Ibs-ft)
155kW
680
1,500
ISUZU A-6SD1TQA(AC/JI) engine,
9.8 L displacement, mechanical controls
                                      Table 4.2-8
              Excavator (lower power output)—1997 Mitsubishi/CAT 320 BL
Engine Characteristic
Rated Speed (rpm)
Peak Torque (Nm)
Peak Power (kW)
Low Idle Speed (rpm)
Peak Torque Speed (rpm)
Other Engine Descriptors
Value
1,800
Nm (4731bs-ft)
95 kW
800
1,200
Mitsubishi/CAT 3066T engine, 6.4 L
displacement
   4.2.2.3 Data Collection Process

   The data collection process for both the microtrip-based and the day-in-the-life duty cycles
was based on collecting engine operational data in the field by mechanical and electronic means.
Engine speed data were measured by instrumenting the engine of each piece of equipment with a
tachometer to measure engine speed in revolutions per minute (rpm).  The torque was measured
either mechanically by linear transducer or as transmitted across the engine's control area
                                        4-118

-------
	Technologies and Test Procedures for Low-Emission Engines

network as a fuel-based torque signal. The mechanical torque measurement utilized rack
position to determine the load being demanded of the engine.  To calibrate the voltage signal
from the linear actuator the engine rack position versus actual fuel rate and engine-out torque
were determined based on laboratory evaluation of the same model engine. Once a map of
engine speed, load, actual torque, and fuel rate was compiled, the in-field load was determined
based on rack position and engine speed, as measured by the tachometer.

   Data loggers were used to record field data during operation and the data loggers were
equipped with flash memory media. The data loggers recorded engine parameters only during
operation, so data gathering did not occur while the engine was stopped. Data collection rates
varied from cycle  to cycle from a rate of 3.33 Hz to 5 Hz. Using cubic spline interpolation, the
data were  then reduced to 1 Hz format for the purpose of cycle generation.

   4.2.2.4 Cycle Creation Process

   The basic methodology of comparing extracted segments to the full body of data were used
for both duty cycle types. The major difference is in how the activity was defined for each.  The
microtrip-based activity specified the type of work performed by various implements for a given
piece of nonroad equipment in an effort to effectively incorporate the different types of operation
through which the equipment could be exercised over its lifetime. The day-in-the-life approach
was meant simply to characterize the nature of the full range of activity seen by the equipment
during its  typical operation over the period of evaluation.  The body of data for neither approach
was meant to be all encompassing to the extent that no other activity  would be expected from
that piece  of equipment over its lifetime.  The microtrip approach represents the broadest sweep
in the compilation of nonroad operation.  The resulting duty cycles in each case do represent the
most representative set of data from the full body of data collected.

   4.2.2.4.1 Microtrip Cycle Creation

   The contractor that conducted the in field testing and data reduction was Southwest Research
Institute (SwRI) with significant input from the Engine Manufacturers Association (EMA) and
direction from the United States Environmental Protection Agency (EPA). The methodology
used for creating the microtrip-based cycles involved extracting the actual data by comparing the
running window of actual data to the full body of data that was collected for each type of
activity. This involved a chi-squarew analysis comparing observed to expected data. The
observed data set was the data being evaluated for inclusion in the cycle for the  specific active
window.  The expected data set was represented by the full body of data from the given activity.
The chi-square comparison involved assessing the following for each window of operation:

   •   Rate of change in speed (dSpeed)
   •   Rate of change in torque (dTorque)
        i -Ei) / Ei where Oi is the Observed frequency in the ith interal and Ei is the Expected frequency in the ith
interval based on the frequency distribution of the entire population for the given quantity.

                                         4-119

-------
Regulatory Impact Analysis
   •   Power
   •   Rate of change in power (dPower)
   •   Speed and torque
   •   Torque and dSpeed
   •   Speed and dTorque
   •   Duration and magnitude of change in power

   The specific steps involved in cycle generation were the following:

   1. Separate the raw vehicle data into data files by vehicle activity.
   2. Load first activity file.
   3. Calculate power. Add to raw data file.
   4. Normalize speed using the FTP process and manufacturer's specified rated speed.
       Normalize torque, and power using measured peak values and create a scalar-normalized
       data file.
   5. Calculate the time derivative of normalized speed, torque, and power.
   6. Calculate the duration and magnitude  of all increases, decreases, and steady-state periods
       from the normalized power data.x  Count occurrences of duration and magnitude of
       changes in power for selected ranges.
   7. Count occurrences of power and rates of change of speed, torque, and power for selected
       ranges.  Count occurrences of speed  and torque, change in speed at selected torque levels,
       change in torque at selected speed levels, and duration and magnitude of changes in
       power for selected ranges.  The relative frequencies of occurrence (RFO) were collected
       within the specified ranges of activity (e.g. normalized range of speed of 20 units).
   8. Characteristic graphs of each activity was created for each piece of equipment. Several
       formats were used to characterize the various analysis of the equipment operation:
       - Scatter plots of normalized speed and load data
       - RFO data for deltaY speed versus normalized torque
       - RFO data for normalized speed versus delta normalized torque
       - RFO plots of magnitudes and duration of delta power
   9. The analysis of steps 1-8 was conducted by SwRI for each activity for each duty  cycle.
   10. The scalar normalized speed data (based on manufacturer specified rated speed) and
       normalized torque (or load - based on the peak torque available at the given speed) was
       used to generate the final set of activity comparisons for extracting the "actual"  data for
       the microtrip from the full body  of activity data collected for the specific application.

   Microtrip Weightings

   The microtrips of the agricultural tractor cycle, backhoe loader cycle, and crawler cycle were
weighted based on feedback from the engine manufacturers on the amount of time each
   xSteady-state is defined as any instantaneous change in normalized speed or normalized torque with a
magnitude less than 2%.

   YDelta is used to describe the instantaneous rate of change of the specified quantity.

                                          4-120

-------
	Technologies and Test Procedures for Low-Emission Engines
application was expected to operate using a given implement performing a set function over the
lifetime of that piece of equipment. The microtrip weighting for the Agricultural Tractor cycle
may be seen in Figure 4.2-2 to Figure 4.2-4. The cycle creation was based on linking the
microtrips with transition points between each activity segment.
                          Planter
                       Bedder
                         5%
                       Plow
                       18%
                     Field Cultivator
                             15%
   Figure 4.2-2
Agricultural Tractor
   Turnaround
 5%  4%  Idle6%
               ixboard Ripper
                     18%
                     29%
       Tandem Offset Disc
                                   Figure 4.2-3
                                Backhoe Loader
                             Grade and Level
                                 Trenching
                                 37%
                                      4-121

-------
Regulatory Impact Analysis
                                       Figure 4.2-4
                                   Crawler Tractor
                                              Idle
                            Pit Activity   	8%
                             34%
                                      Road Bed Preparation
                                                  47%
   In generating the duty cycles and conducting the analyses, relative frequency of occurrence
of various parameters as reported by the contractor were compared with the full set of real-world
data. Figure 4.2-5 shows the difference in the full set of real-world data collected versus the
microtrip, for one activity type. As can be seen in this figure, the difference in the total data set
and the identified microtrip was relatively small, based on the relative frequency of occurrence.
                                         4-122

-------
                      Technologies and Test Procedures for Low-Emission Engines
                                        Reading
                            RFO Differences from Activity to Microtrip
       f
                                                                  Normalized Torque Range
                 Normalized Speed Range
                                  Figure 4.2-.5
                    Example of Microtrip vs. Data Set for Tractor Activity
   Cycle Creation

   Each of the microtrip-based duty cycles were created based on the statistical analysis
previously described.  The linked component microtrips were then reduced to 1 Hz data from the
original 3.33 Hz signal using a cubic spline interpolation. The duty cycle was then speed and
torque normalized, based on the maximum available power/torque mapping.  These duty cycles
were the first set of cycles that were used for creating the composite nonroad transient duty
cycle.

   4.2.2.4.2 Day-in-the-Life Duty Cycle Generation

   In generating the day-in-the-life data, a similar chi-square analysis was used to compare RFO
data from the running window of data with the full body of data. The distinction lies in that this
was not done for multiple activity types for each piece of equipment. The analysis was
conducted using a nineteen-minute window incremented at one-minute intervals. The approach
used for data reduction, while similar, also varied in that the bin increments used for the day-in-
the-life duty cycles was 100 rpm and 200 Ib-ft for torque versus the normalized 20 percent
windows from the microtrip approach. The steps taken by SwRI are as follows.

                                         4-123

-------
Regulatory Impact Analysis
    1. Define "bins" sized at 100 rpm for speed by 200 ft-lb for torque.
    2. Sort entire data file (e.g. 376,768 observations ~ 26 hours) into bins.
    3. Compute a frequency table to indicate the number of observations contained in each bin.
    Similar to the RFO bins from the microtrip analysis.
    4. Increment within data file by 1 minute, and sort the next 19 minutes
    5. Compute the chi-square statistic for comparison with frequency distribution of the
    population data file.
    6. The approach to analyzing each nineteen-minute "window" of activity was repeated at
    one-minute increments for the entire body of data.
    7. The window of activity that best represented the full body of data for that piece of
    equipment was selected as the most typical duty cycle.
    8. Four iterations on the analysis was conducted to develop a typical 1 duty cycle, a typical 2
    duty cycle, a high transient speed2 duty cycle, and a high transient torque duty cycle for each
    application.
    9. For each window of activity, the data used were the actual, contiguous data from the body
    of data for that piece of equipment.

    Given the nature of this data-generation process, the detailed analysis  needed for weighting
the microtrips  and determining the time basis for inclusion into a composite cycle was not
needed. The resulting duty cycles were simply the result of the extraction of data from the
complete raw data set, which were subsequently normalized.

    4.2.2.4.3 Excavator Cycle Generation

    Data files for each piece of equipment were appended together in chronological order to form
a data population for that excavator. Each data population contained columns for time of data
acquisition (incremented at 5 Hz), engine speed, and rack position.  Data for engine speed and
rack position were used to compute a column for torque in units of pound-feet (Ib-ft), based on
the rack-to-torque algorithm using correlation information compiled earlier for the corresponding
excavator engine. Tasks of choosing the representative segments to form  a composite excavator
cycle were then initiated based on these two different data populations.

    The in-use data population of each excavator was sorted into two-dimensional intervals or
"bins," and a histogram was compiled based on the frequency of occurrences for speed and
torque pairs within the designated bins. The percent or relative frequency of occurrence (RFO)
is considered a histogram that describes the data population.  By  choosing a segment that closely
matched the characteristic RFO  compilation, it is therefore rationalized that the chosen segment
is indeed representative of the given data population. Using the same bin  intervals as were
applied to create a histogram (RFO) for each data population, a similar histogram was created for
each 380-second candidate segment of data. Each candidate segment overlapped the previous
segment by 320 seconds, as the process for excerpting candidate  segments incremented through
   zHigh transient duty cycles (speed or torque) represent the single most transient speed or torque window of data
(highest number and magnitude of instantaneous changes in speed or torque) from the full body of data.

                                          4-124

-------
_ Technologies and Test Procedures for Low-Emission Engines

the data population using a 60-second step size. Chi-square analyses tested each candidate
segment to rank each segment by comparing its RFO histogram to the RFO histogram created for
its associated data population. The following is the approach used for computing a chi-square
statistic, relative frequency of occurrence distributions to that of the corresponding population
for engine speed and torque values, for each candidate segment:

   1 . Define "bins" for speed expressed in rpm, and torque as Ib-ft
   2. Sort each data population (approximately 38 hours, at 5 Hz) into bins
   3. Compute a relative frequency of occurrence table to indicate the percentage of
   observations contained in each bin
   4. Increment through the data population by 60 seconds, sort the next 380-second segment
   into similar bins, and compute a relative frequency of occurrence table
   5. Compute a chi-squarea statistic for comparing the frequency distribution of the segment to
   that of the population
   6. Repeat Steps 4 and 5 for all such 380-second candidate segments, for an entire data
   population
   7. Sort segments by increasing chi-square rank (low statistic means  good correlation)

   Note: The chi-square statistic is the summation of:
   where O; is the observed frequency in the ith interval of the 380-second sample window, and
   E; is the expected frequency of the ith interval based on the frequency distribution of the
   entire population.

The sliding 380-second "window" was used to determine the distribution of speed-torque
combinations experienced by each type of equipment over the entire range of operating data
collected on each unit. The "window" was advanced by one-minute increments through the data
to determine a most typical segment for each excavator and a second most typical segment for
the lower-powered unit.

   Based on initial torque map information obtained with each engine on the steady-state test
bench, a normalizing process was applied to each of the 5 Hz data segments (part of "data
smoothing").  FTP normalizing methods outlined in the 40 CFR part 86, subpart N, were used
for expressing observed engine speed and torque values for the three selected segments of 5 Hz
data in terms of the percentage of an engine's full load performance and idle speed. The 5 Hz
data for segments chosen to represent the first- and second-most typical segments in the data
population generated with the Caterpillar 320BL excavator were normalized using the rated
speed and torque map information obtained with the Caterpillar 3066T engine mounted on the
steady-state test bench.  Similarly, the 5 Hz data for the segment best representing the typical
operation of the higher powered Hitachi excavator were normalized using torque map
information obtained for the Isuzu A-6SD1T engine on the steady-state test bench.
                                         4-125

-------
Regulatory Impact Analysis
   An averaging method was applied to the three selected segments to convert each segment
from the original 5 Hz to 1 Hz data files.  Each 5 Hz data pair was first normalized and then the
percentage values were averaged. In general, the smoothing technique produced a value for
speed and a value for torque for each one-second interval (1 Hz) by averaging the five values in
the interval of interest.

   After establishing in-use operating engine speed and torque data populations for excavators
rated in both the low and high power ranges, three representative segments were appended
together to form a 20-minute composite excavator cycle.  The first two segments were the most
representative data from the lower and higher powered excavators, respectively. The third
segment represented the second-most typical data from the lower-powered excavator (i.e.,
ranked number two in chi-square analyses for that population).  This resulted in a composite
cycle that was apportioned with two-thirds data gathered from the Caterpillar 320BL excavator
rated in the 100 to 175 hp range, and one-third from data gathered from the Hitachi EX300LC
excavator rated  in the 176 to 300 hp range.  The three segments were then joined into a
composite 20-minute excavator duty cycle by the addition of appropriate transition segments
leading into and linking each segment of transient operation.  A three-second transition joined
Segment 1 and Segment 2, and  similarly another three-second transition joined Segments 2 and
3. A no-load idle condition was appended for 27 seconds at the beginning and end of the cycle.

4.2.3 Composite Cycle Construction

   Having all seven application cycles in hand, including the four cycle variations apiece for the
arc welder, skidsteer loader and rubber-tire loader, we began construction of a transient
composite nonroad duty cycle.  The approach for addressing the weighting of contributions from
each equipment type to the composite cycle was left at equally weighting each contribution.
While consideration was given  to population-weighted or inventory-based weighting factors for
the composite cycle, in the interest of ensuring  a universally applicable cycle, no unique
weighting factors were assigned.  The decision of which data segments to extract from the
component duty cycles was based on uniqueness of operation (avoidance of replicate data in the
composite cycle) and level of transient operation (steady-state operation was not included in the
transient cycle).AA Extracted cycle segments were linked using three second transition periods,
when needed, to ensure smooth transitions within the cycle and to avoid spurious data generation
based on changes in speed and  load that were unrealistic between segments. Transition periods
were deemed necessary when the change in the magnitude of the torque or speed value was
greater than twenty using the normalized data.  The cycle was constructed using the
denormalized segments for each component cycle based on the original engine map for the
engines used to  generate the component cycles.  Once the raw data were available, the
normalization based on the max speed map was conducted.  This was necessary because each
cycle was originally normalized using different procedures (e.g., FTP  speed and torque
   AASteady State Operation is defined as an instantaneous speed or torque change less than 2% of the maximum
magnitude.

                                         4-126

-------
	Technologies and Test Procedures for Low-Emission Engines

normalization or GCSBB speed with FTP torque normalization). The MAP used for normalizing
the raw data remained FTP-based (percent of maximum torque at the given speed) for torque.
The Maximum Speed Determination was used for the speed normalization. Figure 4.2-6
identifies the location of the cycle segments as extracted from the component application duty
cycles, the segment duration, and segment position in the composite duty cycle.
                                       Figure 4.2-6
Supplemental NRTC (Nonroad Transient Composite) Cycle



Application
Number




1




2


3


4

5




6


7






Nonroad
Application




Backhoe Loader




Rubber-Tire Loader


Crawler-Dozer


Aqricultural Tractor

Excavator




Arc Welder


Skid Steer Loader






Application
Duration
(seconds)



206




184


209


150

35




204


185






Application in
Cycle Position
(#seconds)



29-234




235-418


419-627


628-777

778-812




816-1019


1020-1204






Segments from
Application Cycle
(#seconds)



52-86
108-141
174-218
351-442

746-822
531-637

85-206
376-462

265-414

319-338
431-445



1007-1103
544-650

264-365
150-232





Segment
Name


Start/Transition

Reading
Trenching
Loading
Grade/Level

Typical Operation
Hi-Spd Transient

Road Bed Prep
Clearing

AgTractor

LowerHp (128Hp)
HigherHp (208Hp)

Transition

Typical Operation
Hi-Spd Transient

Typical Operation
Hi-Trg Transient

Idle/Transition/End



Segment
Duration
(seconds)

28

35
34
45
92

77
107

122
87

150

20
15

3

97
107

102
83

34



Cumulative
Cycle Time
(seconds)

28

63
97
142
234

311
418

540
627

111

797
812

815

912
1019

1121
1204

1238



Segment in
Cycle Position
(#seconds)

0-28

29-63
64-97
98-142
143-234

235-311
312-418

419-540
540-627

628-777

778-797
798-812

813-815

816-912
913-1019

1020-1121
1122-1204

1215-1238
     GCS Speed or Governed Central Speed is defined as the speed corresponding to the point along the engine's
MAP (maximum allowable power) curve at which power is 50% of maximum measured rated power once the
maximum measured power has been surpassed.
                                         4-127

-------
Regulatory Impact Analysis
4.2.4 Cycle Characterization Statistics

   The characterization of the operational data were also subsequently revisited for purposes of
comparison in addressing composite cycle construction.  The nature of the transient activity is
characterized in a report to EPA by Dyntel.159 The goal of the analysis was to provide an
assessment of the transient nature of nonroad activity between different applications. These
analyses (small bin, large bin, and general cycle) were used to address the comparability of the
resulting composite nonroad diesel transient duty cycle to the component data set that was
collected for each of the component cycles. The size of the bin was simply a reference to the
scale used for the analysis (either coarse or fine). As may be seen in Figure 4.2-7, the composite
nonroad transient duty cycle fit well within the average of all of the original nonroad duty cycles
based on the operational data. The figure is a plot of the nonroad composite cycle characteristics
with the statistics of the remainder of the nonroad diesel  cycles plotted as a mean with the
standard deviation between those statistics from the other cycles  shown.  The ten cycles
represented include:
   •Ag Tractor
   •Crawler
   •Skid Steer Typical 1
   •Wheel Loader High Torque Transient
   •Arc Welder High Torque Transient
• Backhoe
•Arc Welder Typical 2
•Wheel Loader Typical 1
•Excavator
•Skid Steer Loader High Torque Transient
                                     Figure 4.2-7
                Summary of Nonroad Cycles Comparison to NR Composite
                    NRC Compared  to the  10 Cycle
               Means and the 95%  Confidence Limits
IUU
on
fif)
DU
AC\
9D
n





F

[


F



r















             Speed accels/min  Speed decels/min Torque accels/min Torque decels/min   Avg. Speed
                                                                   Avg. Torque
                                 DNRC  Mean+CL  Mean-CL -Mean
                                        4-128

-------
	Technologies and Test Procedures for Low-Emission Engines

4.2.5 Cycle Normalization/Denormalization Procedure

   The actual values for speed and load in rpm and Ibs-ft for each of the application cycles
needed to be converted into normalized values before any application cycle could be used on an
engine, other than the engine originally used to create the application cycle itself.  This process
of normalization entailed converting the actual in-use operating speed and load values of the
"raw" duty cycle, as recorded from the engine used to create the cycle originally, into a
percentage of that engine's maximum achievable speed and load values.  This yields a schedule
of percentage-based speed and load values that can be converted to absolute values for speed
(rpm) and load (Ibs-ft). This conversion depends on applying the normalized percentage values
for speed and load to the maximum achievable power (MAP) for the new test engine.
Multiplying the percentage values of the normalized cycle by the measured speed and  load
maximums of the new engine's MAP curve, in fact, denormalizes the cycle.  This means that the
denormalized speed and load values may be used as commanded values on a test cell
dynamometer to exercise the new engine in exactly the same manner as the original engine was
run for a particular application cycle.  The load values in Ibs-ft for each of the seven types of
application cycles and all their cycle permutations, i,e., Typical, High Transient Speed , etc.,
were all converted to normalized values (and conversely, into denormalized values, at later
times) using the FTP normalization procedure detailed in 40 CFR Part 86. The speed values in
rpm for each type of application cycle were normalized initially in one of three different ways.

   The speed values in each of the original microtrip cycles, the agricultural tractor, backhoe
loader, and crawler-dozer, were all normalized using the FTP procedure. The speed values in
each of the original day-in-the-life  cycles, rubber tire loader, skidsteer loader and arc welder
were all normalized using the governed central  speed procedure (GCS).CC The speed values in
the excavator cycle were normalized, and later denormalized, using the FTP normalization
procedure detailed in 40 CFR Part  86. However, in time and for the construction of EPA's
composite nonroad cycle, all the application cycles were normalized using the Agency's
Maximum Speed determination procedure.

   The Maximum Speed Determination procedure uses the measured speed and load values
from an engine's power curve to determine what is the maximum power that the engine can
attain and at what speed that engine will achieve its maximum power. This value for speed at
maximum power can then be used in lieu of a manufacturer's rated speed number for a particular
engine to conduct a normalization or denormalization of engine or cycle for purposes of running
a duty cycle  on a particular engine. The procedure is based on a spreadsheet calculation and is
discussed in  our analysis of comments associated with the final rule for marine diesel engines
(64 FR 73300, December 29, 1999).160'161 As detailed in Figure 4.2-8, the maximum speed can
be found  below the point on the engine power curve that is the  farthest distance from the point of
   cc GCS is the speed value on the Maximum Achievable Power (MAP) curve of an engine at which the engine's
speed is 50% of the measured rated power for that engine, after measured rated power has been passed on the MAP
curve.

                                         4-129

-------
Regulatory Impact Analysis
origin of the graph of engine's measured speed and power values. That farthest point on the
curve is described as the point of maximum power achievable by the engine under study.

                                      Figure 4.2-8
                           Maximum Test Speed Determination
       40
       20
         0      250     500     750    1000    1250   1500   1750   2000    2250    2500
                                                                       Max Test
                                                                       Speed
rpm
4.2.6 Cycle Performance Regression Statistics

   In assessing the nonroad transient duty cycles, ten nonroad diesel engines were exercised
over the nonregulatory162 nonroad duty cycles to assess emission impacts of each duty cycle, as
well as to determine the ability of typical nonroad diesel engines to pass the existing highway
cycle performance regression statistics. That data may be seen in a report from SwRI with an
accompanying EPA summary of the results in the Memorandum to EPA Air Docket 2001-28
from Cleophas Jackson entitled "Nonroad Duty Cycle Regression Statistics." Subsequent
analysis on the composite nonroad transient cycle was based on test cell data collected from
testing at the National Vehicle and Fuel Emissions Laboratory and Southwest Research Institute,
as well as through the European Commission's Joint Research Center (EC-JRC), and various
engine manufacturers from the United States, Europe, and Japan.

4.2.7 Constant-Speed, Variable-Load Equipment Considerations

   Some nonroad diesel engines operate in equipment that calls for constant engine speeds.
Some examples of engines in this category of nonroad diesel equipment include pumps,
electrical power generator sets (gen sets), pavement saws and cement mixers. While the
operating speed in many cases is not truly constant, it is generally true that the unit's speed will
                                        4-130

-------
	Technologies and Test Procedures for Low-Emission Engines

vary little during operation.  These types of equipment are more tolerant of changes in operating
load than other more closely governed constant-speed nonroad applications. Some pieces of
constant-speed equipment will be governed to a nominal "zero" variation in rpm during
operation for critical operations such as maintenance of electrical power and refrigeration loads.
For those engines designed to operate under less restrictive, more "transient" conditions, the
Agency had proposed an alternative constant-speed, variable-load (CSVL) transient duty test
cycle over which an engine manufacturer might operate their engines.  The CSVL duty cycle
was meant to capture emissions from these infrequent modes of operation.  However, after a
review of comments and a broader look  at the wide range of applications embraced by the
constant-speed, variable-load segment of the nonroad diesel  equipment population, the Agency
has chosen not to adopt a CSVL transient test cycle at this time.  Instead, EPA, with all of its
stakeholders in this regard, will map out a process of engine  testing and analysis to better
characterize constant-speed equipment in-use to design the most appropriate test cycle for the
largest number of constant-speed engines.  Consideration will also be given to addressing the
operation of gen set applications as a potentially unique subset of this category.  EPA
undertakes this process with an eye to initiating a rulemaking which would lead to promulgation
of a transient cycle for constant-speed engines before the Agency's 2007 Nonroad Technical
Review.

   4.2.7.1. Background on Cycle Considered

   The CSVL transient test cycle was derived from EPA's Arc Welder Highly-Transient Torque
nonroad application duty cycle. That cycle was developed on a direct-injection,  naturally-
aspirated, 30kW (40 hp) diesel arc welder engine,  a constant-speed application running at
variable load. The Highly-Transient Torque cycle, one of four arc welder cycles, is comprised of
a single twenty-minute segment of all the real-time operating data collected on that engine.

   While designed to control nonroad engines in a broad range of constant-speed applications,
commenters noted that EPA's proposed  CSVL test cycle had an average speed which was lower
than the speed which many manufacturers considered optimal for their constant-speed engines
in-use. Further, EPA had received comments that many constant speed engines operated near or
at their rated engine rpm during much of that engine's useful life, as with electrical generating
sets in particular. EPA had proposed that these constant-speed engines, when tested in the
laboratory with installed speed governors, be required to meet cycle statistics for engine load but
not for engine speed. This relief was aimed at addressing the twin concerns that many engines
operated at a significantly high percent of their rated speed much of the time in-use and had
different degrees of engine speed variation during that operation.

   Engine manufacturers raised additional  design concerns for constant-speed engines required
to meet emission standards over EPA's proposed cycle. Their  concerns generally focused on the
fact that the cycle had relatively light engine loads and was derived from an arc welder powered
by a naturally-aspirated engine. Commenters questioned the representativeness of the CSVL
cycle for generators, which they claimed was a more common application within the
constant-speed engine  population than was  an arc welder.  A second issue involved the average
load that would be experienced by an engine running on the  CSVL test cycle.  The average load

                                         4-131

-------
Regulatory Impact Analysis
factor of the normalized application cycle was approximately 25% of engine capacity.
Manufacturers of constant-speed engines with significantly higher load factors on their engines
during operation, upwards of 90% of normalized engine load at constant speed, argued that their
engines would not be able to pass cycle-regression statistics for certification without significant
re-tuning of the engines to operate over the CSVL cycle. Several commenters noted that some
nonroad constant-speed engines with high brake-mean effective pressures (BMEP), or high
rated-power constant-speed engines, were narrowly focused on providing higher power
capability at a single speed while meeting emission requirements.  These engines used larger,
less-responsive turbochargers to achieve their requisite higher BMEP. Manufacturers pointed
out that the smaller BMEP engine on which the arc welder cycles were developed was more
responsive to torque changes than their high BMEP engines were designed to encounter.  As
such, these manufacturers felt that their engines would be penalized by the number and
magnitude of torque changes in the CSVL cycle.

   At the same time, however, the Agency shared engine manufacturers' concerns for creating a
duty cycle that achieved emission reductions while appropriately modeling in-use operation of
their engines. EPA  would have find it unproductive to require an approach that lead merely to
improvements in the operation and emissions of the engine under laboratory conditions which,
were in turn unrelated to the engine's in-use operation. Based on the comments the Agency has
received regarding the constant-speed, variable-load duty cycle, we intend to continue to work
with all interested parties to develop a new constant-speed, variable-load duty cycle. The
Agency envisions that any new test cycle would result in comparable stringency for ensuring
effective in-use control, as does the current duty cycle developed for fully transient test
characterization - EPA's NRTC test cycle.

   4.2.7.2. Follow-on Constant-Speed Engine Testing and Analysis

   In consultation with the Engine Manufacturers Association (EMA) and other stakeholders,
the Agency will embark on a process with the nonroad engine and equipment manufacturers that
will result in collection of additional engine operation data that will appropriately characterize
the operation of nonroad diesel engines used in equipment in constant-speed applications.  To
ensure that the data  collected is robust and applicable  to most, if not all, segments of the nonroad
equipment market, and to facilitate global technical regulations and eventual cycle
harmonization, the Agency, manufacturers and other interested parties, in consultation with
non-domestic governmental entities will work together to develop a plan that incorporates the
following elements:

   Define operation of a non-generator, non-transient equipment class:

       - Target Equipment/Application Types00
   °°- When the arc welder application was originally considered for inclusion in the cycle generation effort, EMA
endorsed EPA's choice of the arc welder as a constant speed application.


                                          4-132

-------
	Technologies and Test Procedures for Low-Emission Engines

          •air/gas compressors, pressure washers, water/irrigation pumps, oil field equipment,
          hydro power units, leaf blower/vacuums, shredders, bore/drill rigs, mixing
          equipment, pavement saws, arc welding sets, chippers/shredders/grinders, light
          plants/sign boards, tampers, rammers, and plate compactors, concrete/industrial saws,
          crushers/material handling equipment and refrigeration/AC equipment;
       - Engine Speed Range - anticipate EMA feedback
       - Power range
          • 25 to 175 hp, 175-350 hpEE, and 350 to 750 hp
       - Market Sectors
          • construction, agriculture, maintenance/handling, pumps/welders

   Define sample sizes, duration of "cycle" for application intercomparisons :
       - Number of pieces of equipment in each category
          • Sufficient to discern significant differences in speed and load characteristics
       - Number of hours of operation per application per site
          • Forty or more hours of operating data

   Define data collection parameters:
       - Speed
       -Load
       - Exhaust Temperature
       - Engine Oil Temperature (1st 20 minutes of engine on after 4 hours of engine off)
       - Engine Coolant Temperature (1st 20 minutes of engine on after 4 hours of engine off)

   In addition to ensuring that the sampling plan addresses the issues outlined above, EPA will
seek agreement among the stakeholders on the level of involvement of all parties in the data
collection and generation, data reduction and analysis, and final cycle construction and
assessment efforts. Initially, the logistical questions concerning program timing and duration of
all parts of the data collection and eventual cycle development efforts would have to be charted
and agreed upon by program participants. EPA expects that broad groupings of nonroad engines
from various applications would then be compared between and  among each other to determine
whether particular applications differed in terms of speed and load operating characteristics (see
Figure 4.2-9 below).  One question which is particularly important is whether "constant speed"
applications are similar to one another, but different than either transient or generator-type
applications.  As we move forward with the process of data collection and subsequent cycle
generation, other interested parties, including the state of California, will also be invited to
participate in these efforts. Future engine emission control technologies would need to be
anticipated and considered for their impacts on nonroad equipment emissions.
   EEThe Agency's current data base for cold start operation includes construction equipment in the power range of
150 to 350 hp.

                                         4-133

-------
Regulatory Impact Analysis
                                    Figure 4.2-9

                        Engine  Categories
              with respect to Certification  Cycles
 Measure: speed, load
Measure: speed, load,
em issions
                                                  Measure: on/off
Generator Sets




'Constant Speed;
- Arc Welders ^^^
r"p"iTm p s^4=Ef^ j!
- nil Rigs -f '!
~^jp
- Pavement Saw s-^j
j>
- Irrigation Sets *^CL
jr-
^dUitiiaLyajyuuasj!*

>
'Transient'
- Backhoes
- Dozers
- Others?


Arrows represent examples of comparisons between applications and engine groups, to determine
whether applications differ in terms of speed/load characteristics. An important question is whether the
'constant speed 'applications are similar to each other, but different from generators and
'Transient applications. '
4.2.8 Cycle Harmonization

   4.2.8.1 Technical Review

   One concern raised by the engine manufacturers was that the mapping method used to
generate the real-world torque data introduced an error by no appropriately accounting for the
impact of transient activity of the actual torque signal from the engine.  The basis of the issue
was primarily a torque signal in the field, based on the rack position, that may not have actually
occurred had an in-line torque meter been employed. Two aspects of this warrant review. The
first aspect of actual torque versus inferred torque.  The second aspect of this issue is whether or
not rack position or the demanded load is an appropriate metric for developing duty cycles
representing real-world operation.  To address the second issue in the context of responsiveness
of a nonroad engine, it should be clear that, although feedback torque from the engine provides a
clear signal of what was accomplished by the engine, it is not a fair metric of the demanded load.
Given the fact that a typical operator or driver tends to  demand a desired torque the engine's
response to that demand, though not distinct, is a separate issue. It is this reasoning through
which command cycles are generated. The command cycle represents the speed and load
demanded of the engine.  The engine's responsiveness  can be addressed through performance
statistics.
                                       4-134

-------
	Technologies and Test Procedures for Low-Emission Engines

   Engine manufacturers sought to address the first concern through a playback analysis that
addressed the la correction as an offset to the commanded load signal. The playback approach
would involve rerunning one of the engines (identical engine model) in the test cell over the
defined duty cycle with the calculated la offset to measure torque using an in-line torque meter.
Manufacturers provided the inertia data for their engines either used for cycle development or
anticipated to be included in the testing program. The data provided by members of the Engine
Manufacturers Association (EMA) may be seen in Table 4.2-9 and Table 4.2-10.
                                      Table 4.2-9
                    Nonroad Diesel Engines Used for Cycle Generation
No.
1
2
3
4
5
6
7
8
Engine Mfg
Caterpillar
Caterpillar
Caterpillar
Isuzu
John Deere
Mitsubishi
Perkins
Van mar
Engine Model
3204-D17
3114-D17
3408E - TA
A-6SD1 TQA
6081
3066T
'97 D3.152
'97 4TNE84
Machine Mfg
Caterpillar
Caterpillar
Caterpillar
Hitachi
John Deere
Caterpillar
Lincoln
Daewoo
Machine Model
Cat D4H
Cat 446B
988F-II
EX-300LC
JD 4960
Cat 320 Excavator
97 'Shield-Arc' 250,
K1283
DSL-601
Application
Crawler Tractor
Backhoe Loader
Wheel Loader (2)
Excavator High Power
Ag Tractor
Excavator Low Power
Arc Welder
Skid Steer Loader
Rated Power (Kw)
85 peak
76.8 peak; 70.8
rated
321
161
186
95
28
31
Peak
Torque (N
m)
442
405

834
970
641

121
Rated
Speed
(RPM)
2200
2200
2100
2000
2200
1800
1725
2800
Low Idle
(RPM)
800
800
850
850
850
860
800 (1)
800
                                      Table 4.2-10
                   Engine Inertia Data Used for la Correction Calculation
No.
1
2
3
4
5
6
7
8
Engine Mfg
Caterpillar
Caterpillar
Caterpillar
Isuzu
John Deere
Mitsubishi
Perkins
Yanmar
Engine Model
3204-D17
3114-D17
3408E - TA
A-6SD1 TQA
6081
3066T
'97D3.152
'97 4TNE84
Total Inertia
(Kg-m2)
1.7899
0.9770
2.8637
7.5303
2.4400
0.9160
0.1083
Total Inertia
(N-m-s2)
1.7899
0.9770
2.8637
7.5303
2.4400
0.9160
0.1083
0.2317 2.3629
Engine Inertia
(N-m-s2 = kg-m2)
0.2249
0.5550
1.3147
2.8263
0.5000
0.2160
0.1083

Flywheel Inertia
(N-m/s2 = kg-m2)
1.5650
0.4220
1.5490
4.7040
1.9400
0.7000


   The correction that was undertaken by EPA and Southwest Research Institute (SwRI) used
the following methodology. The original 3 Hz data set was used to correct the torque data rather
than interpolated 1 Hz data to ensure the raw data were corrected to avoid error propagation
within the 1 Hz scalar data.

   1.  Apply the la correction to calculate the new torque command.
   2.  Apply original technique to create 1 Hz raw command cycles using the cubic spline
   interpolation for the those cycles that were originally collected at 3.33  Hz.
   3.  Each resultant correct raw data duty cycle was then normalized using the Maximum Speed
   determination method (See Section 4.2.3).
                                         4-135

-------
Regulatory Impact Analysis
   4. Cycle segments for the Composite Nonroad Transient duty cycle were then reassemble
   from the component duty cycles.

   The result of the correction, as conducted by SwRI, was that there were very small
modifications to the most severe torque excursions. The peaks and valleys were trimmed
slightly. The overall change in the cycle resulted in less than 0.5% correction, typically.

   4.2.8.2 Global Harmonization  Strategy

   4.2.8.2.1 The Need for Harmonization

   Given the increasingly global marketplace in which nonroad engines are sold, alignment of
standards and procedures helps facilitate introduction of cleaner technology at lower across in
multiple markets.  Given the nature  of the nonroad diesel market with a large number of very
diverse product offerings and in some cases, small niche market volumes, the ability to design
once for different markets helps reduce the costs, especially of the lower volume equipment
models. While alignment of limit values may be a key component of harmonized regulations,
alignment of test procedures, measurement protocols, and other aspects of certification and
testing procedures help reduce the testing burden manufacturers will face when selling and
distributing their products in multiple markets.  Much of the development of new procedures and
test methods has originated in the United States, Europe, and Japan. While other markets tend to
adopt emission limits and procedures as a part of a more global  process on a different time
frame. Given the nature of regulatory and technological development, allowing the leading
markets for which new technology will need to be introduced to have comparable protocols
simply reduces the costs those markets will  be forced to absorb. In any effort to utilize
procedures in multiple regulatory arenas, care should be taken to include an assessment of
equivalence and appropriateness.  In so doing, both Europe and the United States conducted an
assessment of real-world operation of nonroad diesel equipment. The data-collection effort in
the United States started in 1995.  The subsequent data-collection effort in Europe confirmed
that,  as expected, nonroad diesel activity in  Europe was  comparable.

   In moving forward with a single test cycle for both Europe and the United States, and
potentially a global nonroad diesel cycle, the basic framework for the cycle was agreed upon. In
addition to the work initiated by the Agency in compiling a nonroad transient duty cycle, it was
important to ensure that concerns about global suitability be addressed. The context used for this
assessment in Europe was the existing European Transient Cycle (ETC).  While this duty cycle
was developed for highway diesel applications, it was seen as an adequate basis for which
European industry and government  staff could assess EPA's proposed Nonroad Transient Duty
Cycle. Representatives from Japan's government and industry have periodically participated in
this process as well; however, no  such framework for comparison was requested for the
evaluation process from any representative from Japan.  Throughout the development of the duty
cycle, industry representatives from the United States, Europe, and Japan have provided detailed
technical input. In Table 4.2-11 shows early results presented by Deutz exercising a nonroad
diesel engine over the EPA-generated Nonroad Transient Duty Cycle indicating an ability to
pass  cycle performance criteria with only a  slight problem with  the Torque Intercept statistic.

                                         4-136

-------
                     Technologies and Test Procedures for Low-Emission Engines
                                    Table 4.2-11
                          Initial Deutz Data Submission for
               EPA Nonroad Diesel Transient Duty Cycle (Nov. 13, 2000)
                                              |    Speed   |    Torque   |   Power    |
Standard error of estimate
(SE)

Slope of the regression line
(m)

Regression coefficient
(r2)

Y intercept of the
regression line (b)
measured
NRTC
ETC
tolerance

measured
NRTC
ETC
tolerance

measured
NRTC
ETC
tolerance

measured
NRTC
ETC
tolerance


red:
green:
56,48 rpm
24,29 rpm
max 100 rpm

1,010
0,990
0,95 to 1,03

0,996
0,993
min 0,9700

18,01 rpm
17,67 rpm
+/- 50 rpm
7,58%
6,59%
max 13%

0,925
0,963
0,83 to 1,03

0,958
0,980
min 0,88

30,10Nm
5,80 Nm
+/- 20 Nm

out of tolerance
to
7,15%
5,67%
max 8 %

0,968
0,976
0,89 to 1,03

0,973
0,981
min 0,91

3,82 kW
0,62 kW
+/- 4 kW

   4.2.8.2.2 Harmonization Methodology

   The composite Nonroad Transient (NRTC) duty cycle developed by the Agency was used as
the reference cycle for conducting subsequent development and testing work.  It was originally
introduced to the global regulatory community and engine industry in Geneva in June 2000.
After an on-going dialogue with industry in the United States and Europe, additional
modifications were suggested by the European Commission based on manufacturer concerns
with their ability to meet test cell performance statistics with this duty cycle. In September 2001,
it was decided by a joint European, American, and Japanese government and industry workgroup
that the Joint Research would use the then "candidate" cycle to conduct additional changes
commensurate with the goal of not allowing the instantaneous transient speed and torque
changes to be greater than those experienced within the European Transient Cycle (ETC).  Using
a Bessel filtering algorithm, the cycle was then modified by the EC-JRC to meet the ETC target
of 23% of torque events faster than 4 seconds. The two cycles may be seen on a time basis in
Figures 4.2-11 and 4.2-10.  The average load and average speed of each cycle are shown in
Table 4.2-12. The speed characteristics of the original cycle were similar to the speed
characteristics of the ETC.  This is not an indication that the speed trace was identical, but rather
that the maximum instantaneous speed changes of the NRTC were similar to the maximum
instantaneous speed changes of the ETC.FF
     Memorandum to EPA Air Docket A-2001-28 from Cleophas Jackson, Report from the JRC entitled
"Contribution to the NRTC Development Based on Test Data Supplied by Engine Manufacturers," February 26,
2001.
                                         4-137

-------
Regulatory Impact Analysis
                             Figure 4.2-10
            EPANonroad Transient Test Cycle as of March 2001

                       Draft Nonroad Transient Duty Cycle
                              Time (seconds)
                                     Table 4.2-12
                             Comparison of Cycle Averages
Duty Cycles
EPANRTC
JRC Modified NRTC
Average Normalized Speed
63%
68%
Average Normalized Torque
47%
39%
                                        4-138

-------
	Technologies and Test Procedures for Low-Emission Engines

   The following figures 4.2-12 through 4.2-16 describe the JRC Modified NRTC with respect
to speed and load and the transient nature of the cycle. This will be contrasted with the same
characteristics of the EPA- generated NRTC. The JRC modified NRTC was also known as the
San Antonio cycle or the JRC.
                                  Figure 4.2-11
                JRC Nonroad Transient Test Cycle after Bessel Filtering

                      Joint EPA-EU Nonroad Transient Cycle, March, 2002
                                                                    o *-  *-
                                     Time (seconds)
                                        4-139

-------
Regulatory Impact Analysis
                                      Figure 4.2-12
              Average Speed    Average Load
                                                              March EPA Cycle
                                                              JRC NRTC
    100
                               Figure 4.2-13
                    Average Speed Changes of the EPA NRTC

                          Speed Changes EPA March NRC
                              Time (seconds)
                                 4-140

-------
Technologies and Test Procedures for Low-Emission Engines
                4-141

-------
Regulatory Impact Analysis
                                  Figure 4.2-14
                    Average Speed Changes of JRC Modified NRTC
                              JRC_NRC Speed Changes
                                    Tim e (seconds)
                                 Figure 4.2-13
                     Average Speed Changes of the EPA NRTC
                                     4-142

-------
  Technologies and Test Procedures for Low-Emission Engines
              Figure 4.2-15
Average Load Changes of JRC Modified NRTC
           JRC_NRC Load Changes
                  Tim e (seconds)
                    4-143

-------
Regulatory Impact Analysis
                                     Figure 4.2-16
                   Average Load Changes of the EPA-Generated NRTC
                                    March EPA_NRC Load Changes
                                              Time (seconds)
   Given the modifications in the duty cycle, it was critical to assess the impact on the emission
signature of the cycle.  Table 4.2-13 shows that the emission signature, based on tests at the
National Vehicle and Fuel Emissions Laboratory and at Southwest Research Institute as of May
2001, were relatively unchanged.
                                        4-144

-------
                     Technologies and Test Procedures for Low-Emission Engines
                                   Table 4.2-13
        Emissions and Cycle-Regression Performance Summary as Presented to
       the Workgroup on June 1, 2001, at the Joint Research Center in Ispera, Italy
Caterpillar 3508
Heavy Duty
850 hp
Sep-OQ
Mar-01
JRC
NOx
Mean Standard Dev.
10.30
10.14
11.198
0.02
0.03
0.03
PM
Mean Standard Dev.
0.20
0.20
0.20
0.004
0.002
0.004
Speed
SE
Mean
79
90
68
Std dev.
1.41
2.12
0.71
M
Mean
1.03
1.01
1.03
R2
Std dev. Mean
0 0.949
0.01 0.939
0.00 0.962
Std dev.
0.001
0.002
0.001
B
Mean
-35
-9
-33
Std dev.
2.83
3.54
1.41

Torque
SE
Mean Std dev.
15 0
15 0
12 0

M
Mean Std dev.
0.8 0
0.83 0.007
0.91 0.007

R2

^^EJ
Mean Std dev. Mean
0.734
0.734
0.765
0.004 184
0.001 188.5
0.001 56


Std dev.
0
3.54
1.41
Power
SE
Mean
14
14
11


Std dev.
0
0
0

«v]
Mean
0.88
0.9
0.95

R2
Std dev. Mean
0 0.801
0 0.804
0 0.823


Std dev.
0.283
0.002
0

El
Mean
29.6
29.5
6.1


Std dev.
0.283
1.273
0.141

Cummins ISB
Medium Duty
^^^^
Sep-00
Mar-01
JRC-Max Spd
JRC-ETC Pk Spd
NOx

Mean Standard Dev.
3.76
3.79
4.06
4.09
0.01
0.03
0.03
0.01
PM

Mean Standard Dev.
0.08
0.08
0.08
0.08
0.001
0.003
0.002
0.009
Speed
SE
Mean
54.7
68
66
50

Std dev.
24.62
18.67
6.22
8.15

M
Mean
0.987
0.98
0.98
0.98

R2
Std dev. Mean
0.011 0.987
0.01 0.982
0.00 0.978
0.00 0.991

Std dev.
0.010
0.008
0.005
0.003

B
Mean
30.0
32
34
37

Std dev.
3.1'
14.48
5.23
6.68

Torque
SE
Mean Std dev.
69.7 2.06
67.5 3.12
43.5 0.14
48.4 2.63

M
Mean Std dev.
0.955 0.011
0.96 0.008
0.981 0.002
0.985 0.00306

R2

B
Mean Std dev. Mean
0.930
0.933
0.960
0.946
0.005 30.0
0.007 26.7
0.001 12.0
0.005 11.6


Std dev.
3.11
2.64
0.354
1.386
Power
SE
Mean
14.8
14.9
9.9
10.0


Std dev.
0.35
0.61
0.21
0.68

M
Mean
0.979
0.981
0.994
0.999

R2
Std dev. Mean
0.009 0.943
0.007 0.943
0.002 0.961
0.002 0.958


Std dev.
0.003
0.005
0.002
0.005

Hi]
Mean
4.5
4.2
1.6
1.6


Std dev.
0.361
0.404
0.141
0.265
   As noted earlier, EPA modified the cycle between September 2000 and March 2001 to
address concerns related to the Arc Welder duty cycle segment of the NRTC.  The modified EPA
version was provided to JRC in early 2001, for its subsequent analysis; however, not knowing
the impact of the changes, all three cycles were tracked until the September 2000 version was
eventually dropped.

   In subsequent data submitted by engine manufacturers through December 5, 2001, the
validity of the cycle from an emission signature and test cell feasibility perspective was
evidenced. Data submitted by Yanmar, Daimler Chrysler, Deere, Caterpillar, and Cummins to
the JRC summary and analysis effort gave clear indication that the duty cycle could be run
across multiple power ranges with good cycle performance results and consistent emission
signature.GG  The cycle performance regression statistics would be defined based on nonroad
engines, rather than adopting the highway performance statistics without review.  The concern
raised by Daimler Chrysler was that the cycle-regression statistics needed to be sufficiently
     Memorandum from Cleophas Jackson to EPA Air Docket A-2001-28, # II-A-170 "JRC December 5, 2001,
Report on Cycle Performance."

                                         4-145

-------
Regulatory Impact Analysis
                                                                                          In
stringent to ensure an accurate and repeatable emission signature was achieved.™1 With the
conclusion of the international workgroup's efforts, EPA considered the cycle to be complete
an effort to facilitate the use of the cycle as a global nonroad transient duty cycle, it has been
introduced into GRPE as a candidate cycle for the global compendium. The ISO procedure
8178-11 is being drafted to  address test cell procedures for exercising an engine over the duty
cycle. New limit values for the cycle performance regression statistics were developed as a part
of this process and may be seen in Table 4.2-1411.
                                        Table 4.2-14
                             NRTC Cycle-Regression Statistics163

Standard Error of
Estimate of Y on X
Slope of the regression
line, m
Coefficient of
determination, r2
Y intercept of the
regression line, b
Soeed [roml
100 rpm
0.95 to 1.03
min 0.970
±50 rpm
Toraue [N-ml
13% of power map
maximum engine torque
0.83-1.03 (hot)
0.77-1.03 (cold)*
min 0.8800 (hot)
min 0.8500 (cold)*
± 20 N-m or ± 2.0% of
max engine torque,
whichever is greater
Power TkWl
8% of power map
maximum
0.89-1.03 (hot)
0.87-1. 03 (cold)a
min 0.9 100 (hot)
min 0.8500 (cold)
± 4 kW or ± 2.0% of max
power, whichever is
greater
    1 Under consideration by ISO workgroup.
4.2.9 Cold-Start Transient Test Procedure

    Nonroad diesel engines typically operate in the field by starting and warming to a point of
stabilized hot operation at least once in a workday. Such "cold-start" conditions may also occur
at other times over the course of the workday, such as after a lunch break. We have observed
that certain test engines, which generally had emission-control technologies for meeting Tier 2 or
Tier 3  standards, had elevated emission levels for about 10 minutes after  starting from a cold
condition.  The extent and duration of increased cold-start emissions will likely be affected by
changing technology for meeting Tier 4 standards, but there is no reason  to believe that this
effect will lessen. In fact, cold-start concerns are especially pronounced for engines with
catalytic devices for controlling exhaust emissions, because many require heating to a "light-off
    ™ Memorandum from Cleophas Jackson to EPA Air Docket A-2001-28, ######Nonroad Transient Duty Cycle
Development Report, Cornetti, G., Hummel, R., and Jackson, C.

    11 The deletion point criteria for engine manufacturers to use in deriving these cycle performance statistics may
be found in regulations at 40 CFR Part 1039, subpart F and Part 1065.530.  See also cycle performance criteria
discussions in Memorandum from Cleophas Jackson to EPA Air Docket A-2001-28, ###### Nonroad Transient
Duty Cycle Development Report, Cornetti, G., Hummel, R., and Jackson, C. and Memorandum from Matthew Spears
to EPA Air Docket A-2001-28, ##### "Test Point Omission Criteria for Determining Cycle Statistics".
                                           4-146

-------
	Technologies and Test Procedures for Low-Emission Engines

or peak-efficiency temperature to begin working.  EPA's highway engine and vehicle programs,
which increasingly involve such catalytic devices, address this by specifying a test procedure
that first measures emissions with a cold engine, then repeats the test after the engine is warmed
up, weighting emission results from the two tests for a composite emission measurement.

   In the proposal, we described an analytical approach that led to a weighting of 10 percent for
the cold-start test and 90 percent for the hot-start test.  Manufacturers pointed out that their
analysis of the same data led to a weighting of about 4 percent for cold-start testing and that a
high cold-start weighting would affect the feasibility of the proposed emission standards.
Manufacturers also expressed a concern that there would be a big test burden associated with
cold-start testing.

   Unlike steady-state tests, which always start with hot-stabilized engine operation, transient
tests come closer to simulating actual in-use operation, in which engines may start operating
after only a short cool-down (hot-start) or after an extended soak (cold-start).  The new transient
test and manufacturers' expected use of catalytic devices to meet Tier 4 emission standards make
it imperative to address cold-start emissions in the measurement procedure." We are therefore
adopting a test procedure that requires measurement of both cold-start and hot-start emissions
over the transient duty cycle, much like for highway diesel engines.  We acknowledge that
limited data are available to establish an appropriate cold-start weighting. For this final rule, we
are therefore opting to establish a cold-start weighting of 5 percent.  This is based on a typical
scenario of engine operation involving  an overnight soak and a total of seven hours of operation
over the course of a workday. Under this scenario, the 20-minute cold-start portion  constitutes 5
percent of total engine operation for the day. Section 4.1.2.3.3 above addresses the feasibility of
meeting the emission  standards with cold-start testing. Regarding the test burden associated with
cold-start testing, we believe that manufacturers will be able to take steps to minimize the burden
by taking advantage of the provision that would allow for forced cooling to reduce total testing
time.

   We believe the 5-percent weighting is based on a reasonable assessment of typical in-use
operation and it addresses the need to design engines to control  emissions under cold-start
operation. We believe cold-start testing with these weighting factors will be sufficient to require
manufacturers to take steps to minimize emission increases under cold-start conditions.  Once
manufacturers apply technologies and strategies to minimize cold-start emissions, they will be
achieving the greatest degree of emission reductions achievable for those conditions. A higher
weighting factor for cold-start testing will likely not be more effective in achieving  in-use
emission control.

   However, given our interest in  controlling emissions under cold-start conditions and the
relatively small amount of information  available in this area, we intend to revisit the cold-start
weighting factor for transient testing in the future as additional data become available.
    JJNote that the cold-start discussion applies only to engines that are subject to testing with transient test
procedures. For example, this excludes constant-speed engines and all engines over 750 hp.

                                          4-147

-------
Regulatory Impact Analysis
Additionally, as the composite transient test represents a combination of variable-speed and
constant-speed operations, we would consider operating data from both of these types of engines
in evaluating the cold-start weighting.  We will apply the same cold-start weighting, as well,
when we adopt a transient duty cycle specifically for engines certified only for constant-speed
operation.

   The planned data-collection effort will focus on characterizing cold-start operation for
nonroad diesel equipment. The objective will be to reassess, and if necessary, develop a
weighting factor that accounts for the degree of cold-start operation so that in-use engines
effectively control emissions during these conditions.  As we move forward with this
investigation, other interested parties, including the State of California, will be invited to
participate. We are interested in pursuing a joint effort, in consultation with other national
government bodies, to ensure a robust and portable data set that will facilitate common global
technical regulations.  This effort will require consideration of at least the following factors:

   •   What types of equipment will we investigate?
   •   How many units of each equipment type will we instrument?
   •   How do we select individual models that will together provide an accurate cross-section
       of the type of equipment they represent?
   •   When will the program start and how long will it last?
   •   How should we define a cold-start event from the range of in-use operation?

   We expect to complete our further evaluation of the cold-start weighting in the context of the
2007 Technology Review, if not sooner.  In case changes to the regulation are necessary, this
timing will allow enough time for manufacturers to adjust their designs  as needed to meet the
Tier 4 standards.

4.2.10 Applicability of Component Cycles to Nonroad Diesel Market

   In the 1997-1998 time frame, we started to pursue  application-specific operating duty cycles
that could be normalized for laboratory testing of nonroad diesel engines. With a standardized
set of operating duty cycles, we would have a basis upon which to compare the brake-specific
emission rates of nonroad engines, both within and across power categories, or bands. These
cycles became the component cycles of the NRTC cycle. The choice of the seven nonroad
component application duty cycles was based on the frequency of finding engines of that
particular mode of operation in the nonroad population and summing those with
engines/equipment doing related work.  Agricultural tractors were seen to have operations
generally similar to combines and off-highway trucks in addition to tractors.  Arc welders
represented the broad group of constant-speed applications. The backhoe-loader group included
most of the lawn/garden/commercial turf tractors, commercial lifts and sweepers.  The
crawler/dozer application matched with other dozer, grader and scraper applications. Rubber-
tire loaders were found to be similar to industrial and rough terrain forklifts, aircraft support and
forestry equipment.  Skidsteer loaders were seen, at the time, as a unique application/category.
Finally, excavators and cranes were grouped together as similar applications. In time, the seven
base nonroad equipment applications, agricultural tractor, arc welder, backhoe loader,

                                          4-148

-------
                     Technologies and Test Procedures for Low-Emission Engines
crawler-dozer, excavator, rubber-tire loader and skidsteer loader were characterized for their
daily operations and engine duty cycles were constructed for each type of work.

   4.2.10.1 Market Representation of Component Cycles

   The determination of which cycles best represent the nonroad equipment population in the
United States was aided by an analysis of the our nonroad equipment population database.164
Our source of data placed the total 1995 nonroad equipment population figure at 7,100,113 units
in the United States. The population broke out into at least 59 different equipment applications,
or specific work categories.  Agricultural tractors held the largest percentage by far at
approximately 34% of units. Constant-speed applications like generating sets, A/C and
refrigeration units comprised a further 14%. Of the remaining pieces of the nonroad equipment,
another 11% of the total population were constant-speed engines like welders, air compressors
and irrigation rigs. Commercial lawn and garden equipment made up an additional 7.5% of all
units, with combines, backhoe and skidsteer loaders at 12%, each application adding a further
4% to the total population.  In the approximately 20% of units remaining, rubber-tire loaders and
crawler-dozers constituted 6% of all nonroad units, each contributing 3% to the nonroad
population. Excavators and cranes comprised a little more than 2% of the total equipment
population. The seven component application classesalone covered 51% of all nonroad
equipment units.  When "related" nonroad applications were grouped with the original seven
applications, over 95% of the nonroad equipment population was represented by the component
applications.

   4.2.10.2 Inventory Impact  of Equipment Component Cycles

   When EPA created an emission distribution from its database according to a list of the seven
nonroad applications used to create the NRTC duty cycle, those seven base  applications
accounted for  59 percent of regulated nonroad engine emissions (see Table 4.2-16).

                                      Table 4.2-16
                   Emissions Attributable to Base Nonroad Applications
Application
Ag tractor
Welder
Backhoe/loader
Crawler
Excavator
R/T Loader
Skid/steer
Total
Emission Distribution by
Application
34%
1%
6%
7%
3%
6%
2%
59%
                                         4-149

-------
Regulatory Impact Analysis
   4.2.10.3 Power and Sales Analysis

   The nonroad equipment market is broad and varies in both range of power available and
application, or intended use, of each piece of equipment. EPA's database was the source for the
distribution of nonroad applications between the various engine power bands. Agricultural
tractors, while accounting for fully a third of the nonroad equipment population, are built
generally to smaller engine displacement specifications and so constituted only 20% of total
nonroad power. With similar equipment applications included, the equipment equipped with
engines that have power or displacement similar to that of agricultural tractors approaches 30
percent. Backhoe loaders, crawler dozers and rubber-tire loaders together accounted for 12
percent of the total power in the nonroad population and, with similar applications included,
accounted for approximately 35 percent of total nonroad power. The last three cycle component
applications—excavators, skidsteer loaders and arc welders, with arc welders and like equipment
generally falling under 50 hp—constitute only 8 percent of total nonroad power.  However,
because small constant-speed engines exist in numerous applications, they also constitute a large
number of discrete units in the nonroad population.  This helps to explain their relatively large
contribution (18%) as a group of similar applications to total nonroad power. Taking the sum of
power represented by all applications similar to the seven component equipment applications
found in the NRTC cycle, we have represented equipment operations and engine  displacements
and, by analogy, in-use operations of 91%  of nonroad equipment units.

   4.2.10.4 Broad Application Control

   Aggregating all those equipment classifications whose operating characteristics were similar
to the seven NRTC component cycles for their emission contributions, we found that the
composite nonroad cycle covered emissions from almost 96% of the documented applications in
the nonroad equipment population (see Table 4.2-17).
                                         4-150

-------
                     Technologies and Test Procedures for Low-Emission Engines
                                     Table 4.2-17
               Similarities Among Various Nonroad Equipment Applications
Application
Ag tractor
Welder
Backhoe/loader
Crawler
Excavator
R/T Loader
Skid/steer
Total
Other Applications with
Similar Operating Characteristics
Combine Off-Hwy Truck
Off-Hwy Tractor
Air Compressors Irrigation Sets
Gas Compressors Leaf Bio w/Vacs
Generators Lt Plants/Signal
Pumps Board
Bore/Drill Rigs Oil Fid Equip.
Cement Mixers Plate Compactors
Chippers/Grinders Pressure Washers
Concrete/Ind. Saw Refrigeration/AC
Crush/Proc. Equip Shredder
Hydr. Power Unit
Aerial Lifts Lawn/Grdn. Tractor
Comm. Turf Rear Eng. Rider
Scrub/Sweeper Specialty carts
Front Mowers Terminal Tractor
Graders Scrapers
R/T Dozer Trenchers
Cranes
Aircraft Support Rough Trn Fork.
Forest Equip
Forklifts


Emission
Distribution
38.4%
25.2%
13.5%
5.7%
2.4%
6.7%
3.6%
95.5%
Cycle
characterization
Heavy -load operation along
governor/lug curve
Transient loads at tightly
governed rated speeds
Widely varying loads and
speeds, weighted toward
lighter operation; most like
highway operation
Widely varying loads and
speeds, weighted toward
heavier operation
Transient loads at loosely
governed rated speed
Stop and go driving with
widely varying loads.
Widely varying loads at
different nominally
constant-speed points

4.2.11 Final Certification Cycle Selection Process

   Figure 4.2-18 outlines the process by which a manufacturer of a particular nonroad diesel
engine might approach certification using the nonroad transient and steady-state test
requirements (NTE certification requirements have been deliberately omitted from this
discussion to simplify the presentation).
                                        4-151

-------
Regulatory Impact Analysis
                                             Figure 4.2-18
                    NR Diesel Engine Transient and Steady-State Testing Requirements
                             Nonroad
                           Manf Seeks
                           Engine Duty
                              Cycle
                           Certification,
   NRTC Cycle
  (NR Transient
 Composite) with
Cold Start porttion
                                                   Ramped Modal
                                                  Cycle option for
                                                   6-mode cycle,
                                                      "6-2"
                        Option for 19 kW
                            or less	Option for 19 kW
                                           or less



Six (6)
Mode Cycle
"6-2"
Weighting




Ramped Modal
Cycle option
for 8-mode
cycle,"C-l"

                                                                               Ramped Modal
                                                                              Cycle option for
                                                                               5-mode cycle,
                                                                                  "0-2"
                                         TRU Outy Cycle
                                          (option for
                                         transportation
                                        refrigeration units
                                             only)














Five (5)
Mode Cycle
"0-2"
Weighting
4.3  Steady-State Testing

   Recognizing the variety of both power classes and work applications to be found within the
nonroad vehicle and engine population, EPA will retain current Federal steady-state test
procedures for nonroad engines. The steady-state duty cycle applicable in each of the following
                                             4-152

-------
	Technologies and Test Procedures for Low-Emission Engines

categories: 1) nonroad engines 25 hp and greater; 2) nonroad engines less than 25 hp; and 3)
nonroad engines having constant-speed, variable-load applications, (e.g., generator sets) will
remain, respectively, the 8-mode cycle, the 6-mode cycle, and the 5-mode cycled
Manufacturers are required to meet emission standards under steady-state conditions in addition
to meeting any emission standards under transient test cycle requirements.  Steady-state test
cycles are needed so that testing for certification will reflect the broad range of operating
conditions experienced by these engines. A steady-state test cycle represents an important type
of modern engine operation, in power and speed ranges that are typical in-use. The mid-to-high
speeds and loads represented by present steady-state testing requirements are the speeds and
loads at which these engines are designed to operate for extended periods for maximum
efficiency and durability.  Manufacturers would perform each steady-state test following all
applicable test procedures detailed in regulations at 40 CFR Part 1039, subpart F,  e.g.,
procedures for engine warm-up and exhaust emissions measurement.  The testing must be
conducted with all emission-related engine control variables in the maximum NOx-producing
condition which could be encountered for a 30 second or longer averaging period at a given test
point. Details concerning the three steady-state procedures for nonroad engines and equipment
can be found in regulations at 40 CFR  1039.505 and in Appendices I-III to Section 1039 which
follow that section, one for each cycle.

4.3.1 Ramped Modal Cycle

    4.3.1.1 Introduction and Background

    In response to manufacturers' concerns for the potential of some PM trap-equipped diesel
engines to exhibit highly variable emissions under current emission test cycles, EPA has
developed ramped modal versions of its steady-state certification duty cycles. These ramped
modal cycle emission tests will reliably and consistently report steady-state emissions from PM
trap and other emission control hardware-equipped nonroad engines.

    For all the laboratory- based steady-state testing currently  specified in 40 CFR Part 89, EPA
has determined that any certification steady-state test cycle may be run as a ramped modal cycle
(RMC). A RMC consists of the same series of steady-state test modes, but they are connected to
one another by gradual ramps in engine speed and/or torque. However, the mode order is
rearranged so as to alternate between high- and low-torque modes.  In a RMC, the steady-state
modes are connected with linear speed and torque transitions.  The difference is that these
transitions are sampled as part of the test.  In other words, emissions sampling would start at the
beginning of a RMC and would not stop until the last mode of the cycle is completed.

    Instead of using weighting factors for each steady-state mode, a RMC specifies different time
durations for each mode.  Time durations are proportioned to weight each mode and transition to
reflect the exact original ISO steady-state test weighting factors upon which the certification
         three certification steady-state test cycles are similar to test cycles found in International Standard ISO
8178-4:1996 (E) and remain consistent with the existing 40 CFR Part 89 steady state duty cycles.

                                          4-153

-------
Regulatory Impact Analysis
testing is based.  The information and test cycle tables needed to run a certification steady-state
test cycle as a RMC are given in 40 CFR Section 1039.505(a)(2). Refer to 40 CFR Part 1039,
subpart F for the procedures required for transforming and running a particular test cycle on a
specific engine.

   Because a RMC weights individual modes by the amount of time spent at each mode, we
considered the effect of a RMC's total test time on emissions. Based on the RMC data presented
in this section, we concluded that if insufficient time was spent in an individual mode, the mode
would not adequately represent the steady-state condition that was intended. This effect was
exaggerated when engines with aftertreatment systems were tested.  By inspecting data from
individual modes, we determined that emissions differences between a RMC and its respective
certification steady-state test cycle occurred primarily when exhaust temperatures between the
two cycles differed greatly.

   As mentioned earlier in this section, the modes in the RMC are intentionally arranged to
alternate between high- and low-torque modes.  This results in more moderate and repeatable
aftertreatment temperatures overall. However, in some cases, more time in  certain modes would
have helped to achieve exhaust temperatures over a RMC that were more representative of
exhaust temperatures for typical steady-state cycles.

   The appropriate total time for the RMC was in part determined from testing of a diesel
engine equipped with both a NOX adsorption catalyst and PM trap exhaust emission controls,
which will be described in this section. Based on the number of modes in a given steady-state
cycle, we determined that twenty minutes is an appropriate total time for a RMC that has five or
fewer steady-state modes. Twenty minutes is also an appropriate minimum time for collecting
an adequate PM sample from an engine certified to a PM standard less than 0.05 g/kW-hr. For
which has six to ten modes, thirty minutes is an appropriate total time.  Thirty minutes ensures
that the lightly weighted modes on the RMC have adequate time to approach the same exhaust
temperatures achieved when the test is run as a steady-state test.  For a RMC with ten to fourteen
modes, forty minutes is an appropriate total test time.  A forty-minute length ensures that a
sufficient amount of the total test time was spent at steady-state rather than in transition from one
mode to the next. For all of the RMCs, these times ensure that less than 10% of the total time is
spent in transition from one steady-state mode to the next.

   There are a number of advantages to running a steady-state test as a RMC.  The current
procedure for conducting a steady-state test allows emission sampling periods as  short as the last
minute of each mode.165  Discrete aftertreatment regeneration events, NOX and SOX regeneration
for NOX adsorption catalysts, forced PM regeneration for PM traps, etc., typically cause
short-duration  sharp increases in NOX, HC and PM emissions. Thus, it may be challenging to
gather good, repeatable emissions from the current steady-state procedures since a regeneration
event may or may not be sampled in a given mode. For sampling low concentrations  of PM, this
inconsistency is exaggerated because the short sample time per mode may not provide enough
PM sample to weigh in a repeatable way. Furthermore, without specific start and  stop times for
sampling each mode, an anticipated regeneration event may be intentionally or unintentionally
                                         4-154

-------
	Technologies and Test Procedures for Low-Emission Engines

included or excluded. With a RMC, this variability is removed by requiring emissions sampling
over the entire cycle.

   There are other advantages to running a steady-state test as a RMC.  The RMC reduces the
number of sampling system starts and stops.  This is significant at low emission standards when
considering that a previous mode's emissions may be incorrectly included in the next mode due
to an unavoidable dead volume in a sampling system.  The longer sampling period of a RMC
also increases the mass of the PM sampled. This is extremely significant because the PM
standard already  approaches the minimum detection limits for many current PM microbalances.

   The RMC also enables the use of batch sampling systems, such as bag samplers. This is an
advantage because batch sampling systems are generally capable of quantifying lower levels of
pollutants with less uncertainty than continuous sampling systems at low emission
concentrations. This may be due to:

1.  Gas analyzer zero-drift over time can be a much larger percentage of the measured value for
   continuous measurements at continuous low average emission concentrations. This is much
   less of an issue with batch measurements at low concentrations, since they can conduct a
   zero and span operation immediately preceding the concentration measurement.

2.  Zero-drift and transient response of the NOX analyzer from engines using high-capacity
   NOx-adsorption catalysts can be a significant challenge for continuous measurement systems.
   For some modes of operation, NOX emissions are truly at, or very close to, zero during
   adsorption with a rapid spike in NOX emissions during regeneration.  Covering the full
   dynamic range requires:

   a.  automatic range switching to allow measurement on a low-concentration analyzer range
       when NOX is near zero during adsorption and switching to a higher range to catch the
       NOX  spike during regeneration, accepting the uncertainty introduced from loss of data
       during the short duration needed to accomplish range switching; or

   b.  operating on a single higher concentration analyzer range and accepting the uncertainty
       and increased zero-drift introduced at low concentrations during adsorption; or

   c.  operating on a single lower concentration analyzer range and accepting loss of data that is
       "clipped" when the analyzer signal saturates during regeneration.

   Batch-sampled NOX can be measured using a single analyzer range appropriate for the
   measured concentration and the same  sample can be measured repeatedly over more than one
   range using the same analyzer.  Thus,  repeat measurements may be utilized to ensure an
   accurate measurement at the lowest possible range.

3.  During a continuous measurement, each instantaneous emission concentration measurement
   has a level of uncertainty associated with it that propagates from each collected data point to
   the final  integrated concentration.  By contrast, a batch-sampled emission measurement is

                                         4-155

-------
Regulatory Impact Analysis
   typically a stabilized average of repeated measurements of a near-constant concentration
   within the bag or other grab-sample container.

During EPA testing of the first pre-production prototype light-duty diesel vehicle (Toyota
Avensis D-Cat) with a NOX adsorption catalyst system, continuous and bag-sampled NOX agreed
to within 4% at very low integrated mass concentrations, but the coefficient of variance for the
continuous NOX measurement was approximately four times the coefficient of variance for the
bag-sampled NOX measurement, which was likely due to a combination of the above effects.

   Use of a RMC can also significantly reduce the cost of steady-state testing. Not only is the
per-test cost anticipated to be lower with the RMC, but the lower thermal-load on CVS and
air-handling systems due to less sustained high-load operation during testing may reduce the cost
for construction of test facilities. The RMC can typically be accomplished in much less time,
further reducing total cost.

   4.3.1.2 Comparison of Steady-State vs. RMC Testing

   4.3.1.2.1 Manufacturer's testing

   An engine manufacturer provided paired and unpaired emissions data to EPA comparing the
13-mode highway SET (supplemental emissions test)  to a RMC developed from the highway
SET.  The paired data contain 34-39 test replicates representing 29 light-heavy, medium-heavy,
and heavy-heavy-duty highway engine families in the range of 250 - 500 hp certified to the 2004
model year heavy-duty on-highway emission standards. The engines were not equipped with
exhaust aftertreatment, but were equipped with high-pressure, electronically controlled fuel
injection systems and cooled EGR systems. The unpaired data are for 10 engine families built
from one basic engine platform for a heavy-heavy-duty engine of approximately 15 liters
displacement. The paired data are summarized in Figures 4.3-1 and 4.3-2. The results of a F-test
comparison of the unpaired SET data to the RMC data are presented in table 4.3-1. Emissions
results did not differ significantly  between the SET and the RMC. Further, when comparing the
uncertainty of the RMC to the SET, it met the F-test criteria at a 90% confidence level using the
test equivalency criteria as per an  EPA letter to the Engine Manufacturers Association, dated
December 12, 2002 regarding guidance on test procedures for heavy-duty on-highway and non-
road engines (page 3, item I).166
                                         4-156

-------
                      Technologies and Test Procedures for Low-Emission Engines
 ; 2.5
  1.5
3? 1
  0.5

1 t






D 13 Mode-NOx
Q RMC-NOx
I 95% Conf. Int.
I 90% Conf. Int.

n





1 3-mode SET NOx 1 3-mode
Ramped-Modal-Cycle
                                      NOx
Figure 4.3-1:  A comparison of SET and RMC NOX emissions
based on paired data from 29 engine families certified to a 2.5
g/bhp-hr NOX and 0.1 g/bhp-hr PM standard.
  0.12
 •0.1
.c
a;
VI
.20.08-
• 13 Mode-PM
S RMC-PM
D 13 Mode-HC
U RMC-HC
I 95% Conf. Int.
I 90% Conf. Int.
                13-mode     13-mode    13-mode     13-mode
                SET PM Ramped-Modal-Cycle SET HC Ramped-Modal-Cycle
                             PM                   HC
Figure 4.3-2:  A comparison of SET and RMC PM and total HC
emissions based on paired data from 29 engine families certified to a
2.5 g/bhp-hr NOX and 0.1 g/bhp-hr PM standard.
                                          4-157

-------
Regulatory Impact Analysis
   Table 4.3-1: F-test comparison of the RMC to the SET steady-state
   test. NOX and HC emissions were measured using continuous
   analyzers. Note that the ability to use batch-sampling for NOX and
   HC would further reduce the standard-deviation for the RMC. The
   PM measurement for the SET also used a single, flow-weighted PM
   filter sample. Using one filter-sample per mode would likely have
   further increased the variability in the SET steady-state tests.

Mean Emissions
(SET)
OSET

Mean Emissions
(RMC)
O~RMC
NOY
2.029
0.056

1.931
0.070
PM
0.0754
0.0080

0.078
0.0057
HC
0.072
0.014

0.072
0.013
CO
0.329
0.062

0.372
0.091
CO,
507
13

510
17
                                F-test
17 •
r <)n%'
*RMr-
Pass at 90%
Confidence
Interval?
2.44
1.56
Pass


2.44
0.516
Pass


2.44
0.776
Pass


2.44
2.18
Pass


2.44
1.69
Pass


   4.3.1.2.2 EPA testing over the 8-mode C-l cycle and its RMC derivative (with and without
   exhaust aftertreatment)

   EPA has determined that its 8-mode C-l test cycle (40 CFR Part 89) may be run as a RMC.
The RMC version of this cycle consists of the  same series of eight steady-state test modes but
the modes are connected to one another by linear speed and torque transitions. That is,
emissions sampling would start at the beginning of this RMC and would not stop until its last
"mode" was completed. As well, the mode order from the 8-mode C-l cycle is rearranged in this
RMC to alternate between high-load and low-load modes.  Instead of using weighting factors for
each steady-state mode, the RMC specifies different time durations for each mode.  Time
durations are proportioned to weight each mode exactly as the original C-l weighting factors.
The information needed to run an 8-mode C-l  test cycle as a RMC is given in 40 CFR,
§1039.505.  The procedures required for transforming and  running this test cycle with a specific
engine are found on 40 CFR Part 1039 subpart F.

   To compare the emission levels between a steady-state 8-mode C-l test and the
corresponding RMC test, four engines ranging from 42 to 400 brake-horsepower (bhp) were
tested at Southwest Research Institute (SwRI)  and at EPA's National Vehicle and Fuel Emissions
Laboratory  (NVFEL).  Table 4.3-2 below contains a summary of the specifications of these
                                         4-158

-------
	Technologies and Test Procedures for Low-Emission Engines

engines. The testing was performed with engines having various exhaust configurations. The
Yanmar engine had no exhaust aftertreatment while the Kubota engine was tested both with and
without a DOC. The DDC engine was tested with a continuously-regenerating trap (CRT)
system that used a platinum-catalyzed DOC located upstream of a non-catalyzed PM trap.

                             Table 4.3-2: Engine properties
Engine
Yanmar 4TNE84
Kubota VI 903E
DDC Series 60
Cummins ISB
Model
Year
2002
2001
1998
2000
Power
(bhp)
48
42
400
180
Fuel
Inj.
DI
IDI
DI
DI
Displ.
(L)
1.99
1.9
12.7
5.9
Air Induction
Naturally
Aspirated
Naturally
Aspirated
Turbocharged
Turbocharged
Configurations
tested
No exhaust
aftertreatment
With and without
DOC
With CRT
(passive
regeneration)
With CDPF + NOX
adsorption catalyst
system
The Cummins ISB engine was tested with a system which combined a catalyzed diesel
particulate filter (CDPF) with a NOx adsorption catalyst.167 The engine was also equipped with a
high-pressure common-rail fuel injection system and cooled low-pressure-loop EGR.. The test
configuration of the ISB engine was that of a 180 b-hp rated nonroad engine and EPA developed
the engine's test calibration values.

   Table 4.3-3, below, summarizes the engine operating conditions for the 8-mode C-l cycle
and for the RMC derived from that cycle. The RMC contains a "split idle mode" (the idle
condition occurs twice versus once in the 8-mode C-l).  Note also that it is possible to run the 8-
mode C-l cycle with different lengths of time-in-mode.  A period of five-minutes duration per
steady-state mode is allowable under current regulations in 40 CFR Part 89 and there is no limit
on maximum time-in-mode.  Different  exhaust sampling periods are  also allowed, having a
minimum length of 60 seconds and no maximum length. Thus, for the 8-mode C-l  steady-state
cycle, the minimum time-in-mode under current regulations would be a period of four minutes of
stabilization with one minute of sampling per mode.  The maximum time for stabilization and
sampling are left undefined.

   All of the engines were tested using a twenty minutes long RMC derived from the 8-mode
C-l cycle. The EPA-modified Cummins ISB was also tested using a thirty minutes long RMC
cycle. The length of time spent in each mode for the 8-mode C-l test cycles varied by engine.
The Yanmar and Kubota  engines were  tested over the 8-mode C-l test cycle at mode lengths of
                                        4-159

-------
Regulatory Impact Analysis
ten minutes each.  Gaseous emissions and PM emissions were sampled for the last five minutes
of each ten-minute mode.  The DDC and the modified Cummins ISB engines were tested over
the 8-mode C-l cycle at mode lengths totaling ten minutes each. Their gaseous and PM
emissions were sampled for the last three minutes of each ten-minute mode. The modified
Cummins ISB engine was also tested using a five minutes long mode length over the 8-mode C-l
cycle. For those tests having a five minutes long mode length, the first four minutes were used
for stabilization and the last minute was used for emissions sampling to model the minimum time
specifications found in 40 CFR Part 89.

                                Table 4.3-3:
    Engine operating conditions for the steady-state 8-Mode C-l and RMC tests
8-Mode
C-l
Speed
Torque
1
2
3
4
Rated
100
75
50
10
5
6
7
Intermediate
100
75
50
8
Idle
No
load

RMC
Speed
Torque
1
Idle
No
load
2
3
4
Intermediate
100
50
75
5
6
7
8
Rated
100
10
75
50
9
Idle
No
load
   Figures 4.3-3 and 4.3-4 below summarize the emissions results obtained from emission
testing on the DDC Series-60 engine. However, due to the use of a non-standard PM sampling
medium and measurement inconsistencies associated with filter handling during emission
testing, PM data are not available for these tests (PM mass loss was attributed to physical
damage to the sample filters after installation into the sampling cassettes). As shown in these
figures, NOX emissions for both engine-out and CRT-out configurations of this engine over the
RMC and 8-mode C-l test cycles do not differ at the 95% confidence interval.  Differences
between HC and CO emission levels over the two cycles were either negligible or extremely low
during all testing and well under the Tier 4 emission standards.
                                        4-160

-------
                     Technologies and Test Procedures for Low-Emission Engines

            NOx (g/hp-hr)
HC (g/hp-hr)
CO (g/hp-hr)
Figure 4.3-3: Emissions from the DDC Series-60 engine over the steady-state
mode C-l test and the 20-minute RMC test with no exhaust aftertreatment.
            NOx (g/hp-hr)            HC (g/hp-hr)             CO (g/hp-hr)
Figure 4.3-4: Emissions from the DDC Series-60 engine over the steady-state
mode C-l test and the 20-minute RMC test with a CRT.
                                         4-161

-------
Regulatory Impact Analysis
   Figures 4.3-5 and 4.3-6 compare exhaust emissions from the Kubota V1903E engine over
both the 8-mode C-l and RMC cycles without and fitted with a DOC, respectively.  PM
emissions over both test cycles from both of the tested engine configurations did not differ at
either the 95% or 90% confidence interval.  There was however a general trend toward a reduced
coefficient of variance for RMC versus 8-mode C-l PM  emissions and the number of replicates
was insufficient for a rigorous F-test comparison of variance. Differences in mean NOX
emissions in the no exhaust aftertreatment configuration  were small, and did not differ at a 95%
confidence interval, but did differ at a 90% confidence interval.  CO emissions were somewhat
lower over the RMC, possibly due to increased CO oxidation caused by  the somewhat higher
exhaust temperatures of that cycle compared to the 8-mode C-l cycle. In both cases, though, CO
emissions were less than  50% of the Tier 4 standard.
          NOx (g/hp-hr)
HC (g/hp-hr)
CO (g/hp-hr)
PM (gtip-hn
Figure 4.3-5: Emissions for the Kubota V1903E engine with no exhaust
aftertreatment over the steady-state 8-mode C-l test and the 20-minute RMC test.
                                         4-162

-------
                     Technologies and Test Procedures for Low-Emission Engines
          NOx(g/hp-hr)
HC (g/hp-hrt
                                                            r
CO (g/hp-hr)
PM (g/hp-hr:
Figure 4.3-6: DOC-out emission levels obtained from the Kubota V1903E engine
over the steady-state 8-mode C-l test and the 20-minute RMC test.
          NOx (
                          HC (gflip-hr)
                CO ig/hp-hn
                PM (g/hp-hr)
Figure 4.3-7: Engine-out emission levels obtained from the Yanmar 4TNE84
engine over the steady-state 8-mode C-l test and the 20-minute RMC test.
                                         4-163

-------
Regulatory Impact Analysis
   Emissions from the Yanmar 4TNE84 engine operating without exhaust aftertreatment over
both the 8-mode C-l and 20 minutes long RMC test cycles are summarized above in Figure
4.3-7. As can be seen in this figure, the average engine-out NOX emission over the RMC is
within the 95% confidence interval of the NOX data gathered over the steady-state 8-mode  C-l
test,  although the number of test replicates were insufficient to determine a confidence interval
for the RMC for this particular data comparison. With regards to the HC and CO emissions, the
data  showed a slight, statistically significant difference for these emissions of 5% or less
between the two cycles. CO emissions exceeded the Tier 1 standards over the 8-mode C-l cycle
and were unusually high for a diesel engine over both  of these cycles.  This may indicate that a
mechanical problem exists with this particular test engine.
    035
    0.30
    0.25
    0.20
  .a 015
    0 10
    0.05
    0.00
D RMC (20-minutes total)
D RMC (30-minutes total)
DC-1 (5-minute modes)
• C-1 (10-minute mode)
I 05% Confidence Interval

I 90% Confidence Interval
               NOx
                               NMHC
                                                 CO
                                                                 PM
Figure 4.3-8: Emissions from the EPA-modified Cummins ISB engine over the
steady-state 8-mode C-l test cycle and the RMC test cycle. Note that the data
represent the 2 different mode lengths specified for the 8-mode C-l and two
different total test times for the RMC.  PM emissions were only available for the
10 minutes per mode 8-mode C-l and the 20 minutes long RMC results.  Results
are shown for mean-emissions calculated for 7 test replicates. Confidence
intervals were calculated using a 2-sided Student t-test.
   Figure 4.3-8 above compares the emission levels obtained from testing the Cummins ISB
engine on the 8-mode C-l cycle at both five minutes per mode and ten minutes per mode, as
described in 40 CFR Part 89, with the RMC version of that cycle at both twenty minutes and
thirty minutes of total cycle time. The five minutes and ten minutes per mode represent mode
lengths that are currently used in the 8-mode C-l test cycle for emissions testing of nonroad
                                          4-164

-------
	Technologies and Test Procedures for Low-Emission Engines

diesel engines. PM emissions were measured only for the ten minutes per mode 8-mode C-l and
the twenty minutes total time RMC cycles, which among the cycles investigated represented the
largest differences in exhaust temperature and gaseous emissions.  PM emissions were extremely
low due to the use of the CDPF and were approximately 50% of the Tier 4 standards. Mean PM
emissions for these two cycles did not differ at either a 95% or a 90% confidence level. Some
statistically significant differences in mean NOX emissions were found between the various
cycles used, including the two different mode length 8-mode C-l cycles, due to differences in
exhaust temperatures achieved over individual cycles. There were statistically significant
differences in mean NOX emissions between the ten minutes per mode and the five minutes per
mode 8-mode C-l cycles. Likewise,  there were statistically significant differences in mean NOX
emissions between the ten minutes per mode 8-mode C-l and both the twenty and thirty minutes
total time RMCs. Differences in mean NOX emissions between the thirty minutes total time
RMC and the five minutes per mode  8-mode C-l were not statistically significant at either a 95%
or a 90% confidence level. All other emission levels were extremely low over both the RMC
and the 8-mode C-l tests. Mean HC, CO and PM emissions did not differ significantly at either
a 95% or a 90% confidence level for  either of these cycles.

   4.3.1.2.3 Summary of engine test results

   These data confirm that emissions from engines which do not use NOX adsorption catalysts
are relatively insensitive to the choice of the 8-mode C-l test cycle or its RMC counterpart.
Neither are these engine emissions sensitive to the impact of time  spent at any  steady-state
speed-load set-point.  However, the effect of test cycle length and time-in-mode on exhaust
temperatures did have an impact on NOX emissions when an engine was equipped with a NOX
adsorption catalyst system, due to the:

1.  effect of catalyst temperature on the ability to oxidize NO-to-NO2 for NOX  storage
   (kinetically-limited at low temperatures and equilibrium-limited at high temperatures);

2.  effect of thermal-desorption of NOX at high temperatures; and

   difficulty in effectively vaporizing fuel reductant at very low exhaust temperatures.

   Based on NOX emissions and engine exhaust temperature data  from EPA tests of the
modified Cummins ISB engine, a thirty minutes total time 8-mode C-l-based RMC was selected
as comparable to the five minutes per mode 8-mode C-l test cycle for NOX emission and engine
exhaust temperature results.  Furthermore, based on the results of both EPA and engine
manufacturer testing, the Agency has determined that steady-state test procedures should be
modified to include changes necessary to allow repeatable NOX emission results for all steady-
state testing conducted on engines having catalytic exhaust emission controls for NOX emissions.
Steady-state testing for certification will be conducted in the following manner:

1.  The manufacturer may choose either the appropriate laboratory-based certification
   steady-state test cycle or its RMC derivative as found in regulations at 40 CFR, Section
   1039.505. For RMC tests with five or fewer modes, the length of the RMC test cycle will be

                                         4-165


-------
Regulatory Impact Analysis
   twenty-minutes long.  For RMC tests with six to nine modes, the length of that test cycle will
   be thirty-minutes long. For RMC tests with ten or more modes, the length of that test cycle
   will be forty-minutes long.

2.  When testing an engine having an exhaust aftertreatment system which reduces NOx
   emissions, a manufacturer will operate that engine for four to six minutes, then sample
   emissions for one to three minutes in each mode. The sampling time for PM emissions may
   be extended to improve measurement accuracy, using good engineering judgment. If a
   longer sampling time is chosen for PM emissions, the manufacturer must calculate and
   validate cycle performance statistics for the gaseous and PM sampling periods separately.

3.  When testing other engines, a manufacturer will operate those engines for at least five
   minutes, then sample emissions for at least one minute in each mode.

These changes in measurement procedures for nonroad engines have been incorporated into
regulations at 40 CFR Section 1039.505.

4.3.2 Transportation Refrigeration Unit Test Cycle

   Transportation refrigeration units (TRU), a specific application of steady-state engine
operation, are refrigeration systems powered by diesel engines designed to refrigerate perishable
products that are transported in various containers, including semi-trailers, truck vans, shipping
containers, and rail cars. TRU engines are relatively small with most units ranging from 7 to 38
kW (10 to 50 horsepower)LL.

   Engines that are designated as TRU engines at the time of certification are expected to
operate in the field primarily under steady-state conditions.  These engines may from time to
time be subject to minor setpoint performance perturbations; however those changes are not
expected to last for a total duration at any one point of greater than 30 seconds and are not
multiple, highly transient, repetitive changes in  speed or load such as seen in the nonroad
transient duty cycle. These parameters appropriately characterize TRU equipment operation
independent of unit application, whether used for fresh product (chilled to 3°C) or for frozen
goods at the standard maximum rating (-20°C).  So, to that end, EPA has adopted a four-mode
steady-state test cycle designed specifically for engines used in TRU applications.

   The TRU certification test cycle consists of four steady-state modes of operation. Two
modes are to be run at 50% of the manufacturer's declared peak torque value for that engine.
The remaining two modes are to be run at 75% of that same declared peak torque value for that
same engine. One of the modes at 50% load is to be run at the engine manufacturer's speed at
peak, or rated, power,  while the other mode at 50% load  is to be run at the engine manufacturer's
   LL Information on the proposed TRU cycle may be found on and downloaded from the CARD website at
http://www.arb.ca.gov/diesel/dieselrrp.htm.  In particular, see the Technical Bulletin to the Proposed TRU cycle
determination.

                                         4-166

-------
	Technologies and Test Procedures for Low-Emission Engines

"intermediate" test speed. Likewise, one of the modes at 75% load is to be run at the engine
manufacturer's speed at rated horsepower and the remaining mode at 75% load is to be run at the
manufacturer's "intermediate" test speed. All four modes would be weighted equally in
determining a particular mode's contribution to the engine's total test cycle emissions. Early data
submissions in response to California-ARB's call for TRU engine operating data showed that the
majority of TRU engines operated in-use in at least three, if not all four of the test cycle's
modes168.  It was equally clear from comments to the rule that a TRU test cycle was more
representative of refrigeration unit operation than the nonroad cycles currently available to
manufacturers  since TRUs generally did not operate at low idle, high idle, peak torque or rated
power conditions.

   EPA will allow manufacturers to test their engines under a broad definition of intermediate
test speed, similar to recommendations found in ISO-8178-4 steady-state test guidelines. The
intermediate speed shall be the declared maximum torque speed if it occurs between 60% and
75% of rated speed / maximum test speed.  If the declared maximum torque speed is less than
60% of rated speed, then the intermediate speed shall be 60% of the rated speed. If the declared
maximum torque speed is greater than 75% of the rated speed then the intermediate speed shall
be 75% of rated speed. This will enable an engine manufacturer to more closely match the TRU
cycle at engine certification to the operation of their engines in-use.

   The set point value for speed in  a TRU engine is expected to remain consistent without
repetitive transient changes on a 1 hertz basis. The magnitude of any changes in actual speed
from the engine are expected to be under 2% which is consistent with the Agency's treatment of
operation as steady state in the creation of the transient duty cycle. Additionally, the set point
value demanded by the application remains within this 2% steady  state definition.  Should
application demands differ from the steady state condition for speed, the engine shall not be
considered an actual steady state TRU engine. The TRU engine is likewise not allowed to drop,
or drift1^1, from a load set point by more than  15% of torque at the speed for a particular
operating mode before changing its load setpoint.

   As seen below in Figure 4.3-9169, the operation of the typical piece of TRU equipment tested
is relatively consistent, as evidenced by the power factor curve, a surrogate for engine response
to load demand on the unit. Many factors may affect unit "drift" from a set point, but that set
point of operation does not deteriorate significantly over longer periods of time, in minutes.
TRU equipment is responsive to feedback from a broad number of engine operating parameters
and user input  options. Operating temperature, intake air temperature, i.e., ambient, and exhaust
air temperature are some engine  operating parameters to which the engine must be responsive.
User or owner  inputs include tolerances around set point temperatures, minimum time-on, and
minimum time-off for the unit.   Likewise the  condition and age of the engine and container,
especially the insulation, may influence ability to hold a desired temperature, or load170,171.
   MMDrift is restricted to load deterioration not to exceed 15% over a sixty minute duration.

                                         4-167

-------
Regulatory Impact Analysis
                                      Figure 4.3-9
                          TRU Equipment Operation at Pull-Down
                            UNIT DATA DURING PULL DOWN TO -22C
     250000
     2000.00
     1500.00
     100000
     50000
                                                                     100.00
                                                                     80.00
                                                                     60.00
                                                                            -ENRPM
                                                                            -power factor
                                                                            -NumCyls
                                                                            -SAT
         000    2000   4000    6000   SO 00    10000   12000  14000   16000   18000
                                  TIME. MINUTES
   The expectation is that the engine is governed in such a way that this demand is not possible.
If that engine is deemed a steady-state TRU engine at the time of certification, the application
within which the engine is sold, must meet these standards of operation.

   As an additional way of ensuring that TRU certification is limited to those engines for which
it is warranted, we are adding a requirement that any TRU-certified engine must meet
appropriate NTE standards for  any in-use operation.  Practically, this means that TRU engines
are subject to NTE standards based on the normal operation that these engines would experience
in the field.  This is limited neither to later model years nor to any particular range of engine
speeds and loads.  If TRU engine operation is limited as much as manufacturers have described,
the resulting "NTE zone" should be practically limited to a narrow range of speeds and loads
very close to those points represented by the specified duty cycle.
                                          4-168

-------
	Technologies and Test Procedures for Low-Emission Engines

4.4 Not-to-Exceed  Testing

   The Agency's examination of emissions data from heavy duty highway diesel engines , and the
confidential discussions with several heavy duty  diesel engine manufacturers, led EPA  to the
conclusion that the 1.25 emission cap associated with the not-to-exceed zone  requirement is
technologically feasible. This conclusion has not changed since the initiation of the not-to-exceed
concept. The Agency believes the 1.25 factor proposed for the not-to-exceed standard provides
sufficient  room to allow for the uneven nature of the emission maps.  For these reasons, EPA
believes the primary technologies discussed earlier in this chapter will provide the necessary
NMHC+NOx and PM control on the existing steady state, as well as on the transient cycle testing
and not-to-exceed zone testing.

   The goal of the Not-To-Exceed (NTE) limits on nonroad diesel engines remains consistent with
the reasoning for highway heavy duty diesel engines . The NTE helps ensure that emission benefits
are achieved in-use and provides a practical approach for a post-promulgation in-use testing
program.  The NTE established for the highway heavy duty diesel engines has been demonstrated
to be not only feasible, but practical. The NTE approach provides an area under the maximum
allowable  torque curve  of an engine for which an engine may not exceed a specified value for the
regulated pollutants. The NTE zones, limits, and ambient conditions are described in detail below.

   The advantages to adopting an NTE strategy originally adopted for highway diesel engines are
numerous. These include:

   •   Proven design strategy can be utilized by manufacturers
   •   Development costs can be minimized as new test protocols will not need to be refined
   •   Assurance of comparable control effectiveness analogous to existing programs
   •   Demonstrated effectiveness in the heavy duty highway diesel market can be carried forward
       to the nonroad diesel market
   •   Allows for direct comparison of control effectiveness in-use

   The Not-To-Exceed (NTE) provision was initially finalized for HDDEs certified to the 2004 FTP
emission standards with implementation beginning in model year 2007. (See 65 FR 59896, October
6, 2000.)  The NTE approach establishes an area (the "NTE control area") under the torque curve
of an engine where emissions must not exceed a  specified  value for any of the regulated
pollutants.1"™ The NTE requirements apply under engine operating conditions that could reasonably
be expected to be seen in normal vehicle operation and use which occur during the conditions
specified in the NTE test procedure. (See 40 CFR 86.1370.) This test procedure covers a specific
range of engine operation and ambient  operating conditions  (i.e., temperature, altitude, and
   ^ Torque is a measure of rotational force. The torque curve for an engine is determined by an engine
"mapping" procedure specified in the Code of Federal Regulations. The intent of the mapping procedure is to
determine the maximum available torque at all engine speeds. The torque curve is merely a graphical representation
of the maximum torque across all engine speeds.

                                         4-169

-------
Regulatory Impact Analysis
humidity). The NTE control area, emissions standards, ambient conditions and test procedures for
nonroad diesel engines are described in the 40 CFR 1039.515.

   The NTE provisions for nonroad diesel engines mirror the highway diesel program and so a
manufacturer will need to undertake the engine mapping procedure as defined in 40 CFR 1065;
however, speed definitions will need to be determined based on 40 CFR 86.1360(c).  Valid NTE
compliance evaluation will be based on the following factors:

   •   Operating speeds greater than the speed determined by: nlo + 0.15 x (nw-nlo)
   •   Engine load points greater than or equal to 30% of the maximum torque value produced by
       the engine
   •   Brake Specific Fuel Consumption (BSFC) requirements as specified in 40 CFR 86.1370-
       2007 (b)(3)
   •   Exclusion areas for which the NTE requirement does not apply may be found in 40 CFR
       86.1370-2007 (e.g. PM carve-out zones for engines certifying to a PM standard above 0.07
       g / kW-hr)
   •   Control area limits as defined in 40 CFR86.1370-2007 (d) for averaging times that may or
       may not include discrete regeneration events
   •   Corrections for ambient conditions as defined in 40 CFR 86.1370-2007 (e)
   •   Cold temperature exclusions as adopted in 40  CFR 86.1370-2007 (f)
   •   Engines equipped with NOx and NMHC aftertreatment systems (both single and multi-bed
       systems) with warm-up provisions as defined in 40 CFR 86.1370-2007 (g)

   The NTE requirements will not apply during engine start-up conditions 40 CFR 86.1370-2007
(g).  In addition, with the  application of advanced exhaust emission control devices, an exhaust
emission control device warm-up provision is a necessary criterion for the NTE to address the
impact of thermal inertia on aftertreatment efficiency for the catalytic reduction strategies. Until the
exhaust gas temperature on the outlet side of the exhaust emission control device(s) achieves 250
degrees Celsius, the engine is not subject to the NTE as discussed in  40 CFR 86.1370-2007 (g).
Control of cold-start emissions is expected to happen through the nonroad transient cycle cold-start
provisions.
   For a more detailed technical description of the application of the NTE Zone to diesel engines,
please  see the Regulatory Impact Analysis: Heavy  -Duty Engine and Vehicle Standards and
Highway Diesel Fuel Sulfur Control Requirements EPA420-R-00-026.
                                        4-170

-------
	Technologies and Test Procedures for Low-Emission Engines

Chapter 4 References

1.  Control of Air Pollution from New Motor Vehicles: Heavy-duty Engine and Vehicle
Standards and Highway Diesel Sulfur Control Requirements; Final Rule, 66 FR 5002, January
18,2001.

2.  Highway Diesel Progress Review, United States Environmental Protection Agency, June
2002, EPA 420-R-02-016.  Copy available in Docket OAR-2003-0012-0919.

3.  Highway Diesel Progress Review Report 2, United States Environmental Protection Agency,
March 2004, EPA 420-R-04-004.  Copy available in Docket OAR-2003-0012-0918.

4.  Final Regulatory Impact Analysis: Control of Emissions of Air Pollution from Highway
Heavy-Duty Engines, Air Docket OAR-2003-0012-0949.

5.  Regulatory Impact Analysis: Control of Emissions of Air Pollution from Highway Heavy-
Duty Engines, Air Docket OAR-2003-0012-950.

6.  Final Regulatory Impact Analysis: Control of Emissions from NR Diesel Engines, Air
Docket, OAR-2003-0012-0952.

7.  Nonroad Diesel Emissions Standards Staff Technical Paper, Air Docket OAR-2003-0012-
0951.

8.  Exhaust and Crankcase Emission Factors for Nonroad Engine Modeling - Compression-
Ignition, EPA420-P-02-016, NR-009B, Air Docket A-2001-28.

9.  Onishi, S. et al, "Active Thermo-Atmosphere Combustion (ATAC) - A New Combustion
Process for Internal Combustion Engines," SAE 790840.

10. Najt, P. and Foster, D.  "Compression-Ignited Homogeneous Charge Combustion," March
1983, SAE 830264.

11. Dickey, D. et al, "NOx Control in Heavy-Duty Diesel Engines - What is the Limit,"
February, 1998, SAE 980174.

12. Kimura, S. et al, "Ultra-Clean Combustion Technology Combining a Low-Temperature and
Premixed Combustion Concept for Meeting Future Emission Standards," SAE 2001-01-0200.

13. Kimura, S. et al, "An Experimental Analysis of Low-Temperature and Premixed
Combustion for Simultaneous Reduction of NOx and Particulate Emissions in Direct Injection
Diesel Engines," International Journal of Engine Research, Vol 3 No.4,  pages 249-259, June
2002.

14. Gray, A. and Ryan, T., "Homogenous Charge Compression Ignition (HCCI) of Diesel Fuel,"
May, 1997 SAE 971676.
                                       4-171

-------
Regulatory Impact Analysis
15.  Stanglmaier, R. et al, "HCCI Operation of a Dual-Fuel Natural Gas Engine for Improved
Fuel Efficiency and Ultra-Low NOx Emissions at Low to Moderate Engine Loads," May, 2002
SAE 2001-01-1897.

16.  Stanglmaier, R. and Roberts, C. "Homogenous Charge Compression Ignition (HCCI):
Benefits, Compromises, and Future Engine Applications," SAE 1999-01-3682.

17.  "Demonstration of Advanced Emission Control Technologies Enabling Diesel-Powered
Heavy-Duty Engines to Achieve Low Emission Levels", Manufacturers of Emission Controls
Association, June 1999 Air Docket A-2001-28.

18.  See Table 2-4 in "Nonroad Diesel Emission Standards - Staff Technical Paper", EPA
Publication EPA420-R-01-052, October 2001. Copy available in EPA Air Docket A-2001-28.

19.  "Demonstration of Advanced Emission Control Technologies Enabling Diesel-Powered
Heavy-duty Engines to Achieve Low Emission Levels: Interim Report Number 1 - Oxidation
Catalyst Technology, copy available in EPA Air Docket A-2001-28. "Reduction of Diesel
Exhaust Emissions by Using Oxidation Catalysts," Zelenka et al, SAE Paper 90211, 1990. See
Table 2-4 in "Nonroad Diesel Emission Standards - Staff Technical Paper", EPA Publication
EPA420-R-01-052, October 2001, copy available in EPA Air Docket A-2001-28.

20.  "Demonstration of Advanced Emission Control Technologies Enabling Diesel-Powered
Heavy-Duty Engines to Achieve Low Emission Levels", Manufacturers of Emission Controls
Association, June 1999 Air Docket A-2001-28.

21.  Miller, R. et al, "Design, Development and Performance of a Composite Diesel Particulate
Filter," March 2002, SAE 2002-01-0323.

22.  "Wall Flow Monoliths," DieselNet Technology Guide, http://www.dieselnet.com/
tech/dpf_wallflow.html

23.  "Ceramic Fibers and Cartridges," DieselNet Technology Guide, http://www.dieselnet.com/
tech/dpf_fiber.html

24.  Hori, S. and Narusawa, K. "Fuel Composition Effects on SOF and PAH Exhaust Emissions
from DI Diesel Engines," SAE 980507.

25.  "Demonstration of Advanced Emission Control Technologies Enabling Diesel-Powered
Heavy-Duty Engines to Achieve Low Emission Levels", Manufacturers of Emission Controls
Association, June 1999 Air Docket A-2001-28.

26. "Demonstration of Advanced Emission Control Technologies Enabling Diesel-Powered
Heavy-Duty Engines to Achieve Low Emission Levels", Manufacturers of Emission Controls
Association, June 1999 Air Docket A-2001-28.
                                       4-172

-------
	Technologies and Test Procedures for Low-Emission Engines

27. Hawker, P., et al, Effect of a Continuously Regenerating Diesel Paniculate Filter on Non-
Regulated Emissions and Particle Size Distribution, SAE 980189.

28. Application of Diesel Particulate Filters to Three Nonroad Engines - Interim Report, January
2003. Copy available in EPA Air Docket A-2001-28.

29. "Nonroad Diesel Emission Standards - Staff Technical Paper", EPA Publication EPA420-R-
01-052, October 2001. Copy available in EPA Air Docket A-2001-28.

30. Engelhard DPX catalyzed  diesel particulate filter retrofit verification,
www.epa.gov/otaq/retrofit/techlist-engelhard.htm. a copy of this information is available in Air
Docket A-2001-28.

31. "Particulate Traps for Construction Machines, Properties and Field Experience," 2000, SAE
2000-01-1923.

32. Letter from Dr. Barry Cooper, Johnson Matthey, to Don Kopinski, U.S. EPA, Air Docket A-
2001-28.

33. EPA Recognizes Green Diesel Technology Vehicles at Washington Ceremony, Press
Release from International Truck and Engine Company, July 27, 2001, Air Docket A-2001-28.

34. Nino, S. and Lagarrigue, M. "French Perspective on Diesel Engines and Emissions,"
presentation at the 2002 Diesel Engine Emission Reduction workshop in San Diego, California,
Air Docket A-2001-28.

35. Highway Diesel Progress Review, United States Environmental Protection Agency, June
2002, EPA 420-R-02-016, Air  Docket A-2001-28.

36. "Nonroad Diesel Emissions Standards Staff Technical Paper", EPA420-R-01-052, October
2001, Air Docket A-2001-28.

37. Allansson, et al, European Experience of High Mileage Durability of Continuously
Regenerating Diesel Particulate Filter Technology. SAE 2000-01-0480.

38. LeTavec, Chuck, et al,  "EC-Diesel Technology Validation Program Interim Report," SAE
2000-01-1854; Clark, Nigel N., et al, "Class 8 Trucks Operating On Ultra-Low Sulfur Diesel
With Particulate Filter Systems: Regulated Emissions," SAE 2000-01-2815; Vertin, Keith, et al,
"Class 8 Trucks Operating On Ultra-Low Sulfur Diesel With Particulate Filter Systems: A Fleet
Start-Up Experience," SAE 2000-01-2821.

39. Vertin,  Keith, et al, "Class 8 Trucks Operating On Ultra-Low Sulfur Diesel With Particulate
Filter Systems: A Fleet Start-Up Experience," SAE 2000-01-2821.

40. Allanson, R. et al, "Optimising the Low Temperature Performance and Regeneration
Efficiency of the Continuously Regenerating Diesel Particulate Filer (CR-DPF) System," March
2002, SAE 2002-01-0428.

                                        4-173

-------
Regulatory Impact Analysis
41.  Jeuland, N., et al, "Performances and Durability of DPF (Diesel Paniculate Filter) Tested on
a Fleet of Peugeot 607 Taxis First and Second Test Phases Results," October 2002, SAE 2002-
01-2790.

42. "Regulatory Impact Analysis: Heavy-Duty Engine and Vehicle Standards and Highway
Diesel Fuel Sulfur Control Requirements," EPA420-R-00-026, December 2000, Docket A-2001-
28, Document No. II-A-01.

43.  Koichiro Nakatani, Shinya Hirota, Shinichi Takeshima, Kazuhiro Itoh, Toshiaki Tanaka, and
Kazuhiko Dohmae, "Simultaneous PM and NOx Reduction System for Diesel Engines.", SAE
2002-01-0957, SAE Congress March 2002.

44.  Allanson, R. et al, "Optimising the Low Temperature Performance and Regeneration
Efficiency of the Continuously Regenerating Diesel Particulate Filer (CR-DPF) System," March
2002, SAE 2002-01-0428.

45.  Flynn, P. et al, "Minimum Engine Flame Temperature Impacts on Diesel and Spark-Ignition
Engine NOx Production," SAE 2000-01-1177, March 2000.

46.  Regulatory Impact Analysis: Control of Emissions of Air Pollution from Highway Heavy-
Duty Engines, Air Docket OAR-2003-0012-950.

47.  Final Regulatory Impact Analysis: Control of Emissions from NR Diesel Engines, Air
Docket,  OAR-2003-0012-0952.

48.  Stanglmaier, Rudolf and Roberts, Charles "Homogenous Charge Compression Ignition
(HCCI): Benefits, Compromises, and Future Engine Applications". SAE 1999-01-3682.

49.  Kimura, Shuji, et al, "Ultra-Clean Combustion Technology Combining a Low-Temperature
and Premixed Combustion Concept for Meeting Future Emission Standards", SAE 2001-01-
0200.

50.  Diesel Emission Control-Sulfur Effects Program, Phase I Interim Data Report No. 1,
August,  1999, www.ott.doe.gov/decse Copy available in Air Docket A-2001-28.

51.  Kawanami, M., et al, Advanced Catalyst Studies of Diesel NOx Reduction for Highway
Trucks,  SAE 950154.

52.  Hakim, N. "NOx Adsorbers for Heavy Duty Truck Engines - Testing and Simulation,"
presentation at Motor Fuels: Effects on Energy Efficiency and Emissions in the Transportation
Sector Joint Meeting of Research Program Sponsored by the U.S. Department of Energy, Clean
Air for Europe and Japan Clean Air, October 9-10,  2002. Copy available in EPA Air Docket A-
2001-28.

53.  Koichiro Nakatani, Shinya Hirota, Shinichi Takeshima, Kazuhiro Itoh, Toshiaki Tanaka, and
Kazuhiko Dohmae, "Simultaneous PM and NOx Reduction System for Diesel Engines.", SAE
                                        4-174

-------
	Technologies and Test Procedures for Low-Emission Engines

2002-01-0957, SAE Congress March 2002.

54.  Schenk, C., McDonald, J. and Olson, B. "High Efficiency NOx and PM Exhaust Emission
Control for Heavy-Duty On-Highway Diesel Engines," SAE 2001-01-1351.

55.  Gregory, D. et al, "Evolution of Lean-NOx Traps on PFI and DISI Lean Burn Vehicles",
SAE 1999-01-3498.

56.  McDonald, J., et al, "Demonstration of Tier 2 Emission Levels for Heavy Light-Duty
Trucks," SAE 2000-01-1957.

57.  Brogan, M, et al, Evaluation of NOx Adsorber Catalysts Systems to Reduce Emissions of
Lean Running Gasoline Engines, SAE 962045.

58.  Gregory, D. et al, "Evolution of Lean-NOx Traps on PFI and DISI Lean Burn Vehicles",
SAE 1999-01-3498.

59.  Sasaki, S., Ito, T., and Iguchi, S., "Smoke-less Rich Combustion by Low Temperature
Oxidation in Diesel Engines," 9th Aachener Kolloquim Fahrzeug - und Motorentechnik 2000.
Copy available in Air Docket A-2001-28.

60.  Whiteacre,  S. et al., "Systems Approach to Meeting EPA 2010 Heavy-Duty Emission
Standards Using a NOx Adsorber Catalyst and Diesel Particle Filter on a 151 Engine," SAE
2004-01-0587, March 2004.

61.  Highway Diesel Progress Review Report 2, United States Environmental Protection Agency,
March 2004, EPA 420-R-04-004, Air Docket OAR-2003-0012-0918.

62.  Brogan, M, et al, Evaluation of NOx Adsorber Catalysts Systems to Reduce Emissions of
Lean Running Gasoline Engines, SAE 962045.

63.  Gregory, D. et al, "Evolution of Lean-NOx Traps on PFI and DISI Lean Burn Vehicles",
SAE 1999-01-3498.

64.  Highway Diesel Progress Review, United States Environmental Protection Agency, June
2002, EPA 420-R-02-016, Air Docket A-2001-28.

65.   Kato, N. et al, "Thick Film ZrO2 NOx Sensor for the Measurement of Low NOx
Concentration," February 1998, SAE 980170.

66.  Kato, N. et al, "Long Term Stable NOx Sensor with Integrated In-Connector Control
Electronics," March 1999, SAE 1999-01-0202.

67.  Sasaki, S., Ito, T., and Iguchi, S., "Smoke-less Rich Combustion by Low Temperature
Oxidation in Diesel Engines," 9th Aachener Kolloquim Fahrzeug - und Motorentechnik 2000.
Copy available in Air Docket A-2001-28.
                                       4-175

-------
Regulatory Impact Analysis
68. Diesel Emission Control - Sulfur Effects (DECSE) Program Phase II Summary Report: NOx
Adsorber Catalysts, October 2000. Copy available in Air Docket A-2001-28.

69. Memo from Byron Bunker to Docket A-99-06, "Estimating Fuel Economy Impacts of NOx
Adsorber De-Sulfurization," December 10, 1999. Copy available in Air Docket A-2001-28.

70. Jobson, E. et al, "Research Results and Progress in LeaNOx II - A Cooperation for Lean
NOx Abatement," SAE 2000-01-2909.

71. Asanuma, T. et al, "Influence of Sulfur Concentration in Gasoline on NOx Storage -
Reduction Catalyst," SAE 1999-01-3501.

72. Guyon, M. et al, "NOx-Trap System Development and Characterization for Diesel Engines
Emission Control,"  SAE 2000-01-2910.

73. Dou, Danan and Bailey, Owen, "Investigation of NOx Adsorber Catalyst Deactivation,"
SAE 982594.

74. Guyon, M. et al, "Impact of Sulfur on NOx Trap Catalyst Activity - Study of the
Regeneration Conditions", SAE 982607.

75. Dearth, et al, "Sulfur Interaction with Lean NOx Traps: Laboratory and Engine
Dynamometer Studies", SAE 982595.

76. Guyon, M. et al, "NOx-Trap System Development and Characterization for Diesel Engines
Emission Control,"  SAE 2000-01-2910.

77. Dou, D and Bailey, O.,"Investigation of NOx Adsorber Catalyst Deactivation," SAE
982594.

78. Dearth, et al, "Sulfur Interaction with Lean NOx Traps: Laboratory and Engine
Dynamometer Studies", SAE 982595.

79. Dearth, et al, "Sulfur Interaction with Lean NOx Traps: Laboratory and Engine
Dynamometer Studies", Figure 5 SAE 982595.

80. Dearth, et al, "Sulfur Interaction with Lean NOx Traps: Laboratory and Engine
Dynamometer Studies", SAE 982595.

81. Dou, D and Bailey, O.,"Investigation of NOx Adsorber Catalyst Deactivation," SAE
982594.

82. Heck, R. and Farrauto, R. Catalytic Air Pollution Control - Commercial Technology, page
64-65. 1995 Van Nostrand Reinhold Publishing.

83. Heck, R. and Farrauto, R. Catalytic Air Pollution Control - Commercial Technology,
Chapter 6. 1995 Van Nostrand Reinhold Publishing.

                                        4-176

-------
	Technologies and Test Procedures for Low-Emission Engines

84.  Asanuma, T. et al, "Influence of Sulfur Concentration in Gasoline on NOx Storage -
Reduction Catalyst,"  SAE 1999-01-3501.

85.  Diesel Emission Control - Sulfur Effects (DECSE) Program Phase II Summary Report: NOx
Adsorber Catalysts, October 2000. Copy available in Air Docket A-2001-28.

86.  Tanaka, H., Yamamoto, M., "Improvement in Oxygen Storage Capacity," SAE 960794.

87.  Yamada, T., Kobayashi, T., Kayano, K., Funabiki M., "Development of Zr Containing TWC
Catalysts", SAE 970466.

88.  McDonald, Joseph, and Lee Jones, U.S. EPA, "Demonstration of Tier 2 Emission Levels for
Heavy Light-Duty Trucks," SAE 2000-01-1957.

89.  Dearth, et al, "Sulfur Interaction with Lean NOx Traps: Laboratory and Engine
Dynamometer Studies", SAE 982595.

90.  Letter from Barry Wallerstein, Acting Executive Officer, SCAQMD, to Robert Danziger,
Goal Line Environmental Technologies, dated December 8, 1997, www.glet.com Air Docket A-
99-06 item II-G-137.

91.  Reyes and Cutshaw, SCONOx Catalytic Absorption System, December 8, 1998,
www.glet.com Air Docket A-99-06 item II-G-147.

92.  Danziger, R. et al 21,000 Hour Performance Report on SCONOX, 15 September 2000 EPA
Docket A-99-06 item IV-G-69.

93.  Table from May  11, 2002 edition of the Frankfurter Allgemeine Zeitung listing Direct
Injection Gasoline Vehicles for sale in Europe; the table has been edited to  indicate which
vehicles are lean-burn (i.e., would use a NOx adsorber catalyst) and which  are stoichiometric-
burn (i.e., would use a conventional 3-way catalyst, indicated by A = 1).  Copy available in Air
Docket A-2001-28.

94.  Schenk, Charles "Summary of NVFEL Testing of Advanced NOx and  PM Emission Control
Technologies" memo to EPA Docket A-99-06, item IV-A-29.

95. "Regulatory Impact Analysis: Heavy-Duty Engine and Vehicle Standards and Highway
Diesel Fuel Sulfur Control Requirements," EPA420-R-00-026, December 2000, Docket A-2001-
28, Document No. II-A-01.

96.  Schenk, C., McDonald, J., and Laroo, C., "High-Efficiency NOx and PM Exhaust Emission
Control for Heavy-Duty On-Highway Diesel Engines - Part Two" SAE 2001-01-3619, Air
Docket A-2001-28.

97.  Schenk, C., McDonald, J., and Laroo, C., "High-Efficiency NOx and PM Exhaust Emission
Control for Heavy-Duty On-Highway Diesel Engines - Part Two" SAE 2001-01-3619, Air
Docket A-2001-28.

                                       4-177

-------
Regulatory Impact Analysis
98. Schenk, C., McDonald, J., and Laroo, C., "High-Efficiency NOx and PM Exhaust Emission
Control for Heavy-Duty On-Highway Diesel Engines - Part Two" SAE 2001-01-3619, Air
Docket A-2001-28.

99. Schenk, C. and Laroo, C. " NOx Adsorber Aging on a Heavy-Duty On-Highway Diesel
Engine - Part One," SAE 2003-01-0042. Copy available in Air Docket A-2001-28.

100. Schenk, C. and Laroo, C. " NOx Adsorber Aging on a Heavy-Duty On-Highway Diesel
Engine - Part One," SAE 2003-01-0042. Copy available in Air Docket A-2001-28.

101. Diesel Emission Control Sulfur Effects (DECSE) Program - Phase I Interim Data Report
No. 1, August 1999. Copy available in Air Docket A-2001-28.

102. Diesel Emission Control Sulfur Effects (DECSE) Program - Phase I Interim Data Report
No. 2: NOx Adsorber  Catalysts, October 1999. Copy available in Air Docket A-2001-28.

103. Diesel Emission Control Sulfur Effects (DECSE) Program - Phase I Interim Date Report
No. 3:  Diesel Fuel Sulfur Effects on Particulate Matter Emissions, November 1999. Copy
available in Air Docket A-2001-28.

104. Diesel Emission Control Sulfur Effects (DECSE) Program - Phase I Interim Data Report
No. 4, Diesel Particulate Filters-Final Report, January 2000. Copy available in Air Docket A-
2001-28.

105. Diesel Emission Control - Sulfur Effects (DECSE) Program Phase II Summary Report:
NOx Adsorber Catalysts, October 2000. Copy available in Air Docket A-2001-28.

106. Diesel Emission Control - Sulfur Effects (DECSE) Program Phase II Summary Report:
NOx Adsorber Catalysts, October 2000. Copy available in Air Docket A-2001-28.

107. Details with quarterly updates on the APBF-DEC programs can be found on the DOE
website at the following location http://www.ott.doe.gov/apbf.shtml..

108. Hakim, N. "NOx Adsorbers for Heavy Duty Truck Engines - Testing and Simulation,"
presentation at Motor Fuels: Effects on Energy Efficiency and Emissions in the Transportation
Sector Joint Meeting of Research Program Sponsored by the U.S. Department of Energy, Clean
Air for Europe and Japan Clean Air, October 9-10, 2002. Copy available in EPA Air Docket A-
2001-28.

109. Shoji, A.; Kamoshita, S.; Watanabe, T.; Tanaka, T.; and Yabe, M., "Development of a
Simultaneous Reduction System of NOx and PM for Light-Duty Truck," JSAE 2003-5567.

110. McDonald, J. "Progress in the Development of Tier 2 Light-Duty Diesel Vehicles," SAE
2004-01-1791, March 2004.

111. "Demonstration of Advanced Emission Control Technologies Enabling Diesel-Powered
Heavy-Duty Engines to Achieve Low Emission Levels", Manufacturers of Emissions Controls

                                       4-178

-------
	Technologies and Test Procedures for Low-Emission Engines

Association, June 1999 Air Docket A-2001-28.

112. Fable, S. et al, "Subcontractor Report - Selective Catalytic Reduction Infrastructure Study,"
AD Little under contract to National Renewable Energy Laboratory, July 2002, NREL/SR-5040-
32689. Copy available in EPA Air Docket A-2001-28.

113. Engelhard DPX catalyzed diesel particulate filter retrofit verification,
www.epa.gov/otaq/retrofit/techlist-engelhard.htm. a copy of this information is available in Air
Docket A-2001-28.

114. Engelhard DPX catalyzed diesel particulate filter retrofit verification,
www.epa.gov/otaq/retrofit/techlist-engelhard.htm. a copy of this information is available in Air
Docket A-2001-28.

115. Johnson Matthey CRT filter retroift verification,
http ://www. epa. gov/otaq/retrofit/techlist-j ohnmatt.htm#j m4 a copy of this information is
available in Air Docket A-2001-28.

116.  "Investigation of the Feasibility of PM Filters for NRMM", Report by the European
Association of Internal Combustion Engine Manufacturers and Engine Manufacturers
Association, July, 2002. Copy available in EPA Air Docket A-2001-28, item # II-B-12

117. Sasaki, S., Ito, T., and Iguchi, S., "Smoke-less Rich Combustion by Low Temperature
Oxidation in Diesel Engines," 9th Aachener Kolloquim Fahrzeug - und Motorentechnik 2000.
Copy available in Air Docket A-2001-28.

118. Jeuland, N., et al, "Performances and Durability of DPF (Diesel Particulate Filter) Tested
on a Fleet of Peugeot 607 Taxis First and Second Test Phases Results," October 2002, SAE
2002-01-2790.

119. "Summary of Conference Call between U.S. EPA and Deutz Corporation on September 19,
2002 regarding Deutz Diesel Particulate Filter System", EPA Memorandum to Air Docket A-
2001-28.

120. "Particulate Traps for Construction Machines: Properties and Field Experience" J.
Czerwinski et al, Society of Automotive Engineers Technical Paper 2000-01-1923.

121. "Engine Technology and Application Aspects for Earthmoving Machines and Mobile
Cranes, Dr. E. Brucker, Liebherr Machines Bulle, SA, AVL International Commercial
Powertrain Conference, October 2001. Copy available in EPA Air Docket A-2001-28, Docket
Item # II-A-12.

122. Phone conversation with  Manufacturers of Emission Control Association (MECA), 9 April,
2003 confirming the use of emission-control technologies on nonroad equipment used in coal
mines, refineries, and other locations where  explosion proofing may be required.
                                         4-179

-------
Regulatory Impact Analysis
123. See for example "Diesel-engine Management" published by Robert Bosch GmbH, 1999,
second edition, pages 6-8 for a more detailed discussion of the differences between and IDI and
DI engines.

124. See Chapter 14, Section 4 of "Turbocharging the Internal Combustion Engine", N. Watson
and M.S. Janota, published by John Wiley and Sons, 1982.

125.  See Section 2.2 through 2.3 in "Nonroad Diesel Emission Standards - Staff Technical
Paper", EPA Publication EPA420-R-01-052, October 2001. Copy available in EPA Air Docket
A-2001-28.

126. See Table 3-2 in "Nonroad Diesel Emission Standards - Staff Technical Paper", EPA
Publication EPA420-R-01-052, October 2001. Copy available in EPA Air Docket A-2001-28.

127. EPA Memorandum "2002 Model Year Certification Data for Engines <50 Hp", William
Charmley, copy available in EPA Air Docket A-2001-28"

128. See Section 2.2 through 2.3 in "Nonroad Diesel Emission Standards - Staff Technical
Paper", EPA Publication EPA420-R-01-052, October 2001. Copy available in EPA Air Docket
A-2001-28.

129. Ikegami, M., K. Nakatani, S. Tanaka, K. Yamane:  "Fuel Injection Rate Shaping and Its
Effect on Exhaust Emissions in a Direct-Injection Diesel Engine Using  a Spool Acceleration
Type Injection System", SAE paper 970347, 1997.  Dickey D.W., T.W. Ryan III, A.C.
Matheaus: "NOx Control in Heavy-Duty Engines-What  is the Limit?", SAE paper 980174, 1998.
Uchida N,  K. Shimokawa, Y. Kudo, M. Shimoda: "Combustion Optimization by Means of
Common Rail Injection System for Heavy-Duty Diesel Engines", SAE paper 982679, 1998.

130. "Effects of Injection Pressure and Nozzle Geometry on DI Diesel Emissions and
Performance," Pierpont, D., and Reitz, R., SAE Paper 950604, 1995.

131. EPA Memorandum "Documentation of the Availability of Diesel Oxidation Catalysts on
Current Production Nonroad Diesel Equipment", William Charmley. Copy available in EPA Air
Docket A-2001-28.

132. See Table 2-4 in "Nonroad Diesel Emission Standards - Staff Technical Paper", EPA
Publication EPA420-R-01-052, October 2001. Copy available in EPA Air Docket A-2001-28.

133. See Table 2-4 in "Nonroad Diesel Emission Standards - Staff Technical Paper", EPA
Publication EPA420-R-01-052, October 2001. Copy available in EPA Air Docket A-2001-28.

134. "Demonstration of Advanced Emission Control Technologies Enabling Diesel-Powered
Heavy-duty Engines to Achieve Low Emission Levels: Interim Report Number 1 - Oxidation
Catalyst Technology, copy available in EPA Air Docket A-2001-28.  "Reduction of Diesel
Exhaust Emissions by Using Oxidation Catalysts," Zelenka et al, SAE Paper 90211, 1990.  See
Table 2-4 in "Nonroad Diesel Emission Standards - Staff Technical Paper", EPA Publication


                                       4-180

-------
	Technologies and Test Procedures for Low-Emission Engines

EPA420-R-01-052, October 2001, copy available in EPA Air Docket A-2001-28.

135. "The Optimized Deutz Service Diesel Particulate Filter System II", H. Houben et al, SAE
Technical Paper 942264, 1994 and "Development of a Full-Flow Burner DPF System for Heavy
Duty Diesel Engines, P. Zelenka et al, SAE Technical Paper 2002-01-2787, 2002.

136. See Tables 6, 8, and 14 of "Nonroad Emission Study of Catalyzed Particulate Filter
Equipped Small Diesel Engines" Southwest Research Institute, September 2001. Copy available
in EPA Air Docket A-2001-28.

137. See Section 2.2 through 2.3 in "Nonroad Diesel Emission Standards - Staff Technical
Paper", EPA Publication EPA420-R-01-052, October 2001. Copy available in EPA Air Docket
A-2001-28.

138. Highway Diesel Progress Review Report 2, United States Environmental Protection
Agency, March 2004, EPA 420-R-04-004.  Copy available in Air Docket OAR-2003-0012-0918.

139. See Section 3 of "Nonroad Diesel Emission Standards -  Staff Technical Paper", EPA
Publication EPA420-R-01-052, October 2001.  Copy available in EPA Air Docket A-2001-28.

140. See Table 3-2 in "Nonroad Diesel Emission Standards - Staff Technical Paper", EPA
Publication EPA420-R-01-052, October 2001.  Copy available in EPA Air Docket A-2001-28.

141. EPA Memorandum "Summary of Model Year 2001  Certification data for Nonroad Tier 1
Compression-ignition Engines with rated power between 0 and 50 horsepower", William
Charmley, copy available in EPA Air Docket A-2001-28, docket item II-B-08.

142. "Effects of Injection Pressure and Nozzle Geometry on DI Diesel Emissions and
Performance," Pierpont, D., and Reitz, R., SAE Paper 950604, 1995.

143. EPA Memorandum "Documentation of the Availability of Diesel Oxidation Catalysts on
Current Production Nonroad Diesel Equipment", William Charmley. Copy available in EPA Air
Docket A-2001-28.

144. See Table 2-4 in "Nonroad Diesel Emission Standards - Staff Technical Paper", EPA
Publication EPA420-R-01-052, October 2001.  Copy available in EPA Air Docket A-2001-28.

145. "Demonstration of Advanced Emission Control Technologies Enabling Diesel-Powered
Heavy-duty Engines to Achieve Low Emission Levels: Interim Report Number 1 - Oxidation
Catalyst Technology, copy available in EPA Air Docket A-2001-28.  "Reduction of Diesel
Exhaust Emissions by Using Oxidation Catalysts," Zelenka et  al, SAE Paper 90211, 1990.  See
Table 2-4 in "Nonroad Diesel Emission  Standards - Staff Technical Paper", EPA Publication
EPA420-R-01-052, October 2001, copy available in EPA Air Docket A-2001-28.

146. Letter from Marty Barris, Donaldson Corporation, to Byron Bunker, U.S. EPA, March
2000.  A copy is available in Air Docket A-2001-28.
                                       4-181

-------
Regulatory Impact Analysis
147. Hawker, P. et al, "Experience with a New Paniculate Trap Technology in Europe," SAE
970182.

148. Hawker, P. et al, "Experience with a New Paniculate Trap Technology in Europe," SAE
970182.

149. Allansson, et al, "European Experience of High Mileage Durability of Continuously
Regenerating Filter Technology," SAE 2000-01-0480.

150. Letter from Dr. Barry Cooper, Johnson Matthey, to Don Kopinski, U.S. EPA.  A copy is
available in Air Docket A-2001-28.

151. Telephone conversation between Dr. Barry Cooper, Johnson Matthey, and Todd Sherwood,
EPA, Air Docket A-99-06.

152. Letter from Dr. Barry Cooper to Don Kopinski U.S. EPA. A copy is available in Air
Docket A-2001-28.

153. Dou, Danan and Bailey, Owen, "Investigation of NOx Adsorber Catalyst Deactivation,"
SAE 982594.

154.Guyon, M. et al, "Impact of Sulfur on NOx Trap Catalyst Activity - Study of the
Regeneration Conditions", SAE 982607.

155. Though it was favorable to decompose  sulfate at 800°C, performance of the NSR (NOx
Storage Reduction catalyst, i.e., NOx Adsorber) catalyst decreased due to sintering of precious
metal.  - Asanuma, T. et al, "Influence of Sulfur Concentration in Gasoline on NOx Storage -
Reduction Catalyst", SAE 1999-01-3501.

156. Nonroad Test Cycle Development, Starr, M., Southwest Research Institute Contractor
report for the United States Environmental Protection Agency, September 1998

157. Nonroad Data Analysis and Composite Cycle Development, Webb, C., Southwest Research
Institute contractor report to the United States Environmental Protection Agency, September
1997

158. Memorandum from Kent Helmer to Cleophas Jackson, "National Excavator Fleet
Population Estimate", Docket A-2001-28, # II-B-32.

159.Bin Analysis of Nonroad Diesel Transient Duty Cycles; Hoffman, G., Dyntel Corporation;
Ann Arbor, MI, March 2003

160."Maximum Speed Determination Procedure," EPA memo to Docket A-2001-28, # II-B-45.

161. See Chapter 8 of "Summary and  Analysis of Comments: Control of Emissions from Marine
Diesel Engines" (EPA420-R-99-028),  December 1999. This report may be found on and
downloaded from the EPA-OTAQ website at http://www.epa.gov/otaq/marine.htm.

                                       4-182

-------
	Technologies and Test Procedures for Low-Emission Engines

162. See http://www.epa.gov/oms/regs/nonroad/equip-hd/cycles/nrcycles.htm.

163. ISO Report on NRTC Cycle Development "Final Report on NRTC test Procedure, Summer
2002" Docket A-2001-28.

164. Memorandum from Kent Helmer to Cleophas Jackson, "Applicability of EPA's NRTC
cycle to the U.S. Nonroad Diesel Population", Docket A-2001-28, # II-B-34.

165.    Title 40, U.S. Code of Federal Regulations, §89.407(c)(10-ll), July 2003.

166.    Guidance Letter to Engine Manufacturers Association from Greg Green, Division
Director, Certification and Compliance Division, OTAQ to Jed Mandel, Engine Manufacturers
Association, Dated 12 December 2002, "Guidance Regarding Test Procedures for Heavy-Duty
On-Highway and Non-Road Engines."

167.    Schenk, C., Laroo, C., Olson, B., Fisher, L., "Four-Flow Path High-Efficiency NOx and
PM Exhaust Emission Control System for Heavy-Duty On-Highway Diesel Engines" SAE
Technical Paper 2003-01-2305, 2003.

168. Memorandum from Kent Helmer to Cleophas Jackson, "Tabular Summary of California
EPA-ARE TRU engine Operating Data", Docket A-2001-28, # IV-E-35.

169. Memorandum from Kent Helmer to EPA Air Docket A-2001-28, # "Discussion with
Kubota  and Carrier Representatives, March 2, 2004", e-Docket OAR-2003-0012-0996.

170. Memorandum from Kent Helmer to EPA Air Docket A-2001-28, # "Correspondence from
Carrier Representative, March 4, 2004", e-Docket OAR-2003-0012-0998.

171. Memorandum from Kent Helmer to EPA Air Docket A-2001-28, # "Corespondence from
Tom Sem of Thermo King, March 6, 2004," e-Docket OAR-2003-0012-0995.
                                       4-183

-------
CHAPTER 5: Fuel Standard Feasibility
    5.1 The Blendstocks and Properties of Non-Highway Diesel Fuel   	5-1
       5.1.1 Blendstocks Comprising Non-highway Diesel Fuel and their Sulfur Levels	5-1
       5.1.2 Current Levels of Other Fuel Parameters in Non-highway Distillate  	5-4
    5.2 Evaluation of Diesel Fuel Desulfurization Technology	5-7
       5.2.1 Introduction to Diesel Fuel Sulfur Control 	5-7
       5.2.2 Conventional Hydrotreating	5-8
           5.2.2.1 Fundamentals of Distillate Hydrotreating	5-9
           5.2.2.2 Meeting a 15 ppm Cap with Distillate Hydrotreating	5-13
           5.2.2.3 Low-Sulfur Performance of Distillate Hydrotreating 	5-19
       5.2.3 Process Dynamics Isotherming	5-21
       5.2.4 Phillips S-Zorb Sulfur Adsorption   	5-24
       5.2.5 Chemical Oxidation and Extraction	5-25
    5.3 Feasibility of Producing 500 ppm Sulfur NRLM Diesel Fuel in 2007	5-27
       5.3.1 Expected use of Desulfurization Technologies for 2007	5-27
       5.3.2 Lead-time Evaluation	5-28
           5.3.2.1 Tier 2 Gasoline Sulfur Program	5-29
           5.3.2.2 15 ppm Highway Diesel Fuel Sulfur Cap	5-30
           5.3.2.3 Lead-time Projections for Production of 500 ppm NRLM Diesel Fuel	5-31
           5.3.2.4 Comparison with the 500 ppm Highway Diesel Fuel Program  	5-35
           5.3.2.5 Small Refiners 	5-35
    5.4 Feasibility of Producing 15 ppm Sulfur NRLM in 2010 and 2012	5-36
       5.4.1 Expected use of Desulfurization Technologies in 2010 and 2012	5-36
       5.4.2 Lead-time Evaluation	5-39
    5.5 Distribution Feasibility Issues	5-40
       5.5.1 Ability of Distribution System to Accommodate the Need for Additional Product
           Segregations That Could Result from This Rule 	5-40
           5.5.1.1 The Diesel Fuel Distribution System Prior to Implementation of the NRLM Sulfur-
               Control Program	5-40
           5.5.1.2 Potential for Additional Product Segregation Under the NRLM Sulfur Program  .5-41
           5.5.1.3 Ability of Fuel Distributors to Handle New Product Segregations that Will Result
               from the NRLM Sulfur Control Program 	5-50
           5.5.1.4 Determining the Boundaries  for the Northeast/Mid-Atlantic Area  	5-55
       5.5.2 Limiting Sulfur Contamination  	5-63
       5.5.3 Handling Practices for Distillate Fuels that Become Mixed in the Pipeline Distribution
           System 	5-65
    5.6 Feasibility of the Use of a Marker in Heating Oil	5-68
    5.7 Impacts on the Engineering and Construction Industry	5-75
       5.7.1 Design and Construction Resources Related to Desulfurization Equipment	5-76
       5.7.2 Number and Timing of Revamped  and New Desulfurization Units  	5-77
       5.7.3 Timing of Desulfurization Projects Starting up in the Same Year	5-78
       5.7.4 Timing of Design and Construction Resources Within a Project  	5-78
       5.7.5 Projected Levels of Design and Construction Resources  	5-80
    5.8 Supply of Nonroad, Locomotive, and Marine Diesel Fuel (NRLM)	5-84
    5.9 Desulfurization Effect on Other Non-Highway Diesel Fuel Properties	5-92
       5.9.1 Fuel Lubricity	5-92
       5.9.2 Volumetric Energy Content	5-95
       5.9.3 Fuel Properties Related to Storage  and Handling  	5-97
       5.9.4 Cetane Index and Aromatics  	5-97
       5.9.5 Other Fuel Properties	5-98
    Appendix 5A: EPA's Legal Authority for Adopting Nonroad, Locomotive, and Marine Diesel Fuel
       Sulfur Controls 	5-101

-------
                                                              Fuel Standard Feasibility
               CHAPTER 5:  Fuel Standard Feasibility
   In this chapter, we present an analysis of the feasibility of complying with the fuel program
adopted in this final rule, including a discussion of the technology used to desulfurize and
distribute ultra low diesel fuel. In Section 5.1, we discuss the sources of the blendstocks which
comprise diesel fuel and summarize their reported sulfur levels. In Section 5.2, we present and
evaluate a wide variety of distillate desulfurization technologies that refiners might use to meet
the 500 and 15 ppm sulfur caps. In Section 5.3, we formally assess the technical  feasibility of
meeting the 500 ppm sulfur cap in 2007, including the sufficiency of the lead time for refiners.
In Section 5.4, we assess the technical feasibility meeting the 15 ppm sulfur cap,  including the
sufficiency of lead time for refiners. In Section 5.5, we  assess the feasibility of distributing 500
and 15 ppm sulfur fuel.  In Section 5.6, we assess the feasibility of using a marker in heating oil.
In Section 5.7, we evaluate the impacts of this program and other sulfur control regulations on
the engineering and construction industry.  In Section 5.8 we assess the impacts of this program
on the supply of NRLM diesel fuel.  In Section 5.9 we discuss how hydro-desulfurization is
expected to affect NRLM diesel fuel properties other than sulfur. Finally, in Chapter 5.10 we
assess how properties other than sulfur will be impacted by desulfurizing NRLM diesel fuel. At
the end of Chapter 5 we include an Appendix summarizing EPA's authority for adopting NRLM
sulfur standards.

5.1 The Blendstocks and Properties of Non-Highway Diesel Fuel

5.1.1  Blendstocks Comprising Non-highway Diesel Fuel and their Sulfur Levels

   The primary sources of sulfur in diesel fuel are the sulfur-containing compounds that occur
naturally in crude oil.A  Depending on the source, crude  oil contains anywhere from fractions of
a percent of sulfur, such as less than 0.05 weight percent (500 ppm) to as much as several weight
percent.1 The average amount of sulfur in crude oil refined in the United States is about one
weight percent.2 Most of the sulfur in crude oil is in the heaviest boiling fractions.  Since most
of the refinery blendstocks that are used to manufacture  diesel fuel come from the heavier
boiling components of crude oil, they contain substantial amounts of sulfur.

   The distillate8 produced by a given refinery is composed of one or more blendstocks from
crude oil fractionation and conversion units at the refinery. Refinery configuration and
equipment, and the types and relative volumes of products manufactured (the product slate) can
   A Additives that contain sulfur are sometimes intentionally added to diesel fuel. For a discussion how the
addition of these additives will be affected under this program, see Section IV.D.5.

   B Distillate refers to a broad category of fuels falling into a specific boiling range. Distillate fuels have a heavier
molecular weight and therefore boil at higher temperatures than gasoline.  Distillate includes diesel fuel, kerosene
and home heating oil. For the purposes of this discussion, we will focus on No. 2 distillate, which comprises the
majority of diesel fuel and heating oil.

                                           5-1

-------
Final Regulatory Impact Analysis
                                       Figure 5.1-1
                          Diagram of a Typical Complex Refinery
  Natural
     Gas
                  Vacuum To\\
                                                               Coker
significantly affect the sulfur content of diesel fuel.  The diagram on the following page
illustrates the configuration and equipment used at a typical complex refinery in the United
States.
   Refineries differ from the model in the preceding diagram depending on the characteristics of
the crude oils refined, and their product slate, as illustrated in the following examples:
   -   Refineries that process lighter crude oils are less likely to have coker and hydrocracker
       units.
   -   Refinery streams that can be used to manufacture diesel fuel can also be used to
       manufacture heating oil, kerosene, and jet fuel.  Much of the distillate product from the
       hydrocracker is often blended into jet fuel rather than diesel fuel; current highway
       regulations generally require that a refinery have a hydrotreater, which is usually not
       necessary if the refinery  produces only high sulfur non-highway diesel fuel.

   On an aggregate basis, most of the distillate manufactured in the United States comes from
the crude fractionation tower (called straight-run or SR).  Most of the remainder comes from the
fluid catalytic cracker (FCC) conversion unit (called light cycle oil or LCO). The remaining

                                           5-2

-------
                                                             Fuel Standard Feasibility
small fraction of diesel fuel volume comes from a coker conversion unit or other units that crack
heavy compounds such as a visbreaker or steam cracker (called other cracked stocks in this
document), or from the hydrocracker conversion unit (called hydrocrackate).

   To comply with the current federal regulatory  requirement on the sulfur content of highway
diesel fuel (500 ppm cap), the blendstock streams  from these process units are typically further
processed to reduce their sulfur content. Desulfurization of highway diesel blendstocks to meet
the 500 ppm cap is accomplished in fixed-bed hydrotreaters that operate at moderate pressures
(500 to 800 psi and higher).3  Nearly all the low-sulfur diesel blendstocks come from such
hydrotreaters. However, a small amount of low-sulfur diesel also comes from hydrocrackers
operating at pressures  of 500 to 3000 psi, although most operate at 1500 to 3000 psi, which
naturally produces distillate fuel with  sulfur levels of 100 ppm or less.

   To comply with applicable non-highway sulfur requirements which range from 2000 to 5000
ppm, or the 40 cetane standard for nonroad, locomotive and marine diesel fuel, some of the
distillate  blendstocks used to produce  non-highway diesel fuel and heating oil are hydrotreated.
A significant amount of hydrocracked distillate is  also blended into non-highway diesel fuel and
heating oil. As discussed in Chapter 7, the use of hydrotreated blendstocks in non-highway
diesel fuel has important implications  for the cost  of desulfurizing NRLM diesel fuel.

   The distillate blendstocks used to produce non-highway diesel fuel and their sulfur content
vary considerably from refinery to refinery. A survey conducted by the American Petroleum
Institute (API) and National Petroleum Refiners Association (NPRA) in 1996 examined the
typical blendstock properties for the U.S. highway and the non-highway diesel pools.4 The
results of this survey for the non-highway distillate pool are in Table 5.1-1.
                                           5-2

-------
Final Regulatory Impact Analysis
                                       Table 5.1-1
                      Average Composition and Sulfur Content of the
                 Non-highway Distillate Pool Outside of California in 1996s
Type of Distillate
Stream
Unhydrotreated
Hydrotreated

Diesel Blendstock
Straight-Run
Light Cycle Oil (LCD)
Coker Gas Oil
Unhydrotreated Subtotal
Hydrotreated Straight -Run
Hydrotreated LCD
Hydrotreated Coker Gas Oil
Hydrocrackate
Hydrotreated Subtotal
Total
Percentage
45
12
1
58
18
10
4
10
42
100
Sulfur Content
(ppm)
2274
3493
2345
-
353
1139
270
115
-
-
   As shown in Table 5.1-1, approximately 42 percent of all blendstocks used to manufacture
non-highway distillate outside of California are hydrotreated to reduce their sulfur content. This
includes hydrocrackate (10 percent of the non-highway distillate pool), which is desulfurized to a
substantial extent as a necessary element of the hydrocracking process and is not further
processed in a hydrotreater. Table 5.1-1  also shows that approximately 58 percent of non-
highway distillate comes from nonhydrotreated blendstocks. As expected, the sulfur levels of
the hydrotreated blendstocks are lower than the nonhydrotreated distillate blendstocks.

   In Chapter 7 of the RIA we use this blendstock information as one of the input parameters for
estimating the relative difficulty and ultimately the cost for desulfurizing diesel fuel. The 1996
data is an  important input for our cost analysis, and we update the mix of blendstocks to 2002
based on changes in relative unit capacities.

5.1.2 Current Levels of Other Fuel Parameters in Non-highway Distillate

   It is useful to review other qualities of high-sulfur distillate, as well as sulfur content. First,
some of the desulfurization technologies affect these other fuel properties.  Second, as discussed
further below, some sulfur compounds are more difficult to treat than others. In some cases,
refiners might try to shift these more difficult compounds to fuels that face less stringent sulfur
standards. Their ability to do this depends, not only on the economics of doing so, but also on
the effect  of such shifts on nonsulfur properties and whether or not these other properties still
meet industry specifications. Thus, it is helpful to evaluate the degree to which current non-
highway distillate fuels meet or exceed applicable industry standards.
                                           5-4

-------
                                                               Fuel Standard Feasibility
   Data on the current distillation characteristics, API gravity, pour point, natural cetane level,
and aromatics content of diesel fuel blendstocks are in Table 5.1-2.

                                        Table 5.1-2
         Average Non-highway Distillate Fuel Property Levels by Geographic Area6 7
                    (Data from 1997 API/NPRA Survey unless specified)
Fuel Parameter
API Gravity
Cetane Number3
Pour Point (°F)
[additized]
Pour Point Depressant
Additive (ppmw)
Distillation
(°F)
T10
T30
T50
T70
T90
PADD1
32.6
N/A
-6
0
434
492
517
545
613
PADD2
34.1
N/A
-8
71
425
476
508
558
604
PADD3
32.6
N/A
0
0
418
457
502
536
598
PADD4
35.6
N/A
6
13
411
443
499
522
591
PADD5
(CA Excluded)
33.8
N/A
12
0
466
517
542
570
616
U.S.
(CA Excluded)
32.8
47
-1
18
419
464
503
539
595
CA
30.8
N/A
4
0
498

556

620
1 From 1997 NIPER/TRW survey data, U.S. average includes California. N/A means not available.
    The American Society for Testing Materials (ASTM) has established requirements that apply
to No. 2 non-highway diesel fuel, as well as for No. 2 distillate fuel (e.g., heating oil).8 The
requirements most relevant to desulfurization are summarized in Table 5.1-3.

                                        Table 5.1-3
              ASTM Requirements that Apply to Non-Highway Distillate Fuels

T-90 Mm °F
T-90 Max °F
Density max (g/cm3) (API Gravity min)
Pour Point max °F
Cloud Point °F
Sulfur max (ppm)
Cetane Number min
No. 2 Diesel Fuel
(Non-highway)
540
640
None
46 to -0.4
5000
40
No. 2 Fuel Oil/Heating
Oil
540
640
0.876 (30.0)
21.2
5000

No. 2 Marine Distillate
(DMA)
=
0.890 (27.5)
21.2

40
                                           5-5

-------
Final Regulatory Impact Analysis
   Comparing Tables 5.1-2 and 5.1-3 shows that the average properties of current non-highway
distillate are within the ASTM requirements, and for some properties, well within requirements.
For example, except for California, the T90 of current non-highway diesel fuel is 25-40°F below
the maximum allowed. The average cetane number of all non-highway distillate is well above
the minimum of 40.  Finally the pour point is well below the maximum allowed for fuel
oil/heating oil and marine distillate fuel.  One exception is that the API gravity of non-highway
distillate fuel in PADDs  1 and 3, which includes the heating oil used in the Northeast, is just
above the  minimum.

   While refiners might try to perform such shifts in blendstocks between fuels, note that we did
not assume refineries would be shifting blendstocks between various distillate fuels to reduce the
compliance costs associated with the NRLM diesel fuel sulfur standards. Instead, we projected
the use of desulfurization techniques that will be sufficient to meet the new sulfur standards
without shifting more difficult-to-treat sulfur compounds to other fuels. This approach appeared
reasonable, given that we were evaluating the potential of over 100 refineries currently
producing non-highway distillate fuel to reduce sulfur in NRLM diesel fuel. The ability to shift
blendstocks between fuels to reduce  costs is very refinery-specific and difficult to estimate on
average across a wide range of refineries.  Also, two primary types of shifts are possible and
both have limits. One approach is to shift the heaviest portion of selected blendstocks such as
LCO to the bunker or residual fuel pool, avoiding the need to desulfurize this material.
However,  the market for these heavy fuels is limited and on a national  basis, this approach is
generally not economically feasible.  The other approach is to shift these difficult-to-treat
streams and portions of streams to heating oil, which can meet less stringent sulfur standards.
This would likely require the addition of additional product tankage and require more refineries
to produce lower-sulfur NRLM diesel fuel. The material being shifted to heating oil could still
require additional desulfurization to  ensure that ASTM and state standards were still being met.
Thus, there would be a cost trade-off, not just a cost reduction.  Again, given the national scale
of this analysis, we decided to avoid the projection of such shifts and limit our analysis to the
desulfurization of current non-highway diesel fuel blendstocks. In this regard, our cost analysis
as presented in Chapter 7 can be viewed as somewhat conservative.

5.2  Evaluation of Diesel Fuel Desulfurization Technology

5.2.1 Introduction to Diesel Fuel Sulfur Control

   As mentioned in Section  5.1, the sulfur in diesel fuel comes from the crude oil processed by
the refinery.  One way to reduce the  amount of sulfur in diesel fuel is therefore to process a crude
oil that is lower in sulfur. Some refiners already do this. Others could switch to low- or at least
lower-sulfur crude oils. However, there is limited capability worldwide to produce low-sulfur
crude oil.  While new oil fields producing light, sweet crude oil are still being discovered, most
of the new crude oil production being brought on-line is heavier, more sour (i.e., higher sulfur)
crude oils.  The incentive to use low-sulfur crude oils has existed for some time and low-sulfur
crude oils have traditionally commanded a premium price relative to higher-sulfur crude oils.
While a few refiners with access to lower-sulfur crude oil might reduce their diesel sulfur levels
                                           5-6

-------
                                                              Fuel Standard Feasibility
this way, it is not feasible for most, let alone all U.S. refiners to switch to low-sulfur crude oils to
meet a tighter diesel fuel sulfur standard. In addition, while helpful, a simple change to a low-
sulfur crude oil may fall short of complying with the 500 ppm sulfur standard, and certainly fall
short of the 15 ppm sulfur standard.  Thus,  changing to a sweeter crude oil was not considered
viable for complying with the nonroad, locomotive, and marine diesel sulfur standards.

   A method to  reduce diesel fuel sulfur much more significantly is to chemically remove sulfur
from the hydrocarbon compounds that comprise diesel fuel. This is usually accomplished
through catalytically  reacting the diesel fuel with hydrogen at moderate to high temperature and
pressure over a fixed bed  of hydrotreating catalyst.  Two specific examples of this process are
hydrotreating and hydrocracking. A modified version of hydrotreating that operates solely in the
liquid state is now available by Process Dynamics.  Another process licensed by Conoco-Phillips
uses a moving bed catalyst to both remove  and adsorb the sulfur using hydrogen at moderate
temperature and  pressure. There are other low-temperature and low-pressure processes being
developed that don't  rely  on hydrotreating, such as chemical  oxidation.  Sulfur can be removed
via these processes up front in the refinery, such as from crude oil, before being processed in the
refinery into diesel fuel. Or, sulfur can be removed from individual refinery streams that are to
be blended directly into diesel fuel. Finally, another method  to moderately reduce diesel fuel
sulfur is to shift sulfur-containing hydrocarbon compounds to other fuels produced by the
refinery.

   After careful review of all these approaches, we expect that the sulfur reduction required by
the 500 ppm sulfur standard will occur through chemical removal via conventional
hydrotreating. For complying with the 15 ppm cap for NRLM diesel fuel, we expect it will be
met primarily through liquid-phase hydrotreating, which is an emerging advanced
desulfurization technology.  This section will begin with a relatively detailed discussion of the
capabilities of these various processes. Refiners may use the other methods to obtain cost-
effective sulfur reductions that will complement the primary  sulfur reduction achieved via
hydrotreating. These other methods, such as FCC feed hydrotreating, adsorption and chemical
oxidation are discussed following the primary discussion of distillate hydrotreating and liquid-
phase hydrotreating.  Another means for aiding the desulfurization of diesel fuel, particularly to
comply with the  15 ppm standard, is undercutting, which removes the most difficult-to-treat
sulfur compounds. Since undercutting can help ease the task of complying with the 15 ppm
standard for any  of the desulfurization technologies, we provide a discussion of undercutting
below.

5.2.2 Conventional Hydrotreating

   Hydrotreating generally combines hydrogen with a hydrocarbon stream at high temperature
and pressure in the presence of a catalyst. Refineries currently employ a wide range of these
processes for various purposes.  For example, naphtha (gasoline-like material that does not meet
gasoline specifications, such as octane level) being fed to the refinery reformer is always
hydrotreated to remove nearly all sulfur, nitrogen and metal contaminants that would deactivate
the noble metal catalyst used in the reforming process.  Similarly, feed to the FCC unit is often
hydrotreated to remove most of the sulfur, nitrogen  and metal contaminants to improve the yield

                                           5-7

-------
Final Regulatory Impact Analysis
and quality of high value products, such as gasoline and distillate, from the FCC unit. Refineries
currently producing highway diesel fuel to the 500 ppm standard hydrotreat their distillate to
remove much of the sulfur present and to improve the cetane. That same unit or another
hydrotreating unit in the refinery also hydrotreats some of the refinery streams used to blend up
non-highway distillate. We expect that nearly all refiners will hydrotreat the  naphtha produced
by the FCC unit to remove most of the sulfur present to comply with the  Tier 2 gasoline sulfur
standards.9

   If the temperature or pressure is increased sufficiently and if a noble metal catalyst is used,
hydrotreating can more dramatically affect the chemical nature of the hydrocarbons, as well as
remove contaminants. For example, through a process called hydrocracking, smaller, lighter
molecules are created by splitting larger, heavier molecules.  In the process, nearly all the
contaminants are removed and olefms and aromatics are saturated into paraffins and naphthenes.
Outside the United States, this process is commonly used to produce distillate from heavier, less
marketable refinery streams. In the United States the hydrocracker is most often used to produce
gasoline from poor quality distillate, such  as LCO from the FCC unit.

   A few refineries also currently hydrotreat their distillate more severely than is typical, but not
as severely as hydrocracking.  Their intent is to remove the sulfur, nitrogen and metallic
contaminants and to also saturate most of the aromatics present. This is done primarily in
Europe to meet very stringent specifications for both sulfur and aromatics applicable to certain
diesel fuels and encouraged by reduced excise taxes. This severe hydrotreating process is  also
used in the United States to "upgrade" petroleum streams that are otherwise too heavy or too low
in quality to be blended into the diesel pool, by cracking some of the material to lower molecular
weight compounds and saturating some of the aromatics to meet the distillation and cetane
requirements.  A different catalyst that encourages aromatic saturation is used instead of one that
simply encourages contaminant removal.

   To meet the 500 ppm and the 15  ppm sulfur standards, we expect refiners to focus as much
as possible on sulfur removal. Other contaminants, such as metals, are already sufficiently
removed by existing refinery processes. While saturation of aromatics generally improves
cetane, the cetane numbers of current nonroad, locomotive, and marine diesel fuels are typically
already sufficient to comply with the applicable ASTM standards. Thus, refiners want to avoid
saturating aromatics to avoid the additional cost of increased hydrogen consumption.
Consequently, we anticipate refiners will choose desulfurization processes that minimize the
amount of aromatics saturation.  Current diesel fuel already meets all applicable specifications;
hydrotreating to remove sulfur should not degrade quality, except possibly lubricity, as discussed
in Section 5.9.1. Thus, with this one exception, there should be no need to improve diesel  fuel
quality as a direct result of this new diesel sulfur standard. Refiners choosing to improve fuel
                                           5-S

-------
                                                               Fuel Standard Feasibility
quality would be focusing on improved profitability, rather than meeting the 15 ppm sulfur
standard.0

    5.2.2.1  Fundamentals of Distillate Hydrotreating

    Almost all distillate hydrotreater designs follow the same broad format. Liquid distillate fuel
is heated to temperatures of 300 to 380°C, pumped to pressures of 500 to 700 psia, mixed with
hydrogen, and passed over a catalyst.  Hydrogen reacts with sulfur and nitrogen atoms contained
in the hydrocarbon molecules, forming hydrogen sulfide and ammonia.  The resulting vapor is
then separated from the desulfurized distillate. The desulfurized distillate is usually simply
mixed with other distillate streams in the refinery to produce diesel  fuel and heating oil.

    The vapor coming off the reactor still contains a lot of valuable  hydrogen, because the
reaction requires the use of a significant amount of excess hydrogen to operate efficiently and
practically. However, the vapor also contains a significant amount  of hydrogen sulfide and
ammonia, which inhibit the desulfurization and denitrogenation reactions and must be removed
from the system. Thus, the hydrogen leaving the reactor is usually mixed with fresh hydrogen
and recycled to the front of the reactor for reaction with fresh distillate feed. This would cause a
build up of hydrogen sulfide and ammonia in the system,  since it has no way to leave the system.
In some cases, the hydrogen sulfide and ammonia are chemically scrubbed from the hydrogen
recycle  stream. In other cases, a portion  of the recycle stream is simply purged from the system
as a mixture of hydrogen, hydrogen sulfide and ammonia. The latter is less efficient since it
leads to higher levels of hydrogen sulfide and ammonia in the reactor, but it avoids the cost of
building and operating a hydrogen sulfide scrubber.

    Current desulfurization processes in the United States generally use only one reactor, due to
the  need to desulfurize diesel fuel only to 500 ppm or slightly lower. However, for diesel
upgrading reactions or for deeper desulfurization reactions, a second reactor can be used.
Instead  of liquid distillate fuel going to the diesel fuel/heating oil pool after the first reactor, it
would be stripped of hydrogen sulfide and ammonia and mixed with fresh hydrogen and sent to
the  second reactor, which is also called a second stage, after the inter stage stripping that occurs.

    Traditional  reactors are cocurrent in nature. The hydrogen is mixed together with the
distillate at the  entrance to the reactor and flow through the reactor together. Because the
reaction is exothermic, heat must be removed periodically. This is sometimes done through the
introduction of fresh hydrogen and distillate fuel in the middle of the reactor. The advantage of
cocurrent design is practical as it eases the control of gas-liquid mixing and contact with the
catalyst. The disadvantage is that the concentration of hydrogen is the highest at the front of the
reactor where the easiest to remove sulfur compounds are highest in concentration and lowest at
the  outlet where the hardest to remove sulfur compounds are highest in concentration. The
   c Refiners can choose to "upgrade" heavy refinery streams that do not meet the cetane and distillation
requirements for highway diesel fuel.  The process for doing so is also called ring opening, since one or more of the
aromatic rings of heavy, aromatic molecules are opened up, improving the value of the stream.  Upgrading the heavy
refinery streams to highway diesel fuel improves the stream's market price by 10-30 c/gal.

                                           5-9

-------
Final Regulatory Impact Analysis
opposite is true for the concentration of hydrogen sulfide.  This increases the difficulty of
achieving extremely low sulfur levels due to the low hydrogen concentration and high hydrogen
sulfide concentration at the end of the reactor.

   The normal solution to this problem is to design a counter-current reactor, where the fresh
hydrogen is introduced at one end of the reactor and the liquid distillate at the other end.  Here,
the hydrogen concentration is highest (and the hydrogen sulfide concentration is lowest) where
the reactor is trying to desulfurize the most difficult (sterically hindered) compounds. The
difficulty of counter-current designs in the case of distillate hydrotreating is vapor-liquid contact
and the prevention of hot spots within the reactor. The SynAlliance (Criterion Catalyst Corp.,
and Shell Oil Co.) has patented a counter-current reactor design called SynTechnology. With
this technology, in a single reactor design, the initial portion of the reactor will follow a co-
current design, while the last portion of the reactor will be counter-current.  In a two reactor
design, the first reactor could be co-current, while the second reactor could be counter-current.

   ABB Lummus estimates that the counter-current design can reduce the catalyst volume
needed to  achieve 97 percent desulfurization by 16 percent relative to a co-current design.10  The
impact of the counter-current design is even more significant when aromatics reduction (or
cetane improvement) is desired in addition to sulfur control.

   Sulfur-containing compounds in distillate can be classified according to the ease with which
they are desulfurized. Sulfur contained in paraffins or aromatics with a single aromatic ring are
relatively easy to  desulfurize. These molecules are sufficiently flexible so the sulfur atom is in a
geometric position where it can make physical contact with the surface of the catalyst.  The more
difficult compounds are contained in aromatics consisting of two aromatic rings, particularly
dibenzothiophenes.  Dibenzothiophene contains two benzene rings that are connected by a
carbon-carbon bond and two carbon-sulfur bonds (both benzene rings are bonded to the same
sulfur atom). This compound is nearly flat in nature and the carbon atoms bound to the sulfur
atom hinder the approach of the sulfur atom to the catalyst surface. Despite this, current
catalysts are very  effective in desulfurizing dibenzothiophenes, as long as only hydrogen is
attached to the carbon atoms bound directly to the sulfur atom.

   Distillate fuel, however, can contain dibenzothiophenes that have methyl or ethyl groups
bound to the carbon atoms, which are in turn bound to the sulfur atom.  These extra methyl or
ethyl groups further hinder the approach of the sulfur atom to the catalyst surface.
Dibenzothiophenes with such methyl or ethyl groups are commonly referred to as being
sterically hindered.  An example of a dibenzothiophene with a single methyl or ethyl group next
to the sulfur atom is 4-methyl dibenzothiophene. An example of a dibenzothiophene with two
methyl or  ethyl groups next to the sulfur atom is 4,6-dimethyl dibenzothiophene.  In 4,6-
dimethyl dibenzothiophene, and similar compounds, the presence of a methyl group on either
side  of the sulfur atom makes it very difficult for the sulfur atom to react with the catalyst
surface to assist the hydrogenation of the sulfur atom.

   Most straight-run distillates contain relatively low levels of these sterically hindered
compounds. LCO contains the greatest concentration of sterically hindered compounds, while

                                           5-10

-------
                                                              Fuel Standard Feasibility
other cracked distillate streams from the coker and the visbreaker contain levels of sterically
hindered compounds in concentrations between straight-run and LCO. Thus, LCO is generally
more difficult to desulfurize than coker distillate, which is in turn more difficult to treat than
straight-run distillate.11 In addition, cracked stocks, particularly LCO, have a greater tendency to
form coke on the catalyst, which deactivates the catalyst and requires its regeneration or
replacement.

   The greater presence of sterically hindered compounds in LCO is related to two fundamental
factors. First, LCO contains much higher concentrations of aromatics than typical straight run
distillate.12 All sterically hindered compounds are aromatics. Second, the chemical equilibria
existing in cracking reactions favors the production of sterically hindered dibenzothiophenes
over unsubstituted dibenzothiophenes. For example, in LCO, methyl substituted aromatics are
twice as prevalent as unsubstituted aromatics.  Di-methyl aromatics are twice as prevalent as
methyl aromatics, or four times more prevalent as unsubstituted aromatics. Generally,
desulfurizing 4-methyl dibenzothiophene using conventional desulfurization is six times slower
than desulfurizing similar non-sterically hindered molecules, while desulfurizing 4,6-dimethyl
dibenzothiophene using conventional desulfurization is 30 times slower.  Slower reactions mean
that either the volume of the reactor must be that much larger, or that the reaction must be
somehow speeded up. The latter implies either a more active catalyst, higher temperature, or
higher pressure. These alternatives are discussed below.

   Because moderate sulfur reduction is often all that is required in  current distillate
hydrotreating,  catalysts have been developed that focus almost exclusively on sulfur and other
contaminant removal, such as nitrogen and metals.  The most commonly used desulfurization
catalyst consists of a mixture of cobalt and molybdenum (Co/Mo). These catalysts interact
primarily with the sulfur atom and encourage the reaction of sulfur with hydrogen.

   Other catalysts have been developed that encourage the saturation (hydrogenation) of the
aromatic rings. As mentioned above, this generally improves the quality of the diesel fuel
produced from this distillate. These catalysts also indirectly  encourage the removal of sulfur
from sterically hindered compounds by eliminating one or both of the aromatic rings contained
in dibenzothiophene. Without one or both of the rings, the molecule is much more flexible and
the sulfur atom can reach the catalyst surface unhindered. Thus, the desulfurization rate of
sterically hindered compounds is greatly increased through the hydrogenation route. The most
commonly used hydrogenation/desulfurization catalyst consists of a mixture of nickel and
molybdenum (Ni/Mo).

   Several important issues related to using the hydrogenation pathway for desulfurization
should be highlighted. As pointed out above, one or both of the aromatics rings are being
saturated, which significantly increases the consumption of hydrogen. It is important that one of
the aromatic rings of a polyaromatic compound is saturated,  as this is the facilitating step
resulting in the desulfurization of a sterically hindered compound. If the mono aromatics
compounds are also saturated, there is only a modest improvement in the desulfurization reaction
rate of the sterically hindered compounds, however, at a large hydrogen cost.  In addition, certain
diesel fuel qualities, such as cetane, improve significantly as more of the aromatic compounds

                                          5-11

-------
Final Regulatory Impact Analysis
are saturated. However, the vendors of diesel desulfurization technology explained to us that if
cetane improvement is not a goal, then the most cost-effective path to desulfurize the sterically
hindered compounds is to saturate the polyaromatic compounds to monoaromatic compounds,
but not to saturate the monoaromatic compounds.  The vendors tell us that because the
concentration of the monoaromatic compounds is at equilibrium conditions within the reactor,
the monoaromatic compounds are being both saturated and unsaturated, which helps to enable
the desulfurization of these compounds. It also means that the concentration of aromatics can be
controlled by the reaction temperature and pressure.

   The vendors also point out a variety of reasons why the cycle length of the catalysts that
catalyze hydrogenation reactions, which likely occur in a second stage, is longer than the first
stage desulfurization catalyst.  First, the temperature at which the hydrogenation reactions occur
to saturate the polyaromatic compounds to monoaromatic compounds, but not to saturate the
monaromatic compounds, is significantly lower than the temperature of the first stage.  The
lower temperature avoids color change problems and reduces the amount of coke formation on
the hydrogenation catalyst. Furthermore, since the first stage has somewhat "cleaned" the diesel
fuel of contaminants such as sulfur, nitrogen and metals, the catalyst in this second
hydrogenation stage is not degraded as quickly.  Because the second stage has a cycle length as
long as or longer than the first stage,  adding the second stage is not expected to shorten the cycle
length of the current distillate hydrotreater.

   If refiners are "upgrading" their diesel fuel by converting heavy, high aromatic,  low cetane,
stocks to 15 ppm sulfur standard,  they are intentionally reacting a lot of hydrogen with the diesel
fuel. The hydrogen reactions with the diesel fuel saturates many or most of the aromatics,
increases cetane number and greatly eases the reduction of sulfur. The lower concentration of
aromatics and improved cetane of the upgraded feedstock then allows the product to be sold as
highway diesel fuel. The much higher sales price of the highway diesel fuel  compared with the
lower value of the feedstock justifies the much larger consumption in hydrogen and the cost of a
larger reactor.

   Up to a certain level of sulfur removal, the CoMo catalyst is generally preferred. It is more
active with respect to desulfurizing non-sterically hindered compounds, which comprise the bulk
of the sulfur in distillate, straight-run or cracked. Below that level, the need to desulfurize
sterically hindered compounds leads to greater interest in NiMo catalysts. Acreon Catalysts had
indicated that NiMo are preferred for deep desulfurization due to this catalyst's ability to saturate
aromatic rings and make the sulfur atom more accessible to the catalyst.  On  the other hand,
Haldor-Topsoe has performed studies indicating that CoMo catalysts may still have an advantage
over NiMo catalysts, even at sulfur levels below 50 ppm.13

   Two-stage processes may also be preferable to achieve ultra-low sulfur levels. Both stages
could emphasize desulfurization or desulfurization could be emphasized in the first stage and
hydrogenation/desulfurization emphasized in the second stage.  In addition to this advantage, the
main advantage of two stages lies in the removal of hydrogen sulfide from the gas phase after the
first  stage. Hydrogen sulfide inhibits desulfurization reactions, as discussed further in the next
section. It can also recombine with nonsulfur-containing hydrocarbon compounds at the end  of

                                          5-12

-------
                                                             Fuel Standard Feasibility
the reactor or even in subsequent piping, effectively adding sulfur to the desulfurized distillate.
Removing hydrogen sulfide after the first stage reduces the hydrogen sulfide concentration at the
end of the second stage by roughly two orders of magnitude, dramatically reducing both
inhibition and recombination.

   In one study, Haldor-Topsoe analyzed a specific desulfurized 50/50 blend of straight run
distillate and LCO at 150 ppm sulfur and found that nearly all the sulfur is contained in sterically
hindered compounds.14 This feed contains more LCO than would be processed in the typical
refinery. A refinery processing less LCO would presumably reach the point where the sulfur
compounds were dominated by sterically hindered compounds at a lower sulfur level.  They also
compared the performance of CoMo and NiMo catalysts on a straight run distillate feed at the
same space velocity.  The NiMo catalyst performed more poorly than the CoMo catalyst above
200 ppm sulfur, and better below that level. This implies that much of the sulfur left at 200 ppm
(and even above this level) was sterically hindered. These two studies indicate that the amount
of sterically hindered compounds can exceed the 15 ppm sulfur cap by a substantial margin.

   In addition to NiMo catalysts, precious metal catalysts are also very effective at desulfurizing
sterically hindered compounds. An example of a precious metal catalyst is the ASAT catalyst
developed by United Catalysts and Sud-Chemie AG, which uses both platinum and palladium.15
They are most commonly used to more severely dearomatize distillate and increase cetane by
opening up the aromatic rings, a process called ring opening.

   5.2.2.2 Meeting a 15 ppm Cap with Distillate Hydrotreating

   Using distillate hydrotreating to meet a 15 ppm sulfur cap on diesel fuel has been
commercially demonstrated. Thus,  meeting the  15  ppm cap is quite feasible using current
refining technology. Assessing the  most reliable and economic means of doing so is more
complicated. Refiners already hydrotreat their highway diesel fuel to meet a 500 ppm  sulfur cap.
These hydrotreaters use a variety of catalysts and have a range of excess capacity. Thus, refiners
are not all starting from the same place. Many refiners will also be producing heating oil, which
must meet only a 5000 ppm cap (lower in some states).  The high-sulfur heating oil may, for
example, provide a place to blend the sterically hindered sulfur-containing compounds. Finally,
the amount of cracked stocks that a  refiner processes into diesel fuel varies widely. Those with a
greater fraction of LCO will face a more difficult task of meeting a 15 ppm cap than those
processing primarily straight-run distillate.

   To understand the types of possible modifications to current distillate hydrotreating to
improve its performance, it is useful to better understand the quantitative relationships  between
the various physical  and chemical parameters involved in hydrotreating. Haldor-Topsoe has
developed the following algebraic expression to describe the rate of desulfurization via both
direct desulfurization and hydrogenation/desulfurization.
                                          5-13

-------
Final Regulatory Impact Analysis
Rate of          =          k x C" x pH2a	  +   k x Cm x pffib
Desulfurization              (1 + Kms x pms)       (1 + KF x CF)
Per Catalyst
Surface Area

Where:
    k, KH2S and KF are various rate constants, which vary only with temperature.
    Cs is the concentration of sulfur in the distillate.
    Pm and Pms are the partial pressures of hydrogen and hydrogen sulfide in the vapor phase.
    KF x CF is the total inhibition due to hydrogen sulfide, ammonia, and aromatics n, m, a, and b
       are various constant exponents.

    The first term represents the rate of direct desulfurization, such as that catalyzed by CoMo.
This reaction rate increased by increasing the partial pressure of hydrogen. However, it is
inhibited by increasing concentrations of hydrogen sulfide, which competes with the distillate for
sites on the catalyst surface.

    The second term represents the rate of desulfurization via hydrogenation of the aromatic ring
next to the sulfur atom. This rate of desulfurization also increases with higher hydrogen partial
pressure. However, this reaction is inhibited by hydrogen sulfide, ammonia, and aromatics. This
inhibition by aromatics leads to the presence of a thermodynamic equilibrium condition that can
prevent the complete saturation of aromatics.  Also, this inhibition makes it more difficult to
desulfurize cracked stocks, which contain high concentrations of both sterically hindered sulfur
compounds and aromatics. While the literature generally expresses a preference for NiMo
catalysts for desulfurizing cracked stocks, Haldor-Topsoe has found situations where this
aromatics inhibition leads them to favor CoMo catalysts even for desulfurizing feeds with a high
concentration of sterically hindered compounds.

    These relationships identify the types of changes that could improve the performance of
current distillate hydrotreaters. First, a more active catalyst can be used. This increases the "k"
terms in the above equations.  Second, temperature can be increased, which also increases the
"k" terms in the above equations. Third, improvements can often be made in vapor-liquid
contact, which effectively increases the surface area of the catalyst. Fourth, hydrogen purity can
be increased. This increases the hydrogen concentration, which the Pm term in the two
numerator terms of the equation.  Fifth, the concentration of hydrogen sulfide in the recycle
stream can be removed by scrubbing.  This decreases the Pms and CF terms in the two
denominator terms of the  equation. Finally, more volume of catalyst can be used, which
increases the surface area proportionally.

    Regarding catalysts, at least two firms have announced  the development of improved
catalysts since the time that most distillate hydrotreaters were built in the United States to meet
the 1993 500 ppm sulfur cap: Akzo Nobel /Nippon Ketjen Catalysts (Akzo Nobel) and Haldor-
Topsoe. Akzo Nobel currently markets four CoMo desulfurization catalysts: KF 752, KF 756
and KF 757, which have been available for several years, and KF 848, which was announced  in
2000.16 KF 752 can be considered to be typical of an Akzo Nobel catalyst  of the 1992-93 time

                                          5-14

-------
                                                             Fuel Standard Feasibility
frame, while KF 756 and 757 catalysts represent improvements. Akzo Nobel estimates that
under typical conditions (e.g., 500 ppm sulfur), KF 756 is 25 percent more active than KF 752,
while KF 757 is more than 50 percent more active than KF 752 and 30 percent more active than
KF 756.17  However, under more severe conditions (e.g., <50 ppm sulfur), KF 757 is 35-75
percent more active than KF 756. KF 848 is 15 - 50 percent more active than KF 757.
Commercial experience exists for both advanced catalysts. KF 756 is widely used in Europe (20
percent of all distillate hydrotreaters operating on January 1, 1998), while KF 757 has been used
in at least three hydrotreaters commercially. KF 757 and KF 842 utilizes what Akzo Nobel calls
STARS technology, .Super Type II Active Reaction Suites. Type II refers to a specific kind of
catalyst site that is particularly good at removing sulfur from sterically hindered compounds.

   In terms of sulfur removal, Akzo Nobel projects that a desulfurization unit producing 500
ppm sulfur with KF 752 will produce 405, 270 and 160 ppm sulfur with KF 756, KF757, and KF
842, respectively.

   In 2001 and 2003, Akzo Nobel announced two new catalysts. In 2001, Akzo announced the
introduction of a highly active catalyst named Nebula, which offers a different way to use
coatings for catalysts.  A typical catalyst is composed of two parts: an active coating containing
metals and a generally inactive substrate. For Nebula, Akzo Nobel concentrated the metal
coatings and omitted the substrate.  Because of the very high metals content, Nebula costs
several times more than conventional catalysts. The higher activity of the Nebula catalyst leads
to an increased tendency for coking, which must be countered by using a high hydrogen partial
pressure, resulting in a higher hydrogen consumption. (The hydrogen consumption is higher
because a higher percentage of the aromatics are saturated to nonaromatic compounds.)
According to Akzo Nobel, a refiner may be able to meet the 15 ppm sulfur standard by simply
replacing a part of or all of its existing catalyst with Nebula and providing significantly more
hydrogen (which may possibly require the addition of a hydrogen plant). Nebula may
significantly reduce the capital investment for meeting the 15 ppm sulfur standard.  In 2003,
Akzo announced that Nebula was modified somewhat to contain 15-20 percent less metals, but
with the same activity as the original Nebula. The updated Nebula catalyst, now called Nebula
20, can better handle heavier feeds.18

    In 2003, Akzo Nobel  announced a new catalyst named KF-760. The KF-760 catalyst is a
CoMo catalyst designed for better denitrogenation of diesel fuel, in addition  to the
desulfurization being sought after. Where the nitrogen content is inhibiting the desulfurization
of the diesel fuel, this catalyst can have 15-20 percent higher activity compared to their
previous best, KF-757, with only a modest increase in hydrogen consumption.19

   Haldor-Topsoe has also developed more active catalysts. Its TK-554 catalyst is analogous to
Akzo Nobel's KF 756 catalyst, while its newer, more active catalyst is termed TK-574. For
example, in pilot plant studies, under conditions where  TK-554 produces 400 ppm sulfur in
straight run distillate, TK  574 will produce 280 ppm. Under more severe conditions, TK-554
will produce 60 ppm, while TK 574 will produce 30 ppm.  Similar benefits are found with a
mixture of straight-run and cracked stocks. Just this year, Haldor Topsoe announced a new line
                                         5-15

-------
Final Regulatory Impact Analysis
of catalysts named Brim.20 The announcement did not include information about the
improvements of this line of catalysts over its previous catalysts.

   UOP projects a similar reduction in sulfur due to an improved catalyst. They estimate that a
hydrotreater producing 500 ppm sulfur distillate today (20 percent LCO, 10 percent light-coker
gas oil) could produce 280 ppm sulfur distillate with a 50 percent more active catalyst.21

   Over the last six years, Criterion Catalyst Company announced two new catalyst
technologies. One was called Century, and the other was called Centinel.22 These two lines of
catalysts were reported to be 45 to 70 percent and 80 percent more active, respectively,  at
desulfurizing petroleum fuel than conventional catalysts used in the mid-90s.  These
improvements have come about primarily through better dispersion of the active metal on the
catalyst substrate. Criterion announced a new line of catalyts in early 2004 named Ascent.23
These catalysts are expected to be at least 20 percent more active than the Centinel  line of
catalysts.24

   Axens catalysts, which is associated with IFF, offers three catalysts designed for deep
desulfurization of distillate fuel.  One is a CoMo catalyst named HR 406 and it is reported to be
40 percent more  active than HR 306, its predecessor. Another catalyst offered by Axens is
named HR 468 and it offers a  mixture of CoMo with NiMo metals.  The third catalyst offered by
Axens is a NiMo catalyst named HR 448.  The NiMo catalyst is recommended for deep
desulfurization at higher pressures, while HR 468 is more recommended for use at lower
pressures.25

   This shows that changing to a more active catalyst, by itself, can reduce sulfur significantly.
Based on the history of the industry, improvements in catalyst performance can be anticipated
over time to result in roughly a 25 percent increase in catalyst activity every four years. Vendors
have informed us that the cost of these advanced catalysts is very modest relative to less active
catalysts.  BP-Amoco projects that a 70 percent improvement in catalyst activity could reduce
sulfur from a current hydrotreater meeting a 500 ppm sulfur specification to 30 ppm.26
Acreon/IFP/Procatalyse is not optimistic, however, that a catalyst change alone will enable
refiners to meet this sulfur level.27 Improved catalysts will, however, reduce the reactor size
needed for achieving the target sulfur level compared to a less active catalyst.

   The second way to improve the hydrotreating of diesel fuel for deeper desulfurization is to
reduce the concentration of hydrogen sulfide, which reduces the inhibition of the desulfurization
and hydrogenation reactions.  Hydrogen sulfide can be removed by chemical scrubbing. Haldor-
Topsoe indicates that decreasing the concentration of hydrogen sulfide at the inlet to a co-current
reactor by three to six volume percent can decrease the average temperature needed to achieve a
specific sulfur reduction by 15-20°C, or reduce final sulfur levels by more than two-thirds. UOP
projects that scrubbing hydrogen sulfide from recycled hydrogen can reduce sulfur  levels from
roughly 285 to 180 ppm  in an existing hydrotreater.

   The third type of improvement to  current distillate hydrotreating is to improve vapor-liquid
contact.  Akzo Nobel estimates that an improved vapor-liquid distributor can reduce the

                                          5-16

-------
                                                             Fuel Standard Feasibility
temperature necessary to meet a 50 ppm sulfur level by 10 °C, which would in turn increase
catalyst life and allow an increase in cycle length from 10 to 18 months. Based on the above
data from Haldor-Topsoe, if temperature were maintained, the final sulfur level could be reduced
by 50 percent. Similarly, in testing of an improved vapor-liquid distributor in commercial use,
Haldor-Topsoe found that the new distributor allowed a 30 percent higher-sulfur feed to be
processed at 25°C lower temperatures, while reducing the sulfur content of the product from 500
to 350 ppm. Maintaining temperature should have allowed an additional reduction in sulfur of
more than two-thirds.  Thus, ensuring adequate vapor-liquid contact can have a major impact on
final sulfur levels.

    The fourth type of improvement possible is to increase hydrogen partial pressure and/or
purity. As  discussed above, this increases the rate of both desulfurization and hydrogenation
reactions. Haldor-Topsoe indicates that increasing hydrogen purity is preferable to a simple
increase in  the pressure of the hydrogen feed gas, since the latter will also increase the partial
pressure of hydrogen sulfide later in the process, which inhibits both beneficial reactions.
Haldor-Topsoe projects that an increase in hydrogen purity of 30 percent lowers the temperature
needed to achieve the same sulfur removal rate by 8 to 9°C.  Alternatively, temperature could be
maintained while increasing the amount of sulfur removed by roughly 40 percent.  Hydrogen
purity can be increased through the use of a membrane separation system or a PSA unit. UOP
projects that purifying hydrogen can reduce distillate sulfur from 180 to 140 ppm from an
existing hydrotreater.

    The fifth type of improvement is to increase reactor temperature. Haldor-Topsoe has shown
that an increase of 14°C while processing a mix of straight run distillate and LCO with its
advanced TK-574 CoMo catalyst will reduce sulfur from 120 ppm to 40 ppm.28  UOP projects
that a 20  °F increase in reactor temperature would decrease sulfur from 140 to 120 ppm. The
downside of increased temperature is reduced catalyst life (i.e., the need to change catalyst more
frequently). This increases the cost of catalyst, as well as affects highway diesel fuel production
while the unit is down for the catalyst change.  Still, current catalyst life currently  ranges from 6
to 60 months, so some refiners  could increase temperature and still remain well within the range
of current industry performance. The relationship between temperature and life of a catalyst is a
primary criterion affecting  its marketability.  Thus, catalyst suppliers generally do not publish
these figures.

    Sixth, additional sulfur can  be removed by increasing the amount of recycle gas sent to the
inlet of the  reactor. However, the effect is relatively small. Haldor-Topsoe indicates that a 50
percent increase in the ratio of total gas/liquid ratio decreases the necessary reactor temperature
only by 6 to 8°C. Or, temperature  can be maintained and the final sulfur level reduced by 35 to
45 percent.

    Overall, Akzo-Nobel projects that current hydrotreaters can be modified short  of a revamp
with a second reactor to achieve 50 ppm sulfur. While this improvement is somewhat greater
than the 50 percent improvement measured by Akzo Nobel at current desulfurization severity, it
indicates that it may be possible to improve current hydrotreaters to produce distillate sulfur
levels in the 50-100 ppm range. Thus, it appears that additional measures would be needed to

                                          5-17

-------
Final Regulatory Impact Analysis
meet a 15 ppm cap.  This leads to the seventh means to realize deeper desulfurization, which is
to increase catalyst volume through the addition of a second reactor.  UOP projects that doubling
the catalysts volume by adding another reactor would reduce sulfur from 120 to 30 ppm. For
each refinery, refiners would need to examine how much additional sulfur control they would be
able to achieve through measures one through six, and then size this second reactor to achieve
the 15 ppm sulfur cap.

   These individual improvements described cannot be simply combined, either additively or
multiplicatively. As mentioned earlier, each existing distillate hydrotreater is unique in its
combination of design, catalyst, feedstock, and operating conditions. While the improvements
described above can be made in many cases, it is not likely that all the improvements mentioned
are applicable to any one unit; the degree of improvement could either be greater than or less
than the benefits indicated.

   Some refiners may therefore have to implement one additional technical change listed by
UOP to be able to meet the 15 ppm standard.  This last technical change is to add a second stage
to current single-stage 500 ppm hydrotreaters. This second stage would consist of a second
reactor, and a high pressure, hydrogen sulfide scrubber between the first and second reactor. The
compressor would also be upgraded to allow the new second reactor to be operated at a higher
pressure. Assuming use of the most active catalysts available in both reactors,  UOP projects that
converting from a one-stage to a two-stage hydrotreater could produce  5 ppm sulfur relative to a
current level of 500 ppm today.

   In addition to these major technological options, refiners may have to debottleneck or add
other more minor units to support the new desulfurization unit.  These units could include
hydrogen plants, sulfur recovery plants, amine plants and sour water scrubbing facilities. All
these units are already operating in refineries but may have to be expanded or enlarged.

    To assess the degree that these measures would be needed, it is useful to examine the
commercial and pilot plant performance of distillate hydrotreating to achieve very low sulfur
levels.

   5.2.2.3  Low-Sulfur Performance of Distillate Hydrotreating

   Data  from both pilot plant studies and commercial  performance are available indicating the
capability of various hydrotreating technologies to reduce distillate sulfur levels to very low
levels.  While many reports of existing commercial operations focus on reducing sulfur to meet a
500 ppm sulfur standard, or somewhat below that sulfur level, studies of achieving lower sulfur
levels (e.g., 10 to 50 ppm) are associated with also reducing aromatic content significantly.  This
combination is related to the  fact that Swedish Class II diesel fuel must meet a tight aromatics
specification in 2005 along with a 10 ppm sulfur standard.  Other European diesel fuel must also
meet a 10 ppm sulfur standard.

   Another study projected the technology and resulting cost to reduce diesel fuel sulfur to
comply with EPA's highway 15 ppm sulfur cap standard and sulfur standards on nonhighway

                                          5-18

-------
                                                              Fuel Standard Feasibility
distillate . The Engine Manufacturers Association retained Mathpro for this study. The
projections of this study are discussed in Chapter 7.  The discussion in this chapter will focus on
the available pilot plant and commercial data demonstrating the achievement of low sulfur
levels. It is worth noting that until the 15 ppm standard was established for highway diesel fuel
in the United States and the announcements by the German government to seek sulfur levels as
low as 10 ppm, there had been little effort by industry to develop technology capable of such a
level across the diesel  pool. Recent advances by catalyst manufacturers demonstrating the
feasibility of producing diesel fuel meeting these levels through pilot plant testing and some
commercial demonstrations should be considered a first-generation of technology, with new and
continual advances expected over time.

       As of mid 2003, Criterion Centinel and SynCat™ catalysts were installed in 37 deep
desulfurization units in operation in the World, including 13 Syn Technology Units. While the
purpose for each unit is to desulfurize distillate to 50 ppm or below, eight of them served as a
first stage of a two stage dearomatization type unit where ULSD was capable of being produced.
(Lummus' licensed SynTechnology).

       The other 24 hydroprocessing units operating with Criterion's Centinel's catalysts are
desulfurizing distillate down to under 50 ppm sulfur, with 6 of these consistently under 15 ppm.

       IFF, using Axens catalysts, offers its Prime D technology for deep desulfurization,
aromatics saturation and cetane improvement.29 Using a NiMo catalyst, IFP's Prime D process
can produce distillate sulfur levels of 10 ppm from straight run distillate and of less than 20 ppm
from distillate containing 20 to 100 percent cracked material using a single-stage reactor.  With a
two-stage process, less than one ppm sulfur can be achieved.

       United Catalysts and Sud-Chemie AG have published data on the performance of their
AS AT catalyst, which uses platinum and palladium.30 The focus of their study was to reduce
aromatics to less than  10 volume percent starting with a feed distillate containing up to 500 ppm
sulfur and at least 100  ppm nitrogen. Starting with a feed distillate containing 400 ppm sulfur
and 127 ppm nitrogen  and 42.5 volume percent aromatics, the ASAT catalyst was able to reduce
sulfur to eight to nine ppm, nearly eliminate nitrogen, and reduce aromatics to two to five
volume percent.  Hydrogen consumption was 800 to 971 standard cubic feet per barrel (SCFB).

       Akzo Nobel has summarized the commercial experience of about a year's worth of
operations of their STARS catalyst for desulfurizing diesel fuel at the BP-Amoco refinery in
Grangemouth, UK.31 The original unit was designed to produce 35,000 barrels per day of diesel
fuel at 500 ppm treating mostly straight-run material, but some LCO was treated as well. Akzo
Nobel's newest and best catalyst (KF 757 at that time) was dense-loaded into the reactor to
produce 45,000 barrels per day diesel fuel at 10-20 ppm (to meet the 50 ppm standard).0 From
the data, it was clear to see that as the space velocity changed, the sulfur level changed inversely
   D   Dense loading is a process of packing a certain volume of catalyst into a smaller space than conventional
catalyst loading.

                                          5-19

-------
Final Regulatory Impact Analysis
proportional to the change in space velocity.  Usually when the space velocity dipped below 1.0,
the sulfur level dropped below 10 ppm. At that refinery, however, it was not necessary to
maintain the sulfur level below 10 ppm.

       Akzo Nobel also has its STARS catalysts operating in four other units in Europe and the
Middle East, three of which are producing diesel fuel with less than 10 ppm sulfur, and another
unit producing diesel fuel with less than 20 ppm sulfur. Three of these units process a blend of
light and heavy straight run feeds, while the other is processing a stream which is predominantly
comprised of cracked stocks.  Additionally, Akzo Nobel is demonstrating its Nebula catalysts
commercially in three different applications in Europe producing diesel fuel ranging from 5 ppm
to 50 ppm.  One of those is for treating cracked stocks in addition to straight run, and the refinery
is meeting a 10 ppm standard  at 650 psi partial pressure.

       Haldor Topsoe has their catalysts in 27 units worldwide, either as standalone
desulfurization units or the first stage of a desulfurization and dearomatization unit, producing
diesel fuel to under 50 ppm sulfur. While most of these are in Europe, some are also in the U.S.
Of these, 17 are producing diesel fuel to under 10 ppm  sulfur; some of these have cracked stocks
while others do not.

       Based on all this laboratory and real world experience, it is clearly feasible to produce
diesel fuel with a sulfur level of 15 ppm or less even if the feedstocks contain a great deal of
cracked stocks. The challenge refiners will face is how to minimize the cost of doing so. To
minimize costs, refiners will have to figure out how to apply the desulfurization/hydrogenation
methods on their own diesel fuels. The specifics, and thus the economics, of accomplishing this
depends on the amount of cracked stocks that the refiner blends into diesel fuel. A few refiners
have the possibility of shifting some of the sterically hindered compounds to fuels complying
with less stringent sulfur standards, such as heating oil. However, our analysis of the feasibility
of desulfurization technology  did not  consider the occurrence of feedstock shifting as necessary
for refiners to meet the diesel  sulfur standards.

5.2.3 Process Dynamics Isotherming

       In the late 1990s, a professor at the University of Arkansas applied some ingenuity in
reaction chemistry to diesel desulfurization.  After conceiving of this process, he started a
company named Process Dynamics.  The reaction technology reacts diesel fuel with hydrogen,
which is totally dissolved in the diesel fuel, in a plug flow reactor. Since the hydrogen gas is
dissolved into the diesel fuel, the reactor needs to be designed only to handle a liquid, instead of
the two phase reactors designed for conventional hydrotreating.  Because only about 75  standard
cubic feet of hydrogen can be dissolved into each barrel of diesel fuel and the hydrogen
consumption for a particular desulfurization step can be much higher than that, this technology
cannot be a once-through process. Process Dynamics solved that limitation by recycling the feed
after a very short residence time in the reactor to recharge the liquid with more hydrogen and to
mix this recycle with some untreated diesel fuel before sending it to the reactor. Thus, the
recycled partially desulfurized diesel fuel acts like a diluent to the fresh feed controlling the
hydrogen consumption, and the diesel fuel is recharged with hydrogen and sent to the reactor to

                                          5-20

-------
                                                              Fuel Standard Feasibility
be desulfurized several times as it is being treated.32 33

       The Process Dynamics Isotherming process has some apparent advantages over
conventional desulfurization.  First, since the hydrogen is already in the liquid phase, the
hydrotreating reaction can occur much more quickly, because, as described by Process
Dynamics, the kinetics of conventional hydrotreating are mass transfer-limited, which is the rate
at which gaseous hydrogen can transfer into the liquid phase. Process Dynamics makes this
point by the following reaction equations for hydrotreating diesel fuel:34

rg = kg (PH2 - P*H2) (rate of hydrogen mass transfer into the liquid phase)

Where:
       rg  = transfer rate of hydrogen gas into diesel fuel.
       kg = hydrogen gas mass transfer rate.
       PH2 = Partial pressure of hydrogen in the gas phase.
       PxH2 = Partial pressure of hydrogen at the catalyst.

and
rs = ks T[S][P>
-------
Final Regulatory Impact Analysis
       There are a two important benefits to the Process Dynamics process because it has a
higher space velocity. One benefit is that the Process Dynamics process requires a smaller
amount of catalyst. By definition, if the same volume of feed can be treated faster than another
process, the amount of catalyst needed is proportionally lower by the inverse proportion of the
space velocity.  The second advantage of having a faster space velocity is that the reactors are
sized much smaller to hold the lower volume of catalyst. Both of these benefits result in lower
costs for the Process Dynamics Isotherming deslfurization process.  The lower catalyst volume
required by Process Dynamics Isotherming costs proportionally less because the Process
Dynamics desulfurization process uses the same catalysts as conventional hydrotreating.
Similarly, the smaller reactor volume reduces the capital costs, although in this case the cost
reduction is not necessarily proportionally less as smaller reactors have a poorer economy of
scale compared with larger reactors.

       The Process Dynamics engineers point out that the Isotherming process also has other
benefits over conventional hydrotreating. When some of the aromatics in diesel fuel are
saturated during the desulfurization process, heat is generated. In the case of conventional
hydrotreating, much of this heat is intentionally quenched away in an attempt to avoid excessive
temperature excursions.  Excessive temperature excursions and local low hydrogen
concentration can lead to coking,  which is a constant problem  with conventional hydrotreating.
However, the higher space velocity of the Process Dynamics process coupled with the fact that
the feed is diluted by  the recycle stream allows for better control of the process temperature.
Furthermore, the ready availability of hydrogen in the liquid phase along with the better
temperature control prevents most of the coking from occurring. Thus, the internally generated
heat can be conserved, instead of being quenched away, and used to heat the process. The
conserved heat means that little to no external heating is required, which provides a savings in
natural gas consumption relative to conventional hydrotreating. However, a small heater is still
needed to heat the feed during start-up.

       Another advantage of the Process Dynamics desulfurization process is that it does not
need a hydrogen gas recycle compressor. Because the hydrogen pumped into solution and going
to the reactor is either used up or remains in solution, there is no residual hydrogen gas to
recycle. Compressors operating at the pressures that diesel fuel desulfurization occurs are
expensive, long leadtime delivery items.  Thus, by omitting the recycle gas compressor and using
smaller reactors, the Process Dynamics desulfurization process not only saves substantial capital
costs compared with conventional hydrotreating, but it also means a somewhat shorter
construction time. The smaller reactors and heater coupled with the fact that a recycle gas
compressor is not needed means that the Process Dynamics process requires a smaller footprint
compared with conventional hydrotreating, facilitating the installation of the Process Dynamics
unit in today's refineries which are often space-limited.

       While aspects of the Process Dynamics Isotherming desulfurization process for diesel
fuel desulfurization are novel compared with conventional  diesel desulfurization, many aspects
of the process are the same. Much of the list of required equipment is the same for the Process
Dynamics process as  for conventional hydrotreating. Table 5.2-1 shows both the similarities and
differences between the two.

                                          5-22

-------
                                                             Fuel Standard Feasibility
                                       Table 5.2-1
 Major Equipment Needed for Process Dynamics Isotherming and Conventional Hydrotreating

Heat Exchangers
Heater
Hydrogen gas compressor
Mixers for dissolving hydrogen into
the diesel fuel
Reactor (s)
Reactor distributor
High-pressure flash drum and
hydrogen separator
Low-pressure separator
Recycle hydrogen compressor
Recycle hydrogen gas scrubber
Process Dynamics Isotherming
Yes
Yes (small and for startup only)
Yes
Yes
Yes (2 - 4 small plug flow)
No
Yes
Yes
No
No
Conventional Hydrotreating
Yes
Yes
Yes (for hydrogen makeup)
No
Yes (1-2 large trickle bed)
Yes
Yes
Yes
Yes
Yes
       Process Dynamics has accumulated some data on the Isotherming desulfurization process
from testing they have done with their pilot plant. Process Dynamics started up a pilot plant in
late 2001. Recently, Process Dynamics installed a commercial demonstration unit of their
technology at a Giant refinery as a revamp to an existing highway hydrotreater to demonstrate
compliance with the highway diesel fuel 15 ppm sulfur standard, which begins in mid 2006.  The
unit was started up in September of 2002 and the Process Dynamics engineers have been
working with the refinery engineers to optimize the unit for the refinery. Since early 2003, the
Process Dynamics demonstration unit has consistently been producing diesel fuel under 15 ppm.

       After successful demonstration of its technology at the Giant refinery, Process Dynamics
is working on signing license agreements for the Process Dynamics desulfurization process.  In
early 2004, Process Dynamics was working on signing four additional license agreements  here in
the U.S.35

5.2.4 Phillips S-Zorb Sulfur Adsorption

       A prospective diesel desulfurization process was announced by Phillips Petroleum  in late
2001.36 This process is an extension of their S-Zorb process for gasoline and thus is called S-
Zorb for diesel fuel. The process is very different from conventional diesel fuel hydrotreating in
which reacts the sulfur with hydrogen over a catalyst to form H2S. The S-Zorb process adsorbs
the sulfur molecule, still  attached to the hydrocarbon, onto a sorbent at a pressure of 275 to 500
pounds per square inch gauge (psig) and at a temperature of 700 to 800° F and in the presence of
hydrogen in the S-Zorb reactor.  The catalyst activity of the sorbent next cleaves the sulfur atom
                                          5-23

-------
Final Regulatory Impact Analysis
from the sulfur-containing hydrocarbon. To prevent the accumulation of sulfur on the catalyst,
the sulfur containing sorbent is continually removed from the reactor. The removed sorbent is
moved over to a receiving vessel by an inert lift gas, at which point the lift gas and the entrained
diesel fuel is removed from the sorbent. The sorbent next drops down into a lockhopper that
facilitates the movement of the sorbent to the regenerator. In the regeneration vessel, the sulfur
is burned off of the sorbent with oxygen and the generated SO2 is sent to the sulfur plant. The
regenerated sorbent then drops down into a reducer vessel where the sorbent is returned back to
its active state.  The sorbent is then recycled back to the reactor for removing more sulfur.
Because the catalyst is continuously being regenerated, Phillips estimates that the unit will be
able to operate four to five years between shutdowns. Because untreated distillate can contain
several percent sulfur, Phillips believes that its S-Zorb process for diesel could be overwhelmed
by the amount of sulfur adsorbing onto the catalyst.  Thus, the S-Zorb process may not be able to
economically treat all untreated distillate streams that are high in sulfur, and is best suited to treat
distillate containing 500 ppm sulfur or less. However, some refiners running sweet crudes and
producing low-sulfur non-highway diesel volumes (from straight-run diesel and hydrocrackate
diesel) may have lower uncontrolled nonhighway sulfur levels. These refiners may be able to
use the S-Zorb process to lower their nonhighway diesel sulfur.

       Phillips' S-Zorb  diesel desulfurization process has been demonstrated in a pilot plant that
started up in early 2002. This pilot plant has provided Phillips data on how the unit will process
varying formulations of diesel fuel or diesel fuel blendstocks. The pilot plant testing data
released by Phillips has  shown that diesel fuels blended with LCO can be desulfurized below 5
ppm.  Phillips has also shown that straight-run diesel fuel can be desulfurized below measurable
levels and a 100 percent LCO stream can be desulfurized down to 10 ppm.

       While the S-Zorb diesel desulfurization process has not been demonstrated commercially,
Phillips has demonstrated the S-Zorb technology for desulfurizing gasoline.  An S-Zorb gasoline
desulfurization unit started up at Phillips' Borger refinery in April of 2001. According to
Phillips, their gasoline desulfurization unit has operated as designed for the past three years. The
successful demonstration of their gasoline desulfurization unit at Borger has interested many
refiners in using S-Zorb gasoline desulfurization process for complying with the Tier 2 gasoline
sulfur program.E Phillips shared with us in late 2003 that they have licensed their S-Zorb for
gasoline processing for installation in 23 refineries in North America. That the Borger S-Zorb
gasoline desulfurization unit has operated as designed and that there are 23 new S-Zorb gasoline
units planned to start up demonstrates that there is agreement within the refining industry that the
S-Zorb process works.

       Most refiners, however, are very conservative and will not be willing to rely only on pilot
plant testing or demonstration of a technology for another fuel as the basis for purchasing a
desulfurization unit that costs tens of millions of dollars. They will want to see a particular
technology operating as a commercial unit for desulfurizing diesel fuel for at least two years
   E Starting this year, many refiners will be starting up their gasoline desulfurization units for complying with the
30 ppm Tier 2 gasoline sulfur standard, which phases in from 2004 to 2006.

                                           5-24

-------
                                                              Fuel Standard Feasibility
before trusting that the technology is reliable.  However, Phillips is not planning to install a
commercial demonstration unit of its S-Zorb diesel fuel desulfurization process, nor is Phillips
planning on installing an S-Zorb for diesel unit for complying with the 15 ppm sulfur highway
diesel fuel standard, which begins to take effect in mid-2006, in any of its refineries.37
Consequently, even though S Zorb for diesel may be capable of desulfurizing diesel fuel to less
than 15 ppm sulfur, it does not appear that it will factor into the mix of technologies used to meet
the NRLM 15 ppm diesel fuel standards.

5.2.5 Chemical Oxidation  and Extraction

       Another desulfurization technology being developed by Unipure and UOP is based on
chemical oxidation. For these companies, the  chemical oxidation desulfurization of diesel fuel is
accomplished by first forming a water  emulsion with the diesel fuel. In the emulsion, the sulfur
atom is oxidized to a sulfone using a strong oxidizing agent, such as catalyzed peroxyacetic acid.
With an oxygen atom attached to the sulfur atom, the sulfur-containing hydrocarbon molecules
become polar and hydrophilic and then move into the aqueous phase. These sulfone compounds
can either be  desulfurized or perhaps be converted to a surfactant that could be sold to the soap
industry at an economically desirable price.  The earnings made from the sales of the surfactant
could offset much of the cost of oxidative desulfurization.

       Unipure has set up a 50 barrel per day pilot plant which started operating in the spring of
2003. UOP is still developing its oxidation technology in the lab. Neither of these oxidation
processes are available for licensing at this time.

       Late in the 1990s, Petrostar had started the development of an oxidation process for
desulfurizing diesel fuel.  This oxidation technology was  similar to that of Unipure's.  However,
sometime in the last year Petrostar abandoned  its work on that technology. Early in 2003,
Lyondell-Citgo announced that they had recently  developed a chemical oxidation desulfurization
technology. This process is similar in  some ways to Unipure's and Petrostar's oxidation
processes, but also different in some pronounced ways. The differences are that instead of the
using expensive peroxyacetic acid to create sulfones, this process uses t-butyl hydroperoxide
oxidant to convert sulfur species in diesel to sulfones (this eliminates the need to recycle a co-
oxidant acid). T- butyl hydroperoxide  is not as corrosive as peroxyacetic acid, thus Lyondell's
process is projected to be constructed from less expensive metallurgy.  Lyondell has pilot plant
success desulfurizing 500 ppm diesel fuel to less than 10 ppm, but abandoned further
development  of this technology in late  2003.

       The best opportunity for oxidation and extraction technologies to penetrate the
desulfurization market my lie with smaller refineries and terminals.  Terminals may find that it is
cheaper to implement some sort of desulfurization technology to handle the overproduction of
off spec downgrade and interface than  it would be to ship it off to the nearest entity equipped to
distill and hydrotreat this material. Many small refineries and terminals don't have access to a
cheap source  of hydrogen and may not have sulfur plants, so having a technology which can treat
their distillate material without the need to install grassroots hydrogen units and sulfur plants
could make the costs associated with desulfurization reasonable to them.

                                           5-25

-------
Final Regulatory Impact Analysis
5.2.6 FCC Feed Hydrotreating

       As described earlier in this section, sulfur can be removed from distillate material early
or late in the refining process. Early in the process, the most practical place to remove sulfur is
before the FCC unit.  The FCC unit primarily produces gasoline, but it also produces a
significant quantity of LCO which makes up 23% of diesel fuel supply in the U.S.

       Many refineries already have an FCC feed hydrotreating unit.  The LCO from these
refineries should contain a much lower concentration of sulfur and fewer sterically hindered
compounds than refineries not hydrotreating their FCC feed.  Adding an FCC feed hydrotreating
is much more costly than distillate hydrotreating. Just on the basis of sulfur removal,  FCC feed
hydrotreating is more costly than distillate hydrotreating, even considering the need to reduce
gasoline sulfur concentrations, as well. This is partly due to the fact that FCC feed hydrotreating
by itself is generally not capable of reducing the level of diesel fuel sulfur to those being
considered in this rule, so post-treating is still necessary.  However, FCC feed hydrotreating
provides other environmental and economic benefits. FCC feed hydrotreating decreases the
sulfur content of gasoline significantly, as well as reducing sulfur oxide emissions from the FCC
unit. It also increases the yield of relatively high value gasoline and LPG from the FCC unit and
reduces the formation of coke on the FCC catalyst.  For individual refiners, these additional
benefits may offset enough of the cost of FCC hydrotreating to make it more economical than
distillate hydrotreating. However, these benefits are difficult to estimate in a nationwide study
such as this. Also, feed hydrotreating is not expected to, by itself, enable a refinery to meet
either the 500 or the 15 ppm standards. Thus, this study will rely on distillate hydrotreating as
the primary means with which refiners will meet the 15 ppm sulfur cap.  For those refiners that
choose FCC feed hydrotreating, their costs will presumably be lower than distillate hydrotreating
and the costs estimated in Chapter 7 can then be considered somewhat conservative in this
respect.

5.3  Feasibility of Producing 500 ppm Sulfur NRLM Diesel Fuel in 2007

5.3.1 Expected use of Desulfurization Technologies for 2007

       To enable our determination of whether it is feasible for the refining industry to meet the
2007 sulfur cap and to estimate the cost of complying with the sulfur standard (see Chapter 7),
we needed to project the mix of available technologies that will be used for compliance. We
considered several different factors for projecting the mix of technologies. First and foremost,
we considered the time refiners will have to choose a new technology, which is important
because of the relatively short lead time before implementation of the 500 ppm standard.
Second, we considered whether the technology will be available for 2007 and, if the technology
is available, how proven it is.  Third, we considered whether the technology is cost-competitive
by comparing it with other technologies.  If a refiner finds that a technology is available at a
lower cost, it is more likely to use that technology. We also considered whether the technology
is available from a vendor that has proven itself to the industry by providing other successful
refining technologies and particularly if the vendor has proven itself in the United States.
                                          5-26

-------
                                                              Fuel Standard Feasibility
Finally, we considered the capability of the vendor to meet the demand of the industry.  We
considered all these issues for each technology but, as described below, some of these issues are
more prominent than others.

       To comply with the 500 ppm sulfur standard in 2007, refiners will have to decide what
technology they will want to use several years before the standard needs to be met.  Several
years are needed to perform a preliminary design, complete a detailed design, purchase the
hardware needed, obtain the air quality permits needed, and then install and start up the
hardware. The timing of this final rule provides refiners three full years to comply with the 500
ppm sulfur standard. Because refiners need about three years to complete the mentioned steps to
have a working new unit, there is little time to shop around for a new desulfurization technology
that is just beginning to prove itself. A thorough review of a newer technology can take months,
so if refiners do not have this extra time, they will tend toward technologies that are more
familiar. See Section 5.3.2 for a more detailed discussion about the lead-time issues for the 2007
standard.

       Of the various technologies we list above for desulfurizing diesel fuel, conventional
hydrotreating is by far the most familiar to refiners.  Refiners are using conventional
hydrotreating to meet the current highway diesel fuel 500 ppm sulfur standard. In the United
States, there are about 90 distillate hydrotreaters with virtually all of them being conventional
hydrotreaters operating since 1993  or before. The one exception is a Process Dynamics
Isotherming commercial demonstration unit that started up recently at a Giant refinery in New
Mexico. Phillips S-Zorb for diesel and the two oxidation and extraction technologies have yet to
accumulate commercial experience. However, refiners usually want to see that a refinery unit
has operated successfully for at least two years to ensure that it will operate with high reliability
and low maintenance requirements.11 The Process Dynamics desulfurization unit that is installed
now and has started to accrue valuable commercial experience will have accumulated somewhat
less than two years of commercial experience by then.

       After considering the above issues, it seems that the short lead time is the central issue of
whether refiners will choose between conventional hydrotreating and other advanced
desulfurization technologies for 2007.  Refiners  do not have the many months needed to
carefully consider the advanced technologies still in development and still at the beginning of the
demonstration stage, so we believe this issue is the most critical one affecting refiners' choice of
desulfurization technologies for 2007.  For these reasons, we believe refiners will default to what
they know will work, which is conventional desulfurization. Since multiple vendors can provide
the preliminary engineering design and any followup support for conventional hydrotreating,
these vendors will be able to serve the refiners needing to install desulfurization units for 2007.
   F Refiners want low-maintenance refining units because they have cut back their engineering staff to reduce
their refining costs for improving their margins, and thus will seek units consistent with that strategy.

                                           5-27

-------
Final Regulatory Impact Analysis
5.3.2 Lead-time Evaluation

       Refiners need sufficient lead time to design, construct, and start up desulfurization
technology to meet the 500 ppm standard if this standard is to be implemented in an orderly way.
If one or more refiners were unable to comply in time, it would have major repercussions for the
refiner and potentially for the regional fuel supply.  If refiners planning on producing 500 ppm
NRLM fuel could not do so in time and could not buy credits, they would have to sell their high-
sulfur distillate fuel as heating oil, export it, or temporarily cease production. As discussed in
Section 5.8, heating oil will no longer be widely distributed in many markets.  Thus, selling large
quantities of heating oil may require distressed pricing and the absorption of trucking costs.
Exportation would be very  costly for refiners not located on an ocean coastline. Temporary
closure would result in serious financial loss.  In addition, users of NRLM diesel fuel would
likely face high fuel prices. Fuel prices respond quickly to supply shortages. Significant price
increases would be expected if refiners were not able to fulfill demand for NRLM diesel fuel
starting June  1, 2007. Thus, providing adequate lead time for refiners to design, construct, and
prove out the necessary new hydrotreaters is critical to avoiding serious economic harm to both
the refiners and the users of NRLM diesel fuel.

       Because of this, we project that refiners will use conventional hydrotreating to meet the
500 ppm standard beginning on June 1, 2007.  Of the 35 refineries projected to produce 500 ppm
NRLM diesel fuel beginning in 2007, 8 are projected to do so by using recently idled highway
diesel fuel hydrotreaters. These refineries are expected to idle their highway hydrotreaters in
response to exiting the highway market or by installing a new grassroots diesel fuel hydrotreater.
The remaining 27 refineries would need to design and construct a new hydrotreater to produce
500 ppm NRLM fuel.0  This is roughly 20 percent of all U.S. refineries currently producing
transportation fuels.  Thus,  the time available between the date  of the final rule and June 1, 2007
must be sufficient across a wide spectrum of refiners and situations.

       We have conducted two lead-time assessments for the refining industry in the past four
years.  One assessment supported the Tier 2 gasoline sulfur program.11  The other assessment was
part of our review of progress being made towards compliance with the 15 ppm sulfur standard
for the highway diesel fuel  program.1 The results of both of these assessments are reviewed
below and then applied to the new NRLM sulfur control program.
    G Without the small-refiner provisions, an additional 20 refineries would have to produce 500 ppm NRLM fuel
by June 1, 2007.

    H Final Regulatory Impact Analysis, Control of Air Pollution from New Motor Vehicles: Tier 2 Motor Vehicle
Emissions Standards and Gasoline Sulfur Control Requirements, U.S. EPA, December 1999.

    1 "Highway Diesel Progress Review," U.S. EPA, June 2002, EPA420-R-02-016.

                                           5-28

-------
                                                               Fuel Standard Feasibility
       5.3.2.1 Tier 2 Gasoline Sulfur Program

       Chapter IV of the Final Regulatory Impact Analysis for the Tier 2 gasoline sulfur
program presented the following table containing the results of its lead-time assessment.

                                        Table 5.3-1
           Lead-time Projections Under the Tier 2 Gasoline Sulfur Program (years)
Project Stage
Scoping Studies
Process Design
Permitting
Detailed Engineering
Field Construction
Start-up/Shakedown
Naphtha/Gasoline Hydrotreating
Time for
Individual Step
0.5-1.0b
0.5
0.25-1.0
0.5-0.75
0.75-1.0
0.25
Cumulative
Time3
0.5
1.0
1.25-2.0
1.5-2.25
2.0-3.0
2.25-3.25
More Major Refinery Modification (e.g., FCC
Feed Hydrotreating)
Time for Individual Step
0.5-1.0b
0.5-0.75
0.25-1.0
0.5-1.0
1.0-1.5
0.25
Cumulative Time3
0.5
1.0-1.25
1.25-2.0
1.5-2.25
2.5-3.5
2.75-3.75
3 Several of the steps shown can overlap.
b Projected to begin before Tier 2 gasoline final rule.
       This table contains lead-time projections for two distinctly different approaches to
gasoline sulfur control.  The first, naphtha hydtrotreating, is more closely related to conventional
distillate hydrotreating.  In fact, several naphtha hydrotreating processes utilize fixed-bed
hydrotreating, which is directly comparable to distillate hydrotreating. The second, FCC feed
hydrotreating, is more complex, extensive, and costly. As discussed earlier in this chapter, some
refiners might use FCC feed hydrotreating to facilitate the production of 500 ppm diesel fuel.
However, this decision was likely tied to their compliance plans for the Tier 2 gasoline sulfur
program, since FCC feed hydrotreating significantly reduces the sulfur content of gasoline, as
well as moderately reducing the sulfur content of LCO. Since refiners will not be able to meet
the sulfur standard using FCC feed hydrotreating, it is highly unlikely that a refiner would just
begin considering FCC feed hydrotreating as the result of this NRLM rule. We will therefore
focus only on the portion of the table that addresses the lead time for naphtha hydrotreating.

       It should also be noted that the cumulative times listed in the table above are not simply
the sum of the times for each step. Some steps overlap, in particular process design and
permitting, permitting and detailed engineering, and detailed engineering and construction. The
relationship between the time necessary for each step in the design and construction of naphtha
and distillate hydrotreaters are examined in detail below.  However, it is useful first to review the
projected lead time related to the 15 ppm highway diesel fuel cap.
                                           5-29

-------
Final Regulatory Impact Analysis
       5.3.2.2 15 ppm Highway Diesel Fuel Sulfur Cap

       The rulemaking implementing the 15 ppm sulfur cap for highway diesel fuel did not
evaluate the lead time required for each individual step of the process. That rule provided 5.5
years of lead time between promulgation and initial implementation.  This amount of lead time
significantly exceeded that considered necessary to design and construct desulfurization
equipment. This amount of lead time was provided, since the timing of the 15 ppm sulfur cap
was set primarily by the availability of highly efficient aftertreatment technology for diesel
engines and not on refiners' ability to meet the 15 ppm standard.

       We reviewed the progress that refiners were making towards complying with the 15 ppm
highway diesel fuel cap in 2002. Part of this review included an assessment of the tasks refiners
had already completed and the length of time needed for those still remaining. The tasks
considered were generally the same as those listed in Table 5.3-1 above, with one exception.
That was the inclusion of the need to develop a corporate strategy towards compliance in the
initial step. This strategy involved a decision regarding the degree to which refiners would
continue marketing highway diesel fuel and if so, whether they would comply with the 15 ppm
standard initially in 2006 or later in 2010. However, diesel fuel can be sold to the highway or
non-highway markets, involving compliance with very different sulfur standards. The flexibility
afforded by the rule's temporary compliance option also gave refiners a choice of when they
chose to comply with the 15 ppm cap. This issue didn't arise in the Tier 2 gasoline rule, since
nearly all gasoline sold  in the United States meets highway quality standards and refiners have
no other market for their gasoline feedstocks.

       The results of the lead-time review are presented in  Table 5.3-2.

                                      Table 5.3-2
         Lead-time Assessment: Progress Review of 15 ppm Highway Diesel Fuel Cap
Project Stage
Strategic Planning
Planning and Front-End Engineering3
Detailed Engineering and Permits
Procurement and Construction
Commissioning and Start -Up
Time Allotted
0.25-2 years
0.5
1.0
1.25-2.5
0.25-0.5
Latest Start Date


Mid-2003
Late 2003 - Early 2004
October 2004
March 2006
              ' Labeled Process Design in Table 5.3-1.
       By grouping several of the process steps shown in Table 5.3-1 this later assessment
reduces the overlap between the various steps considerably.  The primary overlap still remaining
is between detailed engineering and permits and procurement and construction. While
construction cannot begin until permits have been obtained, procurement can proceed.  This is
often essential to any time constrained refining project, due to the long lead times needed to
                                          5-30

-------
                                                              Fuel Standard Feasibility
fabricate specialized equipment.

       Because the progress review was conducted more than a year after the rule was adopted,
we did not add up the times associated with each step to develop a range of cumulative time
requirements. Instead, we focused on the dates by which refiners should have begun each step to
determine if they had indeed begun those steps that should have been started by the date of the
assessment.

       5.3.2.3 Lead-time Projections for Production of 500 ppm NRLM Diesel Fuel

       We utilized the information for gasoline and highway diesel analyses to project the lead
time necessary for a wide spectrum of refiners to start producing 500 ppm NRLM diesel fuel.
Beginning with strategic planning, refiners currently producing high-sulfur diesel fuel/heating oil
will have to decide whether they are going to continue producing high-sulfur heating oil or
produce 500 ppm NRLM diesel fuel.  This would not likely be a difficult choice for many
refiners, as the heating oil market will be too small in their area to support their entire production
of high-sulfur fuel.  For those with a real choice, this step will likely involve discussions between
the refining and marketing divisions of the firm, as well as with  any common carrier pipelines
used by the refiner.  While many refiners prefer to be able to observe their competition's choices
and the relative production volumes and prices of 500 ppm NRLM diesel fuel and high-sulfur
heating oil before making a decision, this is not possible. Given this, it seems reasonable to
allow a relatively short period of time, such as three to six months, to arrive at a corporate
decision to participate either in the NRLM or heating oil markets.

       Scoping and screening studies refer to the process whereby refiners investigate various
approaches to sulfur control. These studies involve discussions with firms supplying
desulfurization and  other refining technology, as well as studies by the refiner to assess the
economic impacts of various approaches to meeting the sulfur standard. In the case of distillate
desulfurization, refiners will likely send samples of their various distillate streams to the firms
marketing desulfurization technology to determine how well each catalyst and associated
hydrotreating technology removes the sulfur from that particular type of distillate (e.g., sulfur
removal efficiency,  yield loss, hydrogen consumption, etc.).

       Under the Tier 2 rule, we projected that six to twelve months were required to evaluate
the various available technologies for naphtha desulfurization.  This extensive period of time was
considered appropriate due to the wide range of technologies available. More importantly,
however, was the fact that many of the new gasoline desulfurization technologies had not been
demonstrated in actual refinery applications by the time of the final rule. Refiners naturally
desire as much demonstrated experience with any new technology as possible before investing
significant amounts of capital in these technologies. We believed that at a minimum, refiners
should have six months after the final rule to assess their situation with respect to the final sulfur
control program and select their vendor and technology. Because the Tier 2 gasoline sulfur
standards phased in over two years, some refiners had more time than others before their new
desulfurization equipment had to be operational. Thus, we expected refiners to take as much
time as they could afford to select the particular desulfurization technology that was optimum for

                                           5-31

-------
Final Regulatory Impact Analysis
their situation. Thus, there was really no upper limit to the amount of time for this step.

       The scoping and screening task refiners face with respect to the 500 ppm NRLM sulfur
cap is both different from and similar to the situation refiners faced with the Tier 2 gasoline
program.  The NRLM program differs because refiners had to choose between a wide variety of
gasoline desulfurization technologies to comply with the Tier 2 sulfur standards. In contrast, we
project above that conventional hydrotreating will likely be the dominant choice for
desulfurizing diesel fuel to 500 ppm in 2007. Furthermore, this is already a well known
technology.  The similarity exists, because refiners will have to consider how to comply with the
15 ppm nonroad diesel fuel cap in 2010 and 15 ppm L&M diesel fuel cap for 2012 when they
design their conventional hydrotreater for 2007. While conventional hydrotreating is well
understood, there are numerous ways to "conventionally hydrotreat" distillate. Variations exist
in operating pressure, hydrogen purity, physical catalyst loading, etc/ To avoid scrapping their
conventional hydrotreaters after just three to five years, we project that the refiners building new
conventional hydrotreating units for 2007 will plan these units to be easily revamped to produce
15 ppm nonroad diesel fuel in 2010 and L&M diesel fuel in 2012. The specific  conventional
hydrotreating design selected for 2007 will therefore have to mesh with their plans for 2010 and
2012. At a minimum, this will involve selection of the operating pressure of the conventional
hydrotreater, provision of physical space for additional equipment, and the capacity of hydrogen
supply and treatment lines. Selecting the operating pressure is likely the most time-critical,
because of the long lead times  involved in procuring pressure vessels. Also, vendors need some
time to assess the deep desulfurization performance of their desulfurization technologies via pilot
plants testing on specific refiners'  diesel fuel samples.

       Fortunately, this process has been underway for some time involving refiners' highway
diesel fuels. By mid-2004, this process should be nearly complete.  In fact, 27 out of the 35
refineries projected to produce 500 ppm NRLM diesel fuel for 2007 have experience producing
highway diesel fuel today under the  500 ppm cap.  Vendors' should have ample capacity to test
refiners' NRLM diesel fuel samples, as well as have developed efficient approaches to translate
test results into specific process designs. Thus, six months should be more than sufficient for
refiners to make the necessary, critical choices about their conventional hydrotreater design.  In
fact, the selection of operating  pressure could be made during the process-design step,
effectively reducing the amount of time to scoping and screening to three months.

       The strategic decision to produce 500 ppm NRLM diesel fuel involves not only
marketing, but an economic assessment of the cost of producing this fuel, both absolutely and
relative to the competition. The scoping and screening studies are also not expensive to conduct.
Refiners do not risk much to conduct them while they are still developing their corporate
strategy. Also, the scoping and screening studies can go on concurrent with the development of
a corporate strategy towards the rule. This means that the time for strategic planning (three to
six months) and the time for scoping and screening (three to six months) can go on concurrently.
   J Many of these issues are uncertainties for refiners installing a new diesel fuel hydrotreater, but would be fixed
for those adapting an existing desulfurization unit or reactor.

                                           5-32

-------
                                                              Fuel Standard Feasibility
       The time required for process design of a conventional distillate hydrotreater should be
no greater than that for a naphtha hydrotreater or the revamp of a diesel fuel hydrotreater (i.e.,
six months in both Tables 5.3-1 and 5.3-2). In fact, the design of the naphtha hydrotreater may
be more complex due to the desire to avoid too great a loss in octane from olefin saturation.
Avoiding octane loss may lead the refiner to treat different parts of the naphtha stream
differently. Octane is not an issue with distillate hydrotreating.  In general, the design of a
grassroots distillate hydrotreater is more complex than that of a revamp.  However, the eventual
revamp in 2010 or 2012 which must follow this 500 ppm step will be is to produce 15 ppm diesel
fuel, a much more  challenging task than producing 500 ppm diesel fuel.  Thus, some extra
planning may be necessary for designing this 500 ppm hydrotreater.  Regardless, six months
should be sufficient for the process design of a 500 ppm NRLM unit.  The cumulative time for
the strategy, scoping, and process-design steps should range from nine to twelve months, as the
choice of distillate hydrotreating is clear.

       Regarding permitting,  we have taken steps to help state and local permitting agencies to
efficiently process refiners' requests for permits related to environmental-related projects such as
these.  Our experience with permits related to naphtha desulfurization indicates that three to nine
months is a more realistic range, as opposed to the three to twelve months projected in the Tier 2
Final Regulatory Impact Analysis.  There, we identified twelve months as being a worst-case
scenario. Experience has confirmed this and we are not aware of any specific situations where
obtaining a permit has taken this long and held up the project completion.

       The detailed design and construction of a distillate hydrotreater could require some
additional time relative to that for a naphtha hydrotreater due to the higher operating pressures
required for distillate hydrotreating. Because fewer firms fabricate higher pressure reactors and
compressors, the lead time for construction and delivery are usually longer.  At the same time,
less time should be required than required for a FCC feed hydrotreater. FCC feed hydrotreating
usually occurs at even higher hydrogen pressures and involves much more cracking of large
molecules into smaller ones. Additional equipment is necessary to handle the significant amount
of gaseous product generated, etc.  Interpolating between the times allocated for the detailed
design and construction of a naphtha hydrotreater and a FCC feed hydrotreater results in six to
nine months to design and twelve to fifteen months to construct a distillate hydrotreater.
Cumulatively, the two steps would take a little more than 1 year and up to 2 years, or 1 to  1.25
years from the time permits were obtained.

       This range is about three months shorter than that projected in Table 5.3-2 for the 15 ppm
highway  diesel fuel rule.  The difference on the high end is due to the fact that 2.5 years for
construction does not appear to be necessary. For this to be typical, all refiners planning to
produce 15 ppm highway diesel fuel would have already been constructing their new or
revamped hydrotreaters by the time of the 2003 precompliance reports. Clearly this was not the
case in the precompliance report results, yet refiners considered themselves  on track to meet the
standard. Thus, the time periods resulting from an interpolation of the naphtha and FCC feed
hydrotreating estimates of Table 5.3-1  appear reasonable for producing 500  ppm NRLM fuel.

       Finally, both the Tier 2 gasoline rule and 15  ppm highway diesel fuel review allocated

                                          5-33

-------
Final Regulatory Impact Analysis
three months for start up for naphtha, FCC feed and highway diesel fuel hydrotreaters.
Allocating the same time period for starting a distillate hydrotreater should therefore be
appropriate.

       Table 5.3-3 presents the results of the above assessment.

                                      Table 5.3-3
                   Lead-time Projections for 500 ppm NRLM Diesel Fuel
Project Stage
Strategic Planning
Scoping and Screening Studies
Process Design
Permitting
Detailed Engineering
Field Construction
Start-up/Shakedown
Time for Individual Step
0.25-0.5
0.25-0.5
0.5
0.25-0.75
0.5-0.75
1.0-1.25
0.25
Cumulative Time
0.25-0.5
0.25-0.5
0.75-1.0
1.0-1.75
1.5-2.25
2.0-3.0
2.25-3.25
       The timing of this final rule should allow some refiners to produce 500 ppm NRLM fuel
as early as July 2006. This coincides with implementation of the 15 ppm highway diesel fuel
cap and the ability to generate early 500 ppm NRLM credits. This analysis indicates that the last
refiners should be able to produce 500 ppm NRLM fuel by July 2007. This is within a month of
implementation of the 500 ppm NRLM cap. If any refiners are in the situation of needing this
last month to produce 500 ppm NRLM fuel, they should be able to purchase early credits from
other refiners and continue producing NRLM fuel until they are able to meet the 500 ppm cap.

       5.3.2.4 Comparison with the 500 ppm Highway Diesel Fuel Program

       The tasks refiners face in meeting the 500 ppm NRLM cap is very similar to the task
refiners faced with meeting the 500 ppm highway diesel fuel cap by October 1, 1993. The
primary difference is that refiners have ten years of experience producing 500 ppm diesel fuel
commercially.  This should only  shorten the time required to prepare for complying with the
standard relative to 1993.  The 500 ppm highway diesel rulemaking was adopted in August 1990
and took effect October 1, 1993.38 Thus, that rulemaking provided 38 months of lead time,
nearly identical to that provided in this final rule for NRLM. Some  price spikes occurred with
the 500 ppm highway diesel fuel standard.  However, these were almost exclusively in
California, where a 10 volume percent aromatics standard was implemented at the same time.
Also, the October implementation coincided with the annual increase in refiners' distillate
production related to winter heating oil use. At that time, the United States was one of the first
nation's to require 500 ppm diesel fuel, so little commercial  experience was available upon
which to base designs. Refiners and technology vendors currently have over ten years of
                                         5-34

-------
                                                            Fuel Standard Feasibility
commercial experience in producing 500 ppm diesel fuel. We have also shifted the
implementation date away from the peak heating oil production season.  Finally, the volume of
highway diesel fuel affected was more than three times as much as that affected by this final
rule, causing greater stress on the engineering and construction industries than we expect to
result from this final rule.

       Many refiners likely to produce 500 ppm NRLM diesel fuel in 2007 also have to invest to
meet the Tier 2 gasoline sulfur standards and the 15 ppm highway diesel fuel cap. However, the
Tier 2 program finishes phasing in in 2006 for most refiners. The 15 ppm highway  diesel fuel
likewise has a 2006 implementation date. This puts them at least one year ahead of the 500 ppm
NRLM standard. This minimum offset of one year should ease the burden on any specific aspect
of the process (e.g., raising capital funds, design personnel, construction personnel,  etc.). The
1993 500 ppm highway diesel fuel cap also occurred in the midst of other fuel-quality
regulations.  The phase 2 gasoline Reid vapor pressure standards and the oxygenated gasoline
programs took affect in 1992, while the reformulated gasoline program began in 1995.  Thus, the
experience with the 500 ppm highway diesel fuel program appears to be a strong confirmation
that the final rule provides sufficient lead time.

       5.3.2.5 Small Refiners

       Small refiners may need more time to comply with a sulfur control program.  Small
refiners generally have a more difficult time obtaining funding for capital projects, and must plan
further in advance of when the funds are needed. We contracted a  study of the refining industry
that assessed the time required for small refiners to obtain loans for capital investments. The
simple survey revealed that small refiners need two to three months longer than large refiners to
obtain funding.  If small refiners are forced to or prefer to seek funding through public means,
such as through bond sales, then the time to obtain funding could be longer yet, by up to one
third longer.39 In addition, because of the more limited engineering expertise of many small
refiners, the design and construction process for these  refineries is relatively more difficult and
time consuming. We also believe the contractors that  design and install refinery processing units
will likely focus first on completing the more expensive upgrade projects for large refiners. This
would also contribute to the additional time for design and construction of desulfurization
hardware for small refiners. The three additional years being provided small refiners should be
sufficient to compensate for these factors.  This additional lead time should provide not only
enough time for these small refiners to construct equipment, but also allow more time for them to
select the most advantageous desulfurization technology.  This additional time  for technology
selection will help to compensate for the relatively poor economy of scale inherent with adding
equipment to a small refinery.

5.4 Feasibility of Producing 15 ppm Sulfur NRLM in 2010 and 2012
                                         5-35

-------
Final Regulatory Impact Analysis
5.4.1 Expected use of Desulfurization Technologies in 2010 and 2012

       Like the 500 ppm sulfur standard for 2007, we considered several criteria to project
which desulfurization technologies will be used to meet the 15 ppm standard for nonroad in 2010
and the 15 ppm L&M standard in 2012.  The criteria we considered included:  (1) the time
refiners will have to choose a new technology, (2) whether the technology will be available for
2010 ane 2012 and, if the technology is available, how proven it is, (3) whether the technology is
cost-competitive by comparing it with other technologies, (4) whether the technology is
available from a vendor that has proven itself to the industry by providing other successful
refining technologies, particularly if the vendor has proven itself in the United States, and (5)
whether the vendor has the capability to meet the industry demands.

       Refiners will have six and eight years to meet the 2010 and 2012 standards, respectively.
Refiners will have from 2 to 4 more years to evaluate the slate of technologies in addition to the
usual amount of time they must have to construct and start up the necessary capital investments.
Refiners are therefore not constrained when making their decisions and this particular issue did
not figure into our judgment regarding projected technologies.

       Next, we considered whether a technology will be available in 2010 and 2012.
Conventional hydrotreating is available, as it has  been used in a variety of applications to meet
very stringent sulfur standards. In addition, many refiners are expected to use conventional
hydrotreating to comply with the highway diesel  15 ppm cap, which applies in 2006. This would
give refiners some experience with this technology before they decide which technology to use.

       Process Dynamics already has a diesel fuel hydrotreating commercial demonstration unit
operating which is a revamp of a 500 ppm highway diesel fuel desulfurization unit (installed
before the existing highway hydrotreater). This unit demonstrates that the technology does
indeed work for treating untreated diesel fuel to 500 ppm, and thus would provide a proven
upgrade path through the revamp of the conventional 500 ppm units installed in 2007 to comply
with the 15 ppm cap standard in 2010 or 2012. A couple more refiners may choose to revamp
their refineries with the Process Dynamics technology for complying with the 15 ppm highway
diesel fuel sulfur standard taking effect in 2006, thus providing several more examples of the
Process Dynamics desulfurization technology being used to revamp 500 ppm treaters to meet the
15 ppm sulfur cap.  Thus, refiners seeking to comply with the 15 ppm sulfur NRLM standard
should be able to see at least one, and probably more examples of the Process Dynamics
Isotherming process operating to desulfurize diesel fuel down to 15 ppm.

       The oxidation and extraction technologies by Unipure and perhaps UOP do not have
units operating now, but Unipure is projecting to  have a commercial demonstration unit
operating by 2006. However, an oxidation and extraction unit that begins operation in 2006 will
not provide two years of operations for interested refiners before they need to choose their
technology for 2010. As a result, it is unlikely to see any significant use by 2010, and use may
be limited to small  refineries and terminals which would take advantage of their lower costs for
smaller installations. Furthermore, without a commercial demonstration unit operating along
with the technology's perceived success, it is difficult to project the penetration into the

                                          5-36

-------
                                                              Fuel Standard Feasibility
desulfurization market even for 2012.

       Another issue refiners will consider is the cost of installing and operating these various
technologies.  Of the oxidation and extraction technologies, Unipure did provide us with
desulfurization cost information based on testing at their laboratory, and that information shows
that it might be cost competitive with conventional hydrotreating. Phillips also has provided us
with diesel fuel desulfurization cost information from their pilot plant, which is backed up by the
success they have had with their commercial gasoline desulfurization unit (see Section 7.2).
That technology  seems to be less expensive than conventional hydrotreating for some refineries;
it appears to be suited primarily for desulfurizing low-sulfur diesel fuel  down to very low sulfur
levels rather than for desulfurizing higher-sulfur feedstocks. Finally,  Process Dynamics
provided us diesel fuel desulfurization cost information based on their pilot plant and their
engineering cost estimates for the commercial demonstration unit at the Giant refinery.  The
Process Dynamics process seems to be less expensive than conventional hydrotreating (see
Section 7.2) and  has been demonstrated to meet a 15 ppm  sulfur standard by revamping a
conventional hydrotreater.

       We next evaluated whether each diesel fuel desulfurization technology vendor is
equipped to provide preliminary engineering and support the installations of its technology to a
significant part of the refining industry. Conventional hydrotreating is provided by numerous
vendors (Akzo Nobel, Criterion, Haldor Topsoe, IFF, and  UOP) the majority of which
manufacture their own line of diesel desulfurization catalysts. Also, these vendors supported the
installation of many diesel fuel hydrotreaters to meet the 500 ppm highway diesel fuel  sulfur
standard, which went into effect in 1993, and will be working with refiners to meet the very
stringent 15 ppm highway diesel fuel sulfur standard, which begins to take effect in 2006.  Thus,
conventional desulfurization technology is poised to make a significant contribution.

       Process Dynamics has only a very small engineering staff, however, they are associated
with Linde Process Plants and Roddy Engineering.  Linde currently licenses several different
technologies, including sulfur and olefins recovery, natural gas processing, hydrogen production,
reforming, air separation. Linde has a large engineering and design department that has been
active for over 30 years. Roddy Engineering has a small engineering staff for additional
engineering support. Thus, Linde and Roddy Engineering are capable of providing substantial
engineering support to Process Dynamics for its IsoTherming desulfurization technology for a
significant penetration into the U.S. refining industry.

       Phillips licenses several different technologies to refiners now, including its S-Zorb
gasoline desulfurization technology and an alkylation technology, and has licensed refining
technologies for  over 60 years. Phillips has a robust research and development staff and also an
engineering staff to  support the licensing of its S-Zorb technology.

       The oxidation and extraction technologies are being developed by two separate entities,
one being Unipure and the other UOP. Unipure is associated with Texaco  and Mustang
engineering.  Thus, Unipure potentially has both research and development and engineering
support for its technology. UOP has substantial capacity for conducting engineering support for

                                          5-37

-------
Final Regulatory Impact Analysis
refiners.

       After evaluating the various criteria for each technology and comparing across
technologies, we developed a projection for the mix of technologies that will be used in 2010
and 2012 for meeting the 15 ppm standards.  Since refiners will have plenty of time to sort
through the various technologies, we believe lead time will have no bearing on refiners ability to
choose an advanced desulfurization technology. Whether a technology will have accumulated at
least two years of commercial experience is an important issue for the S Zorb and oxidation and
extraction technologies as the developers are not expected to have a commercial demonstration
unit operating for at least two years.  Thus, while the Phillips S Zorb, Unipure and UOP
desulfurization technologies might be selected by refiners for 2010, we are not including their
technologies in our projected mix of technologies.

       This leaves conventional hydrotreating and Process Dynamics Isotherming.
Conventional hydrotreating will clearly have the most refining experience due to refiners'
previous experience and also due to production of 15 ppm highway fuel for 2006. However,
Process Dynamics already has one unit operating and perhaps more diesel fuel desulfurization
commercial demonstration units will  be operating for over two years. The Process Dynamics
hydrotreating process is expected to be lower in cost than conventional  hydrotreating providing a
strong incentive to refiners seeking to reduce their capital and operating costs. Also Linde has
research and development and engineering capacity to support Process Dynamics with their
IsoTherming desulfurization processes, though not the same level of support as the multiple
conventional hydrotreating firms. After comparing these various criteria, we believe the lower
cost of Process Dynamics Isotherming will be the most important driver for these technologies.
However, we also believe that some refiners will not be willing to try out a newer desulfurization
technology, especially since they may already have an established relationship with another
vendor. Thus, we believe the Process Dynamics process will be used to a greater extent than
conventional hydrotreating, but still be somewhat market limited.  We project that Process
Dynamics Isotherming will capture 60 percent of the nonroad desulfurization market by 2010,
with conventional hydrotreating capturing the remaining 40 percent of the nonroad
desulfurization market.

       Refiners will have two more years to assess which technology they will use for
complying with the 15 ppm sulfur locomotive and marine standard in 2012. Despite the
additional two years, though, we assume the same penetration of advanced technologies because
of limiting factors for these technologies. Process Dynamics, even when associated with Linde
and Roddy engineering, is expected to be limited by the engineering staff available to them and
the conservative view by some refiners to new technologies. Furthermore,  until a commercial
demonstration unit is operating for Unipure, UOP or Phillips, it did not seem appropriate to
assess potential  market penetration for these advanced technologies. Thus, for 2012 we continue
to assume Processed Dynamic's IsoTherming will meet 60 percent of the desulfurization demand
while extensions of conventional hydrotreating will meet the remaining 40  percent.
                                          5-38

-------
                                                             Fuel Standard Feasibility
5.4.2 Lead-time Evaluation

       More lead time is needed to meet a 15 ppm sulfur standard than a 500 ppm standard.  The
additional time primarily involves the scoping and screening step, as the technology to achieve a
15 ppm sulfur cap is just being demonstrated on a commercial scale and some advanced
technologies promising lower costs are under development. This additional time might be on the
order of a few months, while the 2010 implementation date for 15 ppm nonroad and the 2012
implementation date for 15 ppm L&M fuel provides an additional three and five years of lead
time, respectively.  The amount of lead time available for the  15 ppm NRLM caps should
therefore be more than sufficient for refiners to prepare for producing this fuel.

       Of more interest is the interaction between the timing of the 15 ppm cap on highway
diesel fuel and that for NRLM diesel fuel.  The time periods listed in Table 5.3-3 indicate that
refiners must start their process designs 2.0 to 2.75 years before first producing 15 ppm diesel
fuel and complete these process designs 1.5 to 2.25 years before the implementation date.  This
means that process design should begin by September 1, 2007 to June 1, 2008, and be completed
by March 1 to December 1, 2008.  This would provide refiners planning to produce  15 ppm
nonroad diesel fuel with 15 to 24 months of desulfurization experience from highway diesel fuel
desulfurization units started up in mid-2006 before initiating their process design. Similarly,
refiners producing  15 ppm L&M diesel fuel in 2012 are expected to have 39 to 48 months before
initiating their process design. Given that catalyst cycles last two to three years, refiners could
observe the performance of catalysts used to  produce 15 ppm  highway diesel fuel for one-third
of a cycle to a full cycle before having to begin their process design for desulfurizing nonroad
diesel fuel. Refiners producing L&M diesel fuel in 2012 will  be able to observe the performance
of highway diesel fuel desulfurization catalysts for one to two cycles.  While most of the units
producing highway diesel fuel in 2006 are expected to use conventional hydrotreating, as
discussed above, we also expect the Process Dynamics Isotherming process to have  acquired
significant commercial experience  and perhaps be demonstrated by more refineries choosing to
commercially produce 15 ppm for the highway program.  Thus, refiners planning for 2010 would
be able to observe this newer process for more than three years before selecting their technology
and vendor.  This should be sufficient to overcome uncertainty on the part of most refiners about
its performance.  Overall, the available lead time allows all refiners to take advantage of the
operating performance of the highway units and minimize their costs.

5.5 Distribution Feasibility  Issues

       There are three considerations with respect to the feasibility of distributing NRLM diesel
fuels meeting the sulfur standard's in this final rule. The first pertains to the extent that the
distribution system can reasonably accommodate the additional product segregation which might
result from this final rule, given the existing limitations in the system and the potential cost of
overcome such limitations. The second pertains to whether sulfur contamination can be
adequately managed throughout the distribution system so fuel delivered to the end-user does not
exceed the sulfur requirements in this rule. The third pertains to the ability to handle products
that become mixed in the pipeline distribution system so that they can be made saleable into the
                                          5-39

-------
Final Regulatory Impact Analysis
distillate market. These considerations are evaluated in the following Sections 5.5.1, 5.5.2., and
5.5.3. As discussed in these sections, we have designed the NRLM fuel program to avoid
significant distribution feasibility issues, and therefore have concluded that compliance with the
NRLM diesel sulfur control program will represent a manageable challenge to fuel distributors
that is not unduly burdensome. As a result, these issues are more correctly related to the cost of
compliance rather than feasibility.

5.5.1 Ability of Distribution System to Accommodate the Need for Additional Product
Segregations That Could Result from This Rule

       5.5.1.1 The Diesel Fuel Distribution System Prior to Implementation of the NRLM
       Sulfur-Control Program

       Before 1993, most No. 2 distillate fuel was produced to nearly the same specifications,
shipped fungibly, and used interchangeably for highway diesel engines, nonroad diesel engines,
locomotive and marine diesel engines and heating oil (e.g., furnaces and boilers) applications.
Beginning in 1993, highway diesel fuel  was required to meet a 500 ppm sulfur cap and be
segregated from other distillate fuels as  it left the refinery by the use of a visible level of dye
solvent red 164 in all non-highway distillate. At about the same time, the Internal Revenue
Service (IRS) similarly required non-highway diesel fuel to be dyed red (to a much higher
concentration) prior to retail sale to distinguish it from highway diesel fuel for excise tax
purposes (dyed non-highway fuel is exempt from this tax). This splitting up of the distillate pool
necessitated costly changes in the distribution system to ship and store the now distinct products
separately.

       In some parts of the country where the costs to segregate non-highway diesel fuel from
highway diesel fuel could not be justified, both fuels have been produced to the highway
specifications. Diesel fuel produced to highway specifications but used for non-highway
purposes is referred to as  "spill-over." It leaves the refinery gate and is fungibly distributed as if
it were highway diesel fuel, and is typically  dyed at a point later in the distribution system.  Once
it is dyed it is no longer available for use in highway vehicles, and is not part of the supply of
highway fuel.

       When the 15 ppm highway diesel fuel standard takes effect in 2006, an additional
segregation of the distillate pool is anticipated. Since up to 20 percent of the highway diesel fuel
pool is allowed to remain at 500 ppm until 2010, in some portions of the country as many as
three grades of distillate may be distributed; 15 ppm highway, 500 ppm highway, and high-sulfur
for all non-highway uses. The final highway diesel rule estimated that 500 ppm diesel fuel will
be present in 40 percent of the fungible fuel  distribution system including the Northeast,  parts of
the Midwest and in the area adjacent to the concentration of refineries in PADD 3.  However,
given the results of its refiner's pre-compliance reports which suggests that more than 95 percent
of highway diesel may be manufactured to a 15 ppm sulfur standard, 500 ppm fuel will likely be
restricted to a much smaller portion of the distribution system in 2006.
                                          5-40

-------
                                                              Fuel Standard Feasibility
       5.5.1.2 Potential for Additional Product Segregation Under the NRLM Sulfur
       Program

       The NRLM sulfur-control program is discussed in detail in Section IV of the preamble to
the final rule. Following is a summary of these requirements and a discussion of the potential for
additional product segregation which might result.

       This final rule requires that NRLM fuel comply with a 500 ppm sulfur standard
beginning in 2007. These provisions mirror controls on highway diesel fuel to 500 ppm in 1993.
Refiners and importers can comply with the requirement either by producing NRLM fuel at or
below 500 ppm or, if located outside of the Northeast/Mid-Atlantic Area and  Alaska, by
obtaining  sufficient credits under the averaging banking and trading (ABT) provisions to cover
their continued production of high-sulfur (HS) NRLM through 2010.K  Small  refiners outside of
the Northeast/Mid-Atlantic Area may also continue to produce high HSNRLM until the
HSNRLM small-refiner and credit-use provisions expire in June 1, 2010.

       The 15 ppm sulfur standard for nonroad diesel  fuel takes effect June 1, 2010 and for
L&M diesel fuel takes effect June 1, 2012. The options available to comply with this 15 ppm
requirement parallel those available to  comply with the earlier 500 ppm NRLM requirement.
Refiners and importers can produce nonroad and L&M fuel at or below 15 ppm or, if located
outside of the Northeast/Mid-Atlantic Area and Alaska can obtain sufficient credits under the
averaging banking and trading (ABT) provisions to cover their continued production of 500 ppm
through June  1,  2014. Small refiners outside of the Northeast/Mid-Atlantic Area may also
continue to produce 500 ppm NRLM until the  500 ppm NRLM small-refiner  and credit-use
provisions expire in June 1, 2014.  After June  1 2014, all NRLM diesel fuel must meet a 15 ppm
sulfur standard except for 500 ppm fuel produced in the distribution system due to pipeline
interface mixing and product contamination. Outside of the Northeast/Mid-Atlantic Area and
Alaska, the prescribed marker must be  added to heating oil at the terminal beginning June 1,
2007 and to 500 ppm sulfur L&M diesel fuel produced at a refinery or imported from June 1
2010 through May 31, 2012.

       The application of different sulfur standards to portions of the non-highway distillate
pool based on end-use raises concerns regarding the potential need for additional product
segregation. Currently, distillate fuel for all non-highway uses is typically drawn from a single
pool that meets the most stringent specifications for any non-highway use. For example, it is our
understanding that nearly all  heating oil meets  the cetane specification for nonroad diesel engine
use despite the lack of applicability of a cetane specification for distillate fuel used as heating oil.
This is because fuel  manufactures and marketers have  found that the potential savings from
manufacturing a low cetane heating oil are typically outweighed by the additional costs of
segregating an additional heating-oil-only product throughout the distribution system.
   K The Northeast/Mid-Atlantic Area provisions are discussed in detail in Section IV.D. of the preamble to the
final rule. Our determination of the boundaries of the Northeast/Mid-Atlantic Area is discussed in Section 5.5.1.4.

                                          5-41

-------
Final Regulatory Impact Analysis
       We anticipate that the significant cost of desulfurizing NRLM diesel fuel to meet the new
sulfur standards provides a strong incentive for the fuel distribution system to evaluate whether
the additional costs of distributing non-highway distillate fuels of different sulfur specifications
is economically justified. This situation is analogous to that faced by industry after the 500 ppm
sulfur standard for highway diesel fuel took effect in 1993.

       The IRS requirement that diesel fuel used in NRLM engines be dyed before it leaves the
terminal to indicate its nontaxed status also raises concerns about the potential need for
additional product segregation under the NRLM sulfur program. Fuel that meets highway diesel
specifications but is destined for the NRLM market can leave the terminal undyed provided that
the tax is payed.  Non-highway users of such fuel can then apply to the federal and applicable
state revenue offices for a refund of the highway taxes payed on the fuel. In areas of the country
where only 500 ppm diesel fuel is currently available by pipeline, most bulk plant operators
nevertheless maintain dual tankage  for dyed and undyed 500 ppm diesel fuel to meet the
demands of their customers for highway-tax-free non-highway diesel fuel.  Such bulk plant
operators currently receive dyed diesel fuel by truck from local refineries. Thus, the IRS NRLM
diesel dye requirement may result in a strong incentive for parties in the fuel distribution system
downstream of the terminal  to maintain segregated pools of undyed highway and dyed NRLM
diesel fuel that differ in no other respect than the presence of dye (after implementation of both
the 15 ppm highway diesel requirements in 2007 and the new requirements for NRLM fuel).  We
expect that after the NRLM standards take effect, most bulk plant operators will request that the
terminal (or refinery rack) dye the fuel destined for sale into the NRLM market, so they can
continue their current practice of offering untaxed diesel fuel to their NRLM customers.

       We designed the NRLM sulfur program to minimize the need for additional product
segregation and resulting cost to fuel distributors associated with the need for additional storage
tanks, tank trucks, marker injection equipment, and other hardware and procedural factors.  The
designate and track provisions in this final rule allows the fungible distribution of diesel fuels
that have the same sulfur content through much of the distribution system despite the fact that
they  are destined for different end-uses. Fuel subject to the 500 ppm and 15 ppm NRLM sulfur
standards may be shipped fungibly with highway diesel fuel subject to the same sulfur standard
until the fuel leaves that terminal when red dye must be added to NRLM fuel to comply with IRS
fuel tax requirements. Similarly, high-sulfur and 500 ppm NRLM small-refiner and credit-use
fuel can be shipped fungibly with heating oil meeting the same sulfur specification until the
point when heating oil must be injected with the marker prescribed in this final rule. In addition,
high-sulfur NRLM small-refiner and credit-use fuel (present until 2010) may be commingled
with 500 ppm NRLM diesel fuel.

       The number of possible product segregations that might exist under this rule varies
temporally, geographically,  and based on the location in the fuel distribution system.  The
variation over time is a function of the timing of the implementation dates of the two-step sulfur
control program, and the implementation and sunset dates of the small-refiner and credit use
provisions. In general, the number  of possible segregations is the highest from 2007 - 2010, and
then  begins to decline thereafter as the diesel fuel standards for all highway, nonroad, and L&M
diesel fuel begin to coalesce. The geographic variation is a function of limitations on where

                                          5-42

-------
                                                               Fuel Standard Feasibility
small-refiner and/or credit-use fuel can be used, and where the fuel marker requirements apply.L
In the Northeast/Mid-Atlantic Area, the marker is not required since small-refiner and credit-use
NRLM fuel can not be sold there. In areas outside of the Northeast/Mid-Atlantic Area except
Alaska, the marker is required in heating oil beginning 2007 and in LM diesel fuel produced at a
refinery or imported from 2010-2012. In these areas small-refiner and credit-use NRLM fuel
may be sold.  No marker is required in heating oil used in Alaska. However alternate
requirements apply in Alaska which allow small-refiner NRLM to be sold in Alaska. The
variation in the number of product segregations by location in the fuel distribution system is
primarily a result of: (1) the IRS requirement that off-highway distillate be dye red to indicate its
non-taxed status before leaving the terminal, (2) the requirement that heating oil outside of the
Northeast/Mid-Atlantic Area and Alaska contain the marker specified under this final rule prior
to leaving the terminal, and (3) the provision under this rule that the downstream standard for
L&M outside of the Northeast/Mid-Atlantic Area and Alaska is 500 ppm to account for fuel
generated due to mixing in the pipeline distribution system which can be sold into the
locomotive and marine market after the sale of fuel above 15 ppm is otherwise prohibited.

       Many of the possible product segregations are discretionary, the decision to carry an
additional grade of diesel fuel being based on  an economic evaluation of the associated carrying
costs versus the potential market demand in their area and the additional cost associated with
supplying a single fuel for multiple end-uses which meets the most stringent specifications for
any of these end-uses.  We expect that a substantial part of the fuel distribution system in the
U.S. upstream of the terminal will carry only highway diesel fuel (for sale into both the highway
and NRLM markets).  As noted earlier, this is currently the case due to logistical constraints in
the distribution system. We anticipate that these new NRLM sulfur standards will result in  an
expansion of the area in which only highway diesel fuel is supplied for sale into both the
highway and NRLM markets.  In such cases, the fuel is only differentiated for sale into either the
highway or NRLM markets when it leaves the terminal by the addition of red dye to NRLM fuel
to satisfy IRS requirements.  Other segregations are unavoidable such as the segregation between
15 ppm highway and 15 ppm NRLM downstream of the terminal due to the presence of the IRS
specified red dye in NRLM fuel after it leaves the terminal, and the segregation of heating oil
from NRLM  downstream of the terminal due to the required presence of the marker in heating
oil under this final rule (outside the Northeast/Mid-Atlantic Area and Alaska).

       The following tables list the possible product segregations during various stages by
location in the distribution system. Table 5.5.1.2-1 lists the possible segregations outside of the
Northeast/Mid-Atlantic Area and Alaska. Table  5.5.1.2-2 lists the possible segregations in the
Northeast/Mid-Atlantic Area.  Table 5.5.1.2-3 lists the possible segregations in Alaska. These
tables represent the maximum potential number of product segregations that could result from
this final rule. In most cases there will be fewer actual product segregations particularly in areas
of the country that will receive pipeline shipments of only a single grade of No. 2  diesel fuel for
    L The fuel marker requirements are necessary to support the small-refiner and credit-use provisions. See Section
IV.D of the preamble to the final rule for a discussion of the interactions between the small-refiner and credit-use
provisions and the heating oil marker requirement.

                                           5-43

-------
Final Regulatory Impact Analysis
use in multiple distillate fuel markets. Furthermore, it is important to note that these possible
segregations are not equal in volume. As time goes by, most of the distribution system is
expected to coalesce around a few segregations such that it will look much as it does today.
Table 5.5.4. lists the possible number of product segregations in such areas.  Section 5.5.1.3. in
this RIA discusses the need for fuel distributors to invest in new storage tanks, tank trucks,
injection equipment, and other hardware or to change their operating practices in response to the
new product segregations caused by this rule.
                                          5-44

-------
                                                                        Fuel Standard Feasibility
                                           Table 5.5.1.2.-1
                            Summary of Possible Product Segregations
                      Outside of the Northeast/Mid-Atlantic Area and Alaska
Time Frame
Current
2004
2006-20072
2007-20093
2009-20104
2010-2012
201 2-20 145
2014 &
later6
Refinery Gate
June 1 - May 3 1
500 ppm Hwy
HS NRLM/HO (dyed)
1 5 ppm Hwy
500 ppm Hwy/NRLM7
HS NRLM/HO (dyed)
1 5 ppm Hwy
500 ppm Hwy/NRLM
HS NRLM/HO (dyed)
15 ppm Hwy/NRLM8
500 ppm Hwy/NRLM
HS NRLM/HO (dyed)
15 ppm Hwy /MR
500ppmNR/LM
HSHO
15 ppm Hwy/NRLM
500 ppm NRLM/HO9
HSHO
15 ppm Hwy/NRLM
500 ppm HO9
HSHO
Distribution to Terminal1
June 1 - Aug 1 5
500 ppm Hwy
HS NRLM/HO (dyed)
1 5 ppm Hwy
500 ppm Hwy/NRLM7
HS NRLM/HO (dyed)
1 5 ppm Hwy
500 ppm Hwy/NRLM
HS NRLM/HO (dyed)
15 ppm Hwy/NRLM8
500 ppm Hwy/NRLM
HS NRLM/HO (dyed)
15 ppm Hwy /MR
500ppmNR/LM
HSHO
15 ppm Hwy/NRLM
500 ppm NRLM/HO9
HSHO
15 ppm Hwy/NRLM
500ppmLM/HO9
HSHO
Post Terminal
June 1 - Sept 30
500 ppm Hwy
500 ppmNRLM (dyed)
HS NRLM/HO (dyed)
1 5 ppm Hwy
500 ppm Hwy
SOOppm & HS NRLM/HO (dyed)
1 5 ppm Hwy
500 ppm Hwy
500 ppmNRLM (dyed)
HS NRLM/500ppm NRLM (dyed)
HO (dyed & marked)
1 5 ppm Hwy
15 ppm NRLM8 (dyed)
500 ppm Hwy
500 ppmNRLM (dyed)
HS NRLM/500 ppm NRLM (dyed)
HO (dyed & marked)
1 5 ppm Hwy
ISppmNR(dyed)
500 ppm NR (dyed)
500 ppm L&M (dyed and marked)
HS or 500 ppm HO (dyed and marked)
1 5 ppm Hwy
15 ppm NRLM (dyed)
500 ppmNRLM (dyed)
500 ppm HO9 (dyed & marked)
HS HO (dyed & marked)
1 5 ppm Hwy
15 ppm NRLM (dyed)
500 ppm L&M (dyed)
500 ppm HO9 (dyed & marked)
HS HO (dyed & marked)
1 The term "terminal" is used as shorthand to refer to the point where taxes are paid on highway fuel, dye added to
 NRLM, or marker added to heating oil.
2 The 15 ppm highway diesel program and the 500 ppm NRLM early credit provisions are effective 2006.
3 500 ppm NRLM program effective 2007.
4 15 ppm NRLM early credit generating provisions effective 2009.
5 15 ppm NRLM program effective 2010. HS NRLM small-refiner and credit-use provisions expire. No 500 ppm
 NRLM may be sold except small-refiner, credit, and pipeline interface generated 500 ppm NRLM.
6 500 ppm NRLM small refiner, and credit use provisions expire 2014
7 500 ppm early credit generating NRLM.
8 15 ppm early credit generating NRLM.
9 500 ppm heating oil is not required, but is a fuel grade that some refiners may choose to produce and distributors
                                                 5-45

-------
Final Regulatory Impact Analysis
 transport. Earlier, when 500 ppm NRLM was available, such fuel could have been used for heating purposes.
                                            5-46

-------
                                                                        Fuel Standard  Feasibility
                                           Table 5.5.1.2.-2:
          Summary of Possible Product Segregations In the Northeast/Mid-Atlantic Area
Time
Frame
Current
2004
2006-
20072
2007-
20093
2009-
20 104
2012-
2012
2012 &
later5
Refinery Gate
June 1 - May 3 1
500 ppm Hwy
HS NRLM/HO (dyed)
1 5 ppm Hwy
500 ppm Hwy/NRLM6
HS NRLM/HO (dyed)
1 5 ppm Hwy
500 ppm Hwy/NRLM
HO (dyed)
15 ppm Hwy/NRLM
500 ppm Hwy/NRLM7
HO (dyed)
15 ppm Hwy /MR
500 ppm LM
HO
15 ppm Hwy/NRLM
SOOppm HO9
HSHO
Distribution to Terminal1
June 1 - Aug 1 5
500 ppm Hwy
HS NRLM/HO (dyed)
1 5 ppm Hwy
500 ppm Hwy/NRLM6
HS NRLM/HO (dyed)
1 5 ppm Hwy
500 ppm Hwy/NRLM
HO (dyed)
15 ppm Hwy/NRLM7
500 ppm Hwy/NRLM
HO (dyed)
1 5 ppm Hwy/NR
500 ppm LM
HO
15 ppm Hwy/NRLM
500 ppm HO9
HSHO
Post Terminal
June 1 - Sept 30
500 ppm Hwy
HS NRLM/HO (dyed)
1 5 ppm Hwy
500 ppm Hwy
500 ppm6 & HS NRLM/HO (dyed)
1 5 ppm Hwy
500 ppm Hwy
500 ppm NRLM (dyed)
500 ppm NRLM/HO (dyed)
1 5 ppm Hwy
15 ppm NRLM7 (dyed)
500 ppm Hwy
500 ppm NRLM (dyed)
HO (dyed)
1 5 ppm Hwy
15 ppm NR (dyed)
500 ppm L&M (dyed)
HO (dyed)
1 5 ppm Hwy
15 ppm NRLM (dyed)
500 ppm HO9 (dyed)
HS HO (dyed)
1 Terminal used as shorthand refers to the point where taxes are paid on highway fuel, dye added to NRLM, or marker
 added to heating oil.
2 The 15 ppm highway diesel and the 500 ppm NRLM early credit provisions are effective.
3 500 ppm NRLM program effective 2007 and no HS NRLM may be sold except small-refiner and credit HS NRLM. HS
 NRLM small-refiner and credit-use fuel may not be sold in the Northeast/Mid-Atlantic Area.
4 15 ppm NRLM early credit provisions effective 2009.
5 15 ppm NRLM program effective 2010. No 500 ppm NRLM small refiner, credit use, or pipeline interface generated
 fuels may be sold in the Northeast/Mid-Atlantic Area.
6 500 ppm NRLM credit fuel.
7 15 ppm NRLM credit fuel.
8 500 ppm heating oil is not required, but is a fuel grade that some refiners may choose to produce  and distributors
 transport. Earlier, when 500 ppm NRLM was available, such fuel could have been used for heating purposes.
                                                 5-47

-------
Final Regulatory Impact Analysis
                                           Table 5.5.1.2.-3:
                        Summary of Possible Product Segregations in Alaska
Time Frame
Current
2004
2006-20072
2007-20093
2009-20104
2010-2012
201 2-20 145
20 14 & later6
Refinery Gate
500 ppm Hwy
HSNRLM/HO
1 5 ppm Hwy
500 ppm Hwy/NRLM7
HSNRLM/HO
1 5 ppm Hwy
500 ppm Hwy/NRLM
HS NRLM10
HO
15 ppm Hwy /NRLM8
500 ppm Hwy/NRLM
HS NRLM10
HO
15 ppm Hwy /NR
500 ppm LM
500 ppm NR12
HO
15 ppm Hwy /NRLM
500 ppm NRLM11
HO
15 ppm Hwy /NRLM
500 ppm HO9
HO
Distribution System to
Terminal1
500 ppm Hwy
HSNRLM/HO
1 5 ppm Hwy
500 ppm Hwy/NRLM7
HSNRLM/HO
1 5 ppm Hwy
500 ppm Hwy/NRLM
HS NRLM10
HO
15 ppm Hwy /NRLM8
500 ppm Hwy/NRLM
HS NRLM10
HSHO
1 5 ppm Hwy/NR
500 ppm LM
500 ppm NR12
HO
15 ppm Hwy /NRLM
500 ppm NRLM11
HO
15 ppm Hwy /NRLM
500 ppm HO9
HO
Post Terminal
500 ppm Hwy
HSNRLM/HO
1 5 ppm Hwy
500 ppm Hwy
500 ppm7 & HS NRLM/HO (dyed)
1 5 ppm Hwy
500 ppm Hwy/NRLM
HS NRLM10
HO
15 ppm Hwy /NRLM8
500 ppm Hwy/NRLM
HS NRLM10
HSHO
15 ppm Hwy/NR
500 ppm LM
500 ppm NR12
HO
15 ppm Hwy /NRLM
500 ppm NRLM11
HO
15 ppm Hwy /NRLM
500 ppm HO9
HO
1 Terminal used as shorthand refers to the point where taxes are paid on highway fuel, dye added to NRLM, or marker
 added to heating oil.
2 The 15 ppm highway diesel and the 500 ppm NRLM early credit provisions are effective.
3 500 ppm NRLM program effective 2007.  HS NRLM small-refiner provisions require segregation and tracking of HS
 NRLM.
4 15 ppm NRLM early credit provisions effective 2009.
5 15 ppm NRLM program effective 2010. HS NRLM small-refiner provisions expire in 2010. 500 ppm small-refiner
 NRLM provisions require segregation and tracking of 500 ppm NRLM.
6 500 ppm NRLM small-refiner provisions expire 2014.
7 500 ppm NRLM credit fuel.
8 15 ppm NRLM credit fuel.
9 500 ppm heating oil is not required, but is a fuel grade that some refiners may choose to produce and distributors
 transport. Earlier, when 500 ppm NRLM was available, such fuel could have been used for heating purposes.
10 Segregated HS NRLM small-refiner fuel only.
11 Segregated 500 ppm NRLM small-refiner fuel only.
12 Segregated 500 ppm NR  small-refiner fuel only.
                                                 5-48

-------
                                                                       Fuel Standard Feasibility
                                          Table 5.5.1.2.-4:
 Summary of Possible Product Segregations In Areas of the Country Supplied with only a Single
   Grade of No.2 Diesel Fuel by Pipeline (outside of the Northeast/Mid-Atlantic Area and AK)
Time Frame
Current
2004
2006-20073
2007-20094
2009-20105
201 0-20 146
2014 &
later7
Refinery Gate1
June 1 - May 3 1
500 ppm Hwy
1 5 ppm Hwy
1 5 ppm Hwy
15ppmHwy/NRLM13
15ppmHwy/NRLM
15ppmHwy/NRLM
Distribution to Terminal2
June 1 - Aug 1 5
500 ppm Hwy
1 5 ppm Hwy
1 5 ppm Hwy
15 ppm Hwy /NRLM13
15 ppm Hwy /NRLM
15 ppm Hwy /NRLM
Post Terminal
June 1 - Sept 30
500 ppm Hwy
500 ppmNRLM (dyed)
NRLM/HO (dyed)
1 5 ppm Hwy
15 ppmNRLM (dyed)
500 ppm NRLM8 (dyed)
HS NRLM/HO8 (dyed)
1 5 ppm Hwy
15 ppmNRLM (dyed)
500 ppm Hwy8
500 ppmNRLM8 (dyed)
HS NRLM'VSOOppmNRLM8 (dyed)
HO (dyed & marked)8
1 5 ppm Hwy
15 ppmNRLM10 (dyed)
500 ppm Hwy8
500 ppmNRLM8 (dyed)
HS NRLM8/500 ppmNRLM8 (dyed)
HO (dyed & marked)8
1 5 ppm Hwy
15 ppmNRLM (dyed)
500 ppmNRLM8 (dyed)
500 ppm HO8 (dyed & marked)
HS HO (dyed & marked)8
1 5 ppm Hwy
15 ppmNRLM (dyed)
500 ppm L&M (dyed)9
500 ppm HO8 (dyed & marked)
HS HO (dyed & marked)8
1 Refinery rack sales are covered under the "Post Terminal" segment.
2 The term "terminal" is used as shorthand to refer to the point where taxes are paid on highway fuel, dye added to
 NRLM, or marker added to heating oil.
3 15 ppm highway diesel program and 500 ppm NRLM early credit generating provisions are effective 2006.
4 500 ppm NRLM program effective 2007.
5 15 ppm NRLM early credit generating provisions effective 2009.
6 15 ppm NRLM program effective 2010. HS NRLM small-refiner and credit-use provisions expire.
7 500 ppm NRLM small-refiner and credit-use provisions expire 2014.
8 Refinery rack sales or sales at terminals of segregated interface.
9 Sales at terminals of segregated interface and from transmix processors of fuel produced from transmix.
10 15 ppm early credit generating NRLM.
                                                5-49

-------
Final Regulatory Impact Analysis
       5.5.1.3 Ability of Fuel Distributors to Handle New Product Segregations that Will
       Result from the NRLM Sulfur Control Program

       As noted in Section 5.5.1.1, distribution feasibility concerns related to new product
segregations primarily pertain to the ability of fuel distributors to bear the economic burden of
installing new storage tanks and other equipment.  Thus, the issue is one of cost not feasibility.
Representatives of terminal and bulk plant operators stated that the physical boundaries of some
of their locations and/or the local safety and environmental ordinances under which some of their
facilities operate would prevent them from installing any new storage tanks. Even where the
expansion of tankage facilities is limited by space or other considerations, the issue is still one of
the cost of providing a fuel grade meeting a more stringent standard than necessary and not one
of the feasibility of supplying fuel to a given market.  These considerations and others led us to
structure the NRLM program to minimize the number of additional new product segregations
that would be needed. As  discussed in Section 5.5.1.3, this rule allows fuels of like sulfur
content to be shipped fungibly until they leave the terminal.

       We also structured the fuel marker requirements to minimize the potential impact on
terminal operators. One issue that concerned terminal operators is that they wished to be able to
blend 500 ppm NRLM diesel fuel from high-sulfur heating oil and 15 ppm diesel fuel in order to
avoid the need to install a storage tank for 500 ppm at some of their facilities (while still being
able to serve the 500 ppm NRLM market). The final rule allows the marker to be added as the
fuel leaves the terminal, thereby providing that terminals can blend 500 ppm diesel fuel from 15
ppm highway diesel fuel and high-sulfur heating oil subject to the anti-downgrading provisions
for 15 ppm highway diesel fuel. The primary concern expressed by terminal operators regarding
the potential impact of the fuel marker pertained to the cost of installing new injection equipment
to add the marker to heating oil. The Northeast/Mid-Atlantic Area provisions exclude the area in
which the majority of heating oil will continue to be sold after implementation of this rule,
thereby minimizing this concern. Our determination of the optimal boundaries for the
Northeast/Mid-Atlantic Area is discussed in Section 5.5.1.4.

       The following sections evaluate the potential need for additional product segregation in
each segment of the distribution system from the refinery through to the end-user due to
implementation of the NRLM diesel sulfur standards.  Based on the following  discussion, we
believe the potential impacts of this final rule on the distribution system due to the need for
additional product segregation will be minimal and can be readily accommodated by industry  in
the lead time available.  See Section 7.3 of this RIA for a discussion of the increased distribution
costs that will result from this final rule.

       Refineries:

       Due to economies of scale involved in desulfurization, we expect that many individual
refineries will choose to manufacture a single grade of diesel fuel, or perhaps two grades in some
cases.  We do not anticipate that individual refineries will produce substantial quantities of all
the different diesel fuel sulfur grades (15 ppm fuel, 500 ppm, and heating oil).  Therefore, we do
not anticipate the need for additional product segregation at refineries. Because this final rule

                                          5-50

-------
                                                               Fuel Standard Feasibility
allows highway and nonroad diesel fuels to be shipped fungibly until NRLM fuel is dyed
pursuant to IRS requirements at the terminal, we do not expect that the NRLM sulfur standards
will require refiners to install new product storage tanks.M

       We do not expect that the fuel marker requirements will cause the need for additional
product segregation at the refinery.N  However, refiners that market heating oil beginning in
2007 and 500 ppm L&M diesel fuel from 2010 through 2012 from their racks outside of the
Northeast/Mid-Atlantic Area and Alaska will have to inject the marker into the fuel sold off their
refinery racks as it is loaded into tank trucks. In the NPRM, we projected that the same
equipment currently used for injection of red dye could be used to inject the fuel marker. We
now recognize that due to concerns about contaminating red dyed fuel which is required to
contain no marker, this will only be possible at refineries  at which the only untaxed fuels that
they carry are the fuels subject to the marker requirement. At other refineries, a completely new
injection system will be needed so that the existing system can continue to be used to inject red
dye into fuels in which this is required by IRS, but which this final rule prohibits from containing
the fuelmarker. Nevertheless, we do not expect that the installation of such equipment represents
a significant concern given that the cost of such equipment is modest, the number of refineries
that will need to install such equipment is limited, and the space requirements and construction
resource requirements are minimal.0

       Pipelines:

       Similar to refiners, we anticipate that most pipelines will carry only one or two of the
sulfur level grades (e.g. 15 ppm, 15 ppm and 500 ppm, or 15  ppm and HS), although in a few
instances they may carry all three.  We expect that the pipelines that we projected will carry 500
ppm fuel under the 2007  highway diesel rule's temporary compliance option (TCO) will be the
same pipelines that elect to carry 500 ppm diesel fuel after the NRLM diesel fuel program starts.
We do not expect that any common carrier pipelines will  carry 500 ppm diesel fuel after the
implementation of the 15 ppm sulfur standard for nonroad diesel fuel in 2010. All product
pipelines are expected to carry 15 ppm highway diesel fuel beginning in 2006. As noted earlier,
the final rule provides for the fungible shipment by pipeline of highway and NRLM fuels that
meet the same sulfur specification. We therefore do not expect the NRLM sulfur standards to
necessitate additional product segregation in the pipeline  distribution system.

       There is no physical separation between product batches shipped by pipeline.  When the
    M There will be no physical differences between highway and NRLM fuel produced by refiners to the same
sulfur specification. The distinction between the two fuels is made only for accounting purposes to ensure
compliance with limitations on the volume of 500 ppm highway diesel fuel that can be produced by refiners (under
the highway diesel final rule) is complied with.


    N Under this final rule, heating oil (beginning 2007) and 500 ppm sulfur L&M diesel fuel (2010-2012) must be
marked before it leaves the terminal in areas outside of Alaska and the Northeast/Mid-Atlantic Area.

    0 See Section 7.4. of this RIA for a discussion of the estimated costs of marker injection equipment.

                                           5-51

-------
Final Regulatory Impact Analysis
mixture that results at the interface between two products that touch each other in the pipeline
can be cut into the one of these products, it is referred to as interface. When the mixture must be
removed for reprocessing, it is referred to as transmix. Given that the pipeline operators will be
able to combine batches of highway and NRLM diesel fuel meeting the same sulfur
specification, we do not expect that the NRLM program will increase the volume of product
downgrade or transmix volumes. To the contrary, there may be some opportunity for improved
efficiency because of the increase in batch sizes shipped by pipeline. This potential benefit
could be significant, given that the volume of NRLM shipped by pipeline represents a sizeable
fraction of the total diesel fuel volume.

       The marker requirements for heating oil (beginning 2007) and for 500 ppm sulfur LM
diesel fuel produced by refiners or imported  (2010-2012) applies prior to leaving the terminal.
Furthermore, these marker requirements do not apply in the Norhteast/Mid-Atlantic Area where
most heating oil is used. Therefore, we do not expect that the marker requirement will result in
an increased need for product segregation in the pipeline or an increase in product downgrade or
transmix volumes.

       We believe the demand for heating oil will be sufficiently large only in the
Northeast/Mid-Atlantic to justify the continued distribution of high-sulfur diesel fuel once
nonroad, locomotive, and marine diesel fuel  is removed from the potential high-sulfur diesel
pool (by implementation of the NRLM sulfur standards). Heating oil will therefore unlikely be
present in pipeline systems that supply areas outside of the Northeast, and Mid-Atlantic states.
The pipelines that we project will handle heating oil after the requirements of this final rule take
effect are those that we earlier projected to carry  500 ppm highway diesel fuel in addition to 15
ppm from 2006-10.

       Under the final rule, all nonroad and  L&M diesel fuel produced must meet a 15 ppm
sulfur standard in 2010 and 2012 respectively.  However, limited quantities of small-refiner, and
credit fuel that could remain at 500 ppm until 2014.  Due to the reduction in the total potential
500 ppm diesel pool in 2010 and again in 2012, it is likely that some pipelines will no longer
find it economical to carry 500 ppm as well as  15 ppm diesel fuel. We are projecting that most
pipelines will elect not to carry 500 ppm diesel fuel and will carry only 15 ppm diesel fuel after
2010. This could result in some overall simplification of the diesel distribution system. We
expect that nonroad and L&M fuel, which is produced by refiners to a 500 ppm standard after
2010, will be distributed by the refiner to the end-user via segregated pathways. Outside of
Alaska and the Northeast/Mid-Atlantic Area, limited volumes of 500 ppm fuel can continue to be
produced as locomotive and marine diesel fuel from interface, and transmix indefinitely. This
fuel can also be sold as heating oil within the Northeast/Mid-Atlantic Area and Alaska.  We
anticipate that such fuel will be distributed directly from the transmix facility or terminal that
produces such fuel to the end-user. Therefore, the presence of such 500 ppm fuels in the
distribution system will not result in the need for additional product segregation in pipelines.

       A limited number of refiners outside  of the Northeast/Mid-Atlantic Area may continue to
produce high-sulfur NRLM until 2010, 500 ppm  nonroad from 2010 to 2014, and 500 ppm L&M
from 2012 to 2014 under the small-refiner and credit-use provisions. We expect most of this fuel

                                          5-52

-------
                                                             Fuel Standard Feasibility
will be distributed via segregated means from the refinery rack to the end-user. However, if such
HS or 500 ppm nonroad or L&M is shipped by pipeline, it can be combined with heating oil
meeting the same sulfur specification up to the point where it is distributed from the terminal.
Therefore, we do not expect the small-refiner or credit provisions to create the need for
additional tankage at any location in the fuel distribution system.

       Terminals:

       The product segregation needs at terminals are directly affected by the range of products
that they receive by pipeline. Thus, the discussion regarding the potential impacts of this final
rule on terminal operators closely parallels the preceding discussion on the potential impacts on
pipeline operators.  The allowance that highway and NRLM diesel fuel meeting the same sulfur
specification may be  shipped fungibly until NRLM diesel fuel must be dyed to indicate its non-
tax status upon leaving  the terminal obviates the need for additional product segregation at the
terminal for NRLM fuel meeting  the sulfur standards in this rule with the exception of a limited
number of small  additional storage tanks needed to handle "downstream flexibility" fuel created
due to interface mixing in pipelines (discussed below).  We expect that terminal operators will
generally store NRLM and highway diesel fuel meeting the same sulfur specification in the same
tank and that NRLM  fuel will be  injected with red dye, and LM diesel fuel produced or imported
injected with the fuel marker (from 2010-2012) and red dye as it is delivered from the tank into
tank trucks.

       Similarly, since the marker is required to be present in heating oil (and L&M diesel fuel
from 2010-2012) after it leaves the terminal, we expect that terminal  operators will store heating
oil and HS NRLM (allowed from 2007-2010) in the same storage tank, and 500 ppm L&M
diesel fuel (2010-2012) and 500 ppm nonroad diesel fuel (allowed until 2014) in the same
storage tank.  Marker will be added to the heating oil and 500 ppm sulfur diesel fuel (2010-2012)
when it is dispensed from the storage tank into tank trucks. A limited number of terminal
operators will need to install new equipment to inject the fuel marker. As discussed in Section
5.5.1.4, we crafted the Northeast/Mid-Atlantic Area provisions to minimize the number of
terminals that will need to install  such equipment. We do not expect that the installation of such
equipment represents a  significant concern given that the cost of such equipment is modest, the
number of terminal that will need to install such equipment is limited, and the space
requirements and construction resource requirements are minimal.

       Some terminals  outside of these Northeast/Mid-Atlantic Area may market limited
quantities of 500 ppm diesel fuel  that was generated during the distribution of 15 ppm diesel fuel
("downstream flexibility fuel"). We expect that such fuel will be marketed directly from the
terminal to the end user. Limited additional tankage will be needed at terminals to handle this
500 ppm product as discussed in Section 7.4.3.

       Bulk Plants:

       Bulk plants are secondary distributors of refined petroleum products.  They typically
receive fuel from refinery racks or terminals by tank truck and distribute off-highway diesel fuel

                                          5-53

-------
Final Regulatory Impact Analysis
in bulk by truck to end users, serving the role of the retailer. Bulk plants are one point in the
distribution system where we anticipate some additional tankage will likely be needed as a result
of this final rule.  However, we project that only a small subset of the bulk plants will be faced
with the choice of adding additional tankage. In most areas of the country, a distinct grade of
heating oil will no longer be carried, and bulk plant operators can simply switch the tank that
they previously devoted to high-sulfur service to 500 ppm NRLM service in 2007 and supply
their HO needs out of this same tank.

       In areas where heating oil is anticipated to remain as a separate grade, we anticipate that
bulk plants will face the choice of adding a new tank and perhaps demanifolding their delivery
truck(s) to distribute dyed 500 ppm NRLM diesel  fuel in addition to high-sulfur heating oil.p In
this context demanifolding refers to the process of separating a single storage tank on a delivery
tank truck (or trucks) to make two compartments.  Some bulk plants that face the choice of
installing the facilities to allow additional product segregation may find  the cost of a new storage
tank and  demanifolding their delivery truck(s) is too high, or may not have the space or
capability to add new tank.  However, such bulk plants have other options.  If they own another
bulk plant facility in the area, they may choose to optimize use of available tankage by carrying
one of the grades at each facility. Even if they do not own another facility, they may be able to
establish a similar arrangement with a terminal or other bulk plant in the area.  They could
choose to supply heating oil only during the winter months, and supply NRLM during the
summer months to both markets. Finally, they could simply choose not  to distribute one of the
fuel grades. For example, either sell NRLM for both uses or sell only heating oil and allow other
fuel distributors in the area to satisfy the NRLM market.  We anticipate that approximately 1,600
bulk plants will face the decision of adding new tankage or finding some other means of
continuing to serve both heating oil and nonroad markets.  This is the number of bulk plants that
we project will be located in the areas of the country where heating oil will continue to be
carried by the fungible distribution system after the NRLM standards take effect and where 500
ppm fuel will also be carried. Of these, we expect no more than 1,000 will choose to install a
new tank. Given the ample lead time to prepare for implementation of the NRLM sulfur
standards, the installation of additional tanks at bulk plants is an economic issue rather than a
feasibility issue. Even where the expansion of tankage facilities is limited by space or other
considerations, the issue is still one of the cost of providing a fuel grade  meeting a more stringent
standard  than necessary and not one of the feasibility of supplying fuel to a given market.

       We do not anticipate that bulk plants will invest to carry a separate 500 ppm grade of
NRfuel in addition to 15 ppm nonroad fuel after 2010.  The majority of the nonroad volume will
meet the 15 ppm sulfur standard. We expect that few, if any, bulk plants will carry 500 diesel
L&M diesel fuel since this market is not a substantial one for bulk plants.  Unless a bulk plant
had existing tankage available or supplied a majority of its fuel to NRLM uses, 500 ppm nonroad
and L&M will therefore likely be limited to refinery and terminal distribution. This is how the
bulk of the distribution of locomotive and marine diesel fuel occurs today.
    p In the Northeast/Mid-Atlantic Area heating oil would be dyed. Outside of the Northeast/Mid-Atlantic Area
and Alaska, heating oil would be dyed and marked. In Alaska, heating oil will neither be dyed or marked.

                                           5-54

-------
                                                            Fuel Standard Feasibility
       5.5.1.4 Determining the Boundaries for the Northeast/Mid-Atlantic Area

       Our goal in adopting the Northeast/Mid-Atlantic Area approach is to minimize the
number of terminals that will need to install new injection equipment and the amount of fuel that
will need to be marked, while preserving to the maximum extent possible the flexibilities for
refiners and importers. The key to balancing these somewhat competing concerns of refiners and
terminal operators is the selection of where to draw the boundary of the Northeast/Mid-Atlantic
Area.
       The Northeast/Mid-Atlantic Area approach was first suggested in comments from the
National Oil Heat Research Alliance (NORA).Q NORA suggested that limiting the small-refiner
and credit-use provisions to Petroleum Administration for Defense Districts (PADDs) 2,3,4 & 5
would make the marker requirement for heating oil unnecessary in PADD 1.  Excluding PADD 1
from the heating oil marker requirement could then eliminate nearly all costs associated with the
marker requirement, and might not impose any limits on refiners who may wish to take
advantage of the small-refiner and credit flexibilities. The definition of the 5 PADDs is
illustrated in Figure 5.5.1.4.-1.

                          Figure 5.5.1.4.-1: Definition of PADDs
                            Petroleum Administration
                               for Defense Districts
                                PAD
                PADD 5:
               West Coast,
 'ADD 4:
Rockies
                                     _
                                     PADD 3: Guff Coast

                                                               PADD 1:
                                                             East Coast
       NORA presented a PADD by PADD analysis of data from the Energy Information
Administration (EIA) regarding the volume of diesel fuel used for heating purposes compared to
the volume of fuel used in other non-highway distillate end-uses which it used to support its
suggested exclusion of PADD 1 from the marker requirement for heating oil. Selected results of
this analysis are presented in Table 5.5.1.4-1.
   Q Comments from John Huber of the National Oil Heat Research Alliance (NORA), Docket ID No. OAR-2003-
0012-0840.
                                         5-55

-------
Final Regulatory Impact Analysis
                                      Table 5.5.1.4-1
                        Ratio of Heating Oil to Other Non-Highway
Area
PADD I (Total)
PADD IA1
PADD IB
PADD 1C
PADD II
PADD III
PADD IV
PADDV
Ratio of Non-Highway Diesel Fuel Used
to Non-Highway Diesel Used for
for Heating Purposes
Other Purposes
3.57
16.73
6.73
0.31
0.34
0.09
0.22
0.31
       ' The sub-regions that make up PADD I are illustrated in Figure 5.5.1.4-3.
       NORA stated that the number of heating oil gallons paying for the application of the
small-refiner and credit provisions in PADD I would be much greater than the potential number
of gallons that might use the provisions.11 NORA stated that this indicated that the application
of the small refiner and credit provisions in PADD I was not a good value. NORA stated that an
evaluation of the cost of the marker requirement versus the potential benefits of the small-refiner
and credit provisions indicates that the application of these provisions should be limited to
PADDs in which the ratio of non-highway diesel fuel used as heating oil to non-highway diesel
fuel used for other purposes, essentially NRLM, was less than 1.

       To assess where to draw the boundaries of the Northeast/Mid-Atlantic Area we evaluated
the area supplied by the pipeline distribution systems that are expected to continue to ship
heating oil after implementation of this rule,  evaluated the magnitude of heating oil demand by
state, evaluated where the terminals are located that are likely to carry heating oil, evaluated the
distribution  area of small refmer(s) for high-sulfur NRLM diesel fuel and refiner expectations
regarding the market for high-sulfur NRLM, and solicited input from the potentially affected
parties.

       The marker requirement for 500 ppm sulfur L&M diesel fuel that will be effective
outside of the Northeast/Mid-Atlantic Area and Alaska from June 1, 2010, through May 31,
2012, was not a significant factor in our evaluation how to define the boundary of the
Northeast/Mid-Atlantic Area.  We expect that locomotive and marine diesel  fuel subject to the
marker requirements will primarily be distributed via segregated pathways from a limited
    R "Paying for" refers to the volume of heating oil bearing the costs related to the marker requirements where
these requirements are needed to make the small refiner and credit provisions enforceable.
                                           5-56

-------
                                                              Fuel Standard Feasibility
number of refineries. Therefore, a significant number of terminals will not need to handle L&M
diesel fuel that is subject to the marker requirement. Thus, the potential cost of installing
injection equipment to add the marker to 500 ppm sulfur L&M diesel fuel which is subject to the
marker requirement will be limited to only a few refineries and terminals (i.e. approximately 15,
see section 7.4.4. of this RIA).

      Area Supplied by Pipelines that are Expected to Continue to Ship Heating Oil, and
      Location of Terminals that Will Carry Heating Oil:

       After implementation of the NRLM program, we expect that the demand for heating oil
outside of the Northeast and Mid-Atlantic States will be insufficient to justify its continued
shipment as a segregated product by pipeline.  Heating oil that is shipped by pipeline into the
Northeast and Mid-Atlantic states primarily originates in the cluster of refineries located in
PADD III (e.g. in Texas and Louisiana) and is shipped on the Colonial and Plantation pipelines
North. The Buckeye/Laurel pipeline receives fuel from these pipelines for shipment North and
West into New York state and Pennsylvania.  Some heating oil shipped by pipeline in this area
will also likely originate from refineries within PADD I and from imports into New York harbor.
No heating oil flows by pipeline from PADD I into PADD II. The Buckeye/Laurel has a
pipeline through Southern Pennsylvania that ends in Pittsburgh and a pipeline in New York that
runs West to Buffalo South of the Lake Erie shore.  The Sun pipeline also runs West from
Philadelphia to Pittsburgh. A simplified illustration of these pipeline systems is presented in
Figure 5.5.1.4-2. We anticipate that the branch lines off of the main pipelines South of North
Carolina may no longer find it economical to distribute a separate grade of heating oil.
                                          5-57

-------
Final Regulatory Impact Analysis
              Figure 5.5.1.4-2:  Simplified Illustration of the Pipeline Distribution
                  System that Supplies the Northeast and Mid-Atlantic States*
                                                         Legend

                                                        Colonial Pipeline

                                                        Plantation Pipeline

                                                        Buckeye-Laurel
                                                        & Sun Pipelines
*A11 branch lines are not shown in this figure, and in some cases a more complex local system is condensed into a single
line. The location of the lines are approximate. Product flows from the South to the end of the lines.


       Magnitude of Heating Oil Demand:


       Figure 5.5.1.4-3 shows the residential heating oil use in PADD I by state and by the sub-
districts in PADD Is
    s Energy Information Administration Fuel Oil and Kerosene Sales 2002.

                                             5-58

-------
                                                               Fuel Standard Feasibility
                 Figure 5.5.1.4-3: Residential Use of Heating Oil in PADD I
     Residential Use of No. 2 Distillate in PADD I
    PADD I Refining District
  Central Atlantic
  (PADD IB) \
                                New England
                                (PADD IA)
                    Lower Atlantic
                    (PADD  1C)
Sales of No. 2 Distillate Fuel Oil
for Residential Use
  (Thousand Gallons)
                                        PADD IA
                 PADD IB
                                        ME.
                                        N.H.
                                        VT.
                                        MASS.
                                        R.I.
                                        CONN.
       271,855
       167,740
        85,505
       892,675
       135,745
       529,648
                N.Y.
                PA.
                N.J.
                DEL.
                MD
                B.C.
      1,330,288
        351,645
        366,112
         40,038
        178,006
         14,234
 PADD 1C
 W.VA  19,948
       196,518
       113,584
        15,585
        2,225
        3,785
VA.
NC.
sc.
GA
FLA
PADD I Summary
PADD 1 Total  5,192,749
PADD 1A     2,083,168
PADD IB     2,757,936
PADD 1C      351,645
(40%)
(53%)
 (7%)
       The data summary presented by NORA indicated that PADD 1C was more similar to the
other PADDs than to PADDs IA and IB with respect to the volume of heating oil used in relation
to the use of NRLM fuel.  However, a review  of the levels of heating oil by state (in Figure
5.5.1.4-1) reveals that the level of heating oil use in Virginia and North Carolina is more similar
in magnitude to that in the PADD IA and PADD IB states than to the other states in PADD 1C.
This suggests that assigning Virginia, North Carolina, and the areas in PADD IA and IB to the
Northeast/Mid-Atlantic Area but not the remaining states in PADD 1C might best balance the
criteria of excluding areas with high heating oil demand from the marker requirement while
preserving the widest possible area in which refiners could use the small-refiner and credit
provisions.

       However, a review of the pipeline map in Figure 5.5.1.4-2 and the topography of West
Virginia suggests that the Eastern panhandle of West Virginia should also be in the
Northeast/Mid-Atlantic Area.  The topography of West Virginia has dictated that in some ways
the state's Eastern panhandle is more closely linked with the surrounding states than to the rest
                                           5-59

-------
Final Regulatory Impact Analysis
of West Virginia.1  This also suggests that Eastern panhandle may receive its fuel from the
pipelines that serve the northeast and Mid-Atlantic states.  Discussion with the West Virginia
Petroleum Marketers Association confirmed that the counties in the Eastern panhandle of West
Virginia do receive their fuel from sources that draw from the Colonial and Plantation pipelines,
while the  remainder of the state receives its fuel from other sources.u Therefore, we believe that
it is appropriate to assign the counties in the Eastern panhandle of West Virginia to the
Northeast/Mid-Atlantic Area but not the rest of the state.

       We believe that states outside of PADD I should not be assigned to the Northeast/Mid-
Atlantic Area for several reasons. The first reason is that heating oil users are predominately
located in PADD I.  Therefore, assigning areas outside of PADD  1 to the Northeast/Mid-Atlantic
Area would provide relatively little relief with respect the burden of the marker requirement for
heating oil, while substantially eroding the potential benefits of the small refiner and credit
provisions under today' s rule.  Table 5.5.1.4-2 illustrates that the great maj ority of heating oil
use is localized in PADD IA and IB.

                                       Table 5.5.1.4-2
                           Residential Heating Oil Use in the U.S.
Area
U.S. Total
PADD I
PADDIA
PADD IB
PADD 1C
PADD II
PADD III
PADD IV
PADDV
Residential Heating Oil Use1
(thousand gallons)
5,830,179
5,192,749
2,083,168
2,757,936
351,645
473,972
3,138
19,796
140,524
Percent of U.S. Total
-
89.1%
35.7%
47.3%
6.0%
8.1%
0.1%
0.3%
2.4%
1 Energy Information Administration (EIA), Fuel Oil and Kerosene Sales 2002, Table 19, Adjusted Sales for Residential
Use: Distillate Fuel Oil and Kerosene.

       The estimates in Table 5.5.1.4-2 are based on the reported use and do not speak to the
sulfur content of the fuel. A sizeable fraction of the fuel reported as used as heating oil may be
spillover from the highway diesel pool. This is most likely in areas where heating oil is currently
    T West Virginia University: The Sources of the Political Agenda: Geography, History and Economy, and
Political Culture (of West Virginia), http://www.polsci.wvu.edu/faculty/dilger/PS321/CHAP-l.htm#N_3_

    u Phone conversation with the Western Virginia Petroleum Marketers Association.

                                            5-60

-------
                                                              Fuel Standard Feasibility
not distributed by pipeline.  As noted earlier, we anticipate that after implementation of the
NRLM program, heating oil will only be distributed by pipeline to supply the Northeast and
Mid-Atlantic states. Therefore, it is likely that this rule will result in a greater proportion of the
fuel used for heating purposes outside of PADD I to come from the highway diesel and NRLM
pools.  Though used for heating purposes, such spillover would be designated as highway and
NRLM, would meet the applicable sulfur standards, and thus would not be subject to the marker
requirement. The marker requirement is associated with the sulfur content of the fuel rather than
its designation.

       The second reason is that we expect that the heating oil which is sold outside of the
Northeast and Mid-Atlantic states will primarily be distributed directly  from refiner racks. We
expect that the vast majority of terminals that will continue to carry heating oil will be supplied
by the pipeline systems illustrated in Figure 5.5.1.4-2 and by marine shipments into Northern
PADD I and thus will be located adjacent to these sources. Only a few entities, primarily
refiners, would need to install new injection equipment for the heating oil marker if the marker
requirement were to apply only to areas outside of the Northeast and Mid-Atlantic states.

       Limited volumes of heating oil produced from segregated pipeline interface may be sold
at some terminals outside of the Northeast and Mid-Atlantic states.v However, we anticipate that
for many of the terminal operators that occasionally receive such fuel, the number of such fuel
batches will not be great enough to justify the installation of marker injection equipment.
Instead of adding the marker, such terminals would have the option of designating it as NRLM
through May 31, 2010, 500 ppm nonroad through May 31, 2012, 500 ppm NRLM from June  1,
2012 through May 31, 2014, or 500 ppm L&M beyond 2014. Any fuel designated as such could
still be sold  as heating oil.

       The final reason is that we believe that assigning areas outside of Northeast and Mid-
Atlantic states to the Northeast/Mid-Atlantic Area would significantly diminish the intended
relief of the  refinery flexibility provisions. Thus, we believe that implementation of the heating
oil marker requirement outside of the Northeast and Mid-Atlantic states would allow
implementation of refiner flexibilities that would be of substantial value to refiners in reducing
their compliance burden, especially small refiners who might otherwise find the burden of
compliance prohibitive, while resulting in an acceptably small burden to industry.

       Based on our assessment discussed above, the following areas seemed the best candidates
for assignment to the Northeast/Mid-Atlantic Area:  PADD 1 A, PADD IB, Virginia, North
Carolina, and the Eastern Panhandle of West Virginia.  The following section discusses how we
further refined the definition of the Northeast/Mid-Atlantic Area based  on our evaluation of the
distribution  area of small refiners and additional input from the potentially affected industries.
   v We project that the majority of this segregated interface will meet a 500 ppm specification. Under the
provisions of the final rule, such 500 ppm diesel fuel could be sold directly into the NRLM market from 2007 - 2014
and into the locomotive and marine diesel markets after 2014.

                                           5-61

-------
Final Regulatory Impact Analysis
       Input from refiners and other parties on the appropriate boundary of the Northeast/Mid-
Atlantic Area.

       A critical factor in defining the boundary of the Northeast/Mid-Atlantic Area is
evaluating its impact on small refiners' access to the small-refiner provisions.  Our evaluation of
the location of small refiners who will likely use these provisions indicates that one such small
refiner's distribution area, in Northwestern Pennsylvania, is located within the aforementioned
areas.  With the exception of this refinery, our evaluation indicated that assigning these areas to
the Northeast/Mid-Atlantic Area would not interfere with the use of the small-refiner provisions
or significantly reduce the value of the NRLM credit provisions.  We sought input from the
range of potentially affected  parties on this assessment and on how we might accommodate the
needs of the small refiner to have access to the small-refiner provisions while maintaining our
goal of minimizing the potential number of entities that would need to install injection
equipment and the volume of heating oil that would need to be marked.  The parties that we
solicited input from include:  the American Petroleum Institute, the National Petroleum Refiners
Association, the Ad Hoc Coalition of Small Refiners, the Independent Fuel Terminal Operators
Association (IFTOA), the Association of Oil Pipelines, the National Oil-heat Research Alliance,
Colonial Pipeline, Buckeye Pipeline, the American Refining Group, and Marathon-Ashland
Petroleum.

       Based on these discussions, we determined that they the small-refiner flexibilities would
remain intact if the following counties were not assigned to the Northeast/Mid-Atlantic Area:
Chautauqua, Cattaraugus, and Allegany counties in New York, and Erie,  Crawford, Warren
McKean, Potter, Mercer, Venango, Forest, Clarion, Elf,  Jefferson, and Cameron counties in
Pennsylvania. These counties are located between the two arms of the Buckeye/Laurel pipeline
that project West into New York and Pennsylvania (see Figure 5.5.1.4.-2). There are many
terminals along the paths  of these pipelines but none to our knowledge in the aforementioned
counties. Our review also indicates that it would be most consistent with current distribution
patterns to not assign the Pennsylvania border counties of Lawrence and Greene to the
Northeast/Mid-Atlantic Area. Thus, it appears that not assigning these counties to the
Northeast/Mid-Atlantic Area would not substantially increase the burden to terminal operators
and most closely conforms to the current patterns of product distribution.  Input from the all the
parties we contacted was favorable to not assigning these counties to the Northeast/Mid-Atlantic
Area.

       Conclusion:

       Based on the above, we determined that the Northeast/Mid-Atlantic Area defined below
would minimize the number  of terminals that would need to install new injection equipment and
the amount of fuel that would need to be marked, while preserving the benefits of the small-
refiner and credit high-sulfur NRLM provisions.  All the industry representatives we contacted
stated that the definition of the Northeast/Mid-Atlantic Area in the final rule represents the best
balance of the various selection criteria and meets our stated goals in adopting the exclusion-area
approach. The areas excluded from the marker requirement and where the sale of fuel
manufactured under the credit and hardship provision is  prohibited are: North Carolina, Virginia,

                                          5-62

-------
                                                              Fuel Standard Feasibility
Maryland, Delaware, New Jersey, Connecticut, Rhode Island, Massachusetts, Vermont, New
Hampshire, Maine, Washington D.C., New York (except for the counties of Chautauqua,
Cattaraugus, and Allegany), Pennsylvania (except for the counties of Erie, Warren, Me Kean,
Potter, Cameron, Elk, Jefferson, Clarion, Forest, Venango, Mercer, Crawford, Lawrence, Beaver,
Washington, and Greene), and the eight Eastern-most counties in West Virginia (namely:
Jefferson, Berkely, Morgan, Hampshire, Mineral, Hardy, Grant, and Pendleton).  The
Northeast/Mid-Atlantic Northeast/Mid-Atlantic Area is illustrated in Figure 5.5.1.4-l.w

                      Figure 5.5.1.4.-1: Northeast/Mid-Atlantic Area
5.5.2 Limiting Sulfur Contamination

       The physical hardware and distribution practices for NRLM fuel does not differ
significantly from those for current highway diesel fuel. Therefore, we do not anticipate any
new issues with respect to limiting sulfur contamination during the distribution of 500 ppm
NRLM fuel that would not have already been accounted for in distributing highway diesel fuel.
Highway diesel fuel has been required to meet a 500 ppm sulfur standard since 1993. Thus, we
expect that limiting contamination during the distribution of 500 ppm non-highway diesel engine
fuel can be readily accomplished by industry.

       In the highway diesel rule, we acknowledged that meeting a 15 ppm sulfur specification
would pose a substantial new challenge to the distribution system.  Refiners, pipelines and
terminals would have to pay careful attention to and eliminate any potential sources of
    w The Northeast/Mid-Atlantic Area is shaded.
                                          5-63

-------
Final Regulatory Impact Analysis
contamination in the system (e.g., tank bottoms, dead legs in pipelines, leaking valves, interface
cuts, etc.) In addition, bulk plant operators and delivery truck operators would have to carefully
observe recommended industry practices to limit contamination, including things as simple as
cleaning out transfer hoses, proper sequencing of fuel deliveries, and parking on a level surface.
The necessary changes to distribution hardware and practices and the associated costs are
detailed in the RIA to the highway diesel final rule.40

       We are continuing to work with industry to ensure a smooth transition to the 15 ppm
sulfur standard for highway diesel fuel. In November of 2002, a joint industry-EPA Clean
Diesel Fuel Implementation Workshop was held in Houston, Texas.  This workshop was co-
sponsored by a broad cross-section of trade organizations representing the diesel fuel producers
and distributors who will be responsible for compliance with the 15 ppm highway diesel
standard: the National Petroleum Refiners Association (NPRA), the Association of Oil Pipelines
(AOL), the Independent Fuel Terminal Operators Association (IFTOA), the National
Association of Conveniences Stores (NACS), the Society of Independent Gasoline Marketers of
America, and the Petroleum Marketers Association of America (PMAA). The workshop
featured over 20 presentations by industry the topic of distributing 15 ppm diesel fuel, as well as
a questions and answers discussion.41   Some of these presentations contained the results of the
first test programs conducted by the pipeline industry to develop procedures and identify the
changes needed to limit sulfur contamination.  These initial test programs did not resolve all of
industry's concerns related to the ability to limit sulfur contamination during the distribution  of
15 ppm diesel fuel.  However, the results were promising and indicated that with further testing
and development the distribution industry can successfully manage sulfur contamination during
the distribution of 15 ppm diesel fuel.  We understand that the fuel distribution industry is in the
process of conducting such additional work and that there are plans to develop standard industry
practices for each segment of the distribution industry to limit sulfur contamination. We will
keep abreast of developments in this area.

       Due to the need to prepare for compliance with the highway diesel program, we
anticipate that issues related to limiting sulfur contamination during the distribution of 15 ppm
NRLM diesel fuel will be resolved well in advance of the proposed 2010 implementation date
for 15 ppm sulfur standard for nonroad fuel. We are not aware of any additional issues that
might be raised unique to nonroad fuel. If anything we anticipate limiting contamination will
become easier. We expect that 15 ppm nonroad diesel fuel will be distributed in fungible
batches with 15 ppm highway diesel fuel up to the point when it leaves the terminal and nonroad
diesel fuel must be dyed per IRS requirements. The resulting larger batch sizes as a percentage
of the total 15 ppm diesel throughput may make it somewhat easier to limit sulfur contamination
and could reduce losses to product downgrade during transportation by pipeline. We also expect
that the projected absence of high-sulfur diesel fuel and heating oil in many pipeline systems will
lessen the opportunity for sulfur contamination. As a result, if anything the opportunity for
contamination should decline with the expansion of the 15  ppm pool to include nonroad and
L&M in addition to highway diesel fuel.
                                          5-64

-------
                                                               Fuel Standard Feasibility
5.5.3 Handling Practices for Distillate Fuels that Become Mixed in the Pipeline Distribution
System

       The NRLM sulfur program in this rule raises two issues regarding the potential impact on
the current handling practices for diesel fuel that become mixed with other distillate fuels or with
gasoline during transport by pipeline (pipeline interface).  The first pertains to whether there will
be suitable market for the diesel fuel that is recovered from these mixed products.  The second
pertains to whether the requirements in this rule would interfere with the operations of transmix
processors.  As discussed in the following sections, we included provisions in the NRLM
program to address these potential concerns.

       Ensuring a Suitable Market for Diesel Fuel Recovered from Pipeline Interface

       Fuel batches shipped by pipeline abut each other with no physical separation between the
batches.  Consequently, mixing between the fuel batches that abut each other in the distribution
is unavoidable. When the volume in the mixing zone (interface) meets the specifications of one
of the two fuels being shipped next to each other, the interface is  simply added to the batch of
that fuel.  For example, the interface between regular and premium gasoline is added to the
regular grade batch. Or, the interface between jet fuel and heating oil is added to the heating oil
batch.  One interface which is never added to either adjacent batch is a mixture of gasoline and
any distillate fuel, such as jet or diesel fuel. If this interface was added to the distillate batch, the
gasoline content in the interface would result in a violation of the distillate's flash point
specification.  If this interface was added to the gasoline batch, it would cause the gasoline to
violate its end point specification. Therefore, this interface must be shipped to a transmix
processor to separate the mixture into naphtha (a sub-octane gasoline) and distillate.  The 2007
highway diesel fuel program will not change this practice.  Most of the naphtha produced by
transmix processors from gasoline/distillate mixtures is usually blended with  premium gasoline
to produce regular grade gasoline.  The heaviest portion of this naphtha is typically cut into the
distillate fuel produced so as to lessen the impact on octane (and the resulting need to blend in
premium gasoline to make regular gasoline).  The distillate produced is an acceptable high-sulfur
diesel fuel or heating oil, though if the feed material was primarily low-sulfur distillate and
gasoline it will likely also meet the current 500 ppm highway fuel cap.

       The interface between jet fuel and highway diesel can not be cut into jet fuel due to end
point and other concerns.  However, it can usually be cut into  500 ppm diesel fuel as long as the
sulfur level of the jet fuel is not too  high. With the lowering of the highway standard to 15 ppm,
however, this will no longer be possible. We expect that pipelines minimize this interface by
abutting jet fuel and high-sulfur distillate in the pipeline whenever possible.  However, it will be
unavoidable under many circumstances. A substantial part of the pipeline distribution system
currently does not handle high-sulfur distillate.  We expect that the highway program and this
final rule will cause additional pipeline systems to discontinue carrying high-sulfur distillate.
Pipelines that do not carry high-sulfur distillates will generate this interface whenever they ship
jet fuel. Under the highway program and this final rule, we project that pipeline operators will
segregate this interface by cutting it into a separate storage tank.  Because this interface can be
sold as 500 ppm NRLM fuel or heating oil without reprocessing,  and because these markets exist

                                           5-65

-------
Final Regulatory Impact Analysis
nationwide, there is little impact beyond the need for refiners to produce more 15 ppm highway
diesel fuel to offset the downgraded volume, which was considered as part of the refining costs
in the highway diesel rule.

       With control of nonroad diesel fuel to 15 ppm sulfur in 2010, and L&M in 2012, the
opportunities to downgrade interface to another product become increasing limited. Where
limited this will increase costs due to the need to transport the interface to where it can be
marketed or to a facility for reprocessing. In areas with large heating oil markets, such as the
Northeast and the Gulf Coast, the control of NRLM sulfur content will still have little impact on
the sale of this interface. However, in areas lacking a large heating oil market, the sale of this
distillate interface will be more restricted.  Because this interface will be composed of 15 ppm
diesel fuel and jet fuel, we estimate that the distillate interface created should nearly always meet
a 500 ppm cap. Thus, this interface can be added to 500 ppm NRLM batches (as well as heating
oil, where it is present at the terminal) through 2014. After 2014, this 500 ppm interface fuel can
only be sold as L&M fuel or heating oil.

       In Chapter 7 of the Final RIA, we estimate the costs related to handling this interface fuel
during the three time periods (2007-2010, 2010-2014X,  and 2014 and beyond).  We project that
there will be no additional costs prior to 2010, as 500 ppm fuel will be the primary NRLM fuel
and be widely distributed.  Beyond 2010, we estimate that some terminals will have to add a
small storage tank (or dedicate an existing tank) for this fuel, as 500 ppm highway diesel fuel
and the majority of 500 ppm nonroad  disappears from the distribution system. In many places,
this interface will be the primary, if not sole source of 500 ppm fuel, so existing tankage for this
interface will be limited. We have also added shipping costs to transport this fuel to NRLM  and
heating oil users. The volume of this  interface is significant, sometimes a sizeable percentage of
the combined NRLM fuel and heating oil markets.  In the post-2014 period, the volume of this
interface fuel is larger than the combined L&M fuel and heating oil markets in certain PADDs.
Also, the volume of interface received at each terminal  will vary  substantially, depending on
where that terminal is on the pipeline.  The advantage of this is that where the interface
accumulates it may be of sufficient volume to justify marketing as a separate grade of fuel.
Conversely, the potential users of this 500 ppm interface fuel may not be located near the
terminals with the fuel necessitating additional transportation costs.

       Prior to 2014, 500 ppm fuel can be used as NRLM fuel and heating oil.  Additional
storage tanks will be needed in some cases, as this will  be the only source of 500 ppm fuel in the
marketplace.  There will also be additional costs associated with transporting this 500 ppm to an
appropriate end-user.  Starting in 2014, this interface fuel can no longer be sold to the nonroad
fuel market. Since the interface volume does not change, this increases the proportion which
gets sold to the L&M  and heating oil markets. Thus, overall, transportation distances and costs
will likely increase. We also estimate that some fuel will have to be shipped back to refineries
and reprocessed to meet a 15 ppm cap and shipped out  a second time.
   x The costs are not significantly different from 2010-2012 than they are from 2012-2014.

                                          5-66

-------
                                                              Fuel Standard Feasibility
       By allowing the 500 ppm fuel to continue to be sold into the NRLM market until 2014
and into the L&M market thereafter, the final rule removes issues regarding the feasibility of
handling this material. Without these provisions, a substantial portion of this fuel would need to
be returned to the refinery for reprocessing raising significant cost issues, since the material
would need to be transported by truck in many cases and it might be difficult to locate refiners
willing to reprocess all of the volume. As discussed above there will be some additional
transportation costs to deliver such 500 ppm to a suitable market and a limited volume will need
to be reprocessed starting in 2014.  However, as discussed in Chapter 7 of this RIA, we  expect
the associated costs will be modest and can be accommodated by fuel distributors.

       The Potential Impact on Transmix Processors

       There are two issues regarding the potential impact of this rule on transmix processors.
The first pertains to whether a transmix processor should be  subject to the requirements
applicable to all refiners. The second pertains to whether the heating oil marker requirements
will restrict their ability  sell the distillate fuels they produce  into non-heating oil markets

       As discussed above, some pipeline interfaces do not meet the specifications for sale into
any end-use market. In such cases the interface is referred to as transmix and delivered  to a
transmix processor for separation into marketable products.  Transmix processors operate
distillation towers that separate the gasoline/distillate mixture into their component parts:
gasoline  and distillate fuel (as discussed above). Transmix processors possess no facilities with
which to remove sulfur from fuel and it currently would be burdensome for them to install such
equipment. For example, they do not have access to any hydrogen for desulfurization like at a
typical refinery.  Based on these realities, we believe that it would be inappropriate to treat
transmix processors as refiners with respect to compliance with the sulfur standards under this
rule.  Consequently, the  final rule provides that transmix processors may produce fuels for sale
into the NRLM markets that meet the applicable small-refiner provisions as long as they remain
in effect. After the NRLM small-refiner provisions expire in 2014, transmix processors may
continue to sell 500 ppm fuel into the L&M market as discussed above.  This allows 500 ppm
fuel produced by transmix processors to stay in the diesel fuel market and avoids the costs that
would accrue other wise. The final rule also amends the highway program to  allow similar
flexibility for transmix processors.  Consequently, there are no feasibility issues associated with
transmix processors.

       Transmix processors stated that the presence  of a marker in heating oil would limit the
available markets for their reprocessed distillates. The feed material for transmix processors
primarily consists of the interface mixing zone between batches of fuels that abut each other
during shipment by  pipeline where this mixing zone  can not be cut into either of the adjacent
products. If marked heating oil was shipped by pipeline, the source material for transmix
processors fed by pipelines that carry heating oil would contain SY-124.  Transmix processors
stated that it would be prohibitively expensive to  segregate pipeline-generated transmix
containing the marker from that which does not contain the marker prior to processing, and that
they could not economically remove the marker during reprocessing. Thus, in cases where the
marker would be present in a transmix processor's feed material, they would be limited  to

                                          5-67

-------
Final Regulatory Impact Analysis
marketing their reprocessed distillate fuels into the heating oil market.  Since the final rule
requires that the marker be added at the terminal gate (rather than at the refinery gate), the feed
material that transmix processors receive from pipelines will not contain the marker. Hence,
they will not typically need to process transmix containing the heating oil marker, and today's
marker requirement is not expected to significantly alter their operations. There is little
opportunity for marker contamination of non-heating oil fuel to occur at the terminal and further
downstream.  In the rare instances where this might occur, the fuel contaminated would likely
also be a distillate fuel, and thus could be  sold into the heating oil market without need for
reprocessing.

5.6 Feasibility of the Use of a Marker in Heating Oil

       As discussed in Section IV.D. of the preamble, to ensure that heating oil is not shifted
into the NRLM market, we need  a way to distinguish heating oil from high-sulfur NRLM
produced under the small-refiner and credit provisions. Currently, there is no differentiation
today between fuel used for NRLM uses and heating oil.  Both are typically produced to the
same sulfur specification, and both are required to have the same red dye added prior to
distribution from downstream of the terminal. Based on recommendations from refiners, in the
NPRM, we concluded that the best approach to differentiate heating oil from high-sulfur NRLM
would be to require that a marker be added to heating oil at the refinery gate.  Since the proposal
we received additional information which allows us to rely upon recordkeeping and reporting
provisions to differentiate heating oil from high-sulfur NRLM up the point where it leaves the
terminal  (see Section IV.D. of the preamble to the final rule). The final rule therefore requires
that a marker be added to heating oil before it leaves the terminal, rather than proposed approach
of requiring it to be added at the refinery gate.Y

       Terminal operators suggested that we might also be able to rely on recordkeeping and
reporting downstream of the terminal to differentiate heating oil from high-sulfur NRLM,
thereby eliminating any need for a marker in heating oil.  However, we believe such
recordkeeping and reporting mechanisms would be insufficient to keep heating oil out of the
NRLM market downstream of the terminal under typical circumstances. We can rely on such
measures before the fuel leaves the terminal, because it is feasible to require all the facilities in
the distribution system to send us reports describing their fuel transfers.  As discussed in Section
IV.D of the preamble to the final rule, we can compare these electronic reports to identify parties
responsible for shifting heating oil into the NRLM market. Downstream of the terminal the
parties involved in the fuel distribution system become far too numerous for such a system to be
implemented and enforced (including jobbers, bulk plant operators, heating oil dealers, retailers,
and including farmers.  Reporting errors for even a small fraction would require too many
resources to track down and correct and would eliminate the effectiveness of the system.

       Our proposal envisioned that a fuel marker would be required in heating oil from June 1,
   YHeating oil sold inside the Northeast/Mid-Atlantic Area finalized under today's rule does not need to contain a
marker (see Section IV.D. of today's preamble).

                                          5-68

-------
                                                              Fuel Standard Feasibility
2007 through May 31, 2010, and that the same marker would be required in locomotive and
marine fuel from June 1, 2010 through May 1, 2014.  As a consequence of finalizing a 15 ppm
sulfur standard for locomotive and marine fuel in 2012 we are now requiring the use of a marker
in locomotive and marine fuel from 2010-2012.  However, we are also requiring the continued
use of the marker in heating oil indefinitely (see  Section IV of the preamble to the final rule).

       We proposed and are finalizing that solvent yellow 124 (SY-124) must be added to
heating oil beginning June 1, 2007, and to 500 ppm sulfur L&M diesel fuel produced or
imported from June 1, 2010 through May 31, 2012 at a concentration of 6 milligrams per liter
(mg/1).  The chemical composition of SY-124 is as follows: N-ethyl—[2-[l-(2-
methylpropoxy)ethoxyl]-4-phenylazo]-benzeneamine.z This concentration is sufficient to ensure
detection of SY-124 in the distribution system, even if diluted by a factor of 50. Any fuel found
with a marker concentration of 0.10 milligrams per liter or more will be presumed to be heating
oil from June 1, 2007 through May 31, 2010, and after May 31, 2012. From June  1, 2010
through May  31, 2012, any fuel found to contain a marker concentration of 0.10 milligrams per
liter or less will be considered heating oil if its sulfur content is above 500 ppm, or L&M diesel
fuel if its sulfur content is below 500 ppm.  Below a concentration of 0.10 mg/L, the prohibition
on the use of fuel containing the marker does not apply.

       There are a number of other types of dyes and markers.  Visible dyes are most common,
are inexpensive, and are easily  detected.  Using a second dye in addition to the red dye required
by IRS in all non-highway fuel for segregation of heating oil based on visual identification raises
certain challenges. The marker that we require under today's rule must be different from the red
dye currently required by IRS and EPA and not interfere with the identification of red dye in
distillate fuels.  Invisible markers are beginning to see more use in branded fuels and are
somewhat more expensive than visible markers.  Such markers are detected either by the
addition of a chemical reagent or by their fluorescence when subjected to near-infra-red or
ultraviolet light.  Some chemical-based detection methods are suitable for use in the field.
Others must be conducted in the laboratory due to the complexity of the detection process or
concerns regarding the toxicity of the reagents used to reveal the presence of the marker. Near-
infra-red  and  ultra-violet fluorescent markers can be easily detected in the field using a small
device and after brief training of the operator.  There are also more exotic markers available such
as those based on immunoassay, and isotopic or molecular enhancement. Such markers typically
need to be detected by laboratory analysis.

       We selected SY-124, however,  for a number of reasons:
       1)    There is considerable data and experience with it which indicates there are no
             significant issues with its use.
       2)    It is compatible with the existing red dye
       3)    Test methods exist to quantify its  concentration, even if diluted by a factor of 50
             tol
   z Opinion on Selection of a Community-wide Mineral Oils Marking System, ("Euromarker"), European Union
Scientific Committee for Toxicity, Ecotoxicity and the Environment plenary meeting, September 28, 1999.

                                          5-69

-------
Final Regulatory Impact Analysis
       4)     It is reasonably inexpensive
       5)     It can be produced and provided by a number of sources

       Effective in August 2002, the European Union (EU) enacted the requirement that SY-124
be added at 6 mg/1 to diesel fuel that is taxed at a lower rate in all EU member states.AA  Solvent
yellow 124 is referred to as the "Euromarker" in the EU. The EU has found this treatment rate to
be sufficient for their enforcement purposes while not interfering with the identification of the
various different colored dyes required by different EU member states (including the same red
dye that is required in the U.S.). Despite its name, solvent yellow 124 does not impart a strong
color to diesel fuel when used at a concentration of 6 mg/1. Most often it is reportedly nearly
invisible in distillate fuel given that the slight yellow color imparted is similar to the natural
color of many distillate fuels.BB In the presence of red dye, SY-124 can impart a slight orange
tinge to the fuel.  However, it does not interfere with the visual identification of the presence of
red dye or the quantification of the concentration of red dye in distillate fuel. Thus, the use of
SY-124 at 6 mg/1 in diesel fuel should not interfere with the use of the red dye by IRS to identify
non-taxed fuels.

       Solvent yellow 124 is chemically similar to other additives used in gasoline and diesel
fuel, and EPA has registered it as a fuel additive under 40 CFR part 79. Therefore, we expect
that its products of combustion would not have an adverse impact on emission control devices,
such as a catalytic converter.  Extensive evaluation and testing of solvent yellow 124 was
conducted by the European Commission.  This included combustion testing which showed no
detectable difference between the emissions from marked and unmarked fuel. Norway
specifically evaluated the use of distillate fuel containing solvent yellow 124 for heating
purposes and determined that the presence of the Euromarker did not cause an increase in
harmful emissions from heating equipment. Based on the European experience with solvent
yellow 124, we do not expect that  there would be concerns regarding the compatibility of solvent
yellow 124 in the U.S. fuel distribution system or for use in motor vehicle engines and other
equipment such as in residential furnaces.

       Our evaluation of the process conducted by the EU in selecting the SY-124 for use in the
EU convinced us that SY-124 was also the most appropriate marker to propose for use in heating
oil under the final rule.  We received a number of comments expressing concern about the use of
SY-124.  Based on our evaluation  of these comments (summarized below and in the Summary
and Analysis of Comments), we continue to believe that SY-124 is the most appropriate marker
to specify for use under today's rule.  The final rule therefore requires that, beginning June 1,
2007, SY-124 be added to heating oil, and from June 1, 2020 through May 31, 2012, SY-124 be
added to LM diesel fuel produced  at a refinery or imported at a concentration of 6 mg/1 before
the fuel leaves the terminal, except in  the Northeast/Mid-Atlantic Area and Alaska.
   ^ The European Union marker legislation, 2001/574/EC, document C(2001) 1728, was published in the
European Council Official Journal, L203 28.072001.

   BBThe color of distillate fuel can range from near water white to a dark blackish brown but is most frequently
straw colored.

                                          5-70

-------
                                                              Fuel Standard Feasibility
       The concerns regarding the use of SY-124 primarily pertained to: the potential impact on
jet engines if jet fuel were contaminated with SY-124; the potential health effects of SY-124
when used in fuel for heating purposes, particularly for unvented heaters; the potential cost
impact on fuel distributors and transmix processors; and the potential conflict with IRS red dye
requirements.

       The American Society of Testing and Materials (ASTM), the Coordinating Research
Council (CRC), and the Federal Aviation Administration (FAA) requested that we delay
finalizing the selection of a specific marker for use in this final rule. They requested that
selection of a specific marker should be deferred until testing could be conducted regarding the
potential impact of SY-124 on jet engines. The Air Transport Association stated that we should
conduct an extensive study regarding the potential for contamination, determine the levels at
which the marker will not pose a risk to jet engines, and seek approval of SY-124 as a jet fuel
additive. Other parties, including the Department of Defense (DoD), also stated that we should
refrain from specifying marker under this rule until industry and other potentially affected parties
can recommend an appropriate marker.  Representatives of the heating oil industry expressed a
concern that we had not conducted an independent review regarding the safety/suitability of SY-
124 for use in heating oil.

       We met and corresponded with numerous and diverse parties to evaluate the concerns
expressed regarding the use of SY-124, and to determine whether it might be more appropriate to
specify a different marker for use under today's rule. These parties include IRS, FAA, ASTM,
CRC, various marker/dye manufacturers, European distributors of fuels containing the
Euromarker, marker suppliers, and members of all segments in the U.S. fuel distribution system.

       We believe that concerns related to potential jet fuel contamination have been sufficiently
addressed for us to finalize the selection of SY-124 as the required marker in this rule.cc As
discussed in Section IV.D of the preamble to the final rule, changes in the structure of the fuel
program since the proposal have allowed us to move the point where the marker must be added
to from the refinery gate to the terminal. The vast majority of concerns regarding the potential
for contamination of jet fuel with SY-124 pertained to the shipment of marked heating oil by
pipeline. All parties were in agreement that nearly all the potential for marker contamination of
jet fuel would disappear if the point of marker addition was moved to the terminal. We  spoke
with terminal operators, both large and small, who confirmed that they maintain strictly
segregated distribution facilities for red dyed fuel and jet fuel because of jet fuel contamination
concerns. The same type of segregation practices can be readily adapted regarding the handling
of marked heating oil and jet fuel, and would be equally effective in limiting contamination of jet
fuel with SY-124.  Downstream of the terminal, the only other chance for marker contamination
of jet fuel pertains to bulk plant operators and jobbers that handle marked fuel and jet fuel.  For
the most part, these parties  also currently maintain strict segregation of the facilities used to
transport jet fuel and heating oil (or L&M fuel that will be marked under toady's rule). The one
    GCSee the Summary and Analysis of Comments for a more detailed discussion of our response to concerns about
the possible contamination of jet fuel with the heating oil marker.

                                           5-71

-------
Final Regulatory Impact Analysis
exception is that small bulk plant operators that supply small airports sometimes use the same
tank truck to alternately transport jet fuel and heating oil. In such cases, they flush the tank
compartment prior to transporting jet fuel to remove any residual heating oil left behind after the
tank is drained. We do not expect that bulk plant operators will handle marked L&M diesel fuel.

       The final rule requires that fuel which is required to contain the marker must also contain
red dye.  Therefore, the "white bucket" test that distributors currently use to detect red dye
contamination of jet fuel can also be relied upon to detect marker contamination of jet fuel.
Based on the above discussion, we concluded that the marker requirements under today's rule
would not significantly increase the likelihood  of jet fuel contamination, and that when such
contamination might occur, it could be readily identified without the need for additional testing.
Our finalization of the Northeast/Mid-Atlantic Area in (see Section IV.D.  of the preamble to the
final rule) also minimizes potential concerns regarding the potential that jet fuel  may become
contaminated with the marker since no marker is required in heating oil (or 500 ppm L&M diesel
fuel produced by refiners or imported from 2010-2012) in this area and there is expected to be
little heating oil used outside of the Northeast/Mid-Atlantic Area.

       This final rule requires addition of the marker at the terminal rather than the refinery gate
as proposed. Based on this change, ASTM withdrew its request to delay finalization of the
marker requirements in this rule. However, ASTM stated that some concern remains regarding
jet fuel contamination downstream of the terminal (due to the limited use of the same tank
wagons to alternately transport jet fuel and heating oil discussed above). Nevertheless, ASTM
related that these concerns need not delay finalization of the marker requirements in this rule.
ASTM intends to support a CRC program to evaluate the compatibility of markers with jet fuel.
FAA is also undertaking an effort to identify fuel markers that would be compatible for use in jet
fuel.  We commit to a review of the use of SY-124 in the future based on the findings  of the
CRC and the FAA, experience with the use of SY-124 in Europe,  and future input from ASTM
or other concerned parties. If alternative markers are identified that do not raise concerns
regarding the potential contamination of jet fuel, we will initiate a rulemaking to evaluate the use
of one of these markers in place of SY-124.

       After 2010, today's rule removes the current EPA refinery gate requirement that any
diesel fuel that not meet the specifications for highway diesel fuel must contain visible  evidence
of red dye (40 CFR § 80.520(b)(2)).  This requirement means that diesel fuel which does not
meet highway diesel specifications must currently be dyed before  it is shipped by pipeline from
the refinery. As a result of the implementation of today's rule, we do not expect that any red
dyed fuel will be shipped by pipeline due to the need to comply with EPA requirements after
2010. Based on this change, we expect that today's rule will actually result in an overall
reduction in the potential  for jet fuel to become contaminated with azo dyes such as red dye and
SY-124.

       Since the NPRM,  no new information has been provided which indicates that the
combustion of SY-124 in heating equipment would result in more harmful emissions than when
combusted in engines, or would result in more harmful emissions than combustion of unmarked
heating oil. The European experience with the use of solvent yellow 124 and the evaluation

                                         5-72

-------
                                                              Fuel Standard Feasibility
process it underwent prior to selection by the EU, provides strong support regarding the
compatibility of SY 124 in the U.S. fuel distribution system, and for use in motor vehicle engines
and other equipment such as in residential furnaces. We believe that hypothesized concerns
regarding health impacts from the use of SY-124 do not present sufficient cause to delay
finalization of the marker requirements under today's rule.

       The European Union intends to review the use of Solvent yellow 124 after December
2005, but may undertake the review earlier if any health and safety or environmental concerns
about its use are raised.  We intend to keep abreast of such activities and may initiate our own
review of the use of solvent yellow 124 depending on the European Union's findings, or other
relevant information. There will be nearly four years  of accumulated field experience with the
use of SY-124 in Europe at the time of the review by the EU and nearly 5 years by
implementation of the marker requirement under this rule. This will provide ample time to
identify any new issues with SY-124 and to choose a different marker if warranted.

       Commenters stated that potential health concerns regarding the use of SY-124 might be
exacerbated with respect to its use in unvented space heaters. Commenters further stated that
there are prohibitions against the dying of kerosene (No. 1 diesel) used in such heaters.  No
information was provided to support these concerns, however, and we have no information to
suggest any health concerns exist regarding the use of SY-124 in unvented heaters.
Nevertheless, even if there were such concerns, this rule will not require SY-124 to be used in
the fuel used in unvented heaters. Furthermore, this rule does not require that SY-124 be added
to kerosene. This resolves most of what concern might remain regarding this issue, since
kerosene is the predominate fuel  used in unvented heaters.  However, the DoD stated that diesel
fuel is sometimes used in its tent heaters and expressed concern regarding the presence of SY-
124 in fuel used for this purpose.  We understand that to simplify the DoD fuel distribution
system, it is DoD policy to use a single fuel called JP-8 for multiple purposes where practicable,
including space heating.  Neither JP-8 not diesel fuel used for such a purpose would not be
subject to the heating oil marker requirement in this rule.

       We believe that the concerns expressed regarding the potential impact on distributors and
transmix processors from the presence of SY-124 in heating oil have been addressed by moving
the point of marker addition to the terminal.  Terminal operators  stated that they desire the
flexibility to blend 500 ppm diesel fuel from 15 ppm diesel fuel and heating oil. This practice
would have been prevented by the proposed addition of the marker at the refinery gate.  Under
the final rule, terminal operators will have access to unmarked high-sulfur fuel with which to
manufacture 500 ppm diesel fuel by blending with 15 ppm diesel fuel.00

       Transmix processors stated that the presence of a marker in heating oil would limit the
available markets for their reprocessed distillates. The feed material for transmix processors
   DDTerminals that manufacture 500 ppm diesel fuel by blending 15 ppm and high-sulfur fuel are treated as a
refiner under the final rule.  They must also comply with all applicable designate and track requirements, anti-
downgrading provisions, and other applicable requirements (see Section IV.D of the preamble to the final rule).

                                           5-73

-------
Final Regulatory Impact Analysis
primarily consists of the interface mixing zone between batches of fuels that abut each other
during shipment by pipeline where this mixing zone can not be cut into either of the adjacent
products.  If marked fuel was shipped by pipeline, the source material for transmix processors
fed by pipelines that carry heating oil (or marked L&M diesel fuel) would contain SY-124.
Transmix processors stated that it would be prohibitively expensive to segregate pipeline-
generated transmix containing the marker from that which does not contain the marker prior to
processing, and that they could not economically remove the marker during reprocessing. Thus,
in cases where the marker would be present in a transmix processor's feed material, they would
be limited to marketing their reprocessed distillate fuels into the heating oil market (or the L&M
market from 2010-2012 if the fuel met a 500 ppm sulfur specification).  Since the final rule
requires that the marker be added at the  terminal gate (rather than at the refinery gate), the feed
material that transmix processors receive from pipelines will not contain the marker. Hence,
they will not typically need process transmix containing the marker, and the marker requirement
is not expected to significantly  alter their operations.  There is little opportunity for marker
contamination of fuels that are required  to be marker free to occur at the terminal and further
downstream.  In the rare instances where this might occur, the fuel contaminated would likely
also be a distillate fuel, and thus could be sold into  the heating oil market (or the L&M market
from 2010-2012 if the fuel met a 500 ppm sulfur specification) without need for reprocessing.

       We do not expect that the marker requirement will result in the need for additional fuel
storage tanks or tank trucks in the  distribution system. As discussed in Section VIA of the
preamble to the final rule, we project that implementation of the NRLM sulfur standards will
result in the need for additional storage tanks and tank truck demanifolding at a limited number
of bulk plant facilities.  The marker requirement does not add another criteria apart from the
sulfur content of the fuel which would force additional product segregation.

       As discussed above, industry has expressed concern about the use of the same tank trucks
to alternately transport marked  fuel and jet fuel. We do not expect that the addition of marker to
heating oil (and 500 ppm sulfur diesel fuel produced by refiners or imported from 2010-2012)
will exacerbate these concerns. However, depending on the outcome of the aforementioned
CRC program, the fuel  marker requirements under  today's rule may hasten the current trend to
avoid the use of tank trucks to alternately transport jet fuel and heating oil (or L&M diesel fuel
to the extent that this occurs today).  To the extent that this does occur, we do not expect that it
would result in substantial additional costs since few tank truck operators currently use the same
tank truck compartments to alternately transport heating oil and jet fuel and we are aware of no
instances where tank truck operators currently use the same tank truck compartments to
alternately transport L&M diesel fuel and jet fuel.

       Through our discussions with the IRS, we have confirmed that the presence of SY-124
will not interfere with enforcement of their red dye requirement.™ Although, SY-124 may
impart a slight orange tint to red-dyed diesel fuel, this will not complicate the identification  of
the presence of the IRS red dye. In fact, IRS has determined that the presence of SY-124 may
   EEPhone conversation between Carl Dalton, IRS and Jeff Herzog, EPA February 19, 2004.

                                          5-74

-------
                                                             Fuel Standard Feasibility
even enhance enforcement of their fuel tax program.FF However, as identified in the comments,
implementation of the marker requirement for heating oil arguably may be in conflict with IRS
regulations at 26 CFR 48.4082-l(b), which states that no dye other than the IRS-specified red
dye must be present in untaxed diesel fuel. IRS is evaluating what actions might be necessary to
clarify that the addition of SY-124 to heating oil would not be in violation of IRS regulations.
IRS related that they are investigating a family of markers for potential use in addition to red dye
under their diesel tax program which might be compatible with jet fuel.  IRS stated that the use
of one of the markers in this family under this rule might result in a reduced burden on industry.
Given the changes reflected in the final rule, the marker provisions will not impose a significant
burden.  However, if the IRS program were to develop alternate markers that would be
compatible with jet fuel, we will initiate a rulemaking to evaluate the use of one of these markers
in place of SY-124 for heating oil.

       Commenters also expressed concerns regarding the proprietary rights related to the
manufacture and use of SY-124, and stated that we should adopt a nonproprietary marker if
possible. The proprietary rights related to SY-124 expire several months after implementation of
the marker requirements in this rule.  Therefore, we do not expect that the current proprietary
rights regarding SY-124 are a significant concern.  Commenters also stated that our estimated
cost of SY-124 in the NPRM (0.2 cents per gallon of treated fuel) was high compared to other
markers that cost hundredths of a cent a gallon.  Since the proposal we have obtained more
accurate information which indicates that the current cost of bulk quantities of SY-124 is
approximately 0.03 cents per gallon of treated fuel (see Section 7.4. of this RIA). Based on
conversations with various marker manufacturers, this cost is comparable to or less than other
fuel markers.

5.7  Impacts on the Engineering and Construction Industry

       An important aspect of the feasibility of any fuel quality program is the ability of the
refining industry to design and construct any new equipment required to meet the new fuel
quality standard. In this section we assess the impact of the final NRLM fuel program on
engineering design and construction personnel needs. Specifically, we focus on three types of
workers: front-end designers, detailed designers and construction workers needed to design and
build new desulfurization equipment.  In doing this, we consider the impacts of the Tier 2
gasoline sulfur and the 2007 highway diesel sulfur programs on these same types of personnel.
We compare the overall need for these workers to estimates of total employment in these areas.
In general, it would also be useful to expand this assessment to specific types of construction
workers which might be in especially high demand, such as pipe-fitters and welders. However,
estimates of the number of people currently employed in these job categories are not available.
Thus, it is not possible to determine how implementing the nonroad diesel fuel sulfur cap and
other programs might stress the number of personnel needed in specific job categories.

       To accomplish this task, we first estimated the level of design and construction resources
    Fibid

                                          5-75

-------
Final Regulatory Impact Analysis
related to revamped and new desulfurization equipment. We next projected the number of
revamped and new desulfurization units which would be needed under the final NRLM fuel
program. Then, we developed a schedule for how desulfurization projects due to be completed
at the same time might be spread out during the year. We next developed a time schedule for
when the various resources would be needed throughout each project. Finally, we project the
level of design and construction resources needed in each month and year from 2004 and 2014
and compare this to the number of people employed in each job category.

5.7.1 Design and Construction Resources Related to Desulfurization Equipment

       The number of job-hours necessary to design and build individual pieces of equipment
and the number of pieces of equipment per project were taken from an NPRA technical paper by
Moncrief and Ragsdale.42 Their study was performed to support a recent National Petroleum
Council study of gasoline and diesel fuel desulfurization, as well as other potential fuel quality
changes.43 These estimated job hours are summarized in Table 5.7-1.

                                      Table 5.7-1
               Design and Construction Factors for Desulfurization Equipment


Number of Pieces of Equipment per Refinery

Gasoline2
New
Hydro treater
60

Highway and
Nonroad
Diesel
Treaters
New
Hydro treater
60

Highway and
Nonroad
Diesel
Treaters
Revamp
Existing
Hydro treater
30

Job hours per piece of equipment3
Front End Design
Detailed Design
Direct and indirect construction
300
1200
9150
300
1200
9150
150
600
4575
a Revamped equipment estimated to require half as many hours per piece of equipment. All gasoline treaters for Tier 2
compliance are assumed to be new.

       As discussed in Section 5.3.2, we projected that the lead time for NRLM hydrotreater
modifications can be shortened relative to that required by other fuel programs due to refiners
combining their efforts to comply with this NRLM fuel rule with those for the 2007 highway
diesel fuel program. These tasks include scoping and corporate screening studies, technology
evaluation and permit approvals. We did not, however, reduce the level of E&C personnel
required for the NRLM fuel program to reflect these synergies. Thus, the above resource
requirements are conservative in this regard.  The primary reason for the lack of impact is that
the 2007 implementation date for the 500 ppm NRLM standard is later than the primary 2004-
2006 phase-in period for the Tier 2 gasoline and the 2006 implementation date for the 15 ppm
                                         5-76

-------
                                                             Fuel Standard Feasibility
highway diesel fuel standard.

5.7.2 Number and Timing of Revamped and New Desulfurization Units

       In the Final Regulatory Impact Analysis for the 2007 highway diesel program, we
estimated the number of new and revamped desulfurization units projected for both the Tier 2
and highway diesel fuel programs.44  We subsequently received pre-compliance reports for each
refinery in the country regarding their plans for complying with the highway diesel program.  In
most cases the information was preliminary, but never the less sufficient to provide a better
estimate of the number and timing of new diesel desulfurization units becoming operational, as
shown in Table 5.7-2.  We simplified our highway program analysis by assuming that refineries
who comply early and produce 15 ppm fuel before 2006 will invest to produce  highway fuel in
year 2006.
                                       Table 5.7-2
    Number of Gasoline and Highway Diesel Desulfurization Units Becoming Operational345
Fuel Type and Stage

New gasoline desulfurization units
Highway Diesel Desulfurization Units
(80% revamps, 20% new)
Before
2004
10


2004

37


2005

6


2006

26
96

2007

5


2008

3


2009

4


2010

6
5

a Units become operational on January 1st for gasoline desulfurization and June 1st for highway diesel desulfurization
units.

       The next step was to estimate the types of equipment modifications necessary to meet the
final rule NRLM fuel requirements. This was a complex task, due to the overlap of the highway
and NRLM fuel programs and the fact that refiners' relative production of highway and high-
sulfur distillate fuel varies dramatically. In our assessment of the cost of this rule (see Chapter
7), we separated refineries which produce high-sulfur distillate into three categories and assessed
their need for new or revamped desulfurization equipment separately. These three categories (as
also discussed in Section 7.2.1) are: highway refiners (95% or more of their no. 2 distillate
production meets highway diesel fuel specifications), high-sulfur refiners (5% or less of their no.
2 distillate production meets highway diesel fuel specifications), and mix refiners (producers of
high-sulfur distillate fuel not falling into one of the other categories). In Section 7.2.2.2, we
describe in detail how we projected the number of refiners which would build new hydrotreaters
or revamp existing hydrotreaters by calendar year in response to the final NRLM sulfur program.

       In applying the results of the cost analysis, we assumed that new hydrotreaters designed
to produce 500 ppm NRLM fuel would utilize the level of personnel for a new unit listed in the
table above. In those cases where a refiner produced 15 ppm NRLM fuel in one step, they would
utilize this same level of personnel. However,  when a hydrotreater capable of producing 500
ppm was modified to produce 15 ppm NRLM fuel, either using conventional or Process
                                          5-77

-------
Dynamics technology, we assumed that the personnel levels for a revamp applied.
       Table 5.7-3 presents the results of this analysis for the 63 refineries which we project will
produce 500 ppm and 15 ppm NRLM diesel fuel under the final program.

                                       Table 5.7-3
                   Number and Timing of NRLM Desulfurization Units

Revamped Hydrotreater
New Hydrotreater
2007
0
28
2008


2009


2010
17
24
2011


2012
9
6
2013


2014
14
2
5.7.3 Timing of Desulfurization Projects Starting up in the Same Year

       A worst-case assumption would be that all the units scheduled to start up on January 1 for
gasoline and June 1 for diesel would begin and complete their design and construction at the
exact same time. However, this is not reasonable for a couple of reasons. Our early credit
programs for gasoline, highway and nonroad diesel production will entice some refiners to make
treater modifications ahead of our program startup dates thus shifting E&C workload ahead for
these refiners.  Also, an industry-wide analysis such as this one assumes that all projects take the
same amount of effort and time. This means that each refinery is using every specific type of
resource at exactly the same time as other refineries with the same start-up date. However,  in
reality, refineries' projects will differ in complexity and scope. Even if they all desired to
complete their project on the  same date, their projects would begin over a range of months.
Thus, two projects scheduled to start up at exactly the same time are not likely to proceed
through each step of the design and construction process at the same time. Second, the design
and construction industries will likely provide refiners with economic incentives to avoid
temporary peaks in the demand for personnel.

       For these reasons, we spread out the design  and construction of units expected to start up
in the same calendar year. We assumed that 25 percent of the units would initiate design and
thus, start up each quarter leading up to the date upon which they had to be operating.

5.7.4 Timing of Design and Construction Resources Within a Project

       The next step in this analysis was to estimate how the engineering and construction
resources are spread out during a project. We developed a distribution of each type of resource
across the duration of a project for the Tier 2 gasoline and 2007 highway diesel sulfur programs.
The fractions of total hours expended each month were derived as follows.

       Per Moncrief and Ragsdale, front end design typically takes six months to complete.46 If
25 percent of the refineries scheduled to start up in  a given year start their projects every quarter,
each subsequent group of the refineries starts when the previous group is halfway through their
front end design.  Overall, front end design for the four groups covers a period of 15 months, or
6 months for the first group plus 3 months for each of the three subsequent groups.  In spreading
this work out over the 15 months, we assumed that  the total engineering effort would be roughly
equal over the middle 9 months. The effort during the first and last 3 month period would be

-------
                                                             Fuel Standard Feasibility
roughly two-thirds of that during the peak middle months. The same process was applied to the
other two job categories. The reader is referred to the Final RIA for the 2007 highway diesel
rule for a more detailed description of the methodology used.

       The distribution of resources is summarized in Table 5.7-5.

                                      Table 5.7-5
               Distribution of Personnel Requirements Throughout the Project

Duration per project
Duration for projects starting up in a
given calendar year
Month
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
Front-End Design
6 months
15 months
Detailed Engineering
1 1 months
20 months
Construction
14 months
23 months
Fraction of total hours expended per month from start of that portion of the
project
0.050
0.050
0.050
0.078
0.078
0.078
0.078
0.078
0.078
0.078
0.078
0.078
0.050
0.050
0.050








0.020
0.030
0.040
0.040
0.040
0.050
0.050
0.060
0.065
0.075
0.075
0.075
0.060
0.060
0.050
0.050
0.040
0.040
0.030
0.020



0.030
0.030
0.030
0.040
0.040
0.040
0.040
0.050
0.050
0.055
0.055
0.060
0.060
0.055
0.055
0.050
0.050
0.040
0.040
0.040
0.030
0.030
0.030
       The initiation of each of these three tasks relative to the start-up of the new equipment
and relative to each other was discussed above in Section 5.3.2.3, where we discuss the leadtime
necessary to meet the 2007, 2010 and 2012 NLRM sulfur caps. The following table summarizes
                                         5-79

-------
Final Regulatory Impact Analysis
the relative position of the first month shown in Table 5.7-5 above relative to the June 1 start
date for the two standards.

                                      Table 5.7-6
      Initiation of Activity (Number of Months Prior to Standard Implementation (June 1))

Front End Design
Detailed Engineering
Construction
2007
30
24
24
2010
42
36
36
2012
66
60
60
       As can be seen from Table 5.7-6, we assumed that the design and construction of new
hydrotreaters for the 2007 500 ppm NRLM standard would occur in a somewhat compressed
time frame due to the relatively short lead time available between the promulgation of the
NRLM rule and June 1, 2007.

5.7.5 Projected Levels  of Design and Construction Resources

       We calculated the number of workers in each of the three categories required in each
month by applying the distributions of the various resources per project (Table 5.7-5) to the
number of new and revamped hydrotreaters projected to start up in each calendar year (Tables
5.7-2 and -3) and the number of person-hours required per project (Table 5.7-1). We converted
hours of work into person-years by assuming that personnel were able to actively work 1877
hours per year, or at 90  percent of capacity assuming a 40 hour work week. We then determined
the maximum number of personnel needed in any specific month over the years 2004-2010 for
each job category both before and after the NRLM diesel fuel program. The results are shown in
Table 5.7-6.  In addition to total personnel required, the percentage of the U.S. workforce
currently employed in these areas is also shown. These percentages were based on estimates of
the most recently available employment levels on the Gulf Coast for the three job categories:
1920 front end design personnel, 9585 detailed engineering personnel and roughly  160,000
construction workers (taken from Moncrief and Ragsdale). We assumed that half of all refining
projects occurred on the Gulf Coast.
                                         5-80

-------
                                                              Fuel Standard Feasibility
                                       Table 5.7-7
                         Maximum Monthly Demand for Personnel
Program
Tier 2 Gasoline Sulfur Program
Plus Highway Diesel Fuel
Program
With Final NRLM Program
Parameter
Number of
Workers
Current
Workforce '
Number of
Workers
Current
Workforce '
Front-End
Design
383
(Jan 04)
20%
383
(Jan 04)
20%
Detailed
Engineering
2,720
(Apr 04)
28%
2,720
(April 04)
28%
Construction
17,646
(Nov 04)
11%
17,646
(Nov 04)
11%
1 Based on recent employment in the U. S. Gulf Coast, assuming that half of all projects occur in the Gulf Coast. The year
       and month of maximum personnel demand is shown in parenthesis.
       As can be seen from Table 5.7-7, the final NRLM diesel fuel program has no impact on
the maximum monthly personnel requirements for the front end, detailed design and construction
personnel.

       Table 5.7-8 presents a summary of the average annual personnel demand for the demand
for front end engineering in each year.
                                          5-81

-------
Final Regulatory Impact Analysis
                                      Table 5.7-8
                     Annual Front End Engineering Personnel Demand
Calendar Year
2002
2003
2004
2005
2006
2007
2008
2009
2010
2011
2012
2013
2014
Gasoline + Highway Diesel
Baseline
159
651
97
32
47
55
2
0
0
0
0
0
0
Plus
Final NRLM Program
159
651
97
261
87
320
49
86
23
73
13
0
0
       The impact of the NRLM program on annual front end engineering demand in Table 5.7-
8 reveals that the front end engineers will be needed for the three fuel programs considered here
for over a decade. Prior to this NRLM rule, the peak impact occurs in 2003 and decreases
thereafter. After this NRLM rule, the peak still occurs in 2003, but lesser peaks occur in 2005
2007  related to the design of new hydrotreaters in 2007 and 2010.  Because the level of front
end engineering after 2003 is much less than that in 2003, we do not expect that refiners will
experience any difficulties in obtaining the necessary front end engineering required to meet the
NRLM sulfur caps.

       Table 5.7-9 presents  a summary of the average annual personnel demand for the detailed
end engineering in each year.
                                         5-82

-------
                                                             Fuel Standard Feasibility
                                      Table 5.7-9
                      Annual Detailed Engineering Personnel Demand
Calendar Year
2002
2003
2004
2005
2006
2007
2008
2009
2010
2011
2012
2013
2014
Gasoline + Highway Diesel
Baseline
682
1,315
2,031
400
345
370
193
5
0
0
0
0
0
Plus
Final NRLM Program
682
1,315
2,031
690
1,076
760
1,041
176
273
113
235
17
0
       The impact of the NRLM program on annual detailed engineering demand in Table 5.7-9
reveals that the detailed engineers will be needed for the three fuel programs considered here for
over a decade. Prior to this NRLM rule, the peak impact occurs in 2004 and decreases
thereafter. After this NRLM rule, the peak still occurs in 2004, but lesser peaks occur in 2006
and 2008 related to the design of new hydrotreaters for 2007 and 2010. Because the level of
front end engineering after 2004 is much less than that in 2004, we do not expect that refiners
will experience any difficulties in obtaining the necessary front end engineering required to meet
the 2007 or 2010 NRLM sulfur caps.

       Table 5.7-10 presents a summary of the average annual personnel demand  for
construction workers in each year.
                                         5-83

-------
Final Regulatory Impact Analysis
                                     Table 5.7-10
                         Construction Worker Personnel Demand
Calendar Year
2002
2003
2004
2005
2006
2007
2008
2009
2010
2011
2012
2013
2014
Gasoline + Highway Diesel
Baseline
7,574
5,040
14,778
9,422
249
390
1,474
593
0
0
0
0
0
Plus
Final NRLM Program
7,574
5,040
14,778
11,469
5,326
3,830
7,370
2,596
1,904
1,057
1,632
342
0
       The impact of the NRLM program on annual construction worker demand in Table 5.7-
10 reveals that construction workers will be needed for the three fuel programs considered here
for over a decade. Prior to this NRLM rule, the peak impact occurs in 2004 and decreases
thereafter. After this NRLM rule, the peak still occurs in 2004, from which demand for
construction workers decreases less gradually to 2007. There is another relative peak in 2008,
related to the design of new hydrotreaters 2010. Because the level of front end engineering after
2004 is much less than that in 2004, we do not expect that refiners will experience any
difficulties in obtaining the necessary front end engineering required to meet the NRLM sulfur
caps.

       Thus, we believe that the E&C industry is capable of supplying the refining industry with
the equipment necessary to comply with our final nonroad diesel fuel program. We believe that
this is facilitated by the synergies obtained with highway diesel rule implementation and the later
phase in dates for nonroad compliance.

5.8  Supply of Nonroad, Locomotive, and Marine Diesel Fuel (NRLM)

       We have developed the fuel program in this final rule to minimize the impact on the
distillate fuel supply. For example, the final rule transitions the fuel sulfur level down to 15 ppm
in two steps, providing an estimated six years of leadtime for the final step for nonroad diesel
                                         5-84

-------
                                                              Fuel Standard Feasibility
fuel and eight years for L&M diesel fuel (up to ten years for small refiners). Banking and
trading provisions provide flexibility to refiners and hardship provisions are available for
qualifying refiners. To evaluate the effect of the new fuel standards on supply, we evaluated
four possible cases: (1) whether the new standards could cause refiners to remove certain
blendstocks from the fuel pool, (2) whether the new standards could require chemical processing
that loses fuel in the process, (3) whether the cost of meeting the new standards could lead some
refiners to leave a particular market, and (4) whether the cost of meeting the new standards could
lead some refiners to stop operations altogether (i.e., shut down). In all cases, as discussed
below, we have concluded that the answer is no. Therefore, consistent with our findings made
during the HD2007 rule, we do not expect this rule to cause any supply shortages of nonroad,
locomotive, or marine diesel fuel.

       Blendstock Shift: As mentioned above, we first evaluated whether certain  blendstocks or
portions of blendstocks may need to be removed from the NRLM diesel fuel pool.  Technology
exists to desulfurize any commercial diesel fuel to less than 10 ppm sulfur. Technologies, such
as hydro-dearomatization, have been used on a commercial scale.  More direct, desulfurization
technologies are just now being demonstrated fairly widely as refiners in both the United States
and Europe are producing No. 2 diesel fuel at 15 ppm sulfur or less. Pilot plant studies have
demonstrated that diesel fuels consisting of a wide range of feedstocks and containing high
levels of sulfur can be desulfurized to less than 15 ppm.  Such studies and experience have
reliably demonstrated that at pressures within the range of many current conventional
hydrotreaters, the single most important variable that limits desulfurization to very low sulfur
levels is the length of time the fuel is in contact with hydrogen and the catalyst.  This "residence
time" is primarily a function of reactor volume. Therefore,  we believe there is no technical
reason to remove certain feedstocks from the diesel fuel pool. It may cost more for refiners to
process certain blendstocks, such as light cycle oil, than others.  Consequently, there may be
economic incentives for refiners to move these blendstocks out of the diesel fuel market to
reduce compliance costs. However, that is an economic issue, not a technical issue and is
addressed next.  Thus, this rulemaking should not result in any long-term reduction in the
volume of products derived from crude oil available for blending into diesel fuel or heating oil.

       As mentioned above, certain feedstocks are more expensive to desulfurize than others.
The primary challenge of desulfurizing distillate to sulfur levels meeting the 15 ppm cap is the
presence of sterically hindered compounds, particularly those with  two methyl or ethyl groups
blocking the sulfur atom.GG These compounds are aromatic in nature, and are found in greatest
concentration in light cycle oil (LCO), which itself is highly aromatic.  These compounds can be
desulfurized readily if saturated. However, due to the much higher hydrogen cost of doing so,
could be better economically if this can be avoided. Because these compounds are large in size
and high in molecular weight due to their chemical structure, they distill near the high end of the
diesel range of distillation temperatures.  Thus, it is technically possible to segregate these
compounds from the rest of the LCO via distillation to avoid the need to desulfurize them. One
   GGMeeting a 500 ppm standard can be met without desulfurizing much or any of the sterically hindered
compounds.

                                          5-85

-------
Final Regulatory Impact Analysis
option would be to construct a separate distillation column to keep this stream separate from
other refinery streams, however, this would lead to significant capital costs and operating costs
in the form of heat input.  Another likely more cost-effective option would be to use the existing
FCC fractionator to shift these heavy molecules out of the LCO pool. They would be shifted to
slurry oil, which eventually becomes part of residual fuel. Once there, it would be very difficult
to recover them for blending into heating oil.

       Residual fuel is priced well below diesel fuel. The residual fuel oil market is  also not
growing in  the U.S. and growing only slowly worldwide. We investigated several sources of
price information, including EIA, LCM online and BP publications. According to EIA, spot
heating oil prices averaged roughly 75 cents per gallon from 2000-2003. According to the above
sources, residual fuel averaged 25-35% less, or 48-55 cents per gallon.  Thus, shifting LCO or
heavy LCO to residual fuel would involve a significant long-term reduction in revenue (and
profits), ranging from 20-27 cents per gallon.  Thus, we believe refiners will generally not
attempt to reduce the cost of desulfurizing diesel fuel in this way.

       To evaluate this possibility, using the distillate desulfurization model  described in
Section 7.2 above, we estimated the incremental cost of processing LCO (the worse of the two
blendstocks) into 15 ppm diesel fuel for each domestic refinery. On average, desulfurizing LCO
to 15 ppm sulfur cost 11.4 cents  per gallon.  However, in some cases, this cost reached 15 cents
per gallon.  The model is not able to estimate the cost of processing heavy LCO. In fact, the
quality of LCO and especially heavy LCO is very crude oil dependent.  However, the cost for
heavy LCO could be twice these amounts, since the concentration of both total sulfur and the
most difficult to remove sulfur are concentrated in the heaviest molecules.  Thus, the upper end
of the range of incremental desulfurization costs for heavy LCO could potentially exceed the loss
in revenue from shifting this material to the residual fuel  market. The U.S. residual fuel market
is small relative to the distillate fuel market, flat, and already being fulfilled.  Thus, any
significant shift would likely depress residual  fuel prices  and increase the reduction in profits,
further discouraging the shift. Worldwide, the residual fuel market is growing slowly. Thus, it
is unlikely that large volumes of LCO could leave the NRLM fuel market. However, we cannot
rule out the possibility that some LCO, particularly that produced by capital-strapped refiners,
could be  shifted to residual fuel.

       To estimate the upper limit of this shift, we estimated the volume of heavy LCO
produced by refineries whose LCO processing costs exceeded 12 cents per gallon and which
were not owned by large, integrated oil companies or small refiners. We excluded refineries
located in PADDs 2 and 4, since these refineries face sizeable transportation costs to  get this
material to  a residual fuel market, such as marine. This costly, heavy LCO represents 0.4% of
total NRLM fuel demand, a very small volume.  In this case, we would expect that this loss
could easily be made up by increased imports of 15 ppm  diesel fuel or domestic refiners facing
lower 15 ppm NRLM fuel costs.

       It is possible that refiners could exchange material between the NRLM and heating oil
markets to reduce the cost of meeting a 15 ppm cap, while still maintaining their NRLM fuel
production  volume. In our cost projections, we projected that individual refineries will produce

                                          5-86

-------
                                                              Fuel Standard Feasibility
either 15 ppm, 500 ppm or high-sulfur distillate with their existing slate of blendstocks to avoid
additional tankage and maximize economies of scale for the desulfurization equipment. Thus,
we did not assume that refiners would reduce costs by exchanging feedstocks around, such as
sending LCO to heating oil and straight-run from heating oil to NRLM diesel fuel.  Despite this,
the costs appear to be reasonable.  Thus, some refiners with adequate tankage and access to the
heating oil market may be able to reduce costs with such an exchange of feedstocks. However,
we did not factor these savings into our cost projections. Even if there were such exchanges,
they would not reduce the supply of NRLM diesel fuel.

       Processing Losses: We evaluated whether the new fuel standards might require chemical
processing that results in fuel losses.  Conventional desulfurization processes do not reduce the
energy content of feedstocks, although the feedstock composition may be slightly altered. A
conventional hydrotreater used to produce 15 ppm sulfur diesel converts about 98 percent of its
feedstock to finished diesel fuel.  About 1.5 percent of the remaining two-percent leaves the unit
as naphtha or light-crackate (i.e., gasoline feedstock), while the last 0.5 percent is split about
evenly between liquified petroleum gas (LPG) and refinery fuel gas.  Both naphtha and LPG are
valuable liquids used to  produce other finished products including gasoline.  Refiners can easily
adjust the relative amounts  of gasoline and diesel fuel produced by a unit, especially at the
process level under discussion. This additional naphtha can displace other gasoline or kerosene
blendstocks, which can then be shifted to the diesel fuel pool.  LPG, on the other hand, is used
primarily for space-heating, but depending on where it's produced and how it's cut, can be used
as a feedstock in the petrochemical industry. Because LPG can be used for  space heating, it will
likely displace some volume of heating oil, which in turn could be shifted to the diesel pool.
Currently, heating oil or high-sulfur fuel, has the same basic composition as highway diesel,
other than its sulfur content, and can be used to fuel nonroad, locomotive, and commercial
marine equipment. Thus, the desulfurization process usually has little or no direct impact on a
refinery's net fuel production.  The volume-shift from diesel fuel to fuel gas is very small (0.25
percent) and the gas can be used to reduce consumption of natural gas within the refinery.  This
discussion applies to the full effect of the new fuel standards (i.e., the reduction in sulfur content
from 3000 ppm to 500 ppm and from 500 ppm to 15 ppm). For the first step of fuel standards
the impacts are only about 40 percent of those described above.

       The conversion rate of a given feedstock to light products is reportedly much lower for
the emerging or advanced technologies than for conventional hydrotreaters. For the purposes of
this rulemaking, the newer or advanced technologies are projected to be used only as a second
step  to reduce the fuel to 15 ppm sulfur after it has been reduced from 3000  ppm to 500 ppm
using conventional hydrotreating technology.™1 We project that the Process Dynamics process
might reduce the conversion to light products for the second step by 55 percent.

       Exit the NRLM Diesel Fuel Market: We evaluated whether the compliance costs
   HH While the addition of the Process Dynamics process would facilitate the desulfurization to 15 ppm, the
Process Dynamics unit is expected to be installed as a revamp before the existing conventional hydrotreater handling
the 3000 to 500 ppm step while the conventional hydrotreater would be moved to address the 500 ppm to 15 ppm
step.

                                           5-87

-------
Final Regulatory Impact Analysis
associated with this rulemaking might cause some refiners to consider reducing their production
of NRLM or to leave those markets altogether. As mentioned above, diesel fuel and heating oil
are chemically and physically similar, except for sulfur level. Thus, beginning in mid-2007, a
refiner may shift his high-sulfur distillate from NRLM fuel to the heating oil market and avoid
the need to invest in new desulfurization equipment. Likewise, beginning in mid-2010 or mid-
2012, a refiner may shift part or all of its supply to heating oil.  The result would be a potential
oversupply of heating oil beginning in 2007. We expect such an oversupply of these fuels to
result in a substantial drop in their market price and would consequently increase the cost for a
given refiner to exit the NRLM diesel fuel markets. Furthermore, refiners may be forced to find
new export markets for their excess high-sulfur fuel. Overseas  market prices are often no higher
and are occasionally lower than those in the United States.  We believe these low market
differentials combined with the additional transportation costs will encourage most refiners to
comply with the NRLM program to remain in the domestic low-sulfur fuel markets.

       We addressed this same issue during the development of the highway diesel rule (66 FR
5002). We contracted with  Southwest Research Institute (SwRI) and with Muse, Stancil  &
Company, an engineering firm involved primarily in economic  studies and evaluations
concerning the refining industry to help us assess the potential for refiners to sell their highway
diesel fuel (< 500 ppm) or the blendstocks used to produce it into alternative markets. At that
time, Muse, Stancil & Company found that most refiners had few domestic alternatives for
accommodating highway diesel fuel or its blendstocks.  PADD  I imports significant quantities of
high-sulfur fuel for use as nonroad diesel fuel and heating oil. Muse, Stancil & Company
concluded that PADD I refineries could produce less highway fuel and more high-sulfur fuel and
still avoid over supplying the market by reducing imports. However, refineries in other PADDs
that import little, if any, high-sulfur fuel would be forced to find other, less valuable markets,
including new markets for export, if they exited the highway diesel fuel market.  We concluded
that, at current production levels, refiners faced greater economic losses trying to avoid meeting
the  15 ppm cap than by trying to comply with it, even if the market did not allow them to recover
their capital investment. We believe a similar conclusion can be drawn from an  analysis  of this
final rule for the following six reasons:

       1.      Approximately one-half of what is currently the  U.S. high-sulfur  diesel fuel
              market will have become part of the 500 ppm and 15 ppm markets by the time the
              FID2007 program and the sulfur caps on NRLM fuel have been implemented.
              Within that same time frame we expect few, if any,  of the common carrier
              pipelines, except perhaps those serving the Northeast,  to carry high-sulfur heating
              oil. Therefore, the sale of high-sulfur distillate may be limited to markets that a
              refiner can serve by truck.

       2.      The technology to desulfurize fuel, including refractory feedstocks, to less than
              500 ppm sulfur has been used commercially for over a decade.  The technology to
              reduce fuel to less than 15 ppm sulfur will have been commercially demonstrated
              in mid-2006, a full four years before the 15 ppm sulfur standard for nonroad
              diesel fuel takes effect.

-------
                                                                Fuel Standard Feasibility
       3.     The volume of fuel affected by the 15 ppm nonroad diesel fuel standard in 2010
              and L&M standard in 2012 will be a small fraction of that affected by the
              HD2007 program.  This dramatically reduces the required capital investment.

       4.     Canada, Europe and Japan are implementing rules to reduce sulfur levels in
              highway and nonroad diesel fuel to the 10-15 ppm range, which will effectively
              eliminate these regions as alternative export markets for high-sulfur fuel.

       5.     Refineries outside of the United States and Europe are operating at a lower
              percentage of their capacity than U.S. refineries.11 Capacity utilization rates at
              U.S. refineries are well over 90 percent. Historically, if refinery utilization rates
              approached their maxima, it was usually a strong indication that demand for
              finished products was high. In this environment, product prices usually rose and
              held until the demand pressure was reduced or eliminated. Foreign refinery
              utilization rates as well as wholesale prices tend to be well below domestic rates,
              again, a reflection of lower demand relative to the potential output of finished
              products.  The preceding condition can have at least two effects on the marketing
              decisions domestic refiners may face.  First, if foreign margins are low and U.S.
              market prices high, a foreign refiner can, and most likely will, sell his products
              into the U.S. market, thereby reducing the upward pressure on prices and likely
              reducing domestic refinery  margins.  And, second, it is highly unlikely that a
              domestic refiner will decide to further reduce his margins by adding the cost to
              ship his product into  a foreign market with a less stringent sulfur standard where
              wholesale prices are already lower than in the United States.  Consequently, we
              believe U.S. refiners  will not have a reasonable opportunity to export their high-
              sulfur fuel.

       6.     One measure of the overall  fiscal well-being of a refining operation is its margin.
              Refinery profit margins" during the 1990s were not very encouraging until  about
              1997.  In fact, in  1994, the net margin was less than $0.50 per refined barrel. By
              1997 it had nearly tripled and by 2000 had increased to nearly five times the 1994
              average. Margins leveled out again during 2001 and decreased somewhat during
              2002, but recovered during the last few months of 2002 and in early 2003.
              Current industry  projections into the future indicate the expectation for continued
              high profit margins.
    11 Europe currently imports diesel fuel and is expected to continue to do so. However, European sulfur caps will
be equivalent to those in the United States. Therefore, exporting distillate fuel to Europe is not an option for U.S.
refiners to avoid complying with stringent sulfur caps here.  Likewise, imports from European refiners are not likely.


    "The terms "margin" or the plural "margins" are often used in the petroleum industry in reference to several
different variables including "spread" or "spreads," "net margin" or "cash margin," "gross margin," and "profit
margin."  The numbers these terms represent are all basically a measure of a revenue minus the cost to produce that
revenue, expressed on a per barrel-basis of either crude oil or finished product(s).
                                            5-89

-------
Final Regulatory Impact Analysis
       Once refiners have made their investments to meet the NRLM diesel fuel standards, or
have decided to produce high-sulfur heating oil, we expect the various distillate markets to
operate very similar to current markets.  When fully implemented in 2014, there will be three
distillate fuels in the market, 15 ppm highway and NRLM diesel fuel^ and high-sulfur heating
oil. The resulting options are similar to the current situation in which there are two fuels—500
ppm and high-sulfur distillate. In this case, refiners with the capability of producing 15 ppm
diesel fuel have the most flexibility, since they can sell their fuel to any of the three markets.
Those refiners capable of producing only high-sulfur distillate will be able to participate in only
the heating oil market. Generally, we do not expect one market to provide vastly different profit
margins than the others, as high profit margins in one market will attract refiners from another
via investment in desulfurization equipment.

       Refinery Closure: There are several reasons why we believe refineries will not
completely close down as a result of this final rule.  One reason is that the regulations include a
provision to adjust the sulfur caps for small refiners, as well as any refiner facing unusual
financial hardship.  Another reason is that nonroad, locomotive, and marine diesel fuel is usually
the third or fourth most important product produced by the refinery from a financial perspective.
A total shutdown would mean losing all the revenue and profit from these other products.
Gasoline is usually the most important product, followed by highway diesel fuel and jet fuel. A
few refineries  do not produce either gasoline or highway diesel fuel, so jet fuel and high-sulfur
diesel fuel and heating oil are their most important products. The few refiners in this category
likely face the biggest financial challenge in meeting the requirements in this final rule.
However, those refiners will also presumably be in the best position to apply for the special
hardship provisions, presuming they do not have readily available source of investment capital.
The additional time afforded by these provisions should  allow the refiner to generate sufficient
cash flow to invest in the required desulfurization equipment.  Investment here could also
provide them the opportunity to expand into more profitable (e.g., highway diesel) markets.

       A quantitative evaluation of whether the cost of the fuel program in this final rule could
cause some  refineries to cease operations completely would be very difficult, if not impossible to
perform. A major factor in any decision to shut down is the refiner's current financial situation.
It is very difficult to assess an individual refinery's current financial situation.  This includes a
refiner's debt,  as well as its profitability in producing fuels other than those affected by a
particular regulation.  It can also include the profitability of other operations and businesses
owned by the refiner.

       Such an intensive analysis can be done to some degree in the context of an application for
special  hardship provisions, as discussed above. However, in this case, we may request detailed
financial documents that are not normally available. Prior to such application, as is the case
now, this financial information is usually confidential. Even when it is published, the data
    KK There will also be 500 ppm locomotive and marine diesel fuel produced from transmix in the distribution
system which can be used to satisfy the locomotive and marine demand, although this 500 ppm fuel will be produced
downstream at terminals.

                                           5-90

-------
                                                              Fuel Standard Feasibility
usually apply to more than just the operation of a single refinery.

       Another factor is the need for capital investments other than for this rule.  We can
roughly project the capital needed to meet other new fuel-quality specifications, such as the Tier
2 or highway diesel sulfur standards.  However, we cannot predict investments to meet local
environmental and safety regulations, nor other investments needed to compete economically
with other refiners.

       Finally, any decision to close in the future must be based on some assumption of future
fuel prices. Fuel prices are very difficult to project in absolute terms.  The response of prices to
changes in fuel-quality specifications, such as sulfur content, as is discussed in the next section,
are also very difficult to predict.  Thus, even if we had complete knowledge of a refiner's
financial status and its need for future investments, the decision to stay in business or close
would still depend on future earnings, which are highly  dependent on prices.

       Some studies in this area point to fuel pricing over the past 20 years or so and conclude
that prices will increase only to reflect increased operating costs and will not reflect the cost of
capital. In fact, the rate of return on refining assets has been poor until the late 1990s and until
recently, there has been a steady decline in the number of refineries operating in the United
States. However, this may have been due to circumstances specific to that time period.  The
primary reason is that refinery capacity utilization was less than 80 percent in 1985.

       Current refinery capacity utilization in the United States is generally considered to be at
its maximum sustainable rate. There are no regulatory mandates on the horizon that will
increase production capacity significantly, even if ethanol use in gasoline increases
substantially.LL Consistent with this, refining margins have been much better over the past few
years than during the previous 15 years and the refining industry itself is projecting good returns
for the foreseeable future.

       Conclusion:  Therefore, consistent with our findings made during the HD2007 rule and
the nonroad NPRM, we do not expect this final rule to cause any supply shortages of nonroad,
locomotive, or marine diesel fuel.

5.9 Desulfurization Effect on Other Non-Highway Diesel Fuel Properties

5.9.1  Fuel Lubricity

       Engine manufacturers depend on diesel fuel lubricity properties to lubricate and protect
moving parts within fuel pumps and injection systems for reliable performance. Unit injector
systems and in-line pumps, commonly used in diesel engines, are actuated by cams lubricated
   LL The U.S. Congress is considering legislation that would require the increased use of renewables, like ethanol,
in gasoline and diesel fuel. While the amount of renewables could be considerable, it is well below the annual
growth in transportation fuel use.

                                           5-91

-------
Final Regulatory Impact Analysis
with crankcase oil, and have minimal sensitivity to fuel lubricity. However, rotary and
distributor type pumps, commonly used in light and medium-duty diesel engines, are completely
fuel lubricated, resulting in high sensitivity to fuel lubricity.  The types of fuel pumps and
injection systems used in nonroad diesel engines are the same as those used in highway diesel
vehicles. Consequently, nonroad and highway diesel engines share the same need for adequate
fuel lubricity to maintain fuel pump and injection system durability.

       The state of California currently requires the use of the same diesel fuel in nonroad
equipment as in highway equipment.  Outside of California, highway diesel fuel is often used in
nonroad equipment when logistical constraints or market influences in the fuel distribution
system limit the availability of high-sulfur fuel. Thus, nonroad equipment has been using federal
500 ppm sulfur diesel fuel  and California diesel fuel, some of which may have been treated with
lubricity additives for nearly a decade. During this time, there has been no indication that the
level of diesel  lubricity needed for fuel used in nonroad engines differs substantially from the
level needed for fuel used in highway diesel engines.

       Diesel  fuel lubricity concerns were first highlighted during implementation of the federal
500 ppm sulfur highway diesel program and the state of California's diesel program circa 1993.47
The diesel fuel requirements in the state of California differ from the federal requirements by
substantially restricting the aromatics content of diesel fuel in addition to the sulfur content.
Considerable research remains to better understand which  fuel components are most responsible
for fuel lubricity. Nevertheless, there is evidence that the typical process used to reduce diesel
fuel sulfur content or aromatics content of diesel fuel (i.e., hydrotreating) can reduce fuel
lubricity. Consequently, implementing the sulfur standards in this final rule will likely require
some action to maintain the lubricity of non-highway diesel fuel.

       The potential impacts on fuel lubricity from  NRLM sulfur standards are associated solely
with the additional refinery processing that is necessary to meet these standards.  Although  we
are extending the cetane index/aromatics content specification to NRLM diesel fuel, we do  not
expect this to have a significant impact on fuel lubricity. We require that highway diesel fuel
meet a minimum cetane index level of 40 or, as an alternative, contain no more than 35 volume
percent aromatics.  ASTM already applies a cetane number specification of 40 to NRLM diesel
fuel, which is generally more stringent than the similar 40  cetane index specification. Because
of this, the vast majority of current NRLM  diesel fuel already meets the EPA cetane
index/aromatics specification for highway diesel fuel. Thus, the new requirement will have an
impact only on a limited number of refiners and there will  be little overall impact on other diesel
fuel qualities (including fuel lubricity) associated with producing fuel to meet the
cetane/aromatic requirement.

       Blending small amounts  of lubricity-enhancing additives increases the lubricity of poor-
lubricity fuels  to acceptable levels. These additives currently are available in the market, are
effective, and are in widespread use around the world. Several commenters on our final rule
setting a 15 ppm sulfur standard for highway diesel  fuel  indicated that biodiesel can be used to
increase the lubricity of conventional diesel fuel to acceptable levels. Some testing suggested
that only two volume percent is necessary.  However, more testing may be required to determine

                                          5-92

-------
                                                             Fuel Standard Feasibility
the necessary level of biodiesel for fuels not yet being produced, such as the 15 ppm fuel
required under this final rule.

       In the United States, there is no government or industry standard for diesel fuel lubricity.
Therefore, specifications for lubricity are determined by the market.  Since the beginning of the
500 ppm sulfur highway diesel program in 1993, fuel system producers, engine and engine
manufacturers, and the military have been working with the American Society for Testing and
Materials (ASTM) to develop protocols and standards for diesel fuel lubricity in its D-975
specifications for diesel fuel. ASTM is working towards a single lubricity specification that
would apply to all diesel fuel used in any type of engine.  The ASTM development process has
reached an agreement on the High Frequency Reciprocating Rig (HFRR) lubricity test method
and an initial lubricity level of 520 micron Wear Scar Diameter (WSD) for its lubricity
specification. The specification has been balloted four times in recent years and the current hold
up on the passing of the specification is the lack of an implementation date.  ASTM is hoping to
overcome implementation date issues by allowing an implementation date of 1/1/2005 in the
next ballot or by not putting the specification to a vote until late in 2004.  In light of this, the
California Air Resources Board (CARB) has decided to regulate lubricity starting in August
2004. Initial lubricity levels will require diesel fuel to have a WSD of < 520 microns for the
HFRR. CARB also has provisions in its regulation to lower the required lubricity level to < 460
micron WSD, HFRR pending the outcome of the work being performed by the CRC Diesel
Performance Group. CARB may withdraw this specification if ASTM reaches a consensus and
passes it's lubricity  standard before the CARB implementation date.  We will  follow suit with a
separate lubricity rulemaking similar to CARB's if ASTM does not reach a  consensus on its
lubricity  standard in reasonable time.

       Although ASTM has not yet adopted specific protocols and standards, refiners that
supply the U.S. market have been treating diesel fuel with lubricity additives on a batch to batch
basis, when poor lubricity fuel is expected.  Other evidence of how refiners  are ensuring
adequate fuel lubricity can be found in Sweden, Canada, and the U.S. military. The U.S. military
has found that traditional corrosion inhibitor additives have been highly effective in reducing
fuel system component wear.  Since 1991, the use of lubricity additives in Sweden's 10 ppm
sulfur Class  I fuel and 50 ppm sulfur Class II fuel has resulted in acceptable equipment
durability.48  Since 1997, Canada has required that its 500 ppm sulfur diesel fuel not meeting a
minimum lubricity be treated with lubricity additives.

       The potential need for lubricity additives in diesel fuel meeting a 15  ppm sulfur
specification was evaluated during the development  of EPA's highway diesel  rule. The final
highway  diesel rule did not establish a lubricity  standard for highway diesel fuel. We believe the
issues related to the need for diesel lubricity in fuel used in non-highway diesel engines are not
substantially different from those related to the need for diesel lubricity for highway engines.
Consequently, we are relying on the same industry-based voluntary approach to  ensuring
adequate lubricity in non-highway diesel fuels that we relied upon for highway diesel fuel.
Consistent with the highway diesel final rule, we believe the best approach is to  allow the
industry and the market to address the lubricity issue in the most economical manner. We expect
that a voluntary approach will provide adequate customer protection from engine failures due to

                                          5-93

-------
Final Regulatory Impact Analysis
low lubricity, while providing the maximum flexibility for the industry. We expect that the
American Society for Testing and Materials (ASTM) will finalize a fuel lubricity standard for
use by industry that could be applied to low-sulfur NRLM diesel fuel.

       The degree to which removing the sulfur content from diesel fuel may impact fuel
lubricity depends on the characteristics of the blendstocks used as well as the severity of the
treatment process.  Based on our comparison of the blendstocks and processes used to
manufacture non-highway diesel fuels, we project that the potential decrease in the lubricity of
non-highway diesel fuel that might result from the new sulfur standards will be substantially the
same as that experienced in desulfurizing highway diesel fuel to meet the same sulfur standard.

       A refiner of diesel fuel for use in California and for much of the rest of the United States
as well evaluated the impacts on fuel lubricity of the current federal and California diesel fuel
requirements.49 This refiner concluded that, reducing the aromatics content of diesel fuel
requires more severe hydrotreating than reducing the sulfur content to meet a 500 ppm standard.
Consequently, concerns regarding diesel fuel lubricity have primarily been associated with
California diesel fuel and some California refiners treat their diesel fuel with a lubricity additive
as needed.  The subject refiner stated that outside of California, hydrotreating to meet the current
500 ppm sulfur specification seldom results in a  sufficient reduction in fuel lubricity to require
the use of a lubricity additive. We expect that the same hydrotreating process currently used to
produce highway diesel fuel will be used to reduce the sulfur content of non-highway diesel fuel
to meet the 500 ppm sulfur standard during the first  step under this final rule. We therefore
estimate that there will be only a marginal increase in the use of lubricity additives in NRLM
diesel fuel meeting the 500 ppm sulfur standard for 2007.

       The highway diesel program projected that hydrotreating will be the process most
frequently used to meet the 15 ppm sulfur standard for highway diesel fuel in 2006. However,
we project that the 2010 and  2012 implementation dates for the 15 ppm standard for NRLM
diesel fuel will allow the use of advanced technologies to remove sulfur from 60 percent of the
affected diesel pool.  The use of such developing desulfurization processes is discussed in
Section 5.5. These new processes have less of a tendency to affect other fuel properties than
does hydrotreating.  Therefore, the use of such new  desulfurization technologies might tend to
have less of an impact on fuel lubricity. However, we have no specific information with which
to quantify the impacts of the developing technologies on fuel lubricity.  To provide a
conservatively high estimate of the potential impact of meeting the 15 ppm standard for nonroad
diesel fuel, we assumed that the potential impact on fuel lubricity of the new desulfurization
processes will be the same as that experienced when hydrotreating diesel fuel to meet a 15 ppm
sulfur standard. We therefore assumed, as we did for 15 ppm highway diesel fuel, that all 15
ppm NRLM diesel fuel must be treated with lubricity additives.  The cost associated with the
increased use of lubricity additives in 500 ppm NRLM  diesel fuel and in 15 ppm NRLM diesel
fuel is discussed in Chapter 7.

       Railroads and locomotive manufacturers  have expressed concern that low-sulfur fuel
might damage existing locomotives. Locomotives already use a significant amount of low-sulfur
fuel, especially in California, and there has not been any evidence of sulfur-related problems.

                                          5-94

-------
                                                             Fuel Standard Feasibility
Low-sulfur locomotive diesel fuel meeting the soon to be specified lubricity requirements will
provide adequate protection to these engine and fuel systems.

5.9.2 Volumetric Energy Content

       Some of the projected desulfurization processes for meeting the non-highway diesel
sulfur standards tend to reduce the volumetric energy content (VEC) of the fuel during
processing. Desulfuization also tends to result in a swell in the total volume of fuel.  These two
effects tend to cancel each other out so there is no overall loss in the energy content in a given
batch of fuel that is subjected to desulfurization.  Thus, we do not expect the potential reduction
in VEC that might result from the new sulfur standards to affect the refiners' ability to supply
sufficient quantities  of non-highway diesel fuel. The potential impacts on diesel supply are
discussed in Section 5.8.

       Since a greater volume of fuel must be consumed in the engine to produce the same
amount of power, however, a larger volume of fuel is needed to meet the same level of demand.
The potential increase in the distribution costs associated with a reduction in NRLM diesel VEC
is discussed in Section 7.3.

       The impact of desulfurization on diesel fuel VEC varies depending on the type of
blendstocks and desulfurization process used.  A comparison of the blendstocks used to produce
high-sulfur diesel fuel with those used to produce highway diesel fuel shows that both pools
contain similar fractions  of each type of blendstock.50 Based on this comparison, we believe a
comparable level of severity in the desulfurization process is required to produce NRLM diesel
fuel meeting a given sulfur specification as will be required to produce highway  diesel fuel
meeting the same sulfur specification. Refiners with experience in the use of
hydrodesulfurization to manufacture both 500 ppm and 15 ppm highway diesel fuel provided us
with information that we used to estimate the accompanying reduction in VEC. Using this
information, we estimate that hydrodesulfurization of NRLM diesel fuel to meet a 500 ppm
sulfur standard will result in a reduction in volumetric energy content of 0.7 percent.

       The 15 ppm sulfur standard for nonroad diesel fuel does not start until 2010 and for L&M
diesel fuel until 2012.  The additional lead time allows refiners to take advantage of several less-
expensive desulfurization technologies currently under development to produce diesel fuel
complying with the 15 ppm sulfur standard in addition to conventional hydrotreating. Of the
advanced desulfurization technologies which refiners may consider, we believe that only Process
Dynamics Isotherming will be used extensively (see Section 5.3). We project that Process
Dynamics Isotherming will be used by 60% of the NRLM market, while conventional
hydrotreating will be used by the remaining 40%.  The Process Dynamics engineers estimate that
the Isotherming desulfurization process will have less of an impact on diesel fuel volumetric
energy content than  does hydrodesulfurization. Using the mix of desulfurization technologies
we expect to be available, we estimate that desulfurizing NRLM diesel fuel from 500 ppm to 15
ppm will reduce the  volumetric energy content by an  additional 0.5 percent (0.7% conventional
hydrotreating and 0.4% for IsoTherming). Thus,  reducing the sulfur content of nonroad diesel
fuel from the current maximum 5,000 ppm sulfur cap to the 15 ppm sulfur standard is estimated

                                          5-95

-------
Final Regulatory Impact Analysis
to result in a 1.2 percent reduction in VEC. Table 5.9-1 summarizes the projections for
estimating the impact of the new sulfur standards on VEC, including: (1) the percentage of the
applicable NRLM diesel fuel pool that we expect will be desulfurized using each of the available
desulfurization processes and (2) the projected impact of each desulfurization process on VEC.

                                        Table 5.9-1
                     Projections Used in Estimating the in Reduction in
        Volumetric Energy Content Associated with  Meeting the New Sulfur Standards
Desulfurization Process3
Conventional
De sulfurization
Process Dynamics
Isotherming
Over-all Impact on VEC of
All Desulfurization
Processes Used
Percent of Diesel Pool Desulfurized
Using a Given Process to Meet the
Applicable Sulfur Standard
NRLMb
500 ppm
in 2007
100 %
NA
-
NR
15 ppm
in 2010
40%
60%

L&M
15 ppm
in 2012
40%
60%
-
Reduction in Volumetric Energy Content
Associated with a Given Desulfurization
Process
Reduction in Sulfur Content
HSC to 500 ppm
0.7%
NA
0.7%
500 ppm to 15 ppm
0.7 %
0.4 %
0.5%
a See Section 5.3 regarding the use of conventional hydrodesulfurization , and the Process Dynamics Isotherming process
       to meet the new sulfur standards.
b NR = nonroad diesel fuel, L = locomotive diesel fuel, and M = marine diesel fuel.
0 HS refers to high-sulfur diesel fuel at the current uncontrolled average sulfur level of approximately 3000 ppm.
       It is important to remember that the anticipated reduction in VEC discussed above
applies only to those gallons of NRLM diesel fuel that currently have a high sulfur content. Due
to logistical constraints in the fuel distribution system, much of the fuel used in NRLM engines
meets highway diesel fuel standards (see Section 7.1).  The costs related to the reduction in
NRLM diesel fuel VEC accompanying the new sulfur standards are discussed in Section 7.3.

5.9.3 Fuel Properties Related to Storage and Handling

       In addition to fuel lubricity additives, a range of other additives are also sometimes
required in diesel fuel to compensate for deficiencies in fuel quality.  These additives include
cold flow improvers, static dissipation additives, anti-corrosion additives, and anti-oxidants. The
highway diesel fuel program projected that,  except for an increase in the fuel lubricity additives,
reducing the sulfur content of highway diesel fuel to meet a 15 ppm standard will not result in an
increase in the use of diesel performance additives.  Since that time, we have identified no new
information to alter that projection. Consequently, our estimate of the increase in additive use
resulting from this final rule parallels that under the highway program.  We estimate that the use
of lubricity additives will increase and that the use of other additives will be unaffected.
                                           5-96

-------
                                                             Fuel Standard Feasibility
5.9.4 Cetane Index and Aromatics

       We require that nonroad, locomotive, and marine diesel fuel comply with the current
highway diesel fuel requirements for cetane index or aromatics.  Thus, these non-highway diesel
fuels must meet either a 40 minimum cetane index, or a 35 percent maximum aromatics limit.  In
this section, we present information on what these properties are currently for non-highway
diesel fuel, then we estimate how much they are likely to change when these streams are
desulfurized.

       We have reports of non-highway diesel fuel cetane index values from refinery samples
from 1997 to 2001. The 1997 and 1998 reports were published by the National Institute for
Petroleum and Energy Research (NIPER), Bartlesville, OK, and then this organization changed
their name to TRW Petroleum Technologies, which published the reports for 1999 through 2001.
The reports divided the country into the Eastern, Southern, Central, Rocky Mountain, and
Western Regions. The samples, which averaged about 17 per year, were pooled from the various
regions. The range of cetane index values for the 85  total samples is 39.4  - 57.0. Out of the 85
samples, 5 samples were under the cetane index value of 40 and potentially would not comply
with the cetane index minimum of 40. However, those that were below the 40 cetane index
minimum, were barely below it (i.e., 39.4 versus 40).  Since the aromatics levels were not
provided for these 5 samples, we could not verify if these samples would also not comply with
the aromatics part of the specification.

       As refiners desulfurize their NRLM diesel fuel to comply with the  500 ppm standard in
2007 and then again to comply with the 15 ppm standard in 2010 and 2012, we expect them to
see increased cetane levels in their NRLM diesel fuel. Vendors of the desulfurization
technologies either provided information on the impact that their technologies have on the cetane
index of diesel fuel, or we were able to estimate the impact using changes  to API gravity and the
T-50 distillation point. While the changes in cetane index were provided for the desulfurization
of highway diesel fuel, they apply to NRLM diesel fuel as well, as it is similar in quality and
composition to highway diesel fuel.  The estimated impact of the desulfurization technologies on
cetane index summarized in the following table. As described in Chapter 7, much of the high-
sulfur diesel pool is already hydrotreated (on the order of 50 percent in some PADDs) and will
therefore not be impacted by the first step of fuel control to 500 ppm, so the cetane index is
expressed as a range for the high-sulfur to 500 ppm step.  The lower value of the range reflects
the fact that refiners will have to hydrotreat only half their existing high-sulfur pool to produce
500 ppm sulfur fuel, while the upper value reflects the fact that refiners will have to treat their
entire pool. For conventional hydrotreating, a range in the amount of increase in cetane index
values is also reflected in the 500 ppm to 15 ppm sulfur reduction step, which reflects the
different estimates for the two vendors that provided us the desulfurization information.
                                          5-97

-------
Final Regulatory Impact Analysis
                                      Table 5.9-2
             Impact of Desulfurization Technologies on Diesel Fuel Cetane Index

High-Sulfur to 500 ppm
500 ppm to 15 ppm
Total High-Sulfur to 15 ppm
Conventional Hydrotreating
+2 to +4
+1 to +2
+3 to +6
Process Dynamics Isotherming
+2 to +4
+2
+4 to +6
       As summarized in the above table, conventional hydrotreating improves the cetane index
of diesel fuel by 2 to 4 numbers for the 500 ppm sulfur standard, and 1 to 2 numbers for the 15
ppm sulfur standard incremental to the 500 ppm standard.  If the lowest cetane index values of
non-highway diesel fuel are indeed between 39 and 40 as the NIPER/TRW data suggest, then the
desulfurization of that pool to comply with the 500 ppm sulfur standard, which we expect to be
accomplished using conventional desulfurization technology, is expected to increase the cetane
index to a value above the 40 minimum, thus we do not expect refiners to be constrained by a
cetane index requirement.

       Aromatics should also decrease, although this decrease is expected to occur mostly
through the saturation of polynuclear aromatics to monoaromatics.

5.9.5 Other Fuel Properties

       Desulfurization is expected to impact other qualities of non-highway diesel fuel. The
concentration of nitrogen in current high-sulfur diesel fuel  is on the order of several hundred
parts per million. The desulfurization technologies projected to be used for compliance with the
500 ppm sulfur standard are expected to lower nitrogen levels down to under 100 ppm, although
they may still be above 50 ppm. These same desulfurization technologies are expected to lower
nitrogen levels down to under 10 ppm when achieving compliance with the 15 ppm sulfur
standard.

       Conventional desulfurization and Process Dynamics Isotherming are expected to affect
the distillation temperature of NRLM diesel fuel. For desulfurizing high-sulfur diesel fuel down
to 15 ppm, one vendor of conventional hydrotreating technology estimates that each distillation
point (T-10 - T-90) will experience a 5°F decrease. Consistent with that, API gravity is expected
to increase by 4 numbers, with density decreasing commensurately. Process Dynamics
Isotherming is expected to impact the distillation temperature less than conventional
hydrotreating due to the lower API gravity increase caused by Process Dynamics compared with
conventional hydrotreating.
                                         5-98

-------
                                                             Fuel Standard Feasibility
       Appendix 5A: EPA's Legal Authority for Adopting Nonroad,
       Locomotive, and Marine Diesel Fuel Sulfur Controls

       We are adopting diesel fuel sulfur controls under our authority in section 21 l(c)(l) of the
Clean Air Act. This section gives us the authority to "control or prohibit the manufacture,
introduction into commerce, offering for sale, or sale" of any fuel or fuel additive for use in an
off-highway engine or vehicle (1) whose emission products, in the judgment of the
Administrator, cause or contribute to air pollution which may reasonably be anticipated to
endanger the public health or welfare or (2) whose emission products will impair to a significant
degree the performance of any emission control device or system which is in general use,  or
which the Administrator finds has been developed to a point where in a reasonable time it would
be in general use were the fuel control or prohibition adopted.

       We currently do not have regulatory requirements for sulfur in nonroad, locomotive, or
marine diesel fuel.  Beginning in 1993, highway diesel fuel was required to meet a sulfur cap of
500 ppm and be segregated from other distillate fuels as it left the refinery by the use of a visible
level of dye  solvent red 164 in all non-highway distillate. Any fuel not dyed is treated as
highway fuel.  Beginning in 2006, highway diesel fuel will be required to start meeting a sulfur
cap of 15 ppm.

       We are adopting controls on sulfur levels in off-highway diesel fuel based on both of the
Clean Air Act criteria described above. Under the first criterion, we believe that emission
products of sulfur in nonroad, locomotive, and marine diesel fuel used in these engines
contribute to PM and SOx pollution. As discussed in Chapter 2, emissions of these pollutants
cause or contribute to ambient levels of air pollution that endanger public health and welfare.
Control of sulfur to 15 ppm for NRLM fuel will lead to significant, cost-effective reductions in
emissions of these pollutants, with the benefits to public health and welfare significantly
outweighing the costs. In the proposal and Draft RIA EPA discussed controlling sulfur through
a first step to 500 ppm for NRLM fuel, based on the public health and welfare benefits from such
a fuel control, with a second step to 15 ppm for nonroad fuel, based on technology enablement
for associated nonroad engine  standards. EPA also discussed various alternatives, such as a
second step to 15 ppm for locomotive  and marine fuel as well as a single step to 15 ppm for
NRLM fuel, both based on the public health and welfare benefits from such a fuel sulfur control.

       Adopting a  15 ppm standard for locomotive and marine fuel makes it clear that for
purposes of section 21 l(c)(l)(A) the most appropriate way to view the final fuel control program
adopted in this rule is as a complete program, covering all of NRLM fuel. This is because the
reduction to 15 ppm for nonroad fuel is in essence no different from the reduction to 15 ppm for
locomotive and marine fuel. Basically, the same desulfurization technology is used, the same
per-gallon desulfurization costs are incurred, and the same per gallon emissions reductions and
benefits are achieved from the fuel control. The only significant difference is the magnitude of
total actual reductions and costs, based on the volume of diesel fuel controlled.  Therefore for
purposes of section 21 l(c)(l)(A), EPA has analyzed and justified the reduction of NRLM fuel


                                         5-99

-------
Final Regulatory Impact Analysis
sulfur from current sulfur levels to 15 ppm as a complete program, without drawing any
distinction between nonroad and locomotive and marine fuel.

       Under the second criterion, we believe that sulfur in nonroad diesel fuel will significantly
impair the emission-control systems expected to be in general use in nonroad engines designed
to meet the emission standards adopted in this rule. Chapter 4.1.7 describes the substantial
adverse effect of high fuel-sulfur levels on the emission-control devices or systems for diesel
engines meeting the proposed emission standards.  Controlling sulfur levels in nonroad diesel
fuel to 15 ppm will enable emission-control technology that will achieve additional significant,
cost-effective reduction in emissions of NOx, NMHC and PM pollutants, beyond that achieved
by the fuel control itself. The following sections summarize our analysis of the various issues
related to adopting fuel-sulfur controls for nonroad, locomotive, and marine diesel fuel.

5A.1 Health and Welfare Concerns of Air Pollution Caused by Sulfur in
Diesel Fuel

       At the current unregulated levels of sulfur in this diesel fuel, the emission products from
the combustion of diesel sulfur in these engines can reasonably  be anticipated to endanger public
health and welfare. Sulfur in nonroad, locomotive and marine diesel fuel leads directly to
emissions of SO2 and sulfate PM from the exhaust of diesel vehicles, both of which cause
adverse health and welfare impacts, as described in Chapter 2.  SO2 emissions from nonroad,
locomotive and marine engines are directly proportional to the amount of sulfur in the fuel.  SO2
is oxidized in the atmosphere to SOS which then combines with water to form sulfuric acid
(H2SO4) and further combines with ammonium in the atmosphere to form ammonium sulfate
aerosols.  These aerosols are what is often referred to as sulfate  PM. This sulfate PM comprises
a significant portion of the "secondary" PM that does not come  directly from the tailpipe, but is
nevertheless formed in the  atmosphere from exhaust pollutants.  Exposure to secondary PM may
be different from that of PM emitted directly from the exhaust, but the health concerns of
secondary PM are just as severe as for directly emitted particulate matter, with the possible
exception of the carcinogenicity concerns with diesel exhaust.

       Approximately 1-2% of the sulfur in nonroad, locomotive and marine diesel fuel is not
converted into SO2, but is instead further oxidized into SOS which then forms sulfuric acid
aerosols (sulfate PM) as it leaves the tailpipe. While only a small fraction of the overall sulfur is
converted into sulfate emissions in the exhaust, it nevertheless accounts for approximately 10%
of the total PM emissions from diesel engines today. This sulfate PM is also directly
proportional to the sulfur concentration in the fuel.  The health and welfare implications of
emissions of PM and SO2 and the need for reductions in these emissions are discussed in
Chapter 2.

       The reduction in the sulfur level of nonroad, locomotive, and marine diesel fuel to 15
ppm would achieve in excess of 99 percent reduction in the emissions of SO2 and sulfate PM
emissions from nonroad, locomotive, and marine diesel engines compared with today's levels.
The first step to 500 ppm would achieve about a 90% reduction and the second step to 15 ppm


                                         5-100

-------
                                                              Fuel Standard Feasibility
would achieve in excess of a 99 percent reduction in these pollutants.

       EPA has evaluated the technical feasability of achieving these sulfur levels, including the
cost of the reductions and the impact on fuel supply.  EPA has concluded that these reductions
are feasible in the lead time provided, and should not have an adverse impact on the adequacy of
NRLM fuel supply to meet demand; see RIA Chapter 5.

       EPA also evaluated the emissions reductions achieved by controlling NRLM sulfur levels
and compared them to the benefits and the costs to achieve these reductions. EPA evaluated the
monetary value of many of the public health and welfare benefits that will be achieved by these
reductions in emissions; see RIA Chapter 9. The monetized value of the health and welfare
benefits of the emissions reductions obtained by lowering sulfur in NRLM diesel fuel from
current levels to 15 ppm are expected to significantly exceed the costs of this reduction in sulfur
levels. This is the case for the complete fuel program (going from current levels of sulfur in
NRLM to 15 ppm for NRLM), as well as for each of the two steps used to achieve the complete
fuel program (going from current levels to 500 ppm, and then going from 500 ppm to 15 ppm).
The costs per gallon are also reasonable for going from current sulfur levels to 15 ppm^  EPA
also evaluated the cost per ton of emissions reduced for lowering sulfur in NRLM from current
levels down to 15 ppm, the complete program. The results are comparable to the cost per ton of
the entire engine and fuel program adopted in this final rule, as well as for other control
programs designed to reduce emissions of the same pollutants; see RIA Chapter 8.  The most
appropriate way to evaluate the cost per ton is to consider the complete fuel program adopted in
the final rule, since that is the action we are taking. However, we have also evaluated the cost
per ton considering the two steps separately. The cost per ton of emissions reduced in the first
step to 500 ppm is comparable to other control programs. The cost per ton for the second step,
when considered in isolation, is somewhat high compared to the cost per ton of other control
programs, however the monetized benefits from the reduction in emissions achieved by the
second step are greater than the costs to achieve these reductions.  In sum, EPA concludes that
the entire body of evidence strongly supports the view that controlling sulfur in NRLM fuel to 15
ppm, through a two step process, is quite reasonable in light of the emissions reductions and
benefits achieved, taking costs into consideration.

       The rationales for the two-step approach to fuel sulfur control and the levels  associated
with each step are discussed in Chapters 5 and 12.  Aside from its dramatic and immediate in-use
emission benefits, the proposed sulfur level of 500 ppm for the first step was chosen primarily
due to its consistency with the current highway diesel fuel standard.  The magnitude of the
distribution system costs would virtually prohibit the widespread distribution of any other grades
of diesel fuel, as discussed in Section IV.B of the preamble to the proposed rule. The 15 ppm
level was chosen as the final level for the same reasons, as well as for the reasons discussed
below concerning the need for 15 ppm sulfur fuel to enable the use of advanced emissions
   MM The cost per gallon to go from current levels to 15 ppm is the same cost per gallon to go from current sulfur
levels to 500 ppm plus the cost per gallon to go from 500 to 15 ppm. The cost per gallon for each of the separate
steps is by definition less than the cost for the combined steps of the total fuel program.

                                          5-101

-------
Final Regulatory Impact Analysis
control technology. Consequently, the choice of sulfur level was limited to one of the existing
three grades; 15 ppm, 500 ppm, or uncontrolled. A reduction in the sulfur directly to 15 ppm
was inconsistent with the proposed 2-step approach to diesel fuel sulfur control. Therefore,
given the need to achieve reductions, the 500 ppm level was selected for this temporary first step
of control.

       Section 21 l(c)(2)(A) requires that, prior to adopting a fuel control based on a finding that
the fuel's emission products contribute to air pollution that can reasonably be anticipated to
endanger public health or welfare, EPA consider  "all relevant medical and scientific evidence
available, including consideration of other technologically or economically feasible means of
achieving emission standards under [section 202 of the Act]." EPA's analysis of the medical and
scientific evidence relating to the emissions impact from nonroad, locomotive and marine
engines, which are impacted by sulfur in diesel fuel, is described in more detail in Chapter 2 of
the RIA.

       EPA has also satisfied the statutory requirement to consider "other technologically or
economically feasible means of achieving emission standards under section [202 of the Act]."
This provision has been interpreted as requiring consideration of establishing emission standards
under section 202 prior to establishing controls or prohibitions on fuels or fuel additives under
section 21 l(c)(l)(A).  See Ethyl Corp. v. EPA, 541 F.2d. 1, 31-32 (D.C. Cir. 1976). In Ethyl, the
court stated that section 21 l(c)(2)(A) calls for good faith consideration of the evidence and
options, not for mandatory deference to regulation under section 202 compared to fuel controls.
Id. at 32, n.66.

       EPA recently set emissions  standards for heavy-duty highway diesel engines under
section 202 (66 FR 5002, January 18, 2001). That program will reduce particulate matter and
oxides of nitrogen emissions from heavy duty engines by 90 percent.  In order to meet these
more stringent standards for diesel engines, the program requires a 97 percent reduction in the
sulfur content of diesel fuel. EPA does  not believe it is appropriate to seek further reductions at
this time from these engines. Also, section 21 l(c)(2)(A) refers to  standard setting under section
202 for highway engines or vehicles, and does not refer to standard setting under section 213.  In
any case, EPA is adopting stringent new standards for nonroad diesel engines under section 213.

       The two-step  reduction of sulfur to 15 ppm for nonroad, locomotive and marine diesel
fuel represents an appropriate exercise of the Agency's discretion  under section 21 l(c)(l)(A).
The control of NRLM fuel down to 15 ppm provides significant reductions in emissions of PM
and SO2, producing reductions in excess of 99% of these emissions. The fuel program is cost
effective and produces benefits to public health and welfare whose value significantly outweighs
the costs. These reductions can be achieved in a manner that is technologically feasible, will not
disrupt fuel supply, and is harmonized with the similar fuel controls for highway diesel fuel.
Using two  steps to reduce the level of NRLM sulfur to 15 ppm allows for a short lead time for
implementation,  enabling the environmental benefits to begin as soon as possible.
                                          5-102

-------
                                                              Fuel Standard Feasibility
5A.2 Impact of Diesel Sulfur Emission Products on Emission-Control
Systems

       EPA is restricting  the sulfur content of nonroad diesel fuel nationwide to no more than
15 ppm beginning in 2010, to enable compliance with new emission standards based on the use
of advanced emission control technology that will be available to nonroad diesel engines. It is
apparent that sulfur in nonroad diesel fuel significantly impairs the emission-control technology
of nonroad engines designed to meet the final emission standards.  As discussed in Chapter 4.1,
existing aftertreatment technologies will be capable of achieving dramatic reductions in NOx and
PM emissions from nonroad engines when the standards based on use of advanced aftertreatment
devices take effect in the 2011 and later model years. The aftertreatment technology for PM is
already in an advanced state of development and being tested in  fleet demonstrations in the U.S.
and Europe. The NOx aftertreatment technology is in a less-advanced, but still highly
promising, state of development, and, as discussed in Chapter 4.1, EPA believes the lead time
between now and 2011 will provide sufficient opportunity to adapt this technology for use on
nonroad engines. EPA believes these aftertreatment technologies will  be in general use
beginning in 2011, with the diesel sulfur controls adopted in this rule.

       At today's typical sulfur concentrations, these aftertreatment  technologies cannot be
introduced widely into the  marketplace. Not only does their efficiency at reducing emissions fall
off dramatically at elevated fuel sulfur concentrations, but engine operation impacts and
permanent damage to the aftertreatment systems are also possible.  To ensure regeneration of
the diesel particulate filter  at exhaust temperatures typical of nonroad diesel engines as described
in Chapter 4.1.1.3, we are expecting that  precious group metals  (primarily platinum) will be
used in their washcoat formulations. There are two primary mechanisms by which sulfur in
nonroad diesel fuel can limit the effectiveness or robustness of diesel parti culate filters which
rely on a precious metal oxidizing catalyst.  The first is inhibition of the oxidation of NO to NO2
and the second is the preferential oxidation of SO2 to SO3, forming a precursor to sulfate
particulate matter. With respect to NOx aftertreatment, all the NOx aftertreatment technologies
discussed in Chapter 4.1.2  that EPA believes will generally be available to meet the standards
are expected to utilize platinum to oxidize NO to NO2 to either improve the NOx reduction
efficiency of the catalysts at low temperatures or, as in the case of the NOx adsorber, as an
essential part of the process of NOx storage and regeneration.  This reliance of NO2 as an
integral part of the reduction process means that the NOx aftertreatment technologies, like the
PM aftertreatment technologies, would be significantly impaired by the sulfur in nonroad diesel
fuel. This is because sulfur, in the form of SOx, competes with NOx to be stored by the
aftertreatment device.  The resulting sulfate is harder to break down than the stored NOx, and is
not normally released during the regeneration phase (i.e. SOx is  stored preferentially to NOx by
the device). The sulfur therefore continues to build up, preventing storage of NOx, and
rendering the device ineffective. Further, although this problem can be addressed by adding a
"desulfation" phase to aftertreatment operation, the number of these desulfation events needs to
be minimized in order to prevent damage to the aftertreatment device.

   Current sulfur levels also impair performance and durability  of diesel oxidation catalysts


                                          5-103

-------
Final Regulatory Impact Analysis
(DOCs), which some of the 0-75 hp nonroad engines may utilize to achieve the 2008 emission
standards for PM. See chapter II. A of the preamble and Chapter 4.1.1.2 of this RIA. Although
EPA would not justify its decision to reduce sulfur levels in nonroad diesel fuel to 500 ppm for
this reason alone, it is worth pointing out the benefits to these PM emission control technologies
which result from the reduction.

5A.3 Sulfur Levels that Nonroad Engines  Can Tolerate

       As discussed in Chapter 4, there are three key factors which, taken together, lead us to
conclude that a nonroad diesel sulfur cap of 15 ppm is necessary so the NOx and PM
aftertreatment technology on nonroad engines will function properly and be able to meet the
emission standards.  These factors are the impact of higher sulfur levels on the efficiency and
reliability of the control systems,  and on the engine's fuel economy.

       The efficiency of emission control technologies at reducing harmful pollutants is directly
impacted by sulfur in nonroad diesel fuel. Initial and long term conversion efficiencies for NOx,
HC, CO and diesel PM emissions are significantly reduced by catalyst poisoning and catalyst
inhibition due to sulfur. NOx conversion efficiencies with the NOx adsorber technology in
particular are dramatically reduced in a very short time due to sulfur poisoning of the NOx
storage bed. In addition, total  PM control efficiency is negatively impacted by the formation of
sulfate PM. The formation of sulfate PM is likely to be in excess of the total PM standard ,
unless nonroad diesel fuel sulfur levels are below 15 ppm.  When sulfur is kept at these low
levels, both PM and NOx aftertreatment devices are expected to operate at high levels of
conversion efficiency, allowing compliance with the PM and NOx emission standards.

       The reliability of the emission control technologies to continue to function as required
under all operating conditions for the life of the engine is also directly impacted by sulfur in
nonroad diesel fuel.  As discussed in Chapter 4, sulfur in nonroad diesel fuel can prevent proper
operation and regeneration of both NOx and PM advanced aftertreatment control technologies
leading  to permanent loss in emission control effectiveness and even catastrophic failure of the
systems. For example, if regeneration of a PM filter does not occur, catastrophic failure of the
filter can occur in less than a single tank full of high-sulfur nonroad diesel fuel. For NOx
adsorbers, keeping sulfur levels no higher than 15 ppm is needed to minimize the number of
desulfation events to provide a high efficiency operation over the useful life of the engine.  It is
only through the availability of nonroad diesel fuel with sulfur levels less than 15 ppm that these
technologies can reliably be used to achieve the 90+ % emission reductions of PM and NOx on
which the 2011 and later model year standards are based. We believe that diesel fuel sulfur
levels of 15 ppm are needed and would allow these technologies to operate properly throughout
the life of the vehicle, including proper periodic or continuous regeneration.

       The sulfur content of nonroad diesel fuel will  also impact the fuel economy of nonroad
engines equipped with NOx and PM aftertreatment technologies. As discussed in detail in
Chapter 4.1.7, NOx adsorbers are expected to consume nonroad diesel fuel in order to cleanse
themselves of stored sulfates and  maintain efficiency. The larger the amount of sulfur in
                                          5-104

-------
                                                              Fuel Standard Feasibility
nonroad diesel fuel, the greater this adverse impact on fuel economy.  As sulfur levels increase
above 15 ppm, the fuel economy impact quickly changes from merely noticeable to
unacceptable. Likewise PM trap regeneration is inhibited by sulfur in nonroad diesel fuel.  This
leads to increased PM loading in the diesel particulate filter, increased exhaust backpressure, and
poorer fuel economy.  Thus for both NOx and PM technologies, the lower the fuel sulfur level,
the better the fuel economy of the vehicle.

       As a result of these factors, we find that 15 ppm represents an upper threshold of
acceptable nonroad diesel fuel sulfur levels for use with nonroad engines using generally
available advanced aftertreatment for PM and for NOx.

5A.4 Sulfur Sensitivity of Other Emission Control Devices or Systems

       Section 21 l(c)(2)(B) requires that, prior to adopting a fuel control based on a significant
impairment to vehicle emission-control systems, EPA consider available scientific and economic
data, including a cost benefit analysis comparing emission-control devices or systems which are
or will be in general use that require the proposed fuel control with such devices or systems
which are or will be in general use that do not require the proposed fuel control. As  described
below, we conclude that the aftertreatment technology expected to be used to meet the  nonroad
standards would be significantly impaired by operation on high-sulfur (greater than 15 ppm)
nonroad diesel fuel. Our analysis of the available scientific and economic data can be found
elsewhere in this document, including an analysis of the environmental benefits of the emission
standards (Chapter 3), an analysis of the costs and the technological feasibility of controlling
sulfur to the levels established in the final rule (Chapter 7), and a cost-effectiveness analysis of
the sulfur control and nonroad emission standards (Chapter 8).  Under section 21 l(c)(2)(B), as
just noted, EPA is also required to compare the costs and benefits of achieving emission
standards through emission-control systems that would not be sulfur-sensitive, if any such
systems are or will be in general use.

       We have determined that there are not (and will not be in the foreseeable future) emission
control devices available for general use in nonroad engines that can meet the nonroad emission
standards and would not be significantly impaired by nonroad diesel fuel with high sulfur levels.
NOx and PM emissions cannot be reduced anywhere near the magnitude contemplated by the
final emission standards without the application of aftertreatment  technology. As discussed in
Chapter 4, there are a number of aftertreatment technologies that are currently being developed
for both NOx and PM control with varying levels of effectiveness, sulfur sensitivity, and
potential application to nonroad engines.

       As discussed in Chapter 4.1, all the aftertreatment technologies that could be used to
meet the PM or NOx standards are significantly impaired by the sulfur in diesel fuel. For PM
control, all PM aftertreatment t technology that is capable of meeting the PM aftertreatment-
based Tier 4 standards would  need the level of sulfur control adopted in this rule. In addition,
the NOx aftertreatment technologies evaluated by EPA all rely  on the use of catalytic processes
to increase the effectiveness of the device in reducing NOx emissions. For example  both NOx
                                          5-105

-------
Final Regulatory Impact Analysis
adsorbers and compact SCR would rely on noble metals to oxidize NO to NO2, to increase NOx
conversion efficiency at the lower exhaust temperatures found in diesel motor vehicle operation.
This catalytic process, however, produces sulfate PM from the sulfur in the diesel fuel, and these
NOx aftertreatment devices therefore need the level of sulfur control adopted in this rule in order
for the vehicle to comply with the PM standard.

       In addition, compact SCR is not a technology that would be generally available by the
model year 2011 time frame. SCR systems require refilling with urea on a regular basis in order
to operate.  Significant and widespread changes to the fuel distribution system infrastructure thus
would have to be made, and there is no practical expectation that this would occur, with or
without the low-sulfur standard adopted in this final rule.  While it is feasible and practical to
expect that compact SCR may have a role in specific controlled circumstances, such as certain
centrally fueled fleets, or for generator sets using greater than 750 hp engines, it is not realistic at
this time to expect that the fuel distribution system infrastructure changes needed for widespread
and general use of compact  SCR on nonroad engines will be in place by the model year 2011
time frame. Finally, for NOx control, both NOx adsorbers and compact SCR are significantly
impaired by sulfur in diesel  fuel, and (as explained above) both technologies would need very
large reductions in sulfur from current levels to meet the NOx standard adopted in this final rule.
EPA believes that the requirement of a cost benefit analysis under section 21 l(c)(2)(B) is not
aimed at evaluating emission-control technologies that would each require significant additional
or different EPA fuel control regulations before the technology could be considered generally
available.

       Moreover, it is undisputed that any generally available technology  capable of achieving
the PM aftertreatment-based standards requires 15 ppm sulfur fuel. Thus, 15 ppm sulfur fuel will
be needed in any event.

       In sum, EPA believes that both PM and NOx aftertreatment technologies require  15 ppm
sulfur fuel.

        As described in Chapter 4, EPA anticipates that all the nonroad engine technologies
expected to be used to meet the final nonroad standards will require the use of nonroad diesel
fuel with sulfur levels capped at 15 ppm.  If we do not control diesel sulfur to the finalized
levels, we would not be able to set nonroad standards as stringent as those we are finalizing in
this final rule.  Consequently, EPA concludes that the benefits that would be achieved through
implementation of the  engine and sulfur control programs cannot be achieved through the use of
emission control technology that does not need the sulfur control adopted in this rule, and would
be generally available to meet the emission standards adopted in this rule.

       This also means that if EPA were to adopt emission standards without controlling diesel
sulfur content, the standards would be significantly less stringent than those finalized in this rule,
based on what would be technologically feasible with current or 500 ppm sulfur levels.
                                          5-106

-------
                                                             Fuel Standard Feasibility
5A.5 Effect of Nonroad Diesel Sulfur Control on the Use of Other Fuels or
Fuel Additives

       Section 21 l(c)(2)(C) requires that prior to prohibiting a fuel or fuel additive, EPA
establish that such prohibition will not cause the use of another fuel or fuel additive "which will
produce emissions which endanger the public health or welfare to the same or greater degree"
than the prohibited fuel or additive. This finding is required by the Act only prior to prohibiting
a fuel or additive, not prior to controlling a fuel or additive.  Since EPA is not prohibiting use of
sulfur in nonroad, locomotive or marine fuel, but rather is controlling the level of sulfur in these
diesel fuels, this finding is not required for this rulemaking.  However, EPA does not believe that
the sulfur control will result in the use of any other fuel or additive that will produce emissions
that will endanger public health or welfare to the same or greater degree as the emissions
produced by nonroad diesel with uncontrolled sulfur levels.

       Unlike the case of unleaded gasoline in the past,  where lead performed a primary
function by providing the necessary octane for the vehicles to function properly, sulfur does not
serve any useful function in nonroad, locomotive or marine diesel fuel.  It is not added to diesel
fuel, but comes naturally in the crude oil into which diesel fuel is processed.  Were it not for the
expense of sulfur removal, it would have been removed from diesel fuel years ago to improve the
maintenance and durability characteristics of diesel engines. EPA is  unaware of any function of
sulfur in nonroad, locomotive or marine diesel fuel that might have to be replaced once sulfur is
removed, with the possible exception of lubricity characteristics of the fuel. As discussed in
Chapters 4 and 5, there is some evidence suggesting that as sulfur is removed from diesel fuel
the natural lubricity characteristics of diesel fuel may be  reduced.   Depending on the crude oil
and the manner in which desulfurization occurs some low-sulfur diesel  fuels can exhibit poor
lubricity characteristics. To offset this concern lubricity  additives  are sometimes added to the
diesel fuel. These additives, however, are  already in common use today and EPA is unaware of
any health hazards associated with the use of these additives in diesel fuel, which would merely
be used in larger fractions of the diesel fuel pool. We do not anticipate that their use would
produce emissions which would reduce the large public health and welfare benefits that this rule
would achieve.

       EPA is unaware of any other additives that might be necessary to add to nonroad,
locomotive or marine diesel fuel to offset the existence of sulfur in the fuel. EPA is also
unaware of any additives that might need to be added to nonroad,  locomotive or marine diesel
fuel to offset any other changes to the fuel which might occur during the process of removing
sulfur.
                                         5-107

-------
Final Regulatory Impact Analysis
References to Chapter 5

1. Baseline Submissions for the Reformulated Gasoline Program.

2. Swain, Edward J., Gravity, Sulfur Content of U.S. Crude Slate Holding Steady, Oil and Gas
Journal, January 13, 1997.

3. Montcrieff, Ian T., Montgomery, David W., Ross, Martin T., An Assessment of the Potential
Impacts of Proposed Environmental Regulations on U.S. Refinery Supply of Diesel Fuel,
Charles River Associates, August 2000.

4. Final Report, 1996 American Petroleum Institute /National Petroleum Refiners Association,
Survey of Refining Operations and Product Quality, July 1997.

5. Final Report, 1996 American Petroleum Institute /National Petroleum Refiners Association,
Survey of Refining Operations and Product Quality, July 1997.

6. Final Report, 1996 American Petroleum Institute /National Petroleum Refiners Association,
Survey of Refining Operations and Product Quality, July 1997.

7. Dickinson, Cherl L., Strum, Gene P., Diesel Fuel Oils, 1997, TRW Petroleum Technologies,
November 2001.

8. American Society for Testing and Materials (ASTM), "Standard Specification for Diesel Fuel
Oils", ASTM D 975 and "Standard Specification for Fuel Oils", ASTM D 396. Some pipeline
companies that transport diesel fuel have limits for density and pour point, which are properties
that ASTM D 975 does not provide specifications on.

9. Regulatory Impact Analysis - Control of Air Pollution from New Motor Vehicles, Tier 2
Motor Vehicle Emission Standards and Gasoline Sulfur Control Requirements, Environmental
Protection Agency, December 1999.

10. Hamilton, Gary L., ABB Lummus, Letter to Lester Wyborny, U.S. EPA, August 2, 1999.

11. Mayo, S.W., "Mid-Distillate Hydrotreating: The Perils and Pitfalls of Processing LCO."

12. Peries, J-P., Jeanlouis, P-E, Schmidt, M, and Vance, P.W., "Combining NiMo and CoMo
Catalysts for Diesel Hydrotreaters," NPRA 1999 Annual Meeting, Paper 99-51, March 21-23,
1999.

13. Tippett, T., Knudsen, and Cooper, B., "Ultra Low Sulfur Diesel: Catalyst and Process
Options," NPRA 1999 Annual Meeting, Paper 99-06, March 21-23, 1999.

14.Tippett, T., Knudsen, and Cooper, B., "Ultra Low Sulfur Diesel: Catalyst and Process
Options," NPRA 1999 Annual Meeting, Paper 99-06, March 21-23, 1999.
                                         5-108

-------
                                                            Fuel Standard Feasibility
IS.Tungate, F.L., Hopkins, D., Huang, D.C., Fletcher, J.C.Q., and E. Kohler, "Advanced
distillate Hydroprocessing, AS AT, A Trifunctional HDAr/HDS/HDN Catalyst," NPRA  1999
Annual Meeting, Paper AM-99-38., March 21-23, 1999.

16.Gerritsen, L.A., Production of Green Diesel in the BP Amoco Refineries, Presentation by
Akzo Nobel at the WEFA conference in Berlin, Germany, June 2000.

IT.Gerritsen, L.A., Sonnemans, J.W M, Lee, S.L., and Kimbara, M., "Options to Met Future
European Diesel Demand and Specifications."

18.Conversation with Steve Mayo, Technical Service Development Manager, Akzo Nobel,
January 2003.

19.www.akzonobel-catalysts.com.

20. Brim Technology Bulletin, Haldor Topsoe Incorporated, March 2004.

21.Eng, Odette T., Kennedy, James E., "FCC Light Cycle Oil: Liability or Opportunity?,"
Technical Paper #AM-00-28, presented at the National Petrochemical and Refiners Association
Annual Meeting, March 26-28, 2000.

22.Centinel Hydroprocessing Catalysts: A New Generation of Catalysts for High-Quality Fuels,
Criterion Catalysts and Technologies Company, October 2000.

23. Ascent and Centinel Catalysts, Criterion Catalysts and Technologies Bulletin, March 2004.

24.Conversation with Lee Grannis, Criterion Catalysts, February 2004.

25.Axens catalyst bulletins 021HR-406A and 021HR-468A.

26.Wilson, R., "Cost Curves for Conventional HDS to Very Low Levels," February 2, 1999.

27. Peries, J-P., Jeanlouis, P-E, Schmidt, M, and Vance, P.W., "Combining NiMo and CoMo
Catalysts for Diesel Hydrotreaters," NPRA 1999 Annual Meeting, March 21-23, 1999.

28.Tippett, T., Knudsen, and Cooper, B., "Ultra Low Sulfur Diesel: Catalyst and Process
Options," NPRA 1999 Annual Meeting, Paper 99-06, March 21-23, 1999.

29. "Processes for Sulfur Management," IFF.

SO.Tungate, F.L., Hopkins, D., Huang, D.C., Fletcher, J.C.Q., and E. Kohler, "Advanced
distillate Hydroprocessing, AS AT, A Trifunctional HDAr/HDS/HDN Catalyst," NPRA  1999
Annual Meeting, Paper AM-99-38., March 21-23, 1999.

Sl.Gerritsen, L.A., Production of Green Diesel in the BP Amoco Refineries, Presentation by
Akzo Nobel at the WEFA conference in Berlin, Germany, June 2000.
                                        5-109

-------
Final Regulatory Impact Analysis
32.Patent number 6,123,835, filed June 1998, granted September 2000.

33.Patent number 6,428,686, filed June 2000, granted August 2002.

34. Ackerson, Michael; Skeds, Jon, Presentation to the Clean Diesel Independent Review Panel,
Process Dynamics and Linde Process Plants, July 30, 2002.

35.Conversation with Jon Skeds, Director of Refining, Linde BOC Process Plants LLC, January
2004.

36.Kidd, Dennis, S-Zorb - Advances in Applications of Phillips S-Zorb Technology, Presented at
the NPRA Q & A meeting, October 2000.

37.Conversation with Gary Schoonveld, Fuel Regulation/Asset Development, Conoco-Phillips,
November 2003.

38. 55 FR 34138, August 21, 1990.

39.Refining Industry Profile Study; EPA contract 68-C5-0010, Work Assignment #2-15, ICF
Resources, September 30, 1998.

40.Regulatory Impact Analysis for the Highway Diesel Final Rule, EPA Air Docket A-99-06

41.Presentations from the November 2002 Clean Diesel Fuel Implementation Workshop in
Houston, Texas are available at http://www.epa.gov/otaq/diesel.htmtfpublic Also available at this
website are rulemaking documents and fact sheets related to the highway diesel fuel final rule.
42. Moncrief, Philip and Ralph Ragsdale, "Can the U.S. E&C Industry Meet the EPA's Low
Sulfur Timetable," NPRA 2000 Annual Meeting, March 26-28. 2000, Paper No. AM-00-57.

43. National Petroleum Council, "U.S. Petroleum Assuring Adequacy and Affordability of
Cleaner Fuels", June 2000 pages 118-133.

44. Regulatory Impact Analysis - Control of Air Pollution from New Motor Vehicles: The Tier
2 Motor Vehicle Emissions Standards and Gasoline Sulfur Control Requirements, U.S. EPA,
December 1999, EPA 420-R-99-023.

45. Regulatory Impact Analysis - Control of Air Pollution from New Motor Vehicles: The Tier
2 Motor Vehicle Emissions Standards and Gasoline Sulfur Control Requirements, U.S. EPA,
December 1999, EPA 420-R-99-023.

46. Moncrief, Philip and Ralph Ragsdale, "Can the U.S. E&C Industry Meet the EPA's Low
Sulfur Timetable," NPRA 2000 Annual Meeting, March 26-28. 2000, Paper No. AM-00-57.

47.Chapter IV of the Regulatory Impact Analysis for the Final Highway Diesel Rule contained a
substantial background discussion regarding past experience in maintaining adequate fuel

                                        5-110

-------
                                                             Fuel Standard Feasibility
lubricity in low-sulfur fuels, EPA Air docket A-99-06.

48. Letter from L. Erlandsson, MTC AB, to Michael P. Walsh, dated October 16, 2000. Docket
A-99-06, item IV-G-42.

49. Chevron Products Diesel Fuel Technical Review provides a discussion of the impacts on fuel
lubricity of current diesel fuel compositional requirements in California versus the rest of the
nation. http://www.chevron.com/prodserv/fuels/bulletin/diesel/12%5F7%5F2%5Frf.htm

50. Final Report, 1996 American Petroleum Institute /National Petroleum Refiners Association,
Survey of Refining Operations and Product Quality, July  1997.
                                         5-111

-------
CHAPTER 6: Estimated Engine and Equipment Costs
   6.1 Methodology for Estimating Engine and Equipment Costs 	6-2
   6.2 Engine-Related Costs	6-5
      6.2.1 Engine Fixed Costs	6-5
          6.2.1.1 Engine and Emission-Control Device R&D 	6-5
          6.2.1.2 Engine-Related Tooling Costs	6-17
          6.2.1.3 Engine Certification Costs	6-21
      6.2.2 Engine Variable Costs  	6-25
          6.2.2.1 NOx Adsorber System Costs	6-28
          6.2.2.2 Catalyzed Diesel Paniculate Filter Costs  	6-34
          6.2.2.3 CDPF  Regeneration System Costs	6-38
          6.2.2.4 Diesel  Oxidation Catalyst (DOC) Costs  	6-40
          6.2.2.5 Closed-Crankcase Ventilation (CCV) System Costs	6-42
          6.2.2.6 Variable Costs of Conventional Technologies for Engines under 75 hp and
             over 750 hp  	6-43
          6.2.2.7 Summary of Engine Variable Cost Equations	6-48
      6.2.3 Engine Operating Costs 	6-49
          6.2.3.1 Operating Costs Associated with Oil-Change Maintenance for New and
             Existing Engines  	6-50
          6.2.3.2 Operating Costs Associated with CDPF Maintenance for New CDPF-
             Equipped Engines  	6-54
          6.2.3.3 Operating Costs Associated with Fuel Economy Impacts on New Engines
              	6-55
          6.2.3.4 Operating Costs Associated CCV Maintenance on New Engines	6-60
   6.3 Equipment-Related Costs  	6-60
      6.3.1 Equipment Fixed Costs	6-61
          6.3.1.1 Equipment Redesign Costs  	6-61
          6.3.1.2 Costs Associated with Changes to Product Support Literature	6-67
          6.3.1.3 Total Equipment Fixed  Costs  	6-67
      6.3.2 Equipment Variable Costs 	6-69
      6.3.3 Potential Impact of the Transition Provisions for Equipment Manufacturers . .  6-72
   6.4 Summary of Engine and Equipment Costs  	6-74
      6.4.1 Engine Costs	6-74
          6.4.1.1 Engine Fixed Costs  	6-74
          6.4.1.2 Engine Variable Costs	6-75
          6.4.1.3 Engine Operating Costs	6-75
      6.4.2 Equipment Costs	6-77
          6.4.2.1 Equipment Fixed Costs  	6-77
          6.4.2.2 Equipment Variable Costs	6-77
      6.4.3 Engine and Equipment Costs on a Per Unit Basis 	6-78
   6.5 Weighted Average Costs for Example Types of Equipment	6-82
      6.5.1 Summary of Costs for Some Example Types of Equipment  	6-82
      6.5.2 Method of Generating Costs for a Specific Piece of Equipment 	6-86
      6.5.3 Costs for Specific Examples from the Proposal	6-89
   6.6 Residual Value of Platinum Group Metals  	6-90

-------
                                            Estimated Engine and Equipment Costs
     CHAPTER 6: Estimated Engine and Equipment Costs
   This chapter presents the engine and equipment costs we have estimated for meeting the new
engine emissions standards.  Section 6.1 includes a brief outline of the methodology used to
estimate the engine and equipment costs. Sections 6.2 and 6.3 present the projected costs of the
individual technologies we expect manufacturers to use to comply with the new emissions
standards, along with a discussion of fixed costs such as research and development (R&D),
tooling, certification, and equipment redesign.  Section 6.4 summarizes these costs and Section
6.5 details cost estimates for several example pieces of equipment.  A complete presentation of
the aggregate cost of compliance for engines and equipment is in Chapter 8.

   Note that the costs presented here are for those nonroad engines and equipment that are
mobile nonroad equipment and are, therefore, subject to nonroad engine standards. These costs
would not apply for that equipment that is stationary—some portion of some equipment
segments such as generator sets, pumps, compressors—and not subject to nonroad engine
standards. The reader should know that some nonroad diesel equipment is not covered by
nonroad engine standards. Those nonroad engines that receive permits from local authorities as
stationary source emitters (i.e., some gensets, pumps, compressors, etc.) are not covered by
nonroad engine standards. In most cases, for what are very similar products, some fraction will
be permitted as stationary sources while others  remain mobile sources.

   To maintain consistency in the way our emission reductions, costs, and cost-effectiveness
estimates are calculated, our cost methodology  for engines and equipment relies on the same
projections of new nonroad engine growth as those used in our emissions inventory projections.
Our NONROAD emission inventory model includes estimates of future engine populations that
are consistent with the future engine sales used in our cost estimates.  The NONROAD model
inputs include an estimate of what percentage of gensets sold in the U.S. are "mobile" and, thus,
subject to the nonroad standards, and what percentage are "stationary" and not subject to the
nonroad standards. These percentages vary by  power category and are  documented in "Nonroad
Engine Population Estimates," EPA Report 420-P-02-004, December 2002.  For gensets >750
horsepower, NONROAD assumes 100 percent  are stationary and, therefore, not subject to the
new nonroad standards.  For gensets <750 horsepower, we have assumed other percentages  of
mobile versus stationary. During our discussions with engine manufacturers after the proposal,
it became apparent not only that our estimate for >750 horsepower gensets may not be correct
and many are indeed mobile, but also that some of our estimates for <750 horsepower gensets
may also not be correct and many more than we estimate may indeed be mobile. If true, this
increased percentage of mobile gensets will be  subject to the new nonroad standards.
Unfortunately, we have not received sufficient  data to make a conclusive change to the
NONROAD model and, therefore, for the above described purpose of maintaining consistency,
we have not included the costs or the emissions reductions in our official estimates for this final
rule.  In Chapter 8, Appendix A, we present a sensitivity analysis that includes both an estimate
                                         6-1

-------
Final Regulatory Impact Analysis
of the costs and emissions reductions that would result from including a higher percentage of
gensets as mobile machines and subject to the new standards.

   Note also that the costs presented here do not include potential savings associated with our
engine ABT program or our Transition Program for Equipment Manufacturers.  In addition, we
have assumed that engine companies who are  eligible for the small business engine manufacturer
specific provisions do not take advantage of the unique flexibilities the rule provides for them,
which includes the opportunity to delay compliance with the Tier 4 emission standards for a full
three model years. While we fully expect companies to use them to reduce compliance costs, we
do not factor them into the  cost analysis because they  are voluntary programs.  This  analysis of
compliance costs relates to regulatory requirements that are part of the nonroad Tier 4 final rule.
Unless noted otherwise, all costs are in 2002 dollars.

6.1 Methodology for Estimating Engine and Equipment Costs

   This analysis makes several simplifying assumptions  regarding how manufacturers will
comply with the new emission standards. First, in each power category, we assume  a single
technology recipe, as discussed in Chapter 4.  However, we expect that each manufacturer will
evaluate all possible technology avenues to  determine how to  best balance costs while ensuring
compliance.  As noted, for  developing cost estimates,  we have assumed that the industry does
not use either the transition program for equipment manufacturers or averaging, banking, and
trading, both of which offer the opportunity for significant cost reductions. Given these
simplifying assumptions, we believe the projections presented here probably overestimate the
costs of the different approaches toward  compliance that manufacturers may ultimately take.

   For smaller nonroad engines—those under 75 hp—many of the anticipated emission-control
technologies will be applied for the first time.  Therefore, we have sought input from a large
section of the regulated community regarding  the future costs  of applying these technologies to
diesel engines. Under contract with EPA, ICF Consulting provided questions to several engine
and parts manufacturers regarding costs associated with emission-control technologies for diesel
engines.  The responses to these questions were used as a first step toward estimating the costs
for many of the  technologies we believe  manufacturers will use. These costs form the basis for
our estimated costs for "traditional" engine technologies such as EGR and fuel-injection
systems.1  Note that, while  these technologies  are expected to  be added to <75hp engines for the
first time, they are being added, or will be, to >75hp engines to meet the Tier2/3 standards. We
have used the same methodology to develop the costs for these technologies for <75hp engines
as was used to develop the  costs for >75hp engines.2

   Costs for exhaust emission-control devices (for example, catalyzed diesel particulate filters
(CDPF), NOx adsorbers, and diesel oxidation  catalysts (DOC)) were estimated using the
methodology used in our FID2007 rulemaking. In that rulemaking effort, ICF Consulting, under
contract to EPA, provided surveys to nine engine manufacturers seeking information relevant to
estimating the costs for and types of emission-control  technologies that might be enabled with
low-sulfur diesel fuel. The survey responses were used as the first step in estimating the costs
                                          6-2

-------
                                              Estimated Engine and Equipment Costs
for advanced emission-control technologies anticipated for meeting the HD2007 standards.3 We
then built upon these costs based on input from members of the Manufacturers of Emission
Controls Association. Because the anticipated emission-control technologies are the same as
expected for highway engines, and because the suppliers of the technologies are the same for
nonroad engines as for highway engines, we have used that analysis as the basis for estimating
the costs of these technologies in this rulemaking.

   Costs of control include variable costs (for incremental hardware costs, assembly costs, and
associated markups) and fixed costs (for tooling, R&D, and certification).  For technologies sold
by a supplier to the engine manufacturers, costs are either estimated based on a direct cost to
manufacture the system components plus a 29 percent markup to account for the supplier's
overhead and profit or, when available, based on estimates from suppliers on expected total costs
to the manufacturers (inclusive of markups).4  Estimated variable costs for new technologies
include a markup to account for increased warranty costs. Variable costs are additionally
marked up to account for both manufacturer and dealer overhead and carrying costs. The
manufacturer's carrying cost was estimated to be four percent of the direct costs to account for
the capital cost of the  extra inventory and the incremental costs of insurance, handling, and
storage.  The dealer's carrying cost was estimated to be three percent of their direct costs to
account for the cost of capital tied up in inventory.  We adopted this same approach to markups
in the HD2007 rule, based on industry input.5

   We have  also identified various factors that cause cost impacts to decrease over time, making
it appropriate to  distinguish between near-term and long-term costs. Research in the costs of
manufacturing has consistently shown that, as manufacturers gain experience in production, they
are able to apply innovations to  simplify machining and assembly operations, use lower cost
materials, and reduce the number or complexity of component parts.6  This analysis incorporates
the effects of this learning curve as described in Section 6.2.2.

   Fixed costs for engine R&D are estimated to be incurred over the five-year period preceding
introduction of the engine.A  Fixed costs for tooling and certification are estimated to be incurred
one year ahead of initial production. Fixed costs for equipment redesign8 are estimated to be
incurred over a two-year period preceding introduction of the piece of equipment, while
equipment tooling costs are estimated to be incurred one year ahead of initial production. All
fixed cost expenditures are amortized using a seven percent capital cost to reflect the time value
of money. Engine fixed costs are then "recovered" over a five-year amortization period
including the same seven percent cost of capital.  This is true except where a phase-in of a new
standard occurs in which case the fixed costs are recovered during the phase-in years and then
   A There is one exception to this - for engine R&D conducted to support the new standards for <75 horsepower
engines in the 2008 model year, we have used a four year period (i.e., 2004 through 2007) over which to spread the
R&D expenditures.

     Throughout this analysis we use the term "redesign" to refer to all work needed to complete the equipment
modifications we believe will be necessary to accommodate the engine changes that will result from the new engine
standards.

                                           6-3

-------
Final Regulatory Impact Analysis
during the five years following 100 percent compliance.0 Equipment fixed costs are recovered
over a 10-year amortization period including the same seven percent captial cost; the longer
amortization period for equipment fixed costs reflects the longer product cycle for equipment.
We have also included lifetime operating costs where applicable.  These include costs associated
with the higher cost fuel, expected fuel economy impacts, increased maintenance demands
resulting from the addition of new emission-control hardware, and expected savings associated
with lower oil-change maintenance costs as a result of the low-sulfur fuel.

    A simplified overview of the methodology used to estimate engine and equipment costs is as
follows:

•   For fixed costs (i.e., R&D, redesign, tooling, certification), we estimate the total dollars that
    industry will spend. We then calculate the total dollars that they will recover in each year of
    the program following implementation.  These annual recovered costs represent our estimate
    of fixed costs associated with this final rule. In Section 6.5 and in some engine-related fixed
    cost tables in  Section 6.2.1, we also present an estimate of per-unit fixed costs.  These per-
    unit fixed costs are impacted by the way we have broken up the power categories in this cost
    analysis and by other factors (for example, the engine prices we have estimated) as discussed
    in more detail below.  Because we do not know how manufacturers recover their costs on a
    per-unit basis, we present these per-unit fixed costs for informational purposes only. We do
    not use these per-unit fixed cost estimates in our cost-per-ton calculations; instead, we use
    the annual cost of recovery totals in the  aggregate cost-per-ton calculations presented in
    Chapter 8.

•   For engine variable costs (i.e., emission-control hardware), we first estimate the cost per
    piece of technology/hardware. As described in detail in Section 6.2.2, emission-control
    hardware costs tend to be directly related to engine characteristics—for example, emission-
    control devices are sized according to engine displacement so costs vary by displacement;
    fuel-injection systems vary in  cost according to how many fuel injectors are required so costs
    vary by number of cylinders. This way  we are able to determine a variable cost equation as a
    function of engine displacement or as a function of the number of cylinders. We then
    consider each unique engine's baseline technology package using a database from Power
    Systems Research of all nonroad equipment sold in the United States.7 That database lists
    engine characteristics  for every one of over 4,500 unique equipment models sold in the
    United States and provides the sales of each piece of equipment. Using the baseline engine
    characteristics of each engine, the projected technology package for that engine, and the
    /-i
     We have estimated a "recovered" cost for all engine and equipment fixed costs to provide for a per-unit analysis of
the cost of the final rule. In general, in environmental economics, it is more conventional to simply count the total costs
of the program (i.e., opportunity costs) in the year they occur. However, this approach does not directly estimate a per-
unit production cost since fixed costs occur before the standards take effect and, therefore, prior to the production of new
compliant engines. In our methodology, fixed costs grow at a seven percent rate until they can be "recovered" on
complying units.  Note that the approach used here results in a higher estimate of the total costs of the program since the
recovered costs include a seven percent capital cost to reflect the time value of money. Our intent is to reflect the cost of
capital investments made in emissions control rather than investments made in other activities.

                                            6-4

-------
                                             Estimated Engine and Equipment Costs
   variable cost equations described in Section 6.2, we calculate a variable cost for the engine in
   each of the over 4,500 unique equipment models sold in the United States. This variable cost
   per engine is then multiplied by that engine's projected sales in each year for the years after
   the new standards take effect. We then total the annual costs for all engines to get the
   fleetwide variable costs per year. These fleetwide variable costs per year are then used in the
   cost-per- ton calculations presented in Chapter 8.

•  Note that the cost-per-ton calculation (see Chapter 8 of this RIA  for our cost-per-ton
   analysis) is never impacted by how many power categories we use in our cost analysis. We
   sometimes break up the fleet into more power categories than would seem reasonable given
   the structure of the emission standards. We do this for several reasons: (1) phase-ins of
   standards and/or different levels of baseline versus new standards sometimes force such
   breakouts; and, (2) greater stratification (i.e., breaking up the 75  to 175 hp range and the 175
   to 750 hp range) provides a better picture for use in our estimate  of potential recovery of
   fixed costs.  Importantly, the number of power categories used does not impact the total costs
   estimated as a result of the new emission standards, and these are the total costs used to
   calculate a cost-per-ton number.

   Engine costs are presented first - fixed costs, variable costs, then operating costs. Equipment
costs follow - fixed costs then variable costs. A summation of engine and equipment costs
follows these discussions.  Variable cost estimates presented here represent an expected
incremental cost of the engine or piece of equipment in the model year of introduction. Variable
costs in subsequent years decrease as a result of several factors, as described below. All costs
are presented in 2002 dollars.

6.2 Engine-Related  Costs

6.2.1 Engine Fixed Costs

   6.2.1.1 Engine and Emission-Control Device R&D

   The technologies described in Chapter 4 represent those technologies we believe will be used
to comply with the Tier 4 emission standards. These technologies are also part of an  ongoing
research and development effort geared toward compliance with the  HD2007 standards and, to
some extent, the current and future light-duty diesel vehicle standards in the US and in Europe.
Those engine manufacturers making R&D expenditures toward compliance with highway
emission standards will have to undertake some R&D effort to transfer emission-control
technologies to  engines they wish to sell into the nonroad market.  These R&D efforts will allow
engine manufacturers to develop and optimize these new technologies for maximum emission-
control effectiveness, while continuing to design engines with good performance, durability, and
fuel efficiency characteristics. However, many nonroad engine manufacturers are not part of the
ongoing R&D effort toward compliance with highway emission standards because they do not
sell engines into the highway market.  These manufacturers are expected to learn from the R&D
work that has already occurred and will continue through the coming years through their contact
                                          6-5

-------
Final Regulatory Impact Analysis
with highway manufacturers, emission-control device manufacturers, and the independent engine
research laboratories conducting relevant R&D. Despite these opportunities for learning, we
expect the R&D expenditures for these nonroad-only manufacturers to be higher than for those
manufacturers already conducting R&D in response to the HD2007 rule and the light-duty diesel
requirements in the US and Europe.

   We are projecting that several technologies will be used to comply with the Tier 4 emission
standards.  We are projecting that NOx adsorbers and CDPFs will be the most likely
technologies used to meet the new emission standards for engines over 75 hp and, for engines
between 25 and 75 hp, that CDPFs will be used in 2013 to meet the new PM standard.  The fact
that these technologies are being developed for implementation in the highway market before the
emission standards in this final rule take effect, and the fact that engine manufacturers have
several years to comply with the Tier 4 standards, ensures that the technologies used  to comply
with the nonroad standards will undergo significant development before reaching production.
This ongoing development will likely lead to reduced costs in three ways. First, we expect
research will lead to enhanced effectiveness for individual technologies, allowing manufacturers
to use simpler packages of emission-control technologies than we would predict currently, given
the current state of development. Second, we anticipate that the continuing effort to improve the
emission-control technologies will include innovations that allow lower-cost production. And
finally, we believe manufacturers will focus research efforts on any drawbacks, such as fuel
economy impacts or maintenance costs, in an effort to minimize or overcome any potential
negative effects.

   We anticipate that manufacturers will introduce a combination of primary technology
upgrades to meet the new emission standards. Achieving very low NOx emissions requires basic
research on NOx emission-control technologies and improvements in engine management.
Manufacturers are expected to address the challenge by optimizing the engine and exhaust
emission-control system to realize the best overall performance.  This will entail optimizing the
engine and emission control system for both emissions and fuel economy performance in light of
the presence of the new exhaust emission control devices and their ability to control pollutants
previously controlled only via in-cylinder means or with exhaust  gas recirculation. The NOx
adsorber technology in particular is expected to benefit from re-optimization of the engine
management system to better match the NOx adsorber's performance characteristics. The
majority of the dollars we have estimated for research is expected to be spent on developing this
synergy between the engine and NOx exhaust emission-control systems.  Therefore, for engines
where we project use of both a CDPF and a NOx adsorber (i.e., 75 to 750 hp), we have attributed
two-thirds of the R&D expenditures to NOx control, and one-third to PM control.0
   D In order to avoid inconsistencies in the way our emission reductions, and cost-effectiveness estimates are
calculated, our cost methodology for engines and equipment relies on the same projections of new nonroad engine
growth as those used in our emissions inventory projections. Our NONROAD emission inventory model includes
estimates of future engine populations that are consistent with the future engine sales used in our cost estimates.  The
NONROAD model inputs include an estimate of what percentage of gensets sold in the U.S. are "mobile" and, thus,
subject to the nonroad standards, and what percentage are "stationary" and not subject to the nonroad standards.
These percentages vary by power category and are documented in "Nonroad Engine Population Estimates," EPA

                                            6-6

-------
                                                Estimated Engine and Equipment Costs
    For this analysis, we have estimated two elements to engine R&D: (1) corporate R&D, or
that R&D conducted by manufacturers using test engines to learn how NOx and PM control
technologies work and how they work together in a system; and, (2) engine line specific R&D,
or that R&D done to tailor the corporate R&D knowledge to each particular engine line. To
distinguish between these two R&D elements, here we refer to the former as corporate R&D and
the latter as engine line R&D.

    With respect to the former of these R&D elements—corporate R&D—we begin with our
HD2007 rule. In that rule, we estimated that each engine manufacturer would expend $35
million for R&D toward successfully implementing catalyzed diesel particulate filters (CDPF)
and NOx adsorbers. For this analysis, we express all monetary values in 2002 dollars which
means our FID2007 starting point equates to $36.1  million. For their nonroad R&D efforts on
>75 hp engines - those engines where we project that compliance will require a CDPF and a
NOx adsorber or CDPFs-only (engines >750 hp) - engine manufacturers that also sell into the
highway market will incur some level of R&D effort but not at the level incurred for the highway
rule. In many cases, the engines used by highway manufacturers in nonroad products are based
on the same engine platform as  those engines used in highway products. However, power and
torque characteristics are often different, so manufacturers will need to expend some effort to
accommodate those differences. For these manufacturers, we have estimated that they will incur
an average R&D expense of $3.6 million not including the engine line R&D. This $3.6 million
R&D expense allows for the transfer of learning from highway R&D to their nonroad engines.
For reasons noted above, two-thirds of this R&D is attributed to NOx control and one-third to
PM control for 75 to 750 hp engines; for the portion of this R&D that is allocated to engines
>750 hp, all of this R&D is attributed to PM control.

    For those manufacturers that sell larger engines only into the nonroad market, and where we
project those engines to add a CDPF and a NOx  adsorber (75 to 750 hp) or a CDPF-only (>750
hp), we believe they will incur a corporate R&D expense approaching that incurred by highway
manufacturers for the highway rule although not quite at the same levelE. Nonroad
Report 420-P-02-004, December 2002. For gensets >750 horsepower, NONROAD assumes 100 percent are
stationary and, therefore, not subject to the new nonroad standards. For gensets <750 horsepower, we have assumed
other percentages of mobile versus stationary.  During our discussions with engine manufacturers after the proposal,
it became apparent not only that our estimate for >750 horsepower gensets may not be correct and many are indeed
mobile, but also that some of our estimates for <750 horsepower gensets may also not be correct and many more
than we estimate may indeed be mobile. If true, this increased percentage of mobile gensets will be subject to the
new nonroad standards. Unfortunately, we have not received sufficient data to make a conclusive change to the
NONROAD model to include the potentially increased percentages of mobile gensets and, therefore, for the above
described purpose of maintaining consistency, we have not included their costs or their emissions reductions in our
official estimates for this final rule (costs and emissions reductions for the current percentages in the NONROAD
model are included in our estimates for the final rule). Instead, we present a sensitivity analysis in Chapter 8 of the
RIA that includes both an estimate of the costs and emissions reductions that would result from including a higher
percentage of gensets as mobile equipment and subject to the new standards.

     Note that, while >750 hp mobile machine engines are not expected to add a NOx adsorber to comply with the new
engine standards, we have considered that the corporate R&D conducted for engines expected to add both a NOx
adsorber and a CDPF will apply for engines >750 hp given the  general similarity between large engines above and below

                                             6-7

-------
Final Regulatory Impact Analysis
manufacturers will be able to learn from the R&D efforts already underway for both the highway
rule and for the Tier 2 light-duty highway rule (65 FR 6698), and the light-duty and heavy-duty
diesel requirements in Europe. This learning may come from seminars, conferences, technical
publications regarding diesel engine technology (e.g., Society of Automotive Engineers technical
papers), and contact with highway manufacturers, emission-control device manufacturers, and
the independent engine research laboratories conducting relevant R&D. Therefore, we have
estimated an average expenditure of 70 percent of that spent by highway manufacturers in their
highway efforts. This lower number—$25.3 million versus $36.1 million in the highway
rule—reflects the transfer of knowledge to nonroad manufacturers from the many other
stakeholders in the diesel industry.  As noted above, two-thirds of this R&D is attributed to NOx
control and one-third to PM control. This value does not include the engine line R&D.

   Note that the $3.6 million and $25.3 million estimates represent our estimate of the average
corporate R&D expected by manufacturers.  Each manufacturer may have more or less than
these average figures.

   For manufacturers selling smaller engines that we project will add only a CDPF (i.e., 25 to
75 hp engines in 2013), we have estimated that their average R&D will be roughly one-third that
incurred by manufacturers conducting CDPF/NOx adsorber R&D. We believe this is a
reasonable estimate because CDPF technology is further along in its development than is NOx
adsorber technology and, therefore, a 50/50  split is not appropriate.  Using this estimate, the
average corporate R&D incurred by manufacturers that already have been selling engines into
both the highway and the nonroad markets will be $1.2 million not including engine line R&D,
and the average corporate R&D for manufacturers selling engines only into the nonroad market
will be roughly $8.3 million not including engine line R&D. All this R&D is attributed to PM
control.

   For manufacturers selling engines that will add only a DOC or will make only some engine-
out modification (i.e., to meet the PM standard for engines under 75  hp in 2008), we have
estimated that their average corporate R&D  will be roughly one-half the amount estimated for
their CDPF-only R&D. Application of a DOC should require very little R&D effort because
these devices have been used for years and because they require no special fueling strategies or
operating conditions to operate properly. Nonetheless, to avoid underestimating costs,  we have
estimated that the R&D incurred by manufacturers selling any engines into both the highway and
nonroad markets will be roughly $600,000 not including engine line R&D, and the corporate
R&D for manufacturers selling engines only into the nonroad market will be roughly $4.2
million not including engine line R&D. Because these R&D expenditures are strictly for
meeting a PM  standard, they are fully attributed to PM control.

   All these corporate R&D estimates are outlined in Table 6.2-1.
750 hp. We have included additional engine line R&D for all engines, including those >750hp, that is unique from this
corporate R&D estimate.

                                           6-8

-------
                                               Estimated Engine and Equipment Costs
                                        Table 6.2-1
              Estimated Corporate R&D Expenditures by Type of Manufacturer
                          Totals per Manufacturer over Five Years
                                        (SMillion)

Horsepower range
For new standards starting in year:
Manufacturer sells into both highway and
nonroad markets
Manufacturer sells only into the nonroad
market
Manufacturer has already done
CDPF&NOx Adsorber R&D
Manufacturer has not done CDPF&NOx
Adsorber R&D
% Allocated to PM
% Allocated to NOx
R&D for
DOC/engine -out
Engines
075
2011 (175-750hp)
2012 (75-175hp)
2015 (>750hp)
$3.6
$25.3


33%
67%
R&D for CDPF-only
Engines
25
-------
Final Regulatory Impact Analysis
    When certifying engines, manufacturers project the sales of each engine they certify.0  Using
the projected sales information, we were able to determine how many engine sales each
manufacturer expects to have in each of the power categories of interest.  As a result, not every
manufacturer is expected to incur all the R&D costs shown in Table 6.2-1. For example, some
manufacturers do not certify engines under 75 hp. Such a manufacturer will not incur R&D
costs for CDPF-only engines or for those engines expected to add a DOC or make only engine-
out changes. Also, some engine manufacturers produce and  sell engines to specifications
developed by other manufacturers.  Such joint venture manufacturers or wholly owned
manufacturers do not conduct engine-related R&D but simply manufacture an engine designed
and developed by another manufacturer. For such manufacturers, we have assumed no engine
R&D expenditures, given that we believe they will conduct no R&D themselves and will instead
rely on their joint venture partner. This is true unless the parent company has no engine sales in
the power categories covered by the partner company. Under such a situation, we have
accounted for the necessary R&D by  attributing it to the parent company. For example, Perkins
is an engine manufacturer wholly owned by Caterpillar so we have attributed no R&D costs to
Perkins.  However, Perkins sells engines in power categories that Caterpillar does not.  As a
result, we have attributed R&D costs  to Caterpillar for conducting R&D that will benefit Perkins
engines.  We have identified nine manufacturers to whom we have attributed no R&D because of
a joint partner agreement.11  For some of these (such as Perkins), we have attributed R&D costs
to their parent for the engines they will sell, and some are effectively the same company as their
parent (for example, Detroit Diesel and their parent DaimlerChrysler, New Holland and their
parent CNH).  In the end,  it is not important to our analysis to what manufacturer the R&D is
allocated because we have attempted  to estimate the total R&D that will be spent by the entire
industry.

    We have also estimated that some manufacturers will choose not to invest in R&D for the
U.S. nonroad market due to low volume sales that cannot justify the expense.  We have
identified three such manufacturers to whom we have attributed no R&D due to the cost of that
R&D relative to our best estimate of the revenues they receive from engine sales to which  the
new NRT4 standards would apply.1 This is not to say that we believe these manufacturers  will
cease to do business or even choose to leave the market; it only means that, given their low U.S.
     Projected sales information is confidential business information. We cannot present this information here nor can
we present details of calculations that use projected sales data since back calculating could shed light on the projected
sales data.

    TT
     Detroit Diesel and VM Motori were treated as part of DaimlerChrysler; IVECO, New Holland, and CNH were
treated as one; Kirloskar and Kukje were treated as partners of Cummins; Mitsubishi Motors Corporation and Mitsubishi
Heavy Industries are treated as one company; Perkins R&D is attributed to Caterpillar; and, Volvo Construction
Equipment and Volvo Penta AB are treated as one company.

    Estimated engine prices are shown in Table 6.2-3. We multiplied these prices by the manufacturer's projected
sales volume to determine if projected revenues from engine sales will exceed our estimated R&D costs.  If not, we have
assumed that the manufacturer would not invest in the R&D and would instead license the R&D from another
manufacturer.  While this would result in costs to the licensing manufacturer, it would also result in profits to the licensor;
it would therefore not result in increased costs associated with the new emission standards.

                                            6-10

-------
                                               Estimated Engine and Equipment Costs
sales volumes, we believe it is unlikely that they will conduct the necessary R&D themselves.
Instead, they will probably license the technology from another manufacturer, which will serve
to increase their own costs but reduce the net costs incurred by the licensing manufacturer, all
while having no impact on the total costs of the rule. Determining which manufacturers will or
will not invest in R&D is based on projected sales data, so we cannot share the manufacturers'
names.  It is important to note that the total projected sales for all three engine manufacturers
was 77 engines in the 2002 model year.

   Lastly, some certifying manufacturers do not appear to actually make engines. Instead, they
purchase engines from another engine manufacturer and then certify them as their own. We
have identified eight such certifying manufacturers and have attributed no R&D to these eight/

   Excluding the manufacturers we have identified as being in a joint partner arrangement or as
unlikely to invest in R&D, there remain 20 manufacturers expected to invest in CDPF&NOx
Adsorber R&D, 27 manufacturers expected to invest in CDPF-only R&D, and 28 manufacturers
expected to invest in DOC/engine-out R&D. The total estimated corporate R&D expenditures
are shown in Table 6.2-2.

                                        Table 6.2-2
        Estimated Industry-wide Corporate R&D Expenditures for the NRT4 Standards3

Expenditures during Years
Horsepower
Total Industry-wide
Corporate R&D
Expenditures
Corporate R&D for PM
Corporate R&D for NOx
DOC/engine -out
R&Db
2004-2007
075 hp
$121.8
$40.2
$81.6
CDPF-only R&Db
2008-2012
25750 hp engines.
    To this corporate R&D estimate, we have added an engine line R&D element. This engine
line R&D will cover costs for a manufacturer to tailor the knowledge gained through corporate
R&D to each particular engine line in their mix.  Based on confidential comments submitted
during the public comment period and our analysis of them, we have estimated these costs to be
     These eight are: Alaska Diesel Electric; American Jawa; Eastern Tools and Equipment; Escorts, Ltd.; Harvest
Drivemaster USA; International Tractors; Northern Tool and Equipment; Same Deutz-Fahr Group.
                                           6-11

-------
Final Regulatory Impact Analysis
$1 million for each engine line in the 25-75 hp range (to meet the 2013 standards), $3 million for
each engine line from 75-750 hp, and $6 million for those engine lines over 750 hp.  We have
assumed no engine line R&D for <75 hp engines to meet the 2008 standards because we do not
believe that the relatively simple addition of a DOC or the modifications impacting engine-out
emissions will require such a R&D effort.  We have determined the number of engine lines by
considering that, typically, the same basic diesel engine design can be increased or decreased in
size by simply adding or subtracting cylinders. As a result, a four-, six-, or eight-cylinder engine
may be produced from the same basic engine design.  While these engines have different total
displacement, they each have the same displacement per cylinder. Using the PSR database, we
grouped each engine manufacturer's engines into distinct engine lines using increments of 0.5
liters per cylinder.  This way, engines having similar displacements per cylinder are grouped
together and are considered to be one engine line. Table 6.2-3 presents the number of engine
lines for which we have estimated this engine  line R&D expenditure along with the total
industry-wide engine line R&D we have estimated.

                                        Table 6.2-3
       Estimated Industry-wide Engine Line R&D Expenditures for the NRT4 Standards"
Expenditures during Years
Horsepower
Engine Lines
Engine Line R&D per Line
Engine Line R&D Total0
Engine Line R&D for PMC
Engine Line R&D for NOx°
2008-2012
25750 hp
3
$6.0
$7.6
$7.6
-
2006-2014
All
104
-
$117.5
$49.8
$67.7
 a Dollar values are in millions of 2002 dollars.
 b This excludes 16 engine lines - those engine lines considered in the HD2007 rule. We have not included these
 highway engine lines since manufacturers will be conducting engine line R&D to meet the HD2007 standards.
 0 Dollar amounts shown here are those amounts attributable to US sales, as discussed in the main text.
   We have estimated that all engine R&D expenditures—corporate R&D plus engine line
R&D—occur over a five year span preceding the first year any emission-control device is
introduced into the market. The one exception to this being corporate R&D done for the 2008
standards which would be incurred over a four year span beginning today. Those expenditures
are then recovered by the engine manufacturer during any phase-in years and then over a five-
year span following full  introduction of the technology. Since PM standards take effect without
a multi-year phase-in, most PM costs are recovered for five years following the first year of
implementation. Most NOx costs are recovered over the two- or three-year phase-in and then
five years following complete implementation, or a total of seven or eight years. We include a
cost of seven percent when amortizing engine R&D expenditures.

   Our R&D estimates represent the cost to develop advanced aftertreatment-based emission-

                                           6-12

-------
                                                 Estimated Engine and Equipment Costs
control systems enabled by 15 ppm sulfur fuel. We are projecting that manufacturers will need
to do this R&D to sell engines in Europe, Japan, Australia, and Canada because we expect that
similar emission standards will be required in a similar time frame for each of these regions or
countries.8  Therefore, we have attempted to attribute the costs of R&D to the total engine sales
for these regions. Since we do not have sales data for every manufacturer showing what percent
of their engines are sold in the United States relative to these other regions, we have used Gross
Domestic Product (GDP) as a surrogate for sales.^  As a result, we have attributed only a
portion of the R&D expenditures to engine sales within the United States. The United States'
GDP is 42 percent of the total GDP from all the countries that are expected to adopt Tier 4 or
similar emission standards for nonroad diesel engines.L Therefore, we have attributed 42 percent
of the total R&D costs to U.S. sales.M Note that all engine R&D costs for <25 hp engines have
been attributed to U.S. sales since other countries are not expected to have similar standards on
these engines (though, as noted in the preamble for this final rule, the European Commission
may revisit this issue in their 2007 Nonroad standards review).

    The total estimated R&D attributable to US sales associated with the NRT4 engine
standards—corporate R&D presented in Table 6.2-2 and engine line R&D presented in Table
6.2-3—is shown in Table 6.2-4.
    Tf
     We considered using revenue and income data for nonroad engine/equipment companies that might show what
percent of those business metrics were US based versus non-US based. However, we were not able to find information
on all of the more than 50 nonroad diesel engine companies and the more than 600 nonroad diesel equipment companies.
In fact, we were able to locate information on only 10 nonroad engine/equipment companies because many companies
are not publicly traded in the US or do not present revenue and income data on a geographic basis. The results of our
research are contained in a memorandum to the docket (see Charmley, April 7, 2004, EDOCKET OAR-2003-0012-
0927). The limited data set generated by that research shows geographic distribution of revenue and income that is not
inconsistent with our 42 percent distribution.

     According to the Worldbank, in 2000, the European countries of Austria, Belgium, Denmark, Finland, France,
Germany, Greece, Ireland, Italy, Luxembourg, The Netherlands, Portugal, Spain, Sweden, and the United Kingdom had a
combined GDP of $7.8 trillion; Australia's GDP was $0.4 trillion; Canada's GDP was $0.7 trillion; Japan's GDP was
$4.7 trillion; and the U.S.  GDP was $9.9 trillion; for a total GDP of $23.5 trillion (www.worldbank.org).

      This is already factored into the costs shown in Tables 6.2-2 through 6.2-4, but is not factored into the costs shown
in Table 6.2-1.

                                              6-13

-------
Final Regulatory Impact Analysis
                                       Table 6.2-4
                Estimated Total R&D Expenditures for the NRT4 Standards"

Expenditures during Years
Horsepower
Total Industry-wide R&D
Expenditures0
Total R&D for PMC
Total R&D for NOxc
DOC/engine-out
R&Db
2004-2007
075 hp
$230.5
$81.2
$149.3
CDPF-only R&Db
2008-2012
25750 hp
engines are assumed to be recovered between 2015-2019. Delaying implementation dates for
these engines, or a subset of these engines, would not impact our estimated R&D expenditures or
their recovery but would, instead, only affect the timing of their recovery. To weight the costs
between engines in these categories, we have used revenue-weighting rather than a more
simplistic sales-weighting under the belief that manufacturers will attempt to recover more costs
where more revenues occur. Revenue-weighting is simply an estimated price multiplied by a
unit sales figure. The revenue weightings we have used are shown in Table 6.2-5.

   Using this methodology, we have estimated the total R&D expenditures associated with the
new  emission standards to vary from $9 to $57 million per year, with an average of $27 million
per year and a total of $323 million. Total R&D recovery on U.S. sales is estimated at $452
million. All estimated R&D costs are shown in Table 6.2-6. Note that the engine sales numbers
shown in Table 6.2-6 are discussed in greater detail in Chapter 8, where we present aggregate
costs to society.
                                          6-14

-------
                         Estimated Engine and Equipment Costs
                   Table 6.2-5
Revenue Weightings Used to Allocate R&D Cost Recovery
Horsepower
0750
Total
2000
Sales
119,159
132,981
93,914
68,665
112,340
61,851
34,095
2,752
2,785
628,542
Estimated
Engine
Price
$1,500
$2,900
$2,900
$5,200
$5,200
$10,300
$31,000
$80,500
$80,500

Revenue-Weighted Recovery of R&D in the Indicated Years
PM
NOx










2008-2012
N/A
22%
46%
32%






100%
2011-2015
2011-2018





30%
49%
10%
11%
100%
2012-2016
2012-2018



12%
19%
21%
34%
7%
7%
100%
2013-2017
N/A

59%
41%






100%
2015-2019
N/A








100%
100%
                      6-15

-------
Table 6.2-6
Estimated R&D Costs Incurred (Non-Annualized) and Recovered (Annualized) - expressed in $2002
Millions of dollars, except engine sales and per engine costs
                            2003     2004     2005     2006     2007     2008    2009     2010

IT)
1
|



O
Q.
1
S


S
V
Q.
(}
ID


O
O
Q.
V
f2


N-
V
Q.
"{[
0


O
O
Q.
1
1^:


o
o
Q.
f
o


£
Q.
f
o
o



o
£




Q.
?


Estimated US Sales
PM Costs Incurred
NOx Costs Incurred
PM Costs Recovered
NOx Costs Recovered
Per Engine Cost
Estimated US Sales
PM Costs Incurred
NOx Costs Incurred
PM Costs Recovered
NOx Costs Recovered
Per Engine Cost
Estimated US Sales
PM Costs Incurred
NOx Costs Incurred
PM Costs Recovered
NOx Costs Recovered
Per Engine Cost
Estimated US Sales
PM Costs Incurred
NOx Costs Incurred
PM Costs Recovered
NOx Costs Recovered
Per Engine Cost
Estimated US Sales
PM Costs Incurred
NOx Costs Incurred
PM Costs Recovered
NOx Costs Recovered
Per Engine Cost
Estimated US Sales
PM Costs Incurred
NOx Costs Incurred
PM Costs Recovered
NOx Costs Recovered
Per Engine Cost
Estimated US Sales
PM Costs Incurred
NOx Costs Incurred
PM Costs Recovered
NOx Costs Recovered
Per Engine Cost
Estimated US Sales
PM Costs Incurred
NOx Costs Incurred
PM Costs Recovered
NOx Costs Recovered
Per Engine Cost
Estimated US Sales
PM Costs Incurred
NOx Costs Incurred
PM Costs Recovered
NOx Costs Recovered
Per Engine Cost
PM Costs Incurred
NOx Costs Incurred
Total Costs Incurred
PM Costs Recovered
NOx Costs Recovered
Total Costs Recovered
131,507 135,623 139,739 143,855
$2.0 $2.0 $2.0




143,496 147,001 150,506 154,011
$4.3 $4.3 $4.3




100,051 102,097 104,142 106,188
$3.0 $3.0 $3.0




73,162 74,662 76,161 77,660





119,303 121,625 123,946 126,267





66,093 67,507 68,921 70,335
$3.3
$3.6



35,403 35,839 36,275 36,711
$5.5
$6.0



2,902 2,952 3,002 3,052
$1.2
$1.3



2,938 2,989 3,040 3,091
$0.7




$9.3 $9.3 $20.0
$10.8
$9.3 $9.3 $30.8



147,971
$2.0




157,516
$4.3




108,234
$3.0




79,159
$1.5
$1.6



128,588
$2.5
$2.5



71,749
$3.3
$3.6



37,147
$5.5
$6.0



3,102
$1.2
$1.3



3,142
$0.7




$24.0
$14.9
$38.9



152,087


$2.2

$15
161,021
$6.5

$4.6

$29
110,279
$4.6

$3.3

$29
80,659
$1.5
$1.6



130,909
$2.5
$2.5



73,163
$3.3
$3.6



37,583
$5.5
$6.0



3,152
$1.2
$1.3



3,193
$0.7




$25.8
$14.9
$40.8
$10.1

$10.1
156,203


$2.2

$14
164,526
$6.5

$4.6

$28
112,325
$4.6

$3.3

$29
82,158
$1.5
$3.1



133,230
$2.5
$5.1



74,577
$3.3
$7.2



38,019
$5.5
$11.9



3,202
$1.2
$2.5



3,244
$0.7




$25.8
$29.9
$55.7
$10.1

$10.1
160,319


$2.2

$14
168,031
$6.5

$4.6

$27
114,371
$4.6

$3.3

$28
83,657
$1.5
$3.1



135,551
$2.5
$5.1



75,991
$3.3
$7.2



38,455
$5.5
$11.9



3,252
$1.2
$2.5



3,295
$2.2




$27.3
$29.9
$57.2
$10.1

$10.1
164,435


$2.2

$13
171,536
$6.5

$4.6

$27
116,416
$4.6

$3.3

$28
85,157
$1.5
$3.1



137,872
$2.5
$5.1



77,405

$3.6
$4.7
$5.1
$126
38,891

$6.0
$7.7
$8.4
$414
3,302

$1.3
$1.6
$1.8
$1,023
3,346
$1.5

$0.9

$278
$16.7
$19.0
$35.7
$25.0
$15.2
$40.2
168,551


$2.2

$13
175,041
$6.5

$4.6

$26
118,462
$4.6

$3.3

$27
86,656

$1.6
$2.2
$2.2
$50
140,193

$2.5
$3.5
$3.6
$51
78,819

$3.6
$4.7
$5.1
$124
39,327

$6.0
$7.7
$8.4
$410
3,352

$1.3
$1.6
$1.8
$1,007
3,397
$1.5

$0.9

$274
$12.6
$14.9
$27.6
$30.7
$20.9
$51.6
172,667





178,546


$9.1

$51
120,507


$6.5

$54
88,155

$1.6
$2.2
$2.2
$49
142,514

$2.5
$3.5
$3.6
$50
80,233

$3.6
$4.7
$5.1
$121
39,763

$6.0
$7.7
$8.4
$405
3,402

$1.3
$1.6
$1.8
$993
3,448
$1.5

$0.9

$270
$1.5
$14.9
$16.5
$36.2
$20.9
$57.2
176,783





182,051


$9.1

$50
122,553


$6.5

$53
89,654


$2.2
$4.4
$73
144,836


$3.5
$7.1
$74
81,647


$4.7
$10.1
$181
40,199


$7.7
$16.7
$609
3,452


$1.6
$3.5
$1,487
3,499
$1.5

$0.9

$266
$1.5

$1.5
$36.2
$41.9
$78.1
180,899





185,556


$9.1

$49
124,599


$6.5

$52
91,154


$2.2
$4.4
$72
147,157


$3.5
$7.1
$72
83,061


$4.7
$10.1
$178
40,635


$7.7
$16.7
$602
3,502


$1.6
$3.5
$1,465
3,550


$3.1

$861



$38.3
$41.9
$80.2
185,015





189,061


$9.1

$48
126,644


$6.5

$51
92,653


$2.2
$4.4
$70
149,478


$3.5
$7.1
$71
84,475



$5.1
$60
41,071



$8.4
$204
3,552



$1.8
$494
3,601


$2.1

$591



$23.4
$26.7
$50.1
189,131





192,566


$9.1

$47
128,690


$6.5

$50
94,152



$2.2
$23
151,799



$3.6
$24
85,889



$5.1
$59
41 ,507



$8.4
$202
3,602



$1.8
$487
3,652


$2.1

$582



$17.7
$20.9
$38.7
193,247 197,363





196,071 199,576





130,736 132,781





95,652 97,151



$2.2
$23
154,120 156,441



$3.6
$23
87,303 88,717



$5.1
$58
41,943 42,379



$8.4
$200
3,652 3,702



$1.8
$481
3,703 3,754


$2.1 $2.1

$574 $567



$2.1 $2.1
$20.9
$23.1 $2.1

$8.2
$0.0
$11.0
$0.0


$49.6
$0.0
$68.7
$0.0


$35.0
$0.0
$48.5
$0.0


$7.7
$15.6
$10.8
$21.8


$12.5
$25.5
$17.6
$35.7


$16.7
$36.1
$23.4
$50.6


$27.6
$59.7
$38.7
$83.7


$5.8
$12.5
$8.1
$17.6


$10.9
$0.0
$15.3
$0.0

$173.9
$149.3
$323.2
$242.1
$209.5
$451 .5

-------
                                             Estimated Engine and Equipment Costs
   6.2.1.2 Engine-Related Tooling Costs

   Once engines are ready for production, new tooling will be required to accommodate the
assembly of the new engines. In the HD2007 rule, we estimated approximately $1.6 million per
engine line for tooling costs associated with CDPF/NOx adsorber systems. For the Tier 4
standards, we have estimated that nonroad-only manufacturers will incur the same amount -
$1.65 million expressed in 2002 dollars - for each engine line that requires a CDPF/NOx
adsorber system.  These costs are assigned equally to NOx control and PM control. We have
estimated the same tooling costs as in the HD2007 rule because we expect Tier 4 engines to use
the same technologies (i.e., a CDPF and a NOx adsorber).  For those systems requiring only a
CDPF, we have estimated one-half that amount, or $825,000 per engine line. For those systems
requiring only a DOC or some engine-out modifications, we have estimated one-half the CDPF-
only amount, or $412,500 per engine line. Tooling costs for CDPF-only and for DOC engines
are attributed solely to PM control.

   For those manufacturers selling into both the highway and nonroad markets,  we have started
with the same $1.65  million baseline discussed above. For those engines requiring a CPDF/NOx
adsorber system (i.e., those over 75 hp) we have adjusted that $1.65 million baseline by 50
percent. We believe this 50 percent adjustment is reasonable since many nonroad engines over
75 hp are produced on the same engine  line with their highway counterparts. For such lines,
tooling costs will be negligible. For engine lines without a highway counterpart, the $1.65
million tooling cost applies. For highway manufacturers selling into both the highway and the
nonroad markets, we have estimated a 50/50 split of nonroad engine product lines (i.e., 50
percent with highway counterparts and  50 percent without) and therefore applied a 50 percent
factor to the $1.65 million baseline. These tooling costs are split evenly between NOx control
and PM control. For those engine lines requiring  only a CDPF (i.e., those between 25 and 75
hp), we have estimated the same tooling cost as used for nonroad-only manufacturers, or
$825,000. Similarly, the tooling costs for DOC and/or engine-out engine lines has been
estimated to be $412,500. We have used the same tooling costs as the nonroad-only
manufacturers for engines under 75 hp because these engines tend not to have a highway
counterpart.  Tooling costs for CDPF-only and for DOC engines are attributed solely to PM
control.

   We project that engines between 25 and 50 hp will apply EGR systems to meet the new NOx
standards for 2013. For these engines, we have included an additional tooling cost of $41,300
per engine line, consistent with the EGR-related tooling cost estimated for 50 to  100 hp engines
in our Tier 2/Tier 3 rulemaking which specified the same NOx standards. This tooling cost is
applied equally to all engine lines in that power range, regardless of the markets  into which the
manufacturer sells. We have applied this tooling cost equally because engines in this power
range do not tend to  have highway counterparts. We expect EGR systems to be  added to engines
between 25 and 50 hp to meet the new NOx standard, so tooling costs for EGR systems are
attributed solely to NOx control.
                                         6-17

-------
Final Regulatory Impact Analysis
   We have also estimated some tooling costs for >750 horsepower engines to meet the 2011
standards. We have estimated this amount at ten times the amount for 25 to 50 horsepower
engines, or $413,000 per engine line. This cost was not in the proposal since NOx adsorbers
were being projected for all >750 horsepower engines. We have applied this tooling to all
engine lines >750 horsepower, regardless of what markets into which a manufacturer sells, since
such engines clearly have no highway counterpart.  We have attributed this cost to NOx control.

   Tooling costs per engine line and type of manufacturer are summarized in Table 6.2-7.

                                       Table 6.2-7
          Estimated Tooling Expenditures per Engine Line by Type of Manufacturer"

Horsepower range
For new standards starting in
Manufacturer sells into both
highway and nonroad markets
Manufacturer sells only into
the nonroad market
% Allocated to PM
% Allocated to NOx
DOC/engine-
out Engines
0750hp
2011
$413,000
$413,000
0%
100%
EGR Engines
25750hp add cooled EGR in 2011. We
 would expect manufacturers to use a less costly means of control if it allows them to meet the new standard (see
 section 4.1.2 of this RIA for more information regarding our estimates of EGR use).
   As noted, we have applied tooling costs by engine line assuming that engines in the same
line are produced on the same production line.  Typically, the same basic diesel engine design
can be increased or decreased in size by simply adding or subtracting cylinders. As a result, a
four-, six-, or eight-cylinder engine may be produced from the same basic engine design. While
these engines have different total displacement, they each have the same displacement per
cylinder. Using the PSR database, we grouped each engine manufacturer's engines into distinct
engine lines using increments of 0.5 liters per cylinder.  This way, engines having similar
displacements per cylinder are grouped together and are considered to be built on the same
production line. Note that a tooling expenditure for a single engine line may cover engines over
several power categories.  To allocate the tooling expenditure for a given production line to a
specific power range, we have used sales-weighting within that engine line.

   We have applied the above tooling costs to all manufacturers that appear to actually make
engines.  We have not eliminated joint venture manufacturers because these manufacturers still
                                           6-18

-------
                                             Estimated Engine and Equipment Costs
need to invest in tooling to make the engines, even if they do not conduct any R&D. Doing this,
we determined there to be 62 manufacturers expected to invest in tooling for a total of 133
engine lines. Of these, 19 manufacturers sell into both the highway and nonroad markets and
sell a total of 56 engine lines, while 43 manufacturers sell only into the nonroad market and sell
a total of 77 engine lines.  For the same reasons as explained for R&D costs, we have attributed a
portion of the tooling costs to U.S. sales and a portion to sales in other countries expected to
have similar levels of emission control; tooling costs for <25 hp engines are attributed only to
US sales since other countries are not expected to have similar standards on <25 hp engines. All
tooling costs are assumed to be incurred one year before the standard they support and are then
recovered over a five-year period following introduction of the new standard. We include a cost
of seven percent when amortizing engine tooling costs.

   Using this methodology, we estimate the total tooling expenditures attributable to this final
rule at $74 million. Total tooling recovery on U.S. sales is estimated at $91 million. All
estimated tooling costs are shown in Table 6.2-8.
                                          6-19

-------
Final Regulatory Impact Analysis
Table 6.2-8
Estimated Tooling Costs Incurred (Non-Annualized) and Recovered (Annualized) - expressed in $2002
Millions of dollars, except engine sales and per engine costs


LO
'v'
Q.
•^
O



O
Id
Q.
II
V
ID
01


in
£
(J
o
Id


o
o
1
II
LO


LO
|.
II
O
o


8
CO
Q.
II
V
ID



8
CD
Q.
II
V
s
CO


o
V
Q_
II
O
8



Q.
O
IS




Q.
3



Estimated US Sales
PM Costs Incurred
NQx Costs Incurred

PM Costs Recovered

NOx Costs Recovered
Per Engine Cost
Estimated US Sales
PM Costs Incurred
NQx Costs Incurred
PM Costs Recovered

NOx Costs Recovered
Per Engine Cost
Estimated US Sales
PM Costs Incurred
NQx Costs Incurred
PM Costs Recovered

NOx Costs Recovered
Per Engine Cost
Estimated US Sales
PM Costs 1 ncurred
NQx Costs Incurred
PM Costs Recovered
NOx Costs Recovered
Per Engine Cost
Estimated US Sales
PM Costs Incurred
NQx Costs Incurred
PM Costs Recovered

NOx Costs Recovered
Per Engine Cost
Estimated US Sales
PM Costs Incurred
NOx Costs 1 ncurred
PM Costs Recovered

NOx Costs Recovered
Per Engine Cost
Estimated US Sales
PM Costs Incurred

NOx Costs Incurred
PM Costs Recovered
NOx Costs Recovered
Per Engine Cost
Estimated US Sales

PM Costs Incurred
NOx Costs 1 ncurred
PM Costs Recovered
NOx Costs Recovered
Per Engine Cost
Estimated US Sales
PM Costs Incurred

NOx Costs 1 ncurred
PM Costs Recovered
NOx Costs Recovered
Per Engine Cost
PM Costs Incurred
NOx Costs 1 ncurred
Total Costs Incurred
PM Costs Recovered
NOx Costs Recovered
Total Costs Recovered
2007 2008 2009
147,971 152,087 156,203
$5.2


$1.3 $1.3


$8 $8
157,516 161,021 164,526
$5.9

$1.4 $1.4


$9 $9
108,234 110,279 112,325
$4.1

$1.0 $1.0


$9 $9
79,159 80,659 82,158





128,588 130,909 133,230






71,749 73,163 74,577






37,147 37,583 38,019






3,102 3,152 3,202






3,142 3,193 3,244






$15.2

$15.2
$3.7 $3.7

$3.7 $3.7
2010
160,319



$1.3


$8
168,031


$1.4


$9
114,371


$1.0


$9
83,657





135,551






75,991
$11.0
$11.0




38,455
$6.1

$6.1



3,252

$0.5
$0.5



3,295


$0.5



$17.6
$18.1
$35.6
$3.7

$3.7
2011
164,435



$1.3


$8
171,536


$1.4


$8
116,416


$1.0


$9
85,157
$2.8
$2.8



137,872
$4.5
$4.5




77,405


$2.7

$2.7
$69
38,891



$1.5
$1.5
$76
3,302



$0.1
$0.1
$72
3,346




$0.1
$38
$7.3
$7.3
$14.6
$8.0
$4.4
$12.4
2012
168,551



$1.3


$8
175,041
$4.3
$0.5
$1.4


$8
118,462
$3.0

$1.0


$9
86,656


$0.7
$0.7
$16
140,193


$1.1

$1.1
$16
78,819


$2.7

$2.7
$68
39,327



$1.5
$1.5
$75
3,352



$0.1
$0.1
$71
3,397




$0.1
$37
$7.3
$0.5
$7.8
$9.8
$6.2
$16.0
2013
172,667







178,546


$1.0

$0.1
$7
120,507


$0.7


$6
88,155


$0.7
$0.7
$15
142,514


$1.1

$1.1
$16
80,233


$2.7

$2.7
$67
39,763



$1.5
$1.5
$74
3,402



$0.1
$0.1
$70
3,448




$0.1
$37



$7.8
$6.3
$14.2
2014
176,783







182,051


$1.0

$0.1
$6
122,553


$0.7


$6
89,654


$0.7
$0.7
$15
144,836


$1.1

$1.1
$15
81,647


$2.7

$2.7
$66
40,199



$1.5
$1.5
$74
3,452



$0.1
$0.1
$69
3,499
$1.0



$0.1
$36
$1.0

$1.0
$7.8
$6.3
$14.2
2015
180,899







185,556


$1.0

$0.1
$6
124,599


$0.7


$6
91,154


$0.7
$0.7
$15
147,157


$1.1

$1.1
$15
83,061


$2.7

$2.7
$65
40,635



$1.5
$1.5
$73
3,502



$0.1
$0.1
$68
3,550



$0.3
$0.1
$107



$8.1
$6.3
$14.4
2016 2017 2018 2019
185,015 189,131 193,247 197,363







189,061 192,566 196,071 199,576


$1.0 $1.0

$0.1 $0.1
$6 $6
126,644 128,690 130,736 132,781


$0.7 $0.7


$6 $6
92,653 94,152 95,652 97,151


$0.7
$0.7
$15
149,478 151,799 154,120 156,441


$1.1

$1.1
$15
84,475 85,889 87,303 88,717






41,071 41,507 41,943 42,379






3,552 3,602 3,652 3,702






3,601 3,652 3,703 3,754



$0.3 $0.3 $0.3 $0.3

$71 $70 $69 $68



$3.8 $2.0 $0.3 $0.3
$1.9 $0.1
$5.7 $2.2 $0.3 $0.3
Total

$5.2
$0.0

$6.4

$0.0


$10.1
$0.5
$12.4

$0.6


$7.2
$0.0
$8.7

$0.0


$2.8
$2.8
$3.4
$3.4


$4.5
$4.5
$5.5

$5.5


$11.0
$11.0
$13.4

$13.4


$6.1

$6.1
$7.4
$7.4



$0.5
$0.5
$0.6
$0.6


$1.0

$0.5
$1.3
$0.6

$48.4
$25.9
$74.3
$59.1
$31.6
$90.6
                                    6-20

-------
                                               Estimated Engine and Equipment Costs
    6.2.1.3 Engine Certification Costs

    Manufacturers will incur more than the normal level of certification costs during the first few
years of implementation because engines will need to be certified to the new emission standards
using new test procedures.  Consistent with our recent standard setting regulations, we have
estimated engine certification costs at $60,000 per new engine certification to cover testing and
administrative costs.10  The $60,000 certification cost per engine family was used for engines in
the 25 to 75 hp range certifying to the 2008 standards. For 25 to 75 hp engines certifying to the
2013 standards, and for 75-750 hp engines certifying to the appropriate standards, we have added
costs to cover the new test procedures for nonroad diesel engines (i.e., the transient test  and the
NTE);N these costs were estimated at $31,500 per engine family. For engines >750 hp,  the
certification costs used were $87,000 per family since these engines will not be certifying over
the new transient test procedure. For engines <25 hp, we have assumed (for cost purposes) that
all engines will certify to the transient test and the NTE in 2008.  We believe manufacturers may
choose to do this rather than certifying all engines again in 2013 when the transient test and NTE
requirements actually begin for those engines (and the rules explicitly provide the option of
certifying these engines starting in 2008 using these tests). This assumption results in higher
certification costs in 2008 than if these engines certified only to the steady-state standard.
However, we believe manufacturers may choose to do this because it would avoid the need to
recertify all <25 hp engines again in 2013. Certification costs (for engines in all hp ranges)
apply equally to all engine families for all manufacturers regardless of the markets into  which
the manufacturer sells.

    To determine the number of engine families to be certified, we used our certification
database for the 2002 model year. That database provides the number of engine families and the
associated power rating of each. We grouped those power ratings into the nine ranges shown in
Table 6.2-9. We have chosen these nine power categories because: (1) phasing in standards and
having different levels of baseline and complying emission levels force such breakouts; and, (2)
greater  stratification (i.e., breaking up the 75 to 175 hp range and the 175 to 750 hp range)
provides a better picture of cost recovery because it more accurately matches the number of
engine families (certification costs) with the level of engine sales (cost recovery).  Some engine
families will undergo more than one certification process due to the structure of new emission
standards in the final rule.  Table 6.2-9 shows the number of engine families in each  power range
and the year for they are subject to new emission standards, along with the total certification
expenditures for those standards.

    The cost expenditures shown in Table 6.2-9 are estimated to occur one year before the year
shown in the table.  The years  shown in the table coincide with the years for which the new
standards begin, thereby requiring engine certification. Half the 175 to 750 hp  engine families
    N Note that the transport refrigeration unit (TRU) test cycle is an optional duty cycle for steady-state
certification testing specifically tailored to the operation of TRU engines. Likewise, the ramped modal cycles are
available test cycles that can be used to replace existing steady-state test requirements for nonroad constant-speed
engines, generally. Manufacturers of these engines who opt to use one of these test cycles would incur no new costs
above those estimated here and may incur less cost.

                                           6-21

-------
Final Regulatory Impact Analysis
certified for 2011 must again be certified in 2014 when the NOx phase-in becomes 100 percent.
For 25 to 50 hp engines in 2013, half the certification costs are attributed to PM and half are
attributed to NOx, due to the new PM and NOx standards for those engines in that year; all the
certification costs for 50 to 75 hp engine families are attributed to PM because only a new PM
standard applies in that year for those engines.

   Note that these certification costs may overestimate actual costs because they assume all
engines are certified as a result of the new emission standards in this final rule. However, some
engines would have been scheduled for new certification independent of this final rule due to
design changes or power increases among other possible reasons. For such engines, the
incremental certification cost would be those costs associated with the new test procedures and
would not include certification costs associated with the existing test procedure. However, to
remain conservative, here we have applied the full certification costs to all engine families.
Given the magnitude of certification costs relative to other costs  in this final rule, this has little
impact on the costs per ton of emissions reduced or the cost/benefit results.

                                        Table 6.2-9
                           Number of Engine Families, Estimated
                  Certification Costs, and Allocation of Certification Costs"
Power range
0750a
Total families
Total Cert Costs
% Allocated to PM
% Allocated to NOx
Model Year for New Emission Standards
2008
102
132
88






322
$22.5
100%
0%
2011





102
64
9
40
215
$19.5
50%
50%
2012



55
73




128
$11.7
50%
50%
2013

132







132
$12.1
50%
50%


88






88
$8.1
100%
0%
2014





51
32
5

88
$8.0
0%
100%



28
37




64
$5.9
0%
100%
2015








40
40
$3.5
50%
50%
 1 Dollar values are in millions of 2002 dollars.
                                           6-22

-------
                                              Estimated Engine and Equipment Costs
   To estimate recovery of certification expenditures, we have attributed the expenditures to
engines sold in the specific power range and spread the recovery of costs over U.S. sales within
that category.  Expenditures are incurred one year before the emission standard for which the
certification is conducted, and are then recovered over a five-year period following the
certification.  We include a cost of seven percent when amortizing engine certification costs. We
have spread these certification costs only over the engines sold in the United States because U.S.
EPA certification is not presumed to fulfill the certification requirements of other countries.
Total certification expenditures are estimated at $91 million. Recovery of certification costs is
estimated at $111 million.  All estimated certification expenditures and the recovery of those
expenditures are shown in Table 6.2-10.
                                          6-23

-------
Final Regulatory Impact Analysis
Table 6.2-10
Estimated Certification Costs Incurred (Non-Annualized) and Recovered (Annualized) — expressed in $2002
Millions of dollars, except engine sales and per engine costs


io
CM
Q.
O


8
Q-
II
if)


IS
ii
o
If)


o
o
1
II
If)

If)
Q.
II
O
o

o
s
Q-
ii
if)

o
s

II
o

o
IS
Q.
1
8
CD

Q.
O
IS
A



Q.
3



Estimated US Sales
PM Costs 1 ncurred
NQx Costs Incurred
PM Costs Recovered
NOx Costs Recovered
Per Engine Cost
Estimated US Sales
PM Costs 1 ncurred
NQx Costs Incurred
PM Costs Recovered
NOx Costs Recovered
Per Engine Cost
Estimated US Sales
PM Costs 1 ncurred
NOx Costs Incurred
PM Costs Recovered
NOx Costs Recovered
Per Engine Cost
Estimated US Sales
PM Costs Incurred
NQx Costs Incurred
PM Costs Recovered
NOx Costs Recovered
Per Engine Cost
Estimated US Sales
PM Costs 1 ncurred
NQx Costs Incurred
PM Costs Recovered
NOx Costs Recovered
Per Engine Cost
Estimated US Sales
PM Costs Incurred
NQx Costs Incurred
PM Costs Recovered
NOx Costs Recovered
Per Engine Cost
Estimated US Sales
PM Costs Incurred
NQx Costs Incurred
PM Costs Recovered
NOx Costs Recovered
Per Engine Cost
Estimated US Sales
PM Costs 1 ncurred
NOx Costs Incurred
PM Costs Recovered
NOx Costs Recovered
Per Engine Cost
Estimated US Sales
PM Costs Incurred
NQx Costs Incurred
PM Costs Recovered
NOx Costs Recovered
Per Engine Cost
PM Costs 1 ncurred
NQx Costs Incurred
Total Costs 1 ncurred
PM Costs Recovered
NOx Costs Recovered
Total Costs Recovered
2007 2008 2009 2010
147,971 152,087 156,203 160,319
$9.3
$2.3 $2.3 $2.3
$15 $15 $14
157,516 161,021 164,526 168,031
$7.9

$1.9 $1.9 $1.9
$12 $12 $11
108,234 110,279 112,325 114,371
$5.3
$1.3 $1.3 $1.3
$12 $11 $11
79,159 80,659 82,158 83,657





128,588 130,909 133,230 135,551



71,749 73,163 74,577 75,991
$4.7
$4.7


37,147 37,583 38,019 38,455
$2.9
$2.9


3,102 3,152 3,202 3,252
$0.4
$0.4

3,142 3,193 3,244 3,295
$1.7
$1.7

$22.5 $9.7
$9.7
$22.5 $19.5
$5.5 $5.5 $5.5

$5.5 $5.5 $5.5
2011
164,435
$2.3
$14
171,536


$1.9
$11
116,416

$1.3
$11
85,157
$2.5
$2.5



137,872
$3.3
$3.3


77,405

$1.1
$1.1
$29
38,891

$0.7
$0.7
$37
3,302

$0.1
$0.1
$61
3,346
$0.4
$0.4
$254
$5.9
$5.9
$11.7
$7.9
$2.4
$10.2
2012
168,551
$2.3
$14
175,041
$6.0
$6.0
$1.9
$11
118,462
$8.1
$1.3
$11
86,656


$0.6
$0.6
$14
140,193

$0.8
$0.8
$12
78,819

$1.1
$1.1
$29
39,327

$0.7
$0.7
$36
3,352

$0.1
$0.1
$60
3,397
$0.4
$0.4
$250
$14.1
$6.0
$20.1
$9.3
$3.8
$13.1
2013
172,667


178,546


$1.5
$1.5
$16
120,507

$2.0
$16
88,155

$2.2
$0.6
$0.6
$14
142,514
$3.6
$0.8
$0.8
$11
80,233
$5.0
$1.1
$1.1
$28
39,763
$2.8
$0.7
$0.7
$36
3,402
$0.2
$0.1
$0.1
$59
3,448
$0.4
$0.4
$246

$13.9
$13.9
$7.2
$5.3
$12.5
2014
176,783


182,051


$1.5
$1.5
$16
122,553

$2.0
$16
89,654


$0.6
$1.2
$20
144,836

$0.8
$1.7
$17
81,647

$1.1
$2.4
$43
40,199

$0.7
$1.4
$52
3,452

$0.1
$0.2
$74
3,499
$1.7
$1.7
$0.4
$0.4
$243
$1.7
$1.7
$3.5
$7.2
$8.7
$15.9
2015
180,899


185,556


$1.5
$1.5
$16
124,599

$2.0
$16
91,154


$0.6
$1.2
$19
147,157

$0.8
$1.7
$17
83,061

$1.1
$2.4
$42
40,635

$0.7
$1.4
$52
3,502

$0.1
$0.2
$73
3,550
$0.8
$0.8
$478



$7.7
$9.1
$16.7
2016
185,015


189,061


$1.5
$1.5
$16
126,644

$2.0
$16
92,653


$0.6
$1.2
$19
149,478

$0.8
$1.7
$17
84,475

$1.2
$14
41,071

$0.7
$16
3,552

$0.1
$15
3,601
$0.4
$0.4
$236



$5.3
$6.7
$12.0
2017
189,131


192,566


$1.5
$1.5
$15
128,690

$2.0
$15
94,152



$0.5
$6
151,799

$0.9
$6
85,889

$1.2
$14
41,507

$0.7
$16
3,602

$0.1
$15
3,652
$0.4
$0.4
$232



$3.9
$5.3
$9.1
2018 2019
193,247 197,363


196,071 199,576




130,736 132,781



95,652 97,151



$0.5
$6
154,120 156,441

$0.9
$6
87,303 88,717

$1.2
$14
41,943 42,379

$0.7
$16
3,652 3,702

$0.1
$15
3,703 3,754
$0.4 $0.4
$0.4 $0.4
$229 $226



$0.4 $0.4
$3.8 $0.4
$4.2 $0.8
Total

$9.3
$0.0
$11.4
$0.0


$14.0
$6.0
$17.0
$7.4


$13.3
$0.0
$16.3
$0.0


$2.5
$4.7
$3.1
$5.8

$3.3
$7.0
$4.1
$8.5

$4.7
$9.7
$5.7
$11.8

$2.9
$5.7
$3.6
$6.9

$0.4
$0.6
$0.5
$0.8

$3.5
$3.5
$4.2
$4.2

$54.0
$37.2
$91.2
$65.8
$45.4
$111.2
                                    6-24

-------
                                             Estimated Engine and Equipment Costs
6.2.2 Engine Variable Costs

   Engine variable costs are those costs for new hardware required to meet the new emission
standards. In this section, we present our estimates of engine variable costs. Because of the
wide variation of engine sizes in the nonroad market, we have chosen an approach that results
not in a specific cost per engine for engines within a given power range, but rather a set of
equations that can be used to determine the variable costs for any engine provided its
displacement and number of cylinders are known.  As a result, we do not present here a cost of,
say, $50 per engine for engines in the 25 to 50 power range, but instead present cost equations
that can be used to determine the variable costs for an engine having, for example, a 0.5 liter
engine with two cylinders.  We believe this is a more comprehensive approach because it allows
the reader to calculate costs more precisely for whatever engine(s) they are interested in.
Further, variable costs can vary quite significantly within a given power range unless the range is
kept very small.  To state an average variable cost for a range such as 175 to 300 hp is far less
precise than what we present here.  Using the equations presented in this section, we have then
estimated the engine variable costs for certain specific pieces of equipment and for the sales
weighted average piece of equipment. These estimates can be found in Section 6.5.

   The discussion here considers both near-term and long-term cost estimates. We believe there
are factors that cause variable hardware costs to decrease over time, making it appropriate to
distinguish between near-term and long-term costs.  Research in the costs of manufacturing has
consistently shown that as manufacturers gain experience in production, they are able to apply
innovations to  simplify machining  and assembly operations, use lower cost materials, and reduce
the number or complexity of component parts, all of which allows them to lower the per-unit
cost of production. These effects are often described as the manufacturing learning curve.11

   The learning curve is a well documented phenomenon dating back to the 1930s. The general
concept is that unit costs decrease as cumulative production increases. Learning curves are often
characterized in terms of a progress ratio, where each doubling of cumulative production leads to
a reduction in unit cost to a percentage "p" of its former value (referred to as a "p cycle").
Organizational learning, which brings about a reduction in total cost, is caused by improvements
in several areas.  Areas involving direct labor and material are usually the source of the greatest
savings.  Examples include, but are not limited to, a reduction in  the number or complexity of
component parts, improved  component production,  improved assembly speed and processes,
reduced error rates, and improved manufacturing process. These all result in higher overall
production, less scrappage of materials and products, and better overall quality.  As each
successive p cycle takes longer to complete, production proficiency generally reaches a
relatively stable plateau, beyond which increased production does not necessarily lead to
markedly decreased costs.

   Companies and industry sectors learn differently. In a 1984 publication, Button and Thomas
reviewed the progress ratios for 108 manufactured items from 22 separate field studies
representing a variety  of products and services.12 The distribution of these progress ratios is
shown in Figure  6.2-1. Except for one company that saw increasing costs as production
continued, every study showed cost savings of at least five percent for every doubling of

                                          6-25

-------
Final Regulatory Impact Analysis
production volume.  The average progress ratio for the whole data set falls between 81 and 82
percent. Other studies (Alchian  1963, Argote and Epple 1990, Benkard 1999) appear to support
the commonly used p value of 80 percent, i.e., each doubling of cumulative production reduces
the former cost level by 20 percent.
                                     Figure 6.2-1
                            Distribution of Progress Ratios
                         Distribution  of Progress Ratio
   15
   10
 o
 c
        55 57  59  61  63  65  67  69 71  73  75  77  79  81  83 85  87  89  91  93  95  97  99 101 103 105 107
                                        Progress Ratio
 From 22 field studies (n = 108).
   The learning curve is not the same in all industries. For example, the effect of the learning
curve seems to be less in the chemical industry and the nuclear power industry where a doubling
of cumulative output is associated with 11 percent decrease in cost (Lieberman 1984,
Zimmerman 1982).  The effect of learning is more difficult to decipher in the computer chip
industry (Gruber 1992).

   We believe the learning curve is appropriate to consider in assessing the cost impact of diesel
engine emission controls. The learning curve applies to new technology, new manufacturing
operations, new parts, and new assembly operations. Nonroad diesel engines currently do not
use any form of NOx aftertreatment and have used diesel particulate filters only in limited
application. These are therefore new technologies for nonroad diesel engines and will involve
some new manufacturing operations, new parts, and new assembly operations beyond those
anticipated in response to the HD2007 rule.  Since this will be a new product, we believe this is
an appropriate situation for the learning curve concept to apply. Opportunities to reduce unit
labor and material costs and increase productivity (as discussed above) will be great.  We believe
a similar opportunity exists for the new control systems that will integrate the function of the
                                          6-26

-------
                                             Estimated Engine and Equipment Costs
engine and emission-control technologies. While all nonroad diesel engines beginning with Tier
3 compliance are expected to have the basic components of this system—advanced engine
control modules (computers), advanced engine air management systems (cooled EGR, and
variable geometry turbocharging), and advanced electronic fuel systems including common rail
systems— they will be applied in some new ways in response to the Tier 4 standards.
Additionally some new components will be applied for the first time. These new parts and new
assemblies will involve new manufacturing operations.  As manufacturers gain experience with
these new systems, comparable learning is expected to occur with respect to unit labor and
material costs. These changes require manufacturers to start new production procedures, which
will improve with experience.

   We have applied a p value of 80 percent beginning with the first year of introduction of any
new technology.  That is, variable costs were reduced by 20 percent for each doubling of
cumulative production following the year in which the technology was first introduced in a given
power range of engines.  This way, learning is applied at the start of 2013 for engines over 175
hp and in 2014 for engines between 75 to 175 hp because of the one-year difference in their first
year of compliance (i.e., the first year in which new technologies are introduced).  Because the
timing of the emission standards in this final rule follows that of the HD2007 rule, we have used
the first stage of learning done via that rule as the starting point of learning for nonroad engines.
In other words, the first learning phase for highway engines serves as the baseline level of
learning for nonroad engines.  We have then applied one additional learning step from there. In
the HD2007 rule, we applied a second learning step following the second doubling of production
that occurs at the end of the 2010 model year. We could have chosen that point as our baseline
case for nonroad and then applied a single learning curve effect from there.  Instead,  we have
chosen to use as our nonroad baseline the first learning step from the highway rule so that, with
our single nonroad learning step, we have costs consistent with those costs estimated for
highway diesel engines.  In the long term, after applying the nonroad learning curve, our cost
estimates for CDPFs and NOx adsorbers are the same for similar nonroad and highway diesel
engines. This approach is consistent with the approach taken in our Tier 2 light-duty highway
rule and the HD2007 rule for heavy-duty gasoline  engines. There, compliance was being met
through improvements to existing technologies rather than the development of new technologies.
We argued in those rules that, with existing technologies, there is less opportunity for lowering
production costs.  For that reason, we applied only one learning curve effect. The situation is
similar for nonroad engines. Because these will be existing technologies by the time they are
introduced into the market, there would arguably be less opportunity for learning than there will
be for the highway engines where the technologies are first introduced.

   Another factor that plays into our near-term and long-term cost estimates is that for warranty
claim rates. In our HD2007 rule, we estimated a warranty claim rate of one percent.  Subsequent
to that rule, we learned from industry that repair rates can be as much as two to three times
higher during the initial years of production for a new technology relative to later years.13 For
this analysis, we have applied what we have learned in our warranty estimates by using a three
percent warranty claim rate during the first two years and then one percent warranty  claim rate
thereafter. This difference in warranty claim rates, in addition to the learning effects discussed
above, is reflected in the different long-term costs relative to near-term costs.

                                         6-27

-------
Final Regulatory Impact Analysis
   6.2.2.1 NOx Adsorber System Costs

   The NOx adsorber system anticipated for Tier 4 is the same technology as for highway
applications. For the NOx adsorber to function properly, a systems approach that includes a
reductant metering system and control of engine air-fuel ratio is also necessary.  Many of the
new air handling and electronic system technologies developed in order to meet the Tier 2/Tier 3
nonroad diesel engine standards can be applied to accomplish the NOx adsorber control
functions as well. Some additional hardware for exhaust NOx or O2 sensing and for fuel
metering probably will be required. The cost estimates include a DOC for clean-up of
hydrocarbon emissions that occur during NOx  adsorber regeneration events.

   We have used the same methodology to estimate costs associated with NOx  adsorber systems
as was used in our HD2007 rulemaking. The basic components of the NOx adsorber catalyst are
well known and include the following material elements:

•  an oxidation catalyst, typically platinum-based;
   an alkaline earth metal to store NOx, typically barium-based;
   a NOx reduction catalyst, typically rhodium-based;
   a substrate upon which the catalyst washcoating is applied; and,
   a can to hold and  support the substrate.

   Examples of these material costs are summarized in Table 6.2-11 and represent costs to the
engine manufacturers inclusive of supplier markups. The manufacturer costs shown in Table
6.2-11 (as well as Tables 6.2-13 and 6.2-18 for CDPF systems and DOCs, respectively) include
additional markups to account for both manufacturer and dealer overhead and carrying costs.
The application of overhead and carrying costs are consistent with the approach taken in the
HD2007 rulemaking. In that rule, we used an approach to estimating the markup for catalyzed
emission-control technologies based on input from catalyst manufacturers. Specifically, we were
told that device manufacturers could not mark up the cost of the individual components within
their products because those components consist of basic commodities (for example, precious
metals used in the catalyst could not be arbitrarily marked up because  of their commodity status).
Instead, manufacturing entities could mark up costs only where they add a unique value to the
product. In the case of catalyst systems, we were told that the underlying cost of precious
metals, catalyst substrates, PM filter substrates, and canning materials  were well known to both
buyer and seller and no markup or profit recovery for those component costs could be derived by
the catalyst manufacturer. In essence, these are components to which the supplier provides little
value-added engineering. The one component  that was unique to each catalyst manufacturer
(i.e., the component where they add a unique value) was the catalyst washcoat support materials.
This mixture (which is effectively specialized clays) serves to hold the catalytic  metals in place
and to control the surface area of the catalytic metals available for emission control.  Although
the commodity price for  the materials used in the washcoat is almost negligible (i.e., perhaps one
or two dollars), we have estimated a substantial cost for washcoating based on the engineering
value added by the catalyst manufacturer in this step. This is reflected in the costs presented for
NOx adsorber  systems, CDPF systems, and DOCs.  This portion of the cost estimate - the
washcoating - is where the catalyst manufacturer recovers the fixed cost for research and

                                          6-28

-------
                                             Estimated Engine and Equipment Costs
development as well as realizes a profit. To these manufacturer costs, we have added a four
percent carrying costs to account for the capital cost of the extra inventory, and the incremental
costs of insurance, handling, and storage. A dealer carrying cost in included to cover the cost of
capital tied up in extra inventory. Considering input received from industry, we have adopted
this approach of estimating individually the manufacturer and dealer markups in an effort to
better reflect the value each entity adds at various stages of the supply chain.14 Also included is
our estimate of warranty costs for the NOx adsorber system.
                                          6-29

-------
Final Regulatory Impact Analysis
                      Table 6.2-11.  NOx Adsorber System Costs

Horsepower
Engine Displacement (Liter)
Material and Component Costs
Catalyst Volume (Liter)
Substrate
Washcoating and Canning
Platinum
Rhodium
Alkaline Earth Oxide, Barium
Catalyst Can Housing
Direct Labor Costs
Estimated Labor hours
Labor Rate ($/hr)
Labor Cost
Labor Overhead @ 40%
Total Direct Costs to Mfr.
Warranty Cost - Near Term (3% claim rate)
Mfr. Carrying Cost - Near Term
Total Cost to Dealer - Near Term
Dealer Carrying Cost - Near Term
DOC for cleanup - Near Term
Baseline Cost to Buyer - Near Term
Cost to Buyer w/ Highway learning - Near Term
Warranty Cost - Long Term (1% claim rate)
Mfr. Carrying Cost - Long Term
Total Cost to Dealer - Long Term
Dealer Carrying Cost - Long Term
DOC for cleanup - Long Term
Baseline Cost to Buyer - Long Term
Cost to Buyer w/ Highway learning - Long Term
Cost to Buyer w/ Nonroad learning - Long Term
NOx Adsorber Costs ($2002)
9hp
0.39

0.59
$3
$13
$16
$3
$1
$9

2
$30
$45
$18
$109
$9
$4
$122
$4
$105
$231
$206
$3
$4
$116
$3
$99
$219
$195
$176
33 hp
1.50

2.25
$12
$52
$62
$11
$1
$9

2
$30
$45
$18
$210
$17
$8
$235
$7
$132
$375
$326
$6
$8
$224
$7
$125
$356
$310
$273
76 hp
3.92

5.88
$32
$135
$163
$28
$1
$9

2
$30
$45
$18
$431
$34
$17
$482
$14
$192
$688
$589
$11
$17
$459
$14
$182
$656
$561
$485
150hp
4.70

7.05
$38
$162
$195
$34
$1
$9

2
$30
$45
$18
$502
$39
$20
$561
$17
$211
$789
$674
$13
$20
$535
$16
$201
$752
$642
$554
250 hp
7.64

11.46
$62
$263
$318
$55
$1
$13

2
$30
$45
$18
$775
$59
$31
$865
$26
$286
$1,177
$999
$20
$31
$826
$25
$272
$1,123
$952
$816
503 hp
18.00

27.00
$147
$620
$748
$129
$1
$18

2
$30
$60
$24
$1 ,747
$131
$70
$1 ,948
$58
$459
$2,465
$2,064
$44
$70
$1 ,861
$56
$437
$2,354
$1 ,970
$1 ,664
660 hp
20.30

30.45
$166
$700
$844
$145
$1
$18

2
$30
$60
$24
$1 ,957
$146
$78
$2,182
$65
$497
$2,745
$2,295
$49
$78
$2,084
$63
$474
$2,621
$2,191
$1 ,848
1000hp
34.50

51.75
$282
$1,189
$1,434
$246
$1
$18

2
$30
$60
$24
$3,254
$244
$130
$3,628
$109
$734
$4,471
$3,724
$81
$130
$3,466
$104
$700
$4,270
$3,556
$2,985
                                      6-30

-------
                                             Estimated Engine and Equipment Costs
   We have estimated the cost of this system based on information from several reports.15'16'17
The individual estimates and assumptions used to estimate the cost for the system are
documented in the following paragraphs.

   NOx Adsorber Catalyst Volume

   The Engine Manufacturers Association was asked as part of a contractor work assignment to
gather input from their members on likely technology solutions including the NOx adsorber
catalyst.18 The respondents indicated that the catalyst volume for a NOx adsorber catalyst may
range from 1.5 times the engine displacement to as much as 2.5 times the engine displacement
based on current washcoat technology. Based on current lean burn gasoline catalyst designs and
engineering judgment, we have estimated that the NOx adsorber catalyst will be sized on
average 1.5 times the engine displacement. This is consistent with the size of the NOx adsorber
catalyst on the Toyota Avensis diesel passenger car (60 prototypes of a planned 2003 production
car are being tested in Europe), which is sized at 1.4 times engine displacement.19

   NOx Adsorber Substrate

   The ceramic flow-through substrates used for the NOx adsorber catalyst were estimated to
cost $5.27 ($1999) per liter during our HD2007 rule.  This cost estimate was based on a
relationship developed for current heavy-duty  gasoline catalyst substrates.20 We have converted
that value to $5.44 ($2002) using the PPI for Motor Vehicle Parts and Accessories, Catalytic
Converters.21

   NOx Adsorber Washcoating and Canning

   We have estimated a "value-added" engineering and material product, called washcoating
and canning, based on feedback from members of the Manufacturers of Emission Control
Association (MECA).22 By using a value-added component that accounts for fixed costs
(including R&D), overhead, marketing and profits from likely suppliers of the technology, we
can estimate this fraction of the cost for the technology apart from other components that are
more widely available as commodities (e.g, precious metals and catalyst substrates).  Based on
conversations with MECA, we understand this element of the product to represent the catalyst
manufacturer's value added and, therefore, their opportunity for markup. As a result, the
washcoating and canning costs shown in Table 6.2-11 represent costs with manufacturer
markups  included.

   NOx Adsorber Precious Metals

   The total precious metal content for the NOx adsorber is estimated to be 50 g/ft3 with
platinum representing 90 percent of that total and rhodium representing 10 percent.  The costs
for rhodium and platinum used in this analysis are the 2002 average prices of $839 per troy
ounce for rhodium and $542 per troy ounce for platinum, as reported by Johnson Matthey.23

   NOx Adsorber Alkaline Earth Metal - Barium

                                          6-31

-------
Final Regulatory Impact Analysis
   The cost for barium carbonate (the primary NOx storage material) is assumed to be less than
$1 per catalyst as estimated in "Economic Analysis of Diesel Aftertreatment System Changes
Made Possible By Reduction of Diesel Fuel Sulfur Content."

   NOx Adsorber Can Housing

   The material cost for the can housing is estimated based on the catalyst volume plus 20
percent for transition cones, plus 20 percent for scrappage (material purchased but unused in the
final product) and a price of $1.01 per pound for 18 gauge stainless steel as estimated in a
contractor report to EPA and  converted into $2002.24

   NOx Adsorber Direct Labor

   The direct labor costs for  the catalyst are estimated based on an estimate of the number of
hours required for assembly and established labor rates. Additional overhead for labor was
estimated as 40 percent of the labor rate.25

   NOx Adsorber Warranty

   We have estimated both near-term and long-term warranty costs.  Near-term warranty costs
are based on a three percent claim rate and an estimate of parts and labor costs per incident,
while long-term warranty costs are based on a one percent claim rate and an estimate of parts
and labor costs per incident.  The labor rate is assumed to be $50 per hour with four hours
required per claim, and parts  costs are estimated to be 2.5 times the original manufacturing cost
for the component. The calculation of near-term warranty costs for the 9 hp engine shown in
Table 6.2-11 is as follows:

       [($3 + $13 + $16 +  $3 + $1 + $9)(2.5) + ($50)(4hours)](3%) = $9

   NOx Adsorber Manufacturer and Dealer Carrying Costs

   The manufacturer's carrying cost was estimated at 4 percent of the direct costs. This reflects
primarily the costs of capital  tied up in extra inventory, and secondarily the incremental costs of
insurance, handling and storage. The dealer's carrying cost was estimated at 3 percent of the
incremental cost, again reflecting primarily the cost of capital tied up in extra inventory.26

   NOx Adsorber DOC for System  Clean-up

   Included in the costs for the NOx adsorber system are costs for a diesel oxidation catalyst
(DOC) for clean-up of possible excess hydrocarbon emissions that might occur as a result of
system regeneration (removal of stored NOx and reduction to N2  and O2).  The methodology
used to estimate DOC  system costs is consistent with the methodology outlined here for NOx
adsorber systems and is presented below in Section 6.2.2.3. Important to note here is that the
DOC costs shown in Table 6.2-11 are lower in the long term because of the lower warranty
                                          6-32

-------
                                             Estimated Engine and Equipment Costs
claim rate—three percent in the near term and one percent in the long term; learning effects, as
discussed below, are not applied to DOC costs.

   NOx Adsorber Cost Estimation Function

   Using the example NOx adsorber costs shown in Table 6.2-11, we calculated a linear
regression to determine the NOx adsorber system cost as a function of engine displacement.
This way, the function can be applied to the wide array of engines in the nonroad fleet to
determine the total or per engine costs for NOx adsorber hardware. The functions calculated for
NOx adsorber system costs used throughout this analysis are shown in Table 6.2-12. Note that
Table 6.2-11 shows NOx adsorber system costs for engines under 75 hp. We do not anticipate
any engines under 75 hp will apply NOx adsorber systems to comply with the new emission
standards. Nonetheless, the costs shown were used to generate the equations shown in Table
6.2-12. Because of the linear relationship between engine displacement and NOx adsorber
system size (and, therefore, cost), including the costs for these smaller engines does not
inappropriately  shift the cost equation downward.

                                      Table 6.2-12
                       NOx Adsorber System Costs as a Function of
               Engine Displacement (x represents engine displacement in liters)
                                         $2002
Near-Term Cost Function
Long-Term Cost Function
$103(x)+$183
$83(x)+$160
R2=0.9998
R2=0.9997
   Table 6.2-12 shows both a near-term and a long-term cost function for NOx adsorber system
costs. The near-term function incorporates the near-term warranty costs determined using a
three percent claim rate, while the long-term function incorporates the long-term warranty costs
determined using a one percent claim rate. Additionally, the long-term  function incorporates
learning curve effects for certain elements of the NOx adsorber system (i.e., learning effects
were not applied to the DOC portion of the NOx adsorber system, for reasons discussed below).
In the HD2007 rule, we applied two learning effects of 20 percent. Here,  we have assumed one
learning effect of 20 percent as a baseline level of learning; this represents learning done as a
result of the HD2007 rule.  After a single doubling of production (i.e., two years), we have then
applied a single nonroad learning effect of 20 percent. Note that the equations shown in Table
6.2-12 include costs for a clean-up DOC; results generated using the DOC cost estimation
equations presented in Table 6.2-16 should not be added to results generated using the equations
in Table 6.2-12 to determine NOx adsorber system costs.

   6.2.2.2 Catalyzed Diesel Particulate Filter Costs

   As with the NOx adsorber system, the anticipated CDPF system for Tier 4 is the same as that
used for highway applications, except that we are projecting that some form of active
regeneration system will be employed as a backup to the passive regeneration capability of the

                                          6-33

-------
Final Regulatory Impact Analysis
CDPF. For the CDPF to function properly, a systems approach that includes a reductant
metering system and control of engine air-fuel ratio is also necessary.  Many of the new air
handling and electronic fuel system technologies developed in order to meet the Tier 2/Tier 3
nonroad engine standards can be applied to accomplish the CDPF control functions as well.
Nonroad applications are expected to present challenges beyond those of highway applications
with respect to implementing CDPFs.  For this reason, we anticipate that some additional
hardware beyond the diesel particulate filter itself may be required to ensure that CDPF
regeneration occurs. For some engines this may be new fuel control strategies that force
regeneration under some circumstances, while in other engines it might involve an exhaust
system fuel injector to inject fuel upstream of the CDPF to provide necessary heat for
regeneration under some operating conditions.  The cost estimates for such a regeneration system
are presented in Section 6.2.2.3.

   We have used the same methodology to estimate costs associated with CDPF systems used in
our FID2007 rulemaking (although here, for nonroad engines, we have included costs for a
regeneration system that was not part of the cost estimate in the HD2007 rule). The basic
components of the CDPF are well known and include the following  material elements:

   an oxidation catalyst, typically platinum-based;
   a substrate upon which the catalyst washcoating is applied and upon which PM is trapped;
   a can to hold and support the substrate; and,
•  a regeneration system to ensure regeneration under all operating conditions (see Section
   6.2.2.3).

   Examples of these material costs are summarized in Table 6.2-13 and represent costs to the
engine manufacturers inclusive of supplier markups. The total direct cost to the manufacturer
includes an estimate of warranty costs for the CDPF system.  Hardware costs are additionally
marked up to account for both manufacturer and dealer overhead and carrying costs. The
manufacturer's carrying cost was estimated to be four percent of the direct costs accounting for
the capital cost of the extra inventory, and the incremental costs of insurance, handling, and
storage.  The dealer's carrying cost was marked up three percent reflecting the cost of capital
tied up in inventory. Considering input received from industry, we have adopted this approach
of estimating individually the manufacturer and dealer markups in an effort  to better reflect the
value added at each stage of the supply chain.27
                                          6-34

-------
                                             Estimated Engine and Equipment Costs
            Table 6.2-13.  Catalyzed Diesel Particulate Filter (CDPF) System Costs

Horsepower
Average Engine Displacement (Liter)
Material and Component Costs
Filter Volume (Liter)
Filter Trap
Washcoating and Canning
Platinum
Filter Can Housing
Differential Pressure Sensor
Direct Labor Costs
Estimated Labor hours
Labor Rate ($/hr)
Labor Cost
Labor Overhead @ 40%
Total Direct Costs to Mfr.
Warranty Cost - Near Term (3% claim rate)
Mfr. Carrying Cost - Near Term
Total Cost to Dealer - Near Term
Dealer Carrying Cost - Near Term
Savings by removing muffler
Baseline Cost to Buyer - Near Term
Cost to Buyer w/ Highway learning - Near Term
Warranty Cost - Long Term (1% claim rate)
Mfr. Carrying Cost - Long Term
Total Cost to Dealer - Long Term
Dealer Carrying Cost - Long Term
Savings by removing muffler
Baseline Cost to Buyer - Long Term
Cost to Buyer w/ Highway learning - Long Term
Cost to Buyer w/ Nonroad learning - Long Term

9hp
0.39

0.59
$36
$13
$11
$7
$46
$0
2
$30
$60
$24
$198
$12
$8
$218
$7
-$46
$178
$142
$4
$8
$210
$6
-$46
$170
$136
$109
Catalyzed
33 hp
1.50

2.25
$139
$52
$42
$7
$46
$0
2
$30
$60
$24
$370
$24
$15
$409
$12
-$46
$375
$300
$8
$15
$393
$12
-$46
$359
$287
$229
Diesel
76 hp
3.92

5.88
$364
$135
$109
$7
$46
$0
2
$30
$60
$24
$746
$53
$30
$828
$25
-$46
$806
$645
$18
$30
$793
$24
-$46
$770
$616
$493
Particulate Filter (CDPF) Costs
150hp
4.70

7.05
$437
$162
$130
$7
$46
$0
2
$30
$60
$24
$867
$62
$35
$963
$29
-$46
$945
$756
$21
$35
$922
$28
-$46
$903
$722
$578
250 hp
7.64

11.46
$710
$263
$212
$10
$46
$0
2
$30
$60
$24
$1 ,326
$96
$53
$1 ,475
$44
-$46
$1 ,473
$1,178
$32
$53
$1,411
$42
-$46
$1 ,407
$1,125
$900
503 hp
18.00

27.00
$1,673
$620
$499
$14
$46
$0
2
$30
$60
$24
$2,937
$217
$117
$3,271
$98
-$46
$3,323
$2,658
$72
$117
$3,126
$94
-$46
$3,174
$2,539
$2,031
($2002)
660 hp
20.30

30.45
$1 ,886
$700
$563
$14
$93
$0
4
$30
$120
$48
$3,424
$247
$137
$3,808
$114
-$46
$3,876
$3,101
$82
$137
$3,643
$109
-$46
$3,706
$2,965
$2,372

1000hp
34.50

51.75
$3,206
$1,189
$956
$14
$93
$0
4
$30
$120
$48
$5,626
$412
$225
$6,264
$188
-$46
$6,405
$5,124
$137
$225
$5,989
$180
-$46
$6,122
$4,898
$3,918
   CDPF Volume

   During development of our HD2007 rule, the Engine Manufacturers Association was asked
as part of a contractor work assignment to gather input from their members on catalyzed diesel
particulate filters for heavy-duty highway applications.28 The respondents indicated that the
particulate filter volume may range from 1.5 times the engine displacement to as much as 2.5
times the engine displacement based on their experiences at that time with cordierite filter
technologies. The size of the diesel particulate filter is selected largely based on the maximum
allowable flow restriction for the engine. Generically, the filter size is inversely proportional to
its resistance to flow (a larger filter is less restrictive than a similar smaller filter). In the
FID2007 rule and here, we have estimated that the  diesel particulate filter will be sized to be 1.5
times the engine displacement based on the responses received from EMA and on-going research
                                          6-35

-------
Final Regulatory Impact Analysis
aimed at improving filter porosity control to give a better trade-off between flow restrictions and
filtering efficiency.

    CDPF Substrate

    CDPFs can be made from a wide range of filter materials including wire mesh, sintered
metals, fibrous media, or ceramic extrusions. The most common material used for CDPFs for
heavy-duty diesel engines is cordierite. Here we have based our cost estimates on the use of
silicon carbide (SiC) even though it is more expensive than other filter materials. In the FID2007
rule, we estimated that CDPFs will consist of a cordierite filter costing $30 per liter. To remain
conservative in our cost estimates for nonroad applications, we have assumed the use of silicon
carbide filters costing double that amount, or $60 per liter.0 This cost is directly proportional to
filter volume, which is proportional to engine displacement. This $60 value is then converted to
$2002 using the PPI for Motor Vehicle Parts and Accessories, Catalytic Converters.29 The end
result being a cost of $62 per liter.

    CDPF Washcoating and Canning

    These costs were done in a consistent manner as done for NOx adsorber catalyst systems, as
discussed above.

    CDPF Precious Metals

    The total precious metal content for catalyzed diesel particulate filters is estimated to be 30
g/ft3 with platinum as the only precious metal used in the filter.  As done for NOx adsorbers, we
have used a price of $542 per troy ounce for platinum.

    CDPF Can Housins

    The material cost for the can housing is estimated based on the CDPF volume plus 20
percent for transition cones, plus 20 percent for scrappage (material purchased but unused in the
final product)  and a price of $1.01 per pound for 18 gauge stainless steel as estimated in a
contractor report to EPA and converted into $2002.30

    CDPF Differential Pressure Sensor

    We have assumed that the catalyzed diesel particulate filter system will require the use of a
differential pressure sensor to provide a diagnostic  monitoring function of the filter.  A
contractor report to EPA estimated the cost for such a sensor at $45.31 A PPI adjusted cost of
$46 per sensor has been used in this analysis.
     Note that we are being especially conservative with respect to >750 horsepower mobile machines where we
believe that manufacturers may in fact use a wire mesh substrate rather than the SiC substrate we have costed and,
indeed, we have based the level of the 2015 PM standard on this use of wire mesh substrates. We have chosen to remain
conservative in our cost estimates by assuming use of a SiC substrate for all engines.

                                           6-36

-------
                                             Estimated Engine and Equipment Costs
   CDPF Direct Labor

   Consistent with the approach for NOx adsorber systems, the direct labor costs for the CDPF
are estimated based on an estimate of the number of hours required for assembly and established
labor rates.  Additional overhead for labor was estimated as 40 percent of the labor rate.32

   CDPF Warranty

   We have estimated both near-term and long-term warranty costs. Near-term warranty costs
are based on a three percent claim rate and an estimate of parts and labor costs per incident,
while long-term warranty costs are based on a one percent claim rate and an estimate of parts
and labor costs per incident. The labor rate is assumed to be $50 per hour with two hours
required per claim, and parts cost are estimated to be 2.5 times the original manufacturing cost
for the component.

   CDPF Manufacturer and Dealer Carrying Costs

   Consistent with the approach for NOx adsorber systems, the manufacturer's carrying cost
was estimated at 4 percent of the direct costs. This reflects primarily the costs of capital tied up
in extra inventory, and secondarily the incremental costs of insurance, handling and storage. The
dealer's carrying cost was estimated at 3 percent of the incremental cost, again reflecting
primarily the cost of capital tied up in extra inventory.33

   Savings Associated with Muffler Removal

   CDPF retrofits are currently often incorporated in, or are simply replacements for, the
muffler for diesel-powered vehicles and equipment. One  report noted that, "Often, the trap could
be mounted in place of the muffler and had the same dimensions. Thus, rapid replacement was
possible.  The muffling effect was often even better."34 We have assumed that applying a CDPF
allows for the removal of the muffler due to the noise attenuation characteristics of the CDPF.
We have accounted for this savings and have estimated a muffler cost of $46.  The $46 estimate
is an average for all engines; the actual savings will be higher for some and lower for others.

   CDPF System Cost Estimation Function

   Using the example CDPF costs shown in Table 6.2-13, we calculated a linear regression to
determine the CDPF system cost as a function of engine displacement.  This way, the function
can be applied to the wide array of engines in the nonroad fleet to determine the total or per
engine costs for CDPF system hardware. The functions calculated for CDPF system costs used
throughout this analysis are shown in Table 6.2-14.
                                         6-37

-------
Final Regulatory Impact Analysis
                                      Table 6.2-14
                           CDPF System Costs as a Function of
              Engine Displacement (x represents engine displacement in liters)
                                         $2002
Near -term Cost Function
Long-term Cost Function
$146(x)+$75
$112(x)+$57
R2=0.9997
R2=0.9997
   The near-term and long-term costs shown in Table 6.2-14 change due to the different
warranty claim rates and the application of a 20 percent learning curve effect.

   6.2.2.3 CDPF Regeneration System Costs

   The CDPF regeneration system is likely to include an O2 sensor, a means for exhaust air to
fuel ratio control (one or more exhaust fuel injectors or in-cylinder means), a temperature sensor
and possibly a means to control mass flow through a portion of the catalyst system (for example,
for a "dual-bed" system).  Incremental costs for a CDPF regeneration system, along with several
other costs discussed below, were developed by ICF Consulting under contract to EPA.35 The
cost estimates developed by ICF for a CDPF regeneration system are summarized in Table 6.2-
15.

                                     Table 6.2-15.
                  CDPF Regeneration System - Costs to the Manufacturer
ICF Estimated Regeneration System Costs to Manufacturers ($2002)
Horsepower
Displacement (L)
CDPF Regeneration System Costs
20
1
$265
35
2
$279
80
3
$293
150
6
$384
250
8
$408
400
10
$431
650
16
$530
1000
24
$676
   Using these costs, we then estimated costs to the buyer using the same learning curve effects
and warranty claim rate factors discussed above. These results are presented in Table 6.2-16.
                                         6-38

-------
                                             Estimated Engine and Equipment Costs
                                      Table 6.2-16.
                      CDPF Regeneration System - Costs to the User
EPA Estimate of CDPF Regeneration System Costs ($2002)
Horsepower
Dsplacement (L)
CDPF Regeneration System Costs
Vterranty Cost - Near Term (3% claim rate)
Mfr. Carrying Cost (4%) - Near Term
Total Cost to Dealer - Near Term
Dealer Carrying Cost (3%) - Near Term
Total Cost to Buyer - Near Term
Vterranty Cost - Long Term (1%claim rate)
Mfr. Carrying Cost (4%)- Long Term
Total Cost to Dealer - Long Term
Dealer Carrying Cost (3%) - Long Term
Subtotal
Total Cost to Buyer - Long-Term W learning
20
1
$265
$23
$11
$298
$9
$307
$8
$11
$283
$8
$291
$233
35
2
$279
$24
$11
$314
$9
$323
$8
$11
$298
$9
$307
$245
80
3
$293
$25
$12
$330
$10
$340
$8
$12
$313
$9
$323
$258
150
6
$384
$32
$15
$432
$13
$445
$11
$15
$410
$12
$423
$338
250
8
$408
$34
$16
$458
$14
$471
$11
$16
$435
$13
$448
$359
400
10
$431
$35
$17
$484
$15
$498
$12
$17
$460
$14
$474
$379
650
16
$530
$43
$21
$593
$18
$611
$14
$21
$565
$17
$582
$466
1000
24
$676
$54
$27
$756
$23
$779
$18
$27
$721
$22
$742
$594
   As noted above, the CDPF regeneration system is expected to consist of an O2 sensor, a
temperature sensor, and probably a pressure sensor. The costs shown in Table 6.2-16 assume
none of these sensors or other pieces of hardware exist and, more importantly, they assume the
fuel control systems present in the engine are not capable of the sort of precise fuel control that
could perform many of the necessary functions of the regeneration system without any additional
hardware. For this reason, we consider the costs shown  in Table 6.2-16 to be representative of
the costs for an engine with an indirect-injection (IDI) fuel system.  For a direct-injection (DI)
fuel system, we expect that many of the functional capabilities for which costs were generated
will be handled by the existing fuel system.  For example, we are assuming that all DI engines
will either convert to a fuel system capable of late injection or will already have a fuel system
capable of late injection. Late injection is one of the primary means of using fuel strategies to
force a CDPF regeneration event. Our cost estimates associated with  conversion to such fuel
systems are discussed below. Because the regeneration  system costs for DI engines are lower
than for an IDI engine, we have estimated that the regeneration system costs for a DI engine are
half of those presented in Table 6.2-16.

   Also, note that the air handling, electronic, and fuel system hardware used for backup active
CDPF regeneration is expected to be used in common with the NOx adsorber regeneration
system. We have accounted for these  costs here (as a CDPF regeneration system) because
CDPFs are required on a broader range of engines and, for many engines, earlier than are NOx
adsorbers.
                                         6-39

-------
Final Regulatory Impact Analysis
   CDPF Regeneration System Cost Estimation Function

   Using the example regeneration system costs shown in Table 6.2-16, we calculated a linear
regression to determine the CDPF regeneration system cost as a function of engine displacement.
This way, the function can be applied to the wide array of engines in the nonroad fleet to
determine the total costs for CDPF regeneration system hardware.  The functions calculated for
CDPF regeneration system costs used throughout this analysis are shown in Table 6.2-17.

                                     Table 6.2-17
                    CDPF Regeneration System Costs as a Function of
              Engine Displacement (x represents engine displacement in liters)
                                        $2002
IDI Engine
DI Engine
Near -term Cost Function
Long-term Cost Function
Near -term Cost Function
Long-term Cost Function
$20(x) + $293
$16(x)+$223
$10(x)+$147
$8(x)+$lll
R2=0.9916
R2=0.9916
R2=0.9916
R2=0.9916
   Note that these costs—either the IDI or the DI costs, depending on the type of engine—are
incurred for any engine adding a CDPF. The near-term and long-term costs shown in Table 6.2-
17 change due to the different warranty claim rates and the application of a 20 percent learning
curve effect.

   6.2.2.4 Diesel Oxidation Catalyst (DOC) Costs

   The NOx adsorber regeneration and desulfation functions may produce undesirable by-
products in the form of momentary increases in HC emissions or in odorous hydrogen sulfide
(H2S) emissions. We have assumed  that manufacturers may choose to apply a diesel oxidation
catalyst (DOC) downstream of the NOx adsorber technology to control these potential products.
The DOC serves a "clean-up" function to oxidize any HC and H2S emissions to more desirable
products. As discussed below, for our cost analysis we have also projected that engines under 75
hp will add a DOC to comply with the 2008 PM standards, not to serve a "clean-up" function but
rather to serve as the primary means of emission control.

   Our estimates of DOC costs are shown in Table 6.2-18. The individual component costs for
the DOC were estimated in the same manner as for the NOx adsorber systems  and CDPF
systems, as discussed above. However, no  learning effects were applied to DOCs because we
believe DOCs have been manufactured for  a long enough time period such that learning has
already taken place.
                                         6-40

-------
                                            Estimated Engine and Equipment Costs
                                     Table 6.2-18.
                         Diesel Oxidation Catalyst (DOC) Costs

Horsepower
Average Engine Displacement (Liter)
Material and Component Costs
Catalyst Volume (liter)
Substrate
Washcoating and Canning
Platinum (5 g/ft3)
Catalyst Can Housing
Direct Labor Costs
Estimated Labor hours
Labor Rate ($/hr)
Labor Cost
Labor Overhead @ 40%
Total Direct Costs to Mfr.
Warranty Cost - Near Term (3% claim rate)
Mfr. Carrying Cost - Near Term
Total Cost to Dealer - Near Term
Dealer Carrying Cost - Near Term
Total Cost to Buyer - Near Term
Warranty Cost - Long Term (1% claim rate)
Mfr. Carrying Cost - Long Term
Total Cost to Dealer - Long Term
Dealer Carrying Cost - Long Term
Total Cost to Buyer - Long Term
Diesel Oxidation Catalyst Costs ($2002)
9hp
0.39

0.39
$2
$61
$1
$4

0.5
$30
$15
$6
$90
$8
$4
$102
$3
$105
$3
$4
$96
$3
$99
33 hp
1.50

1.50
$8
$76
$5
$4

0.5
$30
$15
$6
$114
$10
$5
$128
$4
$132
$3
$5
$122
$4
$125
76 hp
3.92

3.92
$21
$107
$12
$4

0.5
$30
$15
$6
$166
$14
$7
$186
$6
$192
$5
$7
$177
$5
$182
150hp
4.70

4.70
$26
$117
$14
$4

0.5
$30
$15
$6
$182
$15
$7
$205
$6
$211
$5
$7
$195
$6
$201
250 hp
7.64

7.64
$42
$155
$24
$7

0.5
$30
$15
$6
$248
$20
$10
$277
$8
$286
$7
$10
$264
$8
$272
503 hp
18.00

18.00
$98
$208
$55
$15

0.5
$30
$15
$6
$398
$31
$16
$445
$13
$459
$10
$16
$425
$13
$437
660 hp
20.30

20.30
$110
$220
$63
$17

0.5
$30
$15
$6
$432
$34
$17
$483
$14
$497
$11
$17
$460
$14
$474
1000hp
34.50

34.50
$188
$294
$106
$30

0.5
$30
$15
$6
$638
$49
$26
$713
$21
$734
$16
$26
$680
$20
$700
   DOC Cost Estimation Function

   Similar to what was done for NOx adsorber systems and CDPFs, we used the example costs
shown in Table 6.2-18 to determine a cost function with engine displacement as the dependent
variable.  This way, the function can be applied to the wide array of engines in the nonroad fleet
to determine the total or per unit costs for DOC hardware, whether that hardware be a stand
alone emission-control technology or as part of a NOx adsorber system. The cost functions for
DOCs used throughout this analysis are shown in Table 6.2-19. Note that the NOx adsorber cost
estimation equations shown in Table 6.2-12 include costs for a clean-up DOC; results generated
using the DOC cost estimation equations presented in Table 6.2-19  should not be added to
results generated using the equations in Table 6.2-12 to determine NOx adsorber system costs.
                                         6-41

-------
Final Regulatory Impact Analysis
                                      Table 6.2-19
                              DOC Costs as a Function of
              Engine Displacement (x represents engine displacement in liters)
                                        $2002
Near-term Cost Function
Long-term Cost Function
$18(x)+$116
$18(x)+$110
R2=0.9944
R2=0.9944
   6.2.2.5 Closed-Crankcase Ventilation (CCV) System Costs

   Consistent with our HD2007 rule, we are removing the provision that allows turbocharged
nonroad diesel engines to vent crankcase gases directly to the environment.  Such engines are
said to have an open crankcase system.  We project that this requirement to close the crankcase
on turbocharged engines will force manufacturers to rely on engineered closed crankcase
ventilation systems that filter oil from the blow-by gases before routing them into either the
engine intake or the exhaust system upstream of the CDPF.  We expect these systems to be the
same as those expected for highway engines and have estimated their costs in the same manner
as done in our FID2007 rule. The estimated initial costs of these systems are as shown in Table
6.2-20. These costs are incurred only by turbocharged engines.

                                     Table 6.2-20.
                    Closed Crankcase Ventilation (CCV) System Costs

Horsepower
Average Engine Displacement (Liter)
Cost to Manufacturer
Warranty Cost - Near Term (3% claim rate)
Mfr. Carrying Cost - Near Term
Total Cost to Dealer — Near Term
Dealer Carrying Cost - Near Term
Total Cost to Buyer - Near Term
Warranty Cost - Long Term (1% claim rate)
Mfr. Carrying Cost - Long Term
Total Cost to Dealer - Long Term
Dealer Carrying Cost - Long Term
Cost to Buyer w/ Nonroad Learning - Long Term

9hp
0.39
$28
$5
$1
$34
$1
$35
$2
$1
$30
$1
$25
Closed
33 hp
0.93
$29
$5
$1
$35
$1
$36
$2
$1
$31
$1
$26
Crankcase Ventilation
76 hp
3.92
$34
$6
$1
$41
$1
$42
$2
$1
$37
$1
$31
150hp
4.7
$35
$6
$1
$42
$1
$44
$2
$1
$39
$1
$32
(CCV) System Costs ($2002)
250 hp
7.64
$41
$6
$2
$48
$1
$50
$2
$2
$44
$1
$37
503 hp
18
$59
$7
$2
$69
$2
$71
$2
$2
$64
$2
$53
660 hp
20.3
$64
$8
$3
$74
$2
$76
$3
$3
$69
$2
$57
1000hp
34.5
$89
$10
$4
$103
$3
$106
$3
$4
$96
$3
$79
   CCV Cost Estimation Function

   As discussed above, an equation was developed as a function of engine displacement to
calculate total or per unit CCV costs. These functions are shown in Table 6.2-21. Note that
these costs will be incurred only by turbocharged engines.
                                         6-42

-------
                                            Estimated Engine and Equipment Costs
                                     Table 6.2-21
                               CCV Costs as a Function of
              Engine Displacement (x represents engine displacement in liters)
                                        $2002
Near-term Cost Function
Long-term Cost Function
$2(x) + $34
$2(x) + $24
R2=l
R2=l
   6.2.2.6 Variable Costs of Conventional Technologies for Engines under 75 hp and over
   750 hp

   For the smaller engines, we have projected a different technology mix for complying with the
applicable emission standards. As explained in Chapter 4 of the RIA, we are projecting that
engines will comply either by adding a DOC or by making some engine modifications resulting
in engine-out emission reductions to comply with the 2008 PM standards. For our cost analysis,
we have assumed that all engines will add a DOC. Manufacturers will presumably choose the
least costly approach that provides the necessary emission control.  If engine-out modifications
are less costly than a DOC, the analysis overestimates the costs associated with meeting these
standards. If the DOC proves to be less costly, then our estimate is representative of what most
manufacturers presumably will do.  Therefore, we have assumed that, beginning in 2008, all
engines under 75 hp will add a DOC. Note that, as discussed in Chapter 4, some engines under
75 hp already meet the new PM standards (i.e., such engines will not have to make any changes
nor incur any incremental hardware costs for 2008), which also contributes to the likely
overestimate of costs.  Our cost estimates for DOCs are presented above in Section 6.2.2.4.

   As discussed in Chapter 4, we have also projected that some engines in the 25 to 75 hp range
will have to make  changes to their engines to incorporate more conventional engine technology,
such as electronic  common rail fuel injection, to meet the demands of the newly added CDPF.
These costs were assumed for direct-injection (DI) engines. For indirect-injection (IDI) engines
in this power range, we believe manufacturers will comply not through a fuel system upgrade to
electronic common rail, but through the addition of a CDPF regeneration system to ensure
regeneration of the CDPF.  The costs for CDPF regeneration systems are discussed above in
Section 6.2.2.3.

   In the 25 to 50 hp range, we believe all engines will add cooled EGR to meet NOx standards.
For our cost analysis, this is also true for engines over 750 hp. Note that engines over 750 hp are
also assumed to add the previously discussed emission-control technologies, i.e., a CDPF system
and some sort of CDPF regeneration system.

   We project that manufacturers will add CCV systems to all these engines that are
turbocharged, both large and small. The costs for CCV systems are presented in Section 6.2.2.5.
                                         6-43

-------
Final Regulatory Impact Analysis
    >.2.2.6.1 Electronic Common Rail Fuel-Injection System Costs for DI Engines

   Cost estimates for fuel-injection systems were developed by ICF Consulting under contract
to EPA.36  Table 6.2-22 presents the costs to manufacturers as estimated by ICF for fuel-injection
systems.
                                      Table 6.2-22
                      Fuel-Injection System - Costs to Manufacturers


Horsepower
Displacement (L)
# of Cylinders/Injectors
Type of Fuel System
High Pressure Fuel Pump
Fuel Injectors (each)
Cost for Injectors (total)
Fuel Rail
Computer
Sensors, Wiring, Bearings, etc.
Total Fuel System Cost
Incremental Cost
Fuel System Costs ($2002)
Baseline System
20 hp
1
2
Mech
$340
$16
$32


$68
$440

35 hp 80 hp
2 3
3 4
Mech ER
$340 $350
$16 $25
$48 $100

$300
$82 $189
$470 $939

New System
20 hp 35 hp
1 2
2 3
ECR ECR
$340 $340
$80 $80
$160 $240
$100 $100
$280 $280
$231 $625
$1,111 $1,205
$671 $735

80 hp
3
4
ECR
$350
$80
$320
$100
$280
$639
$1,309
$370
   Mech=Mechanical Fuel Injection; ER=Electronic Rotary Injection; ECR=Electronic Common Rail Injection
   Note that engines in the 50 to 75 hp range (represented in Table 6.2-22 by the 80 hp engine)
are assumed to have electronic rotary fuel-injection systems as a baseline configuration while
smaller engines are assumed to have mechanical fuel injection (see section II. A of the preamble
and section 4.1 of the RIA for more discussion on why this is a valid assumption). On an
incremental basis, the costs for common rail fuel injection are much lower when working from
an electronic rotary baseline because the electronic fuel pump and the computer are already part
of the system.  This explains the large difference in fuel system costs for the 80 hp engine
relative to the 20 and 35 hp engines.

   The costs shown in Table 6.2-22 show consistency for all elements across the power range.
This is because most of the cost elements - fuel pump, costs per injector, and a computer - have
little to no relation to engine size or engine displacement. The primary cost element that
changes for each of the example engines shown is that for the total cost of injectors.  For this
reason, the costs can be more easily understood by separating the per injector cost out from the
rest of the system. This was done for the costs shown in Table 6.2-23, which  also builds on the
manufacturer costs shown in Table 6.2-22 to generate costs to the user in the same manner as
done for other hardware system costs, as discussed above. We have broken out the fuel system
costs this way to make possible a cost equation that applies to all engines. Unlike the other cost
equations we have generated, the cost equation for fuel systems uses the number of injectors

                                          6-44

-------
                                              Estimated Engine and Equipment Costs
(i.e., the number of cylinders) as the dependent variable rather than using engine displacement.
This equation is presented below in Section 6.2.2.6.3.

                                       Table 6.2-23
                     Incremental Fuel System Costs - Costs to the User
EPA Estimated Incremental Fuel System Costs for Dl Engines ($2002)
Horsepower
Number of Cylinders (# of injectors)

Cost to Manufacturer
Warranty Cost - Near Term (3% claim rate)
Mfr. Carrying Cost (4%) -- Near Term
Total Cost to Dealer - Near Term
Dealer Carrying Cost (3%) -- Near Term
Total Cost to Buyer -- Near Term
Warranty Cost - Long Term (1 % claim rate)
Mfr. Carrying Cost (4%)- Long Term
Total Cost to Dealer -- Long Term
Dealer Carrying Cost (3%) -- Long Term
Subtotal
Total Cost to Buyer -- Long-Term w/ learning
20
2
per Injector Remaining System
$65 $551
$8 $44
$3 $22
$75 $617
$2 $19
$78 $636
$3 $15
$3 $22
$70 $588
$2 $18
$72 $605
$58 $484
35
3
perlniector Remaining System
$65 $551
$8 $44
$3 $22
$75 $617
$2 $19
$78 $636
$3 $15
$3 $22
$70 $588
$2 $18
$72 $605
$58 $484
80
4
per Injector Remaining System
$56 $152
$7 $14
$2 $6
$65 $173
$2 $5
$67 $178
$2 $5
$2 $6
$60 $163
$2 $5
$62 $168
$50 $134
Remaining System includes the fuel pump, fuel rail, computer, wiring, and necessary sensors.
   Note that these costs are projected to be incurred only on 25 to 75 hp DI engines.  Note also
that, in determining aggregate variable costs for fuel-injection systems, we have attributed half
of the costs to the Tier 4 standards. We have done this for two reasons: penetration of electronic
fuel systems into the market and user benefits associated with the new fuel systems.  First, we
are projecting that, by 2008, some engines in the 25 to 75 hp range will already be equipped with
electronic fuel systems independent of this rule. This is due to the natural progression of
electronic fuel systems currently available in larger power engines into some of the smaller
power engines. In fact, recent certification data prove that this is already happening, as
discussed in section 4.1.4 of this RIA. During our discussions with some engine companies, they
have indicated that they intend to use electronic fuel  system technologies to comply with the
existing Tier 3 standards in the 50 to  100 hp range.  These manufacturers have informed us that
these electronic fuel systems will also be sold on engines in the 25 to 50 hp range for those
engine product lines built on the same platform as engines over 50 hp. In addition, there are end-
user benefits associated with electronic fuel systems, such as better torque response, lower noise,
easier servicing via on-board diagnostics, and better engine startability. Because we are not able
to predict the precise level of penetration of electronic fuel  systems, nor are we able to quantify
the monetary value of the end-user benefits, we have accounted for these two effects by
attributing half of the costs of the electronic fuel systems to the Tier 4 standards.

   6.2.2.6.2 Cooled EGR System Costs

   Cost estimates for cooled EGR systems were developed by ICF Consulting under contract to
EPA.37  The incremental manufacturer costs for cooled EGR systems are shown in Table 6.2-24.
                                           6-45

-------
Final Regulatory Impact Analysis
                                      Table 6.2-24
                       Cooled EGR System - Costs to Manufacturers
ICF Estimated Cooled EGR System Costs to Manufacturers ($2002)
Horsepower
Displacement (L)
EGR Cooler
EGR Bypass
Electronic EGR Valve
EGR Total Cost to Manufacturer
20
1
$36
$15
$14
$66
35
2
$63
$16
$15
$95
1000
24
$289
$30
$88
$413
   Building on these manufacturer costs, we estimated the costs to the user assuming the
warranty claim rates and learning effects already discussed. These results are shown in Table
6.2-25.  Included in these costs are costs associated with additional cooling that may be needed
to reject the heat generated by the cooled EGR system or other in-cylinder technologies. These
costs were not included in the proposal.  Such additional cooling might take the form of a larger
radiator and/or a larger or more powerful cooling fan. Based on cost estimates from our
Nonconformance Penalty rule (67 FR 51464). In the support document for the NCP rule,38 we
estimated the costs associated with such additional cooling at $130 for a light heavy-duty vehicle
(~200hp) and $300 for a heavy heavy-duty vehicle (~500hp), inclusive of vehicle manufacturer
mark ups. Here, we have used these values to generate a curve with horsepower as the
dependent variable. That curve is $0.60 + $16.7(x), with an R2=l  and where "x" represents
horsepower. Using this curve and the horsepowers shown in Table 6.2-25 we were able to
estimate the costs for additional cooling. The results shown in Table 6.2-25 include a three
percent dealer carrying cost.
                                          6-46

-------
                                             Estimated Engine and Equipment Costs
                                      Table 6.2-25
                         Cooled EGR System - Costs to the User
EPA Estimated Cooled EGR Costs ($2002)
Horsepower
Displacement (L)
EGR System Cost to Manufacturer
Warranty Cost — Near Term (3% claim rate)
Mfr. Carrying Cost (4%) - Near Term
Total Cost to Dealer — Near Term
Dealer Carrying Cost (3%) - Near Term
EGR System Cost to Buyer — Near Term
Warranty Cost — Long Term (1 % claim rate)
Mfr. Carrying Cost (4%)- Long Term
Total Cost to Dealer — Long Term
Dealer Carrying Cost (3%) - Long Term
Subtotal
EGR System Cost to Buyer - Long Term w/ learning
Heat rejection cost to Buyer (incl 3% dealer carrying cost) - Near Term
Heat rejection cost to Buyer (incl 3% dealer carrying cost) - Long Term
Total EGR-related Costs to Buyer - Near-term
Total EGR-related Costs to Buyer — Long-term
20
1
$66
$8
$3
$77
$2
$79
$3
$3
$71
$2
$73
$59
$29
$23
$108
$82
35
2
$95
$10
$4
$109
$3
$112
$3
$4
$102
$3
$105
$84
$38
$31
$151
$115
1000
24
$413
$34
$17
$463
$14
$477
$11
$17
$441
$13
$454
$363
$610
$488
$1 ,087
$851
   Despite the presence of cost data for a 20hp engine in Table 6.2-25, we are projecting that
only engines in the 25 to 50 hp range (in 2013) and engines over 750 hp will need to add cooled
EGR (in 2011), or use some other equally effective approach having presumably similar costs, to
comply with the new engine standards. All the costs associated with these systems have been
attributed to compliance with the new emission standards (i.e., we have not attributed any costs
to user benefits).

   6.2.2.6.3 Conventional Technology Cost Estimation Functions

   In the same manner as already described for exhaust emission-control devices, we were able
to calculate cost equations for cooled EGR systems (inclusive of additional cooling). For fuel
systems, rather than  a linear regression, we simply expressed the fuel system costs as a function
of the number of fuel injectors, and then added on the costs associated with the rest of the
system. The rest of the system includes the fuel pump, the computer, wiring and sensors, which
should not change relative to  engine size or displacement.  This way, the functions could be
applied to the wide array of engines in the nonroad fleet to determine the total costs or per unit
costs for this hardware. The cost estimation functions for these technologies are shown in Table
6.2-26.
                                         6-47

-------
Final Regulatory Impact Analysis
                                      Table 6.2-26
                         Costs for Conventional Technologies as a
           Function of the Indicated Parameter (x represents the dependent variable)
                                         $2002
Technology
Fuel System Costs - DI Only
Near Term
Long Term
Near Term
Long Term
Cooled EGR System (inclusive of
additional cooling)
Near Term
Long Term
Applicable Hp Range
25750hp
25750hp
Dependent
Variable
# of cylinders
displacement
Equation
$78(x) + $636
$58(x) + $484
$67(x)+$178
$50(x)+$134
$43(x) + $65
$33(x)+$48
R2
a
a
1
1
   "Not applicable because a linear regression was not used.
   6.2.2.7 Summary of Engine Variable Cost Equations

   Engine variable costs are discussed in detail in Sections 6.2.2.1 through 6.2.2.6.  For engine
variable costs, we have generated cost estimation equations as a function of engine displacement
or number of cylinders. These equations are summarized in Table 6.2-27'.  Note that not all
equations were used for all engines; equations were used in the manner shown in Table 6.2-27'.
We have calculated the aggregate engine variable costs and present them later in this chapter and
in Chapter 8.
                                          6-48

-------
                                              Estimated Engine and Equipment Costs
                                      Table 6.2-27
                              Summary of Cost Equations for
                 Engine Variable Costs (x represents the dependent variable)
Engine Technology
NOx Adsorber System
CDPF System
CDPF Regen System -
IDI engines
CDPF Regen System -
DI engines
DOC
CCV System
Cooled EGR System
w/ additional cooling
Common Rail Fuel
Injection
(mechanical fuel
system baseline)
Common Rail Fuel
Injection
(electronic rotary fuel
system baseline)
Time Framea
Near term
Long term
Near term
Long term
Near term
Long term
Near term
Long term
Near term
Long term
Near term
Long term
Near term
Long term
Near term
Long term
Near term
Long term
Cost Equation
$103(x) + $183
$83(x) + $160
$146(x) + $75
$112(x) + $57
$20(x) + $293
$16(x) + $223
$10(x) + $147
$8(x) + $lll
$18(x) + $116
$18(x) + $110
$2(x) + $34
$2(x) + $24
$43 (x) + $65
$33(x) + $48
$78(x) + $636
$58(x) + $484
$67(x) + $178
$50(x) + $134
Dependent
Variable (x)
Displacement1"
Displacement
Displacement
Displacement
Displacement
Displacement
Displacement
# of cylinders/
injectors
# of cylinders/
injectors
How Used
>75 hp engines according to
phase-in of NRT4 NOx std.
>25 hp engines according to
NRT4 PM std.
IDI engines adding a CDPF
DI engines adding a CDPF
<25 hp engines beginning in
2008;
25-75 hp engines 2008 thru 2012
All turbocharged engines when
they first meet a Tier 4 PM std.
25-50 hp engines beginning in
2013;
>750hp engines beginning in
2011
25-50 hp DI engines when they
add a CDPF
50-75 hp DI engines when they
add a CDPF
 a Near term = years 1 and 2; Long term = years 3+. Explanation of near term and long term is in Section 6.1.
 b Displacement refers to engine displacement in liters.
6.2.3 Engine Operating Costs

   We are projecting that a variety of new technologies will be introduced to enable nonroad
engines to meet the Tier 4 emission standards. Primary among these are advanced emission-
control technologies and low-sulfur diesel fuel. The technology enabling benefits of low-sulfur
diesel fuel are described in Chapter 4. The incremental cost for low-sulfur fuel is described in
Chapter 7 and is not presented here.  The new emission-control technologies are themselves
expected to introduce additional operating costs in the form of increased fuel consumption and
increased maintenance demands. Operating costs are estimated over the life of the engine and
                                          6-49

-------
Final Regulatory Impact Analysis
are expressed in terms of cents/gallon of fuel consumed. In Section 6.5 we present these lifetime
operating costs as a net present value (NPV) in 2002 dollars for several example pieces of
equipment.

   A note of clarification needs to be added here. In Chapter 8 we present aggregate operating
costs. Every effort is made to be clear what costs are related to (1) the incremental increase in
the cost of fuel (due to the lower sulfur level), and (2) what costs are related to the expected
change in maintenance demands and the expected change in fuel consumption.  The operating
costs discussed in this section are only the latter—maintenance related costs and/or savings and
fuel consumption costs.  Increased costs associated with the lowering of sulfur in nonroad diesel
fuel are discussed in detail in Chapter 7. The cent-per-gallon costs presented in Chapter 7, along
with the cent-per-gallon costs and savings presented here, are then combined with projected fuel
volumes to generate the aggregate costs of the fuel program in this final rule.

   Total operating costs include the following elements: the change in maintenance costs
associated with applying new emission controls to the engines; the change in maintenance costs
associated with low-sulfur fuel such as extended oil-change intervals (extended oil change
intervals results in maintenance savings); the change in fuel costs associated with the
incrementally higher costs for low-sulfur fuel (see Chapter 7), and the change in fuel costs due to
any fuel consumption impacts associated with applying new emission controls to the engines.
This latter cost is attributed to the CDPF and its need for periodic regeneration, which we
estimate may result in a small increase in fuel consumption, as discussed in more detail below.
Maintenance costs associated with the new emission controls on the  engines are expected to
increase, since these devices represent new hardware and therefore new maintenance demands.
Offsetting this cost increase will be a cost savings due to an expected increase in oil-change
intervals, because low-sulfur fuel is far less corrosive than current nonroad diesel fuel.  Less
corrosion corresponds with a slower acidification rate (i.e., less degradation) of the engine
lubricating  oil and therefore more operating hours between oil changes.

   6.2.3.1 Operating Costs Associated with Oil-Change Maintenance for New and Existing
   Engines

   We  estimate that reducing fuel sulfur to 500 ppm will reduce engine wear and oil degradation
to the existing fleet of nonroad diesel engines, as well as locomotive and  marine diesel engines.
Reducing fuel sulfur to 15 ppm will further reduce engine wear and oil degradation. These
improvements provide a savings to users of this equipment. The cost savings will also be
realized by the owners of future nonroad engines that are subject to the emission standards in this
final rule.  As discussed below, these maintenance savings have been estimated to be greater
than 3 cents/gallon when comparing current uncontrolled fuel to 15 ppm  sulfur fuel.

   We  have identified a variety of benefits from the low-sulfur diesel fuel. These benefits are
summarized in Table 6.2-28.
                                          6-50

-------
                                             Estimated Engine and Equipment Costs
                                      Table 6.2-28.
        Engine Components Potentially Affected by Lower Sulfur Levels in Diesel Fuel
Affected Components
Piston Rings
Cylinder Liners
Oil Quality
Exhaust System
(tailpipe)
Exhaust Gas
Recirculation System
Effect of Lower Sulfur
Reduced corrosion wear
Reduced corrosion wear
Reduced deposits, reduced
acid build-up, and less need
for alkaline additives
Reduced corrosion wear
Reduced corrosion wear
Potential Impact on Engine System
Extended engine life and less frequent
rebuilds
Extended engine life and less frequent
rebuilds
Reduce wear on piston ring and
cylinder liner and less frequent oil
changes
Less frequent part replacement
Less frequent part replacement
   The monetary value of these benefits over the life of the equipment will depend upon the
length of time that the equipment operates on low-sulfur diesel fuel and the degree to which
engine and equipment manufacturers specify new maintenance practices and the degree to which
equipment operators change engine maintenance patterns to take advantage of these benefits.
For equipment near the end of its life in the 2008 time frame, the benefits will be quite small.
However,  for equipment produced in the years immediately preceding the introduction of 500
ppm sulfur fuel, the savings will be substantial.  Additional savings will be realized in 2010 with
the introduction of 15 ppm sulfur fuel.

   We estimate the single largest savings will be the impact of lower sulfur fuel on oil-change
intervals.  We have estimated the extension of oil-change intervals realized by 500 ppm sulfur
fuel in 2007 and the additional extension resulting from 15 ppm sulfur fuel in 2010. These
estimates are based on our analysis of publically available information from nonroad engine
manufacturers. Due to the wide range of diesel fuel  sulfur levels that nonroad engines may
currently see around the world, engine manufacturers specify different oil-change intervals as a
function of diesel sulfur levels. We have used these  data as the basis for our analysis. Taken
together, when compared with the relatively high sulfur levels in current nonroad diesel fuel, we
estimate the use of 500 ppm sulfur fuel will enable an oil-change interval extension of 31
percent, while 15 ppm sulfur fuel will enable an oil-change interval extension of 35 percent
relative to current products.39

   We present here a fuel cost savings attributed to the oil-change interval extension in terms of
a cent-per-gallon operating cost. Table 6.2-29 shows the calculation of cent-per-gallon savings
for various power segments of the nonroad fleet, and the locomotive and marine segments, for
both the 500 ppm fuel and the 15 ppm fuel.  The brake specific fuel consumption (BSFC),
average hp, average activity, and average load factor data shown in the table are from our
nonroad model.40 The existing and new NRLM fleets will realize the savings associated with the
                                          6-51

-------
Final Regulatory Impact Analysis
500 ppm fuel for the years 2007 through 2010, and the the savings associated with the 15 ppm
fuel program for the years 2010 and beyond. We estimate that an oil-change interval extension of
31 percent enabled by 500 ppm sulfur fuel results in a weighted savings in fuel operating costs of
2.9 cents/gallon for the nonroad fleet. We project an additional weighted cost savings of 0.3
cents/gallon for the oil-change interval extension enabled by 15 ppm sulfur. Note that the
weighted savings are  determined using the fuel use weightings shown in Table 6.2-29.  For
locomotive and marine engines, these savings are 1 cent/gallon and 0.1 cent/gallon for the 500
ppm step and the 15 ppm step, respectively.

   Thus, for the nonroad fleet as a whole, beginning in 2010, nonroad equipment users can
realize an operating cost savings of 3.2 cents/gallon relative to current engines. For a typical  100
hp nonroad engine, this represents a net present value lifetime savings of more than $500. For
locomotive and marine engines the savings are estimated at 1.1 cents/gallon, which represents a
net present value lifetime savings of more than $2000.
                                          6-52

-------
            Table 6.2-29. Oil-Change Maintenance Savings for Existing and New Nonroad, Locomotive, and Marine Engines ($2002)
Oil Change Savings due to Low S
Rated Power
BSFC
Fuel Density
Population Weighted Avg. Horsepower
Population Weighted Avg. Activity
Population Weighted avg. Load Factor
Sump Oil Capacity
Base Oil Change Interval - 3000 ppm S
Control Oil Change Interval - 500 ppm S
Labor Cost Per Oil Change
Cost of Oil Per Oil Change
Cost of Oil Filter Per Oil Change
Total Cost Per Oil Change
Fuel Consumption in 3000 ppm Oil Interval
Fuel Consumption in 500 ppm Oil Interval
Oil Change Cost/Gallon fuel in 3000 ppm Interval
Oil Change Cost/Gallon fuel 500 ppm Interval
Cost Differential - 3000 to 500 ppm S

Control Oil Change Interval - 15 ppm S
Labor Cost Per Oil Change
Cost of Oil Per Oil Change
Cost of Oil Filter Per Oil Change
Total Cost Per Oil Change
Fuel Consumption in 500 ppm Oil Interval
Fuel Consumption in 15 ppm Oil Interval
Oil Change Cost/Gallon fuel in 500 ppm Interval
Oil Change Cost/Gallon fuel in 15 ppm Interval
Cost Differential - 500 to 15 ppm S
Cost Differential - 3000 to 15 ppm S
Fuel Use Weightings
Units
hp
Ibm/hp-hr
Ibm/gallon
hp
hrs/year
% full load
L
hrs
hrs
$
$
$
$
gallons
gallons
$/gallon
$/gallon
$/gallon

hrs

$
$
$
gallons
gallons
$/gallon
$/gallon
$/gallon
$/gallon
% total
Nonroad Engines
0750hp
0.367
7.1
1282
1130
0.57
124.32
250
327.5
$100.00
$248.65
$70.00
$418.65
9463
12396
$0.04
$0.03
$0.010

337.5
$100.00
$248.65
$70.00
$418.65
12396
12774
$0.03
$0.03
$0.001
$0.011
6.2%
Locomotive

0.367
7.1
1282
1130
0.57
124.32
250
327.5
$100.00
$248.65
$70.00
$418.65
9463
12396
$0.04
$0.03
$0.010

337.5
$100.00
$248.65
$70.00
$418.65
12396
12774
$0.03
$0.03
$0.001
$0.011

Marine

0.367
7.1
1282
1130
0.57
124.32
250
327.5
$100.00
$248.65
$70.00
$418.65
9463
12396
$0.04
$0.03
$0.010

337.5
$100.00
$248.65
$70.00
$418.65
12396
12774
$0.03
$0.03
$0.001
$0.011

(1) Oil-change intervals are from William Charmley memo to docket.41
(2) Labor costs are from IGF Consulting under contract to EPA.42
(3) Oil use estimates are based on sump volumes scaled to engine displacement and, as such,
overstate the cost for some engines while understating the costs for others.
they show differences for each power category. The labor and filter costs are average values over a broad power range and, as such, may

-------
Final Regulatory Impact Analysis
   The savings shown in Table 6.2-29 will occur without additional new cost to the equipment
owner beyond the incremental cost of the low-sulfur diesel fuel although these savings are
dependent on changes to existing maintenance schedules.  Such changes seem likely given the
magnitude of the potential savings. We have not estimated the value of the savings from the
other benefits listed in Table 6.2-28.  Therefore, we believe the 3.2 cents/gallon savings
underestimates actual cost savings as it accounts only for the impact of low-sulfur fuel on oil-
change intervals.

   Operating costs (savings) associated with oil-change maintenance are split evenly  between
NOx and PM control.

   6.2.3.2 Operating Costs Associated with CDPF Maintenance for New CDPF-Equipped
   Engines

   The maintenance demands associated with the addition of new CDPF hardware are discussed
in Section 4.1.1.3.4. To avoid underestimating costs, we have used a maintenance interval of
3,000 hours for engines under 175 hp and 4,500 hours for engines over 175 hp, both of which are
the minimum allowable maintenance intervals specified in our regulations (i.e., manufacturers
are precluded by regulation from requiring more frequent maintenance, and we believe they may
require less frequent maintenance than these minimum allowable maintenance intervals). We
have estimated costs associated with the maintenance at $65 for engines up to 600 hp and $260
per event for engines over 600 hp. The calculations for CDPF maintenance are shown in Table
6.2-30.  Weighting the savings in each power range by the fuel-use weightings shown in the
table, we can calculate the fleet weighted maintenance costs as 0.6 cents/gallon, which will be
incurred only by new engines equipped with a CDPF.  Operating costs associated with CDPF
maintenance are attributed entirely to PM control.

                                      Table 6.2-30
             CDPF Maintenance Costs for New CDPF-Equipped Engines ($2002)
|PM Filter Maintenance Costs
Units
                                                           Nonroad Engines
Rated Power
BSFC
=uel Density
Copulation Weighted Avg. Horsepower
Copulation Weighted Avg. Activity
Copulation Weighted avg. Load Factor
=ilter Maintenance Interval
=ilter Maintenance Cost Materials
=ilter Maintenance Labor
Total Filter Maintenance Cost per event
=uel Use Between Maintenance Interval
Maintenance Cost
=uel Use Weightings
hp
Ibm/hp-hr
Ibm/gallon
hp
hrs/year
% full load
hours
$/event
$/event
$/event
gallons/period
$/gallon
% total
0750hp
0.367
7.1
1282
1130
0.571
4,500
$0
$260
$260
170,326
$0.002
6.2%
Labor costs are from IGF Consulting under contract to EPA.
                                          6-54

-------
                                             Estimated Engine and Equipment Costs
   6.2.3.3 Operating Costs Associated with Fuel Economy Impacts on New Engines

   6.2.3.3.1 What Are the Estimated Fuel Economy Impacts ?

   The high efficiency emission-control technologies expected to be applied to meet the  PM
standards for engines greater than 25 hp and the NOx standards for engines greater than 75 hp
involve wholly new system components integrated into engine designs and calibrations and, as
such, may be expected to change the fuel consumption characteristics of the overall engine
design. After reviewing the likely technology options available to the engine manufacturers, we
believe the integration of the engine and exhaust emission-control systems into a single
synergistic emission-control system will lead to nonroad engines that can meet demanding
emission-control targets with only a small impact on fuel consumption. Technology
improvements have historically eliminated these marginal impacts in the past and it is our
expectation that this kind of continuing improvement will eliminate the modest impact estimated
here. However, because we cannot project the time frame for this improvement to be realized,
we have included this impact in our cost estimates for the full period of the program to avoid
underestimating costs.

       6.2.3.3.1.1 CDPF Systems and Fuel Economy

   Diesel particulate filters are anticipated to provide a step-wise decrease in diesel particulate
(PM) emissions by trapping and oxidizing the diesel PM.  The trapping of the very fine diesel
PM is accomplished by forcing the exhaust through a porous filtering media with extremely
small openings and long path lengths.p  This approach results in filtering efficiencies for diesel
PM greater than 90 percent but requires additional pumping work to force the exhaust through
these small openings.  The impact of this additional pumping work on fuel consumption is
dependent on engine operating conditions. At low exhaust flow conditions (i.e., low engine
load, low turbocharger boost levels), the impact is so small that it typically cannot be measured,
while at very high load conditions, with high exhaust flow conditions, the fuel economy impact
can be as large as one to two percent.44'45 We have estimated that the average impact of this
increased pumping work will be equivalent to an increase fuel consumption of approximately
one percent.46

   Under conditions  typical of much of nonroad engine operation, the soot stored in the PM
filter will be regenerated passively using the heat of the exhaust gas promoted by catalyst
materials. We have performed an analysis of the expected exhaust temperatures for several
typical in-use operating cycles, as described in Section 4.1.3.  That analysis shows that for a
many nonroad engines passive regeneration can be expected.  Under some conditions, including
very low ambient temperatures, or extended low load operation, the exhaust temperature of the
engine may not be hot enough to ensure complete passive regeneration. We believe some
manufacturers will address this situation by employing active backup regeneration systems that
   p
     Typically, the filtering media is a porous ceramic monolith or a metallic fiber mesh. We refer to it as a "filter
trap" in Table 6.2-13.

                                          6-55

-------
Final Regulatory Impact Analysis
provide supplemental heat to initiate regeneration, as discussed in Section 4.1. Also, as
explained in Section 6.2.2.3, we are conservatively costing active regeneration systems for all
engines using a CDPF system.  We have done this because we think it is unlikely that nonroad
engine manufacturers will be able to accurately predict which engines will be operated in a
manner conducive to passive regeneration and which engines will require periodic active
regeneration. There will be no fuel economy impact for nonroad engines that have an active
regeneration technology but experience passive regeneration in use. Examples of current active
PM filter systems that do not benefit from low-sulfur diesel fuel, nor catalytic coatings to
promote regeneration, require additional fuel supplementation of approximately two percent for
active filter regeneration.47 Given the new requirements for clean diesel fuel in this final rule,
the ability to use catalytic coatings to promote soot oxidation, and the fact that many kinds of
nonroad equipment are expected to operate in a way that passive regeneration will occur, we
believe the average fuel economy impact of the backup regeneration systems will be no more
than one percent.

   We have projected that engines between 25 hp to 75 hp will comply with the PM standard of
0.02 g/bhp-hr using a CDPF system including a backup active regeneration system. The NOx
control systems expected in this power category are not advanced catalyst-based systems and, as
such, have limited ability to recover fuel economy through timing advance or other in-cylinder
NOx control strategies, as discussed below.  We therefore project that a two percent fuel
economy impact (i.e., one percent due to backpressure and one percent due to use of backup
regeneration systems) will occur for engines between 25 hp and 75 hp. We believe
manufacturers will overcome this impact in the long term through continuing technology
refinement, as has historically happened. However, to avoid underestimating costs, we have
included this two percent impact for the duration of the program.

   For engines under 25 hp we have projected no need to use CDPF technologies to comply
with the PM standards in the final rule.  We therefore estimate no fuel consumption impact from
the CDPF for this category.

   We believe engines all engines between 75 hp and 750 hp and mobile gensets above 750hp
will use integrated NOx and PM control technologies to comply with the new emission
standards. The advanced catalyst-based emission-control technology that we project industry
will use to meet the new NOx standard offers the  opportunity to improve fuel economy, as
described in the following section.  Based on those projected improvements, we have estimated a
net impact on fuel consumption of one percent for engines between 75 and 750 hp as well as
gensets >750 hp with CDPF technology and NOx technology. Future technology improvements
are likely to recover this fuel consumption impact; however, to avoid underestimating costs, we
have assumed that a one-percent fuel consumption impact persists for the duration of the
emission-control program.

   At  this time we are not setting a NOx standard for nonroad mobile machine engines >750 hp
based on the use  of advanced NOx catalyst based technologies (see Preamble Section II. A).
These engines, like the smaller engines between 25 and 75 hp, are projected to use diesel
particulate filter technologies to meet the Tier 4 PM standards.  Therefore like the 25 to 75 hp

                                          6-56

-------
                                             Estimated Engine and Equipment Costs
engines, we are estimating that nonroad mobile machines above 750 hp will have a two percent
fuel economy impact (i.e., one percent due to backpressure and one percent due to use of backup
regeneration systems). We believe manufacturers will overcome this impact in the long term
through continuing technology refinement, as has historically happened. However, to avoid
underestimating costs, we have included this two percent impact for the duration of the program.

       6.2.3.3.1.2 NOx Control and Fuel Economy

   NOx adsorbers are expected to be the primary technology to reduce NOx emissions for
engines between 75 and 750 hp as well as for mobile gensets above 750 hp. NOx adsorbers
work by storing NOx emissions under fuel-lean operating conditions (normal diesel engine
operating conditions) and then by releasing and reducing the stored NOx emissions over a brief
period of fuel-rich engine operation. This brief periodic NOx release and reduction step is
directly analogous to the catalytic reduction of NOx over a gasoline three-way catalyst.  For this
catalyst function to occur, the engine exhaust constituents and conditions must be similar to
normal gasoline exhaust constituents. That is, the exhaust must be fuel rich (devoid of excess
oxygen) and hot (over 250°C).  Although it is anticipated that nonroad diesel engines, like
highway diesel engines, can be made to operate in this way, it is anticipated that fuel economy
during operation under these conditions will be worse than normal. This increase in fuel
consumption can be minimized by carefully controlling engine air-fuel ratios using the control
systems we anticipate will be used to meet the Tier 3 emission standards.  The lower the engine
air-fuel ratio, the lower the amount of fuel that must be added to reach rich conditions. In the
ideal case where the engine air-fuel  ratio is at the stoichiometric level and additional fuel is
required only as a NOx reductant, the fuel economy penalty is nearly zero. We are projecting
that practical limitations on controlling engine air-fuel ratio will mean that the NOx adsorber
release and reduction cycles will lead to a one percent decrease in the engine fuel economy.48
We estimate that this fuel economy  impact can be regained through optimization of the engine-
PM trap-NOx adsorber system, as discussed below.

   In addition to the NOx release and regeneration event, another step in NOx adsorber
operation may affect fuel  economy.  As discussed earlier, sulfur affects NOx adsorbers even at
the low fuel-sulfur levels we are adopting. As discussed in Chapter 4, this effect can (and must)
be reversed through a periodic "desulfation"  event.  The desulfation of the NOx adsorber is
accomplished in a similar manner to the NOx release and regeneration cycle described above.
However, it is anticipated that the desulfation event will require extended operation of the diesel
engine at rich conditions.49 This rich operation will, like the NOx regeneration event, require an
increase in the fuel consumption rate and will cause an associated decrease in fuel economy.
This loss in fuel consumption is directly proportional to the amount of sulfur in diesel fuel.  The
frequency of desulfation is therefore a function of the fuel sulfur level  and the fuel consumption
rate. Since the desulfation frequency and the associated fuel consumption impacts are
proportional only to fuel rate and to fuel sulfur levels, the projected fuel consumption impacts at
15 ppm sulfur are the same for both highway and nonroad diesel engines. With a 15 ppm fuel
sulfur cap, we are projecting that fuel consumption for desulfation will increase by no more than
one percent, which we believe can be regained through optimization of the engine-CDPF-NOx
adsorber system, as discussed below.

                                          6-57

-------
Final Regulatory Impact Analysis
   While NOx adsorbers impact fuel economy by requiring nonpower-producing fuel
consumption to function properly, they are not unique among NOx control technologies in this
way.  In fact, NOx adsorbers are likely to have a very favorable tradeoff between NOx emissions
and fuel economy compared with our projected technologies for meeting Tier 3 NOx
standards—cooled EGR and injection timing retard.  EGR requires the delivery of exhaust gas
from the exhaust manifold to the intake manifold of the engine and causes a decrease in fuel
economy for two reasons.  The first of these reasons is  that a certain amount of work is required
to pump the EGR from the exhaust manifold to the intake manifold; this necessitates the use of
intake throttling or some other means to accomplish this pumping.  The second of these reasons
is that heat in the exhaust, which is normally partially recovered as work across the turbine of the
turbocharger, is instead lost to the engine coolant through the cooled EGR heat exchanger. In
the end, cooled EGR is approximately 50 percent effective at reducing NOx below the current
Tier 2 NOx levels. Injection  timing retard is another strategy that can be employed to control
NOx emissions. By retarding the introduction of fuel into the engine,  and thus delaying the start
of combustion, both the peak temperature and pressure of the combustion event are decreased;
this lowers NOx formation rates and, ultimately, NOx emissions. Unfortunately, this also
significantly decreases the thermal efficiency of the engine (lowers fuel economy) while also
increasing PM emissions. As an example, retarding injection timing eight degrees can decrease
NOx emissions by 45 percent, but this occurs at a fuel economy penalty of more than seven
percent.50

   Nonroad diesel engines generally rely primarily on charge-air-cooling and injection timing
control (retarding injection timing) to meet Tier 2 NOx+NMHC emission standards. For Tier 3
compliance, we expect that engine manufacturers will use a combination of cooled EGR and
injection timing control to meet the NOx standard. Because of the more favorable fuel economy
trade-off for NOx control with EGR compared with timing control, we forecast that less reliance
on timing control will be needed for Tier 3  than for Tier 2. Fuel economy will therefore not
change even at this lower NOx level.  Similarly for the 25-50 hp engines subject to  a Tier 4 NOx
standard of 3.3 g/hp-hr, we believe the NOx standard will not cause a change in fuel
consumption.  NOx adsorbers have a significantly  more favorable trade-off between NOx
emissions and fuel economy compared with cooled EGR or timing  retard.51 We expect NOx
adsorbers to be able to accomplish a greater than 90 percent reduction in NOx emissions, while
themselves consuming significantly less fuel than that lost through  alternative NOx control
strategies such as retarded injection timing.Q We therefore expect manufacturers to take full
advantage of the NOx control capabilities of the NOx adsorber and project that they will
decrease reliance on the more expensive  (from a fuel economy standpoint) technologies,
especially injection timing retard.  We therefore predict that the fuel economy impact currently
associated with NOx control from timing retard will be decreased by at least three percent. In
other words, through the application of advanced NOx emission-control technologies, which are
     We have estimated the fuel consumption rate for NOx regeneration and desulfation of the NOx adsorber as
approximately 2 percent of total engine fuel consumption. This differs from an EPA contractor report by EF&EE
estimating the total consumption to be approximately 2.5 percent of total fuel consumption.  Additionally the contractor's
estimate of NOx adsorber efficiency ranges from 80 to 90 percent, while we believe over 90 percent control is possible,
as discussed in Chapter 4.

                                           6-58

-------
                                            Estimated Engine and Equipment Costs
enabled by the use of low-sulfur diesel fuel, we expect the NOx trade-off with fuel economy to
continue to improve significantly when compared with current technologies.  This will result in
much lower NOx emissions and potentially overall improvements in fuel economy.
Improvements could easily offset the fuel consumption of the NOx adsorber itself and, in
addition, at least half of the fuel economy impact projected to result from the application of the
CDPF technology.  Consequently, we are projecting a one percent fuel economy impact to result
from this rule for engines between 75 and 750 hp as well as mobile gensets above 750 hp.

       6.2.3.3.1.3 Fuel Economy Impacts for Engines without Advanced Emission-Control
       Technologies (engines under 25 hp)

   The new NOx emission standard for engines under 25 hp is unchanged from the current Tier
2 level. The PM standard, however, decreases by almost 50 percent.  We believe manufacturers
will achieve this significant PM reduction through improvements in combustion system design,
improvements in fuel system design and utilization, and through the use of diesel oxidation
catalysts (DOCs).  DOCs are expected to have no measurable effect on fuel consumption.
However,  changes to the engine designed to reduce PM emissions can lead to a reduction in fuel
consumption, at least for direct-injected diesel engines. The potential range for improved fuel
economy for engines of this size is unknown but experience with changes to engine design that
improve combustion and reduce PM suggest that the improvement may be significant. However,
because of the difficulty in projecting the future ratio of direct-injected and indirect-injected
diesel engines for this portion of the nonroad  market and the first order affect that this ratio has
on average fleet consumption we have not attempted to account for this potential fuel economy
improvement in our cost analysis. We therefore estimate no change in fuel consumption in our
cost analyses for engines under 25 hp.

   6.2.3.3.2 Costs Associated with these Fuel Economy Impacts

   To calculate the costs associated with these fuel economy impacts, we have used a diesel fuel
price, minus taxes, of 60 cents/gallon.  To that, we have added the incremental cost per gallon
for 15 ppm fuel. These incremental fuel  costs are discussed in Chapter 7 as 7.0 cents/gallon.
Using this 67 cent value, we apply the estimated fuel  economy impact of an engine - 1% where
both a CDPF and a NOx adsorber are added, and 2% where a CDPF is added and no NOx
adsorber is present. This results in an increased operating cost for 75-750 hp engines of 0.67
cents/gallon (1% x 67 cents/gallon) for CDPF/NOx adsorber equipped engines and 1.34
cents/gallon for CDPF-only engines (2% x  67 cents/gallon). For 25-75 hp engines, and for >750
hp engines, where we estimate a two percent fuel economy impact, the estimated incremental
cost is 1.34 cents/gallon.  Importantly, these fuel  economy impacts are incurred only on new
engines; existing engines that do not meet the NRT4 standards will not see any fuel economy
impact.

   Operating costs associated with fuel economy impacts are attributed only to PM control.
                                         6-59

-------
Final Regulatory Impact Analysis
   6.2.3.4 Operating Costs Associated CCV Maintenance on New Engines

   For CCV systems, we have used a maintenance interval of 675 hours for all engines and a
cost per maintenance event of $8 to $48 for small to large engines. The 675 maintenance
interval  is chosen as twice the oil-change maintenance interval.  CCV maintenance is assumed to
be done during every other oil-change event; this results in $0 labor cost for CCV maintenance.
The calculation of operating costs associated with CCV maintenance is shown in Table 6.2-31.
The new CCV requirements apply only to turbocharged engines (naturally aspirated engines
already have a closed crankcase requirement) so there are two cent/gallon values shown in Table
6.2-31 within each power range.  The first value is the cent/gallon cost for a turbocharged engine
while the weighted cent/gallon cost within the power range (i.e., weighted by the percentage of
turbocharged engines). Using the fuel use weightings, we can calculate the fleetwide cent/gallon
cost using these latter costs within each power range. The result is a 0.2 cent/gallon cost.

   Operating costs associated with CCV maintenance are attributed evenly to NOx and PM
control.
                                     Table 6.2-31
                          Closed Crankcase Ventilation System
                 Maintenance Costs for New Turbocharged Engines ($2002)
CCV Maintenance Costs
Rated Power
BSFC
=uel Density
Copulation Wfeighted Avg. Horsepower
Copulation Wfeighted Avg. Activity
Copulation Wfeighted avg. Load Factor
CCV Filter Replacement Interval
CCV Filter Replacement Cost
=ilter Maintenance Labor
Total Filter Maintenance Cost per event
=uel Use Between Maintenance Interval
Turbcharged Fleet Fraction
Maintenance Cost for engines adding CCV
Maintenance Cost - weighted for all engines
=uel Use Wfeightings
Units
hp
Ibm/hp-hr
Ibm/gallon
hp
hrs/year
% full load
hours
$/event
$/event
$/event
gallons/period
[%]
$/gallon
$/gallon
% total
Nonroad Enaines
0750hp
0.367
7.1
1282
1130
0.571
675
$48
$0
$48.00
25,549
100%
$0.002
$0.002
6.2%
6.3 Equipment-Related Costs

   Costs of control to equipment manufacturers include fixed costs (those costs for equipment
redesign and for tooling), and variable costs (for new hardware and increased equipment
assembly time). According to the PSR Sales Database for the year 2000,52 there are
approximately 600 nonroad equipment manufacturers using diesel engines in several thousand
different equipment models. We realize that the time needed for equipment manufacturers to
make the necessary changes on such a large number of equipment models will vary significantly
from manufacturer to manufacturer and from application to application. One of the goals of the
transition program for equipment manufacturers is to reduce the potential  for anomalously high
costs for individual equipment models by providing significant additional time (up to seven
                                         6-60

-------
                                             Estimated Engine and Equipment Costs
years) for developing less costly designs or to align the changes with an already scheduled
redesign. To remain conservative in our cost estimates, we have not factored into the analysis
the significant potential cost savings associated with these provisions; Section 6.3.3 explores the
potential cost savings of the transition program for equipment manufacturers.

6.3.1 Equipment Fixed Costs

   6.3.1.1 Equipment Redesign Costs

   The projected modifications to equipment resulting from the new emission standards relate to
the need to package emission-control hardware that engine manufacturers will incorporate into
their engines. As noted in Section 6.2, the additional emission-control hardware is proportional
in size to engine displacement by a 4:1 ratio (1.5 x engine displacement for both the CDPF and
the NOx adsorber, and 1.0 x displacement for the DOC that is part of the NOx adsorber system).
We expect that equipment manufacturers will have to redesign their equipment to accommodate
this new volume of hardware.  Some redesigns will be major in scale, while others will be minor.
For example, redesign may simply involve bolting the new devices onto the existing design, but
in most cases we expect devices to be designed into the piece of equipment in a way that their
presence would not be obvious to the casual observer and, in fact, for some equipment they may
simply replace the existing muffler with no redesign needed. Additionally, a redesign to
accommodate a DOC (1.0 x engine displacement) should be less intensive than a redesign to
accommodate a CDPF/NOx adsorber system. Finally, for engines in the 75-750 hp range where
the final rule phases in new NOx standards, we assume that the redesign effort for those final
pieces of complying equipment (i.e., when the phase-in goes from 50 percent to 100 percent)
will be less costly than the first redesign  effort.

   6.3.1.1.1 Schedule of Equipment Redesigns

   The final rule includes a varying compliance dates for different engines, as shown in Table
6.3-1. For this analysis, because we are assuming no use of the transition program for equipment
manufacturers, we assume that the timing of equipment redesigns will correlate with the timing
of new emission standards (assuming no use banking under the engine ABT program).  This
results in a redesign schedule as shown in Table 6.3-1.  We have noted the percentage of
equipment models we estimate will be redesigned in years for which new emission standards are
implemented. The table also notes the estimated percentage that will be major or minor redesign
efforts.  We  also note what percentage of the redesign costs are allocated to PM and to NOx.
                                         6-61

-------
Final Regulatory Impact Analysis
                                      Table 6.3-1
              Equipment Redesign Assumptions for Equipment Manufacturers
Power
0750 np
Engine
Standard
Dates
2008
2008
2013
2008
2013
2012
2014
2011
2014
2011
2015
Pollutant
Allocation
100%PM
100%PM
50% PM
50% NOx
100%PM
100%PM
50% PM
50% NOx
100% NOx
50% PM
50% NOx
100% NOx
100% NOx
100%PM
Percent of Equipment
Models Undergoing
Minor Redesign
100%
100%

100%


50%

50%
100%

Percent of Equipment
Models Undergoing
Major Redesign


100%

100%
100%

100%


100%
   Note that we have assumed all equipment redesigns for the 75 to 750 hp range are major in
the first year of new emission standards and minor in the last year. The costs associated with
such minor redesign efforts are assumed to be half those associated with major redesign efforts.
We believe this is appropriate because equipment manufacturers will expend less effort to
redesign those pieces equipment needing to add only the NOx adsorber (in those years where the
NOx phase-in schedule changes from 50 percent to 100 percent) for three reasons: (1) these
models will already have been redesigned for the CDPF system and will already incorporate the
necessary electronic systems into their design; (2) equipment manufacturers will presumably
have gained experience during the major redesign phase that should make the minor redesign
phase more efficient; and (3) manufacturers that are aware of the future requirement will be able
to make provisions in the first redesign that account for future needs. Therefore, the second
redesign effort should be less intensive.

   Our equipment redesign cost estimates were developed based on our meetings and
conversations with engine and equipment manufacturers, specific redesign cost estimates
provided by equipment manufacturers for the redesign of equipment to accommodate  engines
meeting the Tier 2 standards, and our engineering judgment as needed. The following section
details our assessment  of costs to equipment manufacturers.
                                         6-62

-------
                                             Estimated Engine and Equipment Costs
   6.3.1.1.2 Costs of Equipment Redesigns

   While developing our equipment redesign cost estimates for the Tier 4 standards, we met
with a wide range of equipment manufacturers. This included equipment manufacturers with
annual revenues less than $50 million and engineering staffs of less than 10 employees,
equipment manufacturers with annual revenues on the order of $200 million and engineering
staffs on the order of 50 employees, and equipment manufacturers with annual revenue well in
excess of $1 billion with annual research and development budgets of more than $100 million
and engineering staffs of more than 500 employees.

   During these meetings and discussions, it became apparent to us that, in spite of the
significant engine technology differences between Tier 2/Tier 3  and Tier 4, the impact on
equipment design and the need for redesign are similar.  That is, for Tier 2, many engines have
added electronic fuel systems, turbocharging, and charge-air-cooling. In addition, many Tier 2
engines rely on retarded fuel injection to lower NOx emissions, which therefore increases heat
rejection and requires the equipment manufacturers to install larger radiators and fans.  The
process of equipment redesign for Tier 2 involved engineering work to accommodate these new
components (for example, charge-air-coolers, turbochargers, larger radiators and fans) and
electronic fuel systems. In many respects, this is similar to what will be required for Tier 4,
where engines still without electronic fuel systems will require them, and equipment
manufacturers will need to integrate aftertreatment systems (as compared with charge-air-
coolers, turbochargers, larger radiators and fans).  However, we believe that equipment redesigns
attributable to Tier 4 are  more likely to  occur early in the design cycle than many design changes
attributable to the Tier 2/3 rules.

   Some companies we  met with before the proposal gave us specific redesign cost information
for the existing nonroad standards and, in some cases, projections for equipment redesigns
necessary to integrate aftertreatment (these data are confidential business information).  We also
received redesign cost estimates from several equipment manufacturers during the Tier2/3
rulemaking regarding their projected costs for the Tier 2 standards (these data are confidential
business information).  The information provided to us through these various channels showed
that there is a very wide range of cost estimates and actual cost data for redesigning nonroad
equipment for the Tier 2  standards. In general, we learned that very large companies tend to
allocate significantly more resources to equipment redesign than the medium or small
companies.
   We have used all this information and data, and our engineering judgment, to develop the
redesign cost estimates presented in Table 6.3-2.  This table presents fixed cost per motive and
nonmotive equipment model  (motive equipment is that with some form of propulsion system
while nonmotive equipment,  such as air compressors, generator  sets, hydraulic power units,
irrigation sets,  pumps, compressors, and welders,  has none) for each power group.  In general,
nonmotive equipment has fewer design demands than does motive equipment - no operator line-
of-sight demands, fewer  serviceability constraints, and almost no impact (collision) concerns.
As a result, we have estimated a lower redesign cost for nonmotive equipment relative to motive
equipment.
                                         6-63

-------
Final Regulatory Impact Analysis
                                      Table 6.3-2
                     Estimated Equipment Redesign Costs Per Model
                                       ($2002)
Power
0750hp
2011
2015
Motive
$53,100
$53,100
$199,125
$371,700
$531,000
$531,000
$796,500
$796,500
$106,200
$796,500
Nonmotive
$53,100
$53,100
$79,650
$106,200
$106,200
$106,200
$106,200
N/A
N/A
N/A
   Using the PSR database we were able to determine the number of equipment models and the
type of equipment model (motive versus nonmotive).  We distinguished motive from nonmotive
using our Nonroad Model definition of stationary applications. Nonmotive applications include
air compressors, generator sets, pumps, hydraulic power units, irrigation sets, and welders.  All
other applications are considered motive. Table 6.3-3 shows the number of equipment models
we have estimated to be redesigned. Note that the models shown in Table 6.3-3 are not
necessarily all models but are instead the unique models that had 2000 model year sales.  The
determination of unique models was based on manufacturer name (i.e., a Caterpillar skid/steer
loader is unique from a Bobcat skid/steer loader) and the market segment to which the model
belonged (i.e., an agricultural tractor is unique from a construction backhoe) and the engine
displacement. Therefore, while a manufacturer may consider two pieces of construction
equipment with the same base engine, one with and one without a turbocharger, to be two
distinct models, we consider that one model for the sake of equipment redesign.
                                         6-64

-------
                                             Estimated Engine and Equipment Costs
                                      Table 6.3-3
                   Number of Motive vs. Nonmotive Equipment Models
                                    to be Redesigned
Power Range
0
-------
                    Table 6.3-4
         Equipment Redesign Expenditures
          Attributable to US Sales ($2002)
Year Incurred
2006
2007
2008
2009
2010
2011
2012
2013
2014
Total to US Sales
0750hp



$ 4,566,600
$ 4,566,600


$ 34,249,500
$ 34,249,500
$ 32,605,524
Total
$ 20,178,000
$ 20,178,000
$
$ 384,975,000
$ 647,607,600
$ 343,596,825
$ 241 ,724,475
$ 195,009,750
$ 34,249,500
$ 800,657,730
                    Table 6.3-5
Expenditures for Changes to Product Support Literature
          Attributable to US Sales ($2002)
Year Incurred
2006
2007
2008
2009
2010
2011
2012
2013
2014
Total to US Sales
0750hp



$ 228,330
$ 228,330


$ 456,660
$ 456,660
$ 575,392
Total
$ 5,495,850
$ 5,495,850
-
$ 7,282,665
$ 13,731,660
$10,938,600
$ 1 1 ,241 ,270
$ 7,208,325
$ 456,660
$27,144,614

-------
                                             Estimated Engine and Equipment Costs
   6.3.1.2 Costs Associated with Changes to Product Support Literature

   Equipment manufacturers are also expected to modify product support literature (dealer
training manuals, operator manuals, service manuals, etc.) due to the product changes resulting
from the new emission standards. For each product line of motive applications, we estimated
that the level of effort needed by equipment manufacturers to modify the support literature will
be about 100 hours—75 hours of junior engineering time, 20 hours of senior engineering time,
and 5 hours of clerical time—which amounts to about $10,620 in $2002. We projected that the
level of effort needed by equipment manufacturers to modify  support literature for each
nonmotive application product line will be about 50 hours (distributed similarly), which is
equivalent to about $5,310.  With the exception of the <25hp  costs, we have attributed only a
portion of the product support literature costs to US sales as described above for equipment
redesign costs. Table 6.3-5 presents the total costs per power category for changes to support
literature.

   6.3.1.3 Total Equipment Fixed Costs

   The annual equipment fixed costs for each power category are shown in Table 6.3-6. As
described above and with the exception of <25 hp expenditures, we have attributed only a
portion of the equipment fixed costs to sales within the United States.  This is appropriate
because we believe these efforts will be needed to sell equipment not only in the United States,
but also in Australia,  Canada, Japan, and the countries of the European Union. As discussed in
Section 6.2.1.1, we have therefore attributed 42 percent of the equipment fixed costs to U.S.
sales.

   The analysis projects that the expenditures will be incurred over a two-year period before the
first year of the emission standards.  The costs were then amortized over ten years at a seven
percent rate beginning with the first year of the engine standard.  The ten-year period for
amortization,  as opposed to the five-year period used for engine costs, reflects the longer product
development cycles for equipment relative to engines.
                                          6-67

-------
                            Table 6.3-6
Recovered (Annualized) Equipment Fixed Costs per Power Category ($2002)
Year Recovered
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
Total
0750hp
$
$
$
$ 593,531
$ 593,531
$ 593,531
$ 593,531
$ 4,889,563
$ 4,889,563
$ 4,889,563
$ 4,889,563
$ 4,889,563
$ 4,889,563
$ 4,296,033
$ 4,296,033
$ 4,296,033
$ 4,296,033
$ 48,895,635
Total
$ 4,514,096
$ 4,514,096
$ 4,514,096
$ 53,068,927
$ 86,376,654
$ 96,954,385
$ 117,689,513
$ 121,985,546
$ 121,985,546
$ 121,985,546
$ 117,471,450
$ 117,471,450
$ 117,471,450
$ 68,916,619
$ 35,608,892
$ 25,031,160
$ 4,296,033
$ 1,219,855,455

-------
                                            Estimated Engine and Equipment Costs
6.3.2 Equipment Variable Costs

   In addition to the incrementally higher cost of new engines estimated in Sections 6.2.1 and
6.2.2, equipment manufacturers will need to purchase hardware to mount the new exhaust
emission-control devices within each newly redesigned piece of equipment. Note that the
redesign costs we have already discussed are for changes in equipment design to accommodate
aftertreatment devices. We assume that there are minimal changes to the variable costs for the
redesigned elements of the equipment (i.e., the redesigned elements cost roughly the same as
before) because they serve the same function and have the same amount of materials. Here, we
estimate the costs associated with the new hardware that will be necessary - new brackets, bolts,
and sheet metal  - for mounting and housing (shrouding) the new aftertreatment devices.

   New brackets and bolts will be required to secure the aftertreatment devices within the piece
of equipment. Additionally, increased labor ($29/hour) and overhead costs (40%) will be
incurred to install these devices.  Table 6.3-7 shows the costs we have used per piece of
equipment ($/machine as shown in the table). Total costs per power range were calculated using
these costs and equipment sales in the year 2000.
                                         6-69

-------
   Final Regulatory Impact Analysis
                                          Table 6.3-7
             Costs for Brackets and Bolts and Associated Labor for Equipment ($2002)
Brackets/bolts/etc.

0750hp
devices added
1

1
1
2
2
2
2
new sets of
brackets/bolts
per device
0

0
2
2
2
2
2
$/set
$2

$0
$2
$5
$5
$11
$11
$/machine
$0

$0
$4
$21
$21
$42
$42
Labor

0750hp
device added
DOC

DOC
DPF
DPF&NOxAds
DPF&NOxAds
DPF&NOxAds
DPF
hrs to install
0

0
0.25
0.5
0.75
1.5
1
subtotal ($)
$0

$0
$7
$14
$22
$43
$29
overhead
$0

$0
$3
$6
$9
$17
$12
Total
$0

$0
$10
$20
$30
$61
$40
   Note to Table 6.3-7: We have assumed the addition of two devices for engines >750hp when only a CDPF is being
   added.  It may have been more appropriate to assume one device but that the number of brackets and bolts needed would
   be twice that for other engines (i.e., four sets rather than two) given the size of the device. Applying two smaller CDPFs
   needing two sets of brackets and bolts leads to the same resultant cost for brackets and bolts.
      Sheet metal costs vary by size of the aftertreatment devices being added which, in turn, vary
   by engine displacement as described in section 6.2.  The amount of sheet metal for the shroud
   was determined using the engine displacement per equipment model information in the 2002
   PSR Sales Database. The volume of the CDPF and NOx adsorber aftertreatment was calculated
   for each unique equipment model (as described in section 6.3.1.1.2) in the PSR database with an
   engine between 75 and 750 hp (1.5 times engine displacement for the CDPF and 1.5  times
   engine displacement for the NOx adsorber). The DOC was assumed to fit in place of the
   muffler. The volume of the aftertreatment was then converted to the volume of a cube and two
   inches were added to each dimension for space between the aftertreatment and the shroud.  Sheet
   metal  was assumed to cover four sides of the aftertreatment with no cover for the bottom or
   equipment facing side of the shroud. Sheet metal was assumed to cost $1.14 per square foot for
   hot rolled steel. The sheet metal cost for each model was multiplied by the total sales for that
   model using the 2000 sales information in the 2002 PSR Sales Database.
                                             6-70

-------
                                             Estimated Engine and Equipment Costs
    Summing these variable costs for each equipment model—sheet metal costs plus costs for
bolts, brackets, and labor—within each power range and then dividing by sales within the power
range gives a rough estimate of the costs we have estimated for a piece of equipment. It is
important to realize that this is weighted value within each power range determined by
calculating a unique cost for each piece  of equipment, multiplying that cost by its sales, and then
totaling those costs within each power range. Table 6.3-8 shows the sales weighted equipment
variable costs within each power range.  A twenty-nine percent manufacturer markup is also
included  in the final cost estimates shown in Table 6.3-8.

                                       Table 6.3-8
           Sales Weighted Variable Costs per Piece of Equipment by Power Range"
                    Totals include a 29% Manufacturer Markup ($2002)
Power Range
0750 hp
Year
2008
2013
2013
2012
2012
2011
2011
2011
2011
Total
$0
$20
$21
$60
$61
$77
$146
$154
$123
                 a These costs do not include the engine variable costs described in section
                 6.2.
   As shown in Table 6.3-8, we have estimated equipment variable costs to be zero for
equipment with engines under 25 hp, under the expectation that an added DOC will replace the
existing muffler and make use of the same bracket/bolt/labor used for the muffler. This is also
expected for engines in the 25 to 75 hp range from 2008 through 2012 when, for our cost
anlaysis, only a DOC is being used by the engine manufacturer for compliance;  additional bolts
and labor costs are included for the addition of a CDPF beginning in 2013.R While we have
assumed the CDPF will simply replace the muffler, there will be additional bracket/bolt/labor
demands due to the greater weight of the CDPF relative to the replaced muffler.
     Note that for costing purposes we have assumed that a DOC is used on all engines under 75 hp to comply with the
2008 standards, although test data show that some engines already meet the new emission standards without a DOC.
                                          6-71

-------
Final Regulatory Impact Analysis
6.3.3 Potential Impact of the Transition Provisions for Equipment Manufacturers

   As discussed in Section III.B of the preamble, we are extending, and in some respects are
expanding, the transition program for equipment manufacturers (TPEM) that was developed in
the 1998 final rule. The TPEM is an important component of this final rule because of the
flexibility it provides for equipment manufacturers.  However, as explained earlier, because the
program is optional, we have not included the potential impacts of TPEM on the estimated costs
of the Tier 4 program.  Nevertheless, this section discusses how the TPEM program may
substantially reduce equipment manufacturer costs.

   The  TPEM can reduce equipment manufacturer costs in two ways. First, it allows equipment
manufacturers to continue to sell limited numbers of equipment with non-Tier 4 engines even
after the Tier 4 standards go into effect. Any engine price increase associated  with the Tier 4
standards will therefore not be incurred by the equipment manufacturer or by the end user during
the time frame the manufacturers use the TPEM.  Second, the TPEM allows manufacturers to
schedule equipment design cycles to coincide with any redesign necessary because of EPA's
emission standards.  We believe this is the most significant cost savings impact of the TPEM.
This is due to the fact that many equipment manufacturers have a several small-volume model
lines.  Using the TPEM program, companies can delay the redesign costs associated with  Tier 4
engines  for up to seven years on a limited number of products.

   We performed a detailed analysis on an equipment manufacturer-by-equipment manufacturer
basis of the more than 6,000 equipment models and 600 equipment manufacturers in an industry-
wide database (the Power Systems Research database).53 This analysis looked at each equipment
manufacturer's product offerings by power category and the estimated 2000 U.S.  sales of each
equipment model.  We used this database to analyze how equipment manufacturers can use
TPEM to maximize the number of equipment models with delayed redesign until the eighth year
of the program (as discussed in Section III.B of the preamble, TPEM provisions allow equipment
manufacturers to sell products with uncertified engines until seven years after  the applicable Tier
4 standard is implemented.). We specifically analyzed the percent-of-production allowance and
the small-volume allowance programs being adopted for the Tier 4 rule (as discussed in the
preamble). The results are shown in Table 6.3-9.  (It should be noted that the newly adopted
technical hardship flexibility provision, which potentially allows an additional 70 percent of
equipment manufacturer's sales in a power category to use non-Tier 4 engines for a limited time
provided an appropriate case-by-case demonstration of extreme technical hardship is made to
EPA, likewise could have associated cost savings.)
                                         6-72

-------
                                             Estimated Engine and Equipment Costs
                                      Table 6.3-9
 Potential Impact of TPEM Program on Equipment Models and Sales (all equipment companies)
Equipment Models/
Equipment Sales
'ercent of all equipment models
hat could use TPEM for full-
>even years
'ercent of equipment sales
hat could use TPEM for full-
>even years
Engine Power Category
<25hp
56%
7%
25< hp <70a
61%
10%
70a750 hp
80%
21%
All Power
Categories
66%
10%
a Note that the power ranges are 25-75 hp and 75-175 hp. This analysis was done using 70 hp as a cut-point. We believe
the results of this analysis would not have been significantly different if the power outpoint had been 75 hp.
   This analysis indicates that if fully utilized by equipment manufacturers, 66 percent of
nonroad diesel equipment models can use the TPEM program to delay an equipment redesign
necessary for the Tier 4 standards for seven years.  Without the TPEM program, equipment
manufactures would need to redesign all their equipment models using a nonroad diesel engine
in the first year of the engine standard implementation. As an example of the flexibility offered
by the TPEM program, Table 6.3-9 indicates that for engines between 25 and 75 hp, 61 percent
of all equipment models in this power range can take advantage of the TPEM (i.e., the percent of
production allowance and the small volume allowance options) to delay an equipment redesign
for seven years. It is important to note that while the TPEM can substantially reduce equipment
redesign costs, it is expected to have a much smaller impact on the emission reductions of the
program. While the TPEM can allow equipment companies to  continue selling products with the
previous tier standards on many equipment models, the total sales that can be impacted by the
TPEM (i.e., the percent of production allowance and the small volume allowance options), which
is also shown in Table 6.3-9, is estimated to be no higher than ten percent for no more than seven
years.

   The analysis presented in Table 6.3-9 is based on the equipment produced by a wide range of
equipment manufacturers, both very large, multi-billion dollar corporations as well  as small
companies who produce a limited number of products. We have performed a similar analysis
using only those equipment companies whose data is contained in the PSR database which we
were able to identify as small businesses. In some respects the  TPEM program, while available
to all equipment manufacturers, was designed specifically to benefit small businesses. Within
the PSR database, we were able to identify 337 small businesses who together produce more
than 2,500 different equipment models.  This data was analyzed as described above for Table
6.3-9. The results are shown in Table 6.3-10.
                                         6-73

-------
Final Regulatory Impact Analysis
                                      Table 6.3-10
     Potential Impact of TPEM Program on Equipment Models and Sales of Small Business
                                Equipment Manufacturers
Equipment Models/
Equipment Sales
'ercent of all equipment models
hat could use TPEM for full-
ieven years
'ercent of equipment sales
hat could use TPEM for full-
>even years
Engine Power Category
<25hp
69%
17%
25< hp <70a
74%
24%
70a750 hp
93%
76%
All Power
Categories
79%
26%
a Note that the power ranges are 25-75 hp and 75-175 hp. This analysis was done using 70 hp as a cut-point. We believe
the results of this analysis would not have been significantly different if the power outpoint had been 75 hp.

   The results in Table 6.3-10 show that in all power categories, the TPEM program provides
more flexibility for small business equipment companies than for the equipment industry as a
whole. In every power category, the number of equipment models which small companies can
delay redesigning for the full seven years is greater than for the industry as a whole, and for the
power categories which will likely require engine aftertreatment (i.e., >25hp), approximately 75
percent or more of the equipment models could delay redesign for a full seven years. The actual
equipment sales for all of the small business equipment companies which could use the TPEM
program under this analysis is 26 percent of the total sales, but in reality this  is less than 3
percent of the total nonroad diesel market, as small business companies have a relatively small
portion of the total  nonroad diesel equipment sales.

6.4 Summary of Engine and Equipment Costs

   Details of our engine and equipment cost estimates were presented in Sections 6.2 and 6.3.
Here we summarize the cost estimates. Section 6.4.1.1 summarizes the total  engine fixed costs.
Section 6.4.1.2 summarizes the engine variable cost equations for estimating engine variable
costs. Section 6.4.1.3 summarizes the engine operating costs. Section 6.4.2.1 summarizes the
total equipment fixed costs and 6.4.2.2 summarizes the estimated equipment  variable costs.
Section 6.4.3 presents these costs on a per unit basis.  Note that all present value costs presented
here are 30-year numbers (the net present values in 2004 of the stream of costs/reductions
occurring from 2007 through 2036, expressed in $2002).

6.4.1 Engine Costs

   6.4.1.1 Engine Fixed Costs

   Engine fixed costs include costs for engine R&D, tooling, and certification. These costs are
discussed in detail in Section 6.2.1. The total estimated engine fixed costs are summarized in
Table 6.4-1. The table also includes 30-year net present values using both a three percent and a
seven percent social discount rate.
                                          6-74

-------
                                             Estimated Engine and Equipment Costs
                                       Table 6.4-1
                         Summary of Engine Fixed Costs ($2002)

Engine R&D
Engine Tooling
Engine Certification
Total
Incurred Costs
(SMillion)
$323
$74
$91
$489
Recovered Cost
($Million)
$452
$91
$111
$653
30 YearNPVof
Recovered Cost
at 3%
($Million)
$336
$70
$84
$490
30 YearNPVof
Recovered Cost
at 7%
($Million
$233
$50
$60
$343
   6.4.1.2 Engine Variable Costs

   Engine variable costs are discussed in detail in Section 6.2.2. For engine variable costs, we
have generated cost estimation equations as a function of engine displacement or number of
cylinders (see Table 6.2-27). Using these equations, we have calculated the costs for each
nonroad diesel engine sold in the year 2000, multiplied that cost by its projected sales during the
30 year period following implementation of the NRT4 program, and then added the future annual
costs for each engine to arrive at annual costs during each of those 30 years. We present those
annual engine variable costs in Chapter 8.  Table 6.4-2 shows the 30-year net present value of
those annual costs assuming a three percent social discount rate and a seven percent social
discount rate.

                                       Table 6.4-2
                    30-Year Net Present Value of Engine Variable Costs
                                        ($2002)

Engine Variable Costs
30 Year NPVat 3%
(SMillion)
$13,562
30 Year NPVat 7%
($Million)
$6,871
   6.4.1.3 Engine Operating Costs

   Engine operating costs are discussed in detail in Section 6.2.3. Table 6.4-3 summarizes
engine operating costs, excluding costs associated with the desulfurization of diesel fuel; these
costs are presented in Chapter 7.
                                          6-75

-------
Final Regulatory Impact Analysis
                                       Table 6.4-3
              Engine Operating Costs Associated with the NRLM Fuel Program
                           (cents/gallon of 15ppm fuel consumed)
Power category
0750 hp
Locomotive/Marine
Oil-Change
Savings
(19.3)
(8.2)
(5.9)
(3.3)
(2.0)
(1.4)
(1.1)
(1.1)
(1.1)
CDPF
Maintenance
0.0
2.3
1.6
0.8
0.2
0.1
0.1
0.2
--
ccv
Maintenance
0.0
0.0
0.1
0.3
0.2
0.1
0.2
0.2
--
CDPF
Regeneration3
0.0
1.3
1.3
0.7
0.7
0.7
0.7
1.3
--
Net Operating
Costsb
(19.3)
(4.6)
(2.9)
(1.5)
(0.9)
(0.5)
(0.1)
0.6
(1.1)
   a A one or two percent fuel consumption increase, a 60 cent/gallon baseline fuel price, and a 7.0 cent/gallon
   incremental fuel cost.
   b The incremental costs for low-sulfur fuel are presented in Chapter 7.
   Engines that make up the existing fleet will realize the oil-change savings shown in Table
6.4-3 while incurring none of the other operating costs, because these engines will not have
CDPF or CCV systems. New engines would incur all the costs and savings shown in Table
6.4-3.

   Table 6.4-3 shows operating costs on a cent-per-gallon basis.  Lifetime engine operating
costs vary by the amount of fuel consumed. We have calculated lifetime operating costs for
some example types of equipment and present those in Section 6.5.  Aggregate operating costs
(the annual total costs) are presented in Chapter 8 and the 30-year net present value of the
NRLM fleet operating costs are shown in Table 6.4-4.
                                           6-76

-------
                                            Estimated Engine and Equipment Costs
                                      Table 6.4-4
           30-Year Net Present Value of NRLM Fleetwide Engine Operating Costs
                                 Excluding Fuel Costs
                                       ($2002)

Engine Operating Costs (a
negative value indicates a
savings)
30 Year NPV at
3%
(SMillion)
-$4,517
30 Year NPV at
7%
(SMillion)
-$2,745
6.4.2 Equipment Costs

   6.4.2.1 Equipment Fixed Costs

   Equipment fixed costs are discussed in detail in Section 6.3.1.  Table 6.4-5 shows the
estimated equipment fixed costs associated with the Tier 4 emission standards. These figures
include estimated costs for equipment redesign and generation of new product support literature.

                                      Table 6.4-5
                       Summary of Equipment Fixed Costs ($2002)

Redesign
Product Literature
Total
Incurred Costs
(SMillions)
$801
$27
$828
Recovered Costs
(SMillions)
$1,180
$40
$1,220
30 Year NPV of
Recovered Cost at
3%
(SMillion)
$819
$28
$847
30 Year NPV of
Recovered Cost at
7%
(SMillion)
$518
$18
$537
   6.4.2.2 Equipment Variable Costs

   Equipment variable costs are discussed in detail in Section 6.3.2. Using the costs presented
there we have calculated the variable costs for the equipment sold in the year 2000 and then
projected those costs over the 30 year period following implementation of the NRT4 program.
We present those annual equipment variable costs in Chapter 8.  Table 6.4-6 shows the 30-year
net present value of those annual costs assuming a three percent and a seven percent social
discount rate.
                                         6-77

-------
Final Regulatory Impact Analysis
                                      Table 6.4-6
                  30-Year Net Present Value of Equipment Variable Costs
                                        ($2002)

Equipment Variable Costs
30 Year NPV at 3%
(SMillion)
$434
30 Year NPV at 7%
(SMillion)
$217
6.4.3 Engine and Equipment Costs on a Per Unit Basis

   For the Nonroad Diesel Economic Impact Analysis Model (NDEIM, see Chapter 10), we
need engine  and equipment costs per unit sold.  These per unit costs serve as inputs to the model
to determine how the cost increases might impact the quantity of units sold. The costs presented
here in Chapter 6 are aggregated in Chapter 8 into annual fleetwide costs during a 30 year period
following implementation of the NRT4 program. The annual fleetwide engine fixed costs by
power category are shown in Table 8.2-1. The costs presented there represent the annual
recovered costs associated with engine R&D, tooling, and certification (note that these costs are
also presented in Tables 6.2-6, 6.2-8, 6.2-10, and 6.3-6. As explained earlier in this chapter, the
recovered engine R&D costs are revenue weighted, meaning that we have attributed the total
industry costs for engine R&D according to our best estimate of revenues from engine sales.
Doing this does not impact the resultant total cost of the new Tier 4 standards and only impacts
how the costs are allocated to each power range. Such an allocation is of importance only when
trying to determine the per unit cost as we are here. Manufacturers may choose to recover their
investments in ways different than we have estimated, although recovering investments based on
revenues seems like the most likely probability.

   Table 6.4-7 shows the per unit costs using this methodology. The values shown in the table
are simply the result of dividing the annual costs by power range shown in Table 8.2-1 by the
engine sales  by power range shown in Table 8.1-1. The costs per unit change from year to year
because engine standards are implemented differently in each power category.  As more engines
across more  power categories phase-in to a new set of engine standards, the engine R&D costs
are recovered according to a different revenue weighting. Note also that tooling costs within
each power range can vary year to year on a per unit basis. This occurs because there are many
engine platforms that span different power ranges.  Therefore, tooling expenditures  done for an
engine platform that spans the 100-175 hp and the  175-300 hp ranges would be recovered only
on the 175-300 hp engines in 2011 and then on both 100-175 hp and 175-300 hp engines
beginning in 2012.  Engine fixed costs per unit become zero after several years because the fixed
costs invested have been completely recovered.

   We can get the engine variable costs per unit in much the same way by dividing the
aggregate engine variable costs by power range shown in Table 8.2-3 by the engine sales by
power range shown in Table 8.1-1. The results are shown in Table 6.4-8. Note that the engine
variable costs per unit continue indefinitely and do not go to zero as do the engine fixed costs
shown in Table 6.4-7. Note also that, by 2020, the engine variable costs are not longer changing

                                         6-78

-------
                                            Estimated Engine and Equipment Costs
due to phase-ins, learning curves, or other factors.

   Equipment fixed and variable costs per unit can be generated in the same way.  Tables 8.3-1
and 8.3-3 present the annual fleetwide equipment fixed and equipment variable costs by power
category. Dividing these costs by sales (Table 8.1-1) results in the per unit costs shown in
Tables 6.4-9 and 6.4-10 for equipment fixed and equipment variable costs per unit, respectively.
                                        6-79

-------
                 Table 6.4-7
 Estimated Engine Fixed Costs per Unit ($2002)

0750hp
2008
$ 38
$ 49
$ 50
$
-
-
-
-
-
2009
$ 37
$ 48
$ 49
$
-
-
-
-
-
2010
$ 36
$ 47
$ 49
$
-
-
-
-
-
2011
$ 35
$ 46
$ 48
$
$
$ 225
$ 527
$ 1,156
$ 570
2012
$ 34
$ 45
$ 47
$ 80
$ 78
$ 220
$ 521
$ 1,138
$ 561
2013
$
$ 74
$ 76
$ 78
$ 77
$ 217
$ 515
$ 1,122
$ 553
2014
$
$ 73
$ 75
$ 108
$ 106
$ 290
$ 735
$ 1,630
$ 545
2015
$
$ 71
$ 73
$ 106
$ 105
$ 285
$ 727
$ 1,606
$ 1,447
2016
$
$ 70
$ 72
$ 104
$ 103
$ 74
$ 220
$ 509
$ 897
2017
$
$ 69
$ 71
$ 29
$ 29
$ 73
$ 218
$ 502
$ 884
2018
$
-
$
$ 28
$ 29
$ 72
$ 216
$ 495
$ 872
2019
$
-
-
-
-
-
-
$
$ 860
2020
$
-
-
-
-
-
-
-
-
2021
$
-
-
-
-
-
-
-
-
2022
$
-
-
-
-
-
-
-
-
2023
$
-
-
-
-
-
-
-
-
2024
$
-
-
-
-
-
-
-
-
                 Table 6.4-8
Estimated Engine Variable Costs per Unit ($2002)

0750hp
2008
$ 129
$ 147
$ 167
$
-
-
-
-
-
2009
$ 129
$ 147
$ 167
$
-
-
-
-
-
2010
$ 123
$ 139
$ 158
$
-
-
-
-
-
2011
$ 123
$ 139
$ 158
$
$
$ 1,981
$ 2,609
$ 4,944
$ 1,973
2012
$ 123
$ 139
$ 158
$ 1,133
$ 1,375
$ 1,981
$ 2,609
$ 4,944
$ 1,973
2013
$ 123
$ 887
$ 837
$ 1,133
$ 1,375
$ 1,536
$ 2,021
$ 3,825
$ 1,543
2014
$ 123
$ 887
$ 837
$ 1,122
$ 1,351
$ 1,937
$ 2,545
$ 4,807
$ 1,543
2015
$ 123
$ 675
$ 636
$ 1,122
$ 1,351
$ 1,937
$ 2,545
$ 4,807
$ 8,335
2016
$ 123
$ 675
$ 636
$ 1,122
$ 1,351
$ 1,937
$ 2,545
$ 4,807
$ 8,335
2017
$ 123
$ 675
$ 636
$ 1,122
$ 1,351
$ 1,937
$ 2,545
$ 4,807
$ 6,734
2018
$ 123
$ 675
$ 636
$ 1,122
$ 1,351
$ 1,937
$ 2,545
$ 4,807
$ 6,734
2019
$ 123
$ 675
$ 636
$ 1,122
$ 1,351
$ 1,937
$ 2,545
$ 4,807
$ 6,734
2020
$ 123
$ 675
$ 636
$ 1,122
$ 1,351
$ 1,937
$ 2,545
$ 4,807
$ 6,734
2021
$ 123
$ 675
$ 636
$ 1,122
$ 1,351
$ 1,937
$ 2,545
$ 4,807
$ 6,734
2022
$ 123
$ 675
$ 636
$ 1,122
$ 1,351
$ 1,937
$ 2,545
$ 4,807
$ 6,734
2023
$ 123
$ 675
$ 636
$ 1,122
$ 1,351
$ 1,937
$ 2,545
$ 4,807
$ 6,734
2024
$ 123
$ 675
$ 636
$ 1,122
$ 1,351
$ 1,937
$ 2,545
$ 4,807
$ 6,734

-------
Estimated Eq

0750hp
2008
$ 15
$ 8
$ 8
$ -
$ -
$ -
$ -
$ -
$ -
2009
$ 15
$ 8
$ 8
$ -
$ -
$ -
$ -
$ -
$ -
2010
$ 14
$ 8
$ 8
$ -
$ -
$ -
$ -
$ -
$ -
2011
$ 14
$ 7
$ 8
$ -
$ -
$ 302
$ 529
$ 1,210
$ 177
2012
$ 14
$ 7
$ 8
$ 109
$ 170
$ 297
$ 523
$ 1,192
$ 175
2013
$ 13
$ 42
$ 44
$ 107
$ 168
$ 291
$ 518
$ 1,175
$ 172
Table 6.4-9
uipment Fixed Costs per Unit ($2002)
2014
$ 13
$ 41
$ 43
$ 132
$ 207
$ 360
$ 642
$ 1 ,451
$ 170
2015
$ 13
$ 40
$ 42
$ 130
$ 204
$ 353
$ 635
$ 1 ,430
$ 1 ,377
2016
$ 12
$ 40
$ 42
$ 128
$ 201
$ 348
$ 628
$ 1,410
$ 1 ,358
2017
$ 12
$ 39
$ 41
$ 126
$ 197
$ 342
$ 622
$ 1 ,390
$ 1 ,339
2018
$ -
$ 32
$ 33
$ 124
$ 194
$ 336
$ 615
$ 1,371
$ 1 ,320
2019
$ -
$ 31
$ 33
$ 122
$ 192
$ 331
$ 609
$ 1 ,353
$ 1 ,302
2020
$ -
$ 31
$ 32
$ 120
$ 189
$ 326
$ 603
$ 1 ,335
$ 1 ,285
2021
$ -
$ 30
$ 32
$ 118
$ 186
$ 65
$ 121
$ 266
$1,114
2022
$ -
$ 30
$ 31
$ 24
$ 37
$ 64
$ 120
$ 263
$1,100
2023
$ -
$ -
$ -
$ 24
$ 37
$ 63
$ 118
$ 259
$ 1 ,085
2024
$ -
$ -
-
$ -
$ -
$ -
$ -
$ -
$ 1 ,072
                  Table 6.4-10
Estimated Equipment Variable Costs per Unit ($2002)

0750hp
2008
$ -
$ -
-
$ -
$ -
$ -
$ -
$ -
$ -
2009
$ -
$ -
-
$ -
$ -
$ -
$ -
$ -
$ -
2010
$ -
$ -
-
$ -
$ -
$ -
$ -
$ -
$ -
2011
$ -
$ -
-
$ -
$ -
$ 58
$ 110
$ 116
-
2012
$ -
$ -
$ -
$ 45
$ 46
$ 58
$ 110
$ 116
-
2013
$ -
$ 20
$ 21
$ 45
$ 46
$ 46
$ 88
$ 92
-
2014
$ -
$ 20
$ 21
$ 48
$ 49
$ 62
$ 117
$ 123
-
2015
$ -
$ 16
$ 17
$ 48
$ 49
$ 62
$ 117
$ 123
$ 123
2016
$ -
$ 16
$ 17
$ 48
$ 49
$ 62
$ 117
$ 123
$ 123
2017
$ -
$ 16
$ 17
$ 48
$ 49
$ 62
$ 117
$ 123
$ 98
2018
$ -
$ 16
$ 17
$ 48
$ 49
$ 62
$ 117
$ 123
$ 98
2019
$ -
$ 16
$ 17
$ 48
$ 49
$ 62
$ 117
$ 123
$ 98
2020
$ -
$ 16
$ 17
$ 48
$ 49
$ 62
$ 117
$ 123
$ 98
2021
$ -
$ 16
$ 17
$ 48
$ 49
$ 62
$ 117
$ 123
$ 98
2022
$ -
$ 16
$ 17
$ 48
$ 49
$ 62
$ 117
$ 123
$ 98
2023
$ -
$ 16
$ 17
$ 48
$ 49
$ 62
$ 117
$ 123
$ 98
2024
$ -
$ 16
$ 17
$ 48
$ 49
$ 62
$ 117
$ 123
$ 98

-------
Final Regulatory Impact Analysis
6.5 Weighted Average Costs for Example Types of Equipment

6.5.1 Summary of Costs for Some Example Types of Equipment

   To better illustrate the engine and equipment cost impacts for this final rule, we have chosen
several types of equipment and present the estimated costs for them using weighted average
inputs—horsepower, displacement, number of cylinders, etc. Using these sales weighted inputs,
we can calculate the costs for these types of equipment in several power ranges and better
illustrate the cost impacts of the new emission standards.  For the weighted average inputs, we
have used the PSR database  and determined the sales weighted averages of various parameters of
interest. These results are shown in Table 6.5-1.  We can use the sales weighted average inputs
shown in Table 6.5-1  along with the engine variable cost equations presented in Table 6.4-2 to
generate the sales weighted average engine variable costs within each power range (doing so will
match the costs presented in Table 6.4-8). For engine fixed costs per unit and equipment fixed
and variable costs per unit, we can use the costs per unit presented in Tables 6.4-7, 6.4-9, and
6.4-10, respectively.

   These results are presented in Table 6.5-2. Costs presented are near-term and long-term
costs for the final standards to which engines in each power category must comply. Long-term
costs include only variable costs and therefore represent costs after all fixed costs have been
recovered.  Note that not all  engines in each power category would incur all the costs shown in
the table. For example, only turbocharged engines will add a CCV system as a result of the
NRT4 final rule—it is important to remember that the costs presented in Table 6.5-2 are sales
weighted averages within each power range. Included in Table 6.5-2 are estimated operating
costs for each power range, again using the sales weighted average inputs shown in Table 6.5-1
along with information presented in Tables 6.2-29 through 6.2-31 and the fuel economy impacts
discussed in section 6.2.3.3.

   We can compare these sales weighted average costs by power range to the typical price of
various types of equipment—construction, agricultural, pumps & compressors, gensets &
welders, refrigeration & A/C, general industrial,  and lawn & garden.  We have estimated the
prices of these equipment using a linear relationship between the price for these types of
equipment and their power.54 Table 6.5-3 shows the resultant equipment prices.  Table 6.5-4
shows the near-term and long-term costs (Table 6.5-2) as a percentage of equipment prices
(Table 6.5-3).
                                         6-82

-------
                                                                  Table 6.5-1
                                  Sales Weighted Average Inputs for Engine & Equipment Costs ($2002)

Sales Weighted Displacement (L)
Sales Weighted # Cylinders
Sales Weighted Hp
% Naturally Aspirated
% Turbo
% Dl
% IDI
0750hp
41 .968
11.8
1335.3
0%
100%
100%
0%
%DI and %IDI refer to the percentage of engines that have a direct injection fuel system and the percentage that have an indirect injection fuel system.

-------
                                                                       Table 6.5-2
                                     Sales Weighted Average Near-Term and Long-Term Costs by Power Category3
                                     ($2002, for the final emission standards to which the equipment must comply)

Near-term costs calculated in the year:
Engine variable costs
Fuel System
EGR
CCV*
CDPF
CDPF regen system
NOx adsorber
DOC
Engine Fixed Costs
R&D
Tooling
Cert
Equipment Variable Costs
Equipment Fixed Costs
Near-term Total Engine & Equipment Costs
Long-term Total Engine & Equipment Costs in the year 2030
Operating Costs (discounted lifetime $)
Fuel Costs
Oil Change Costs (Savings)
System regenerations
CCV maintenance
CDPF maintenance
Total Incremental Operating Costs (Savings)
Baseline Operating Costs (fuel and oil only)
0750 hp
2015

$0
$1 ,451
$91
$6,218
$575
$0
$0

$861
$107
$478
$123
$1,377
$11,280
$6,830

$23,110
-$3,790
$4,430
$620
$500
$24,870
$212,720
a. Near-term costs include both variable costs and fixed costs; long-term
b. For 25 to 75 hp engines, CCV costs in 2013 will be long term because
costs include only variable costs and represent those costs that remain following recovery of all fixed costs.
CCV systems are first required in 2008.

-------
                                                                     Table 6.5-3
                                   Sales Weighted Average Prices for Various Types of Equipment ($2002)

Construction Equipment
Agricultural Equipment
Pumps & Compressors
GenSets & Welders
Refrigeration & A/C
General Industrial
Lawn & Garden
0750hp
$ 976,900
NA
NA
NA
NA
$ 421 ,900
NA
                                                                     Table 6.5-4
                                            Estimated Costs as a Percentage of New Equipment Price

Near-term Cost to Price Ratio
Construction Equipment
Agricultural Equipment
Pumps & Compressors
GenSets & Welders
Refrigeration & A/C
General Industrial
Lawn & Garden
Long-term Cost to Price Ratio
Construction Equipment
Agricultural Equipment
Pumps & Compressors
GenSets & Welders
Refrigeration & A/C
General Industrial
Lawn & Garden
0750hp

1%




3%


1%




2%

* Note that the above percentages include equipment cost estimates that are averaged across all equipment types (i.e, motive and non-motive equipment). Our redesign estimates for non-motive
equipment are lower than for motive equipment (see Table 6.3-2).  Therefore, the near-term percentages for non-motive equipment types (e.g., gensets, pumps, etc.), are skewed slightly high just
as the near-term percentages for motive equipment types are skewed slightly low. As a result, the long-term percentages, that represent the percentages after all fixed costs like engine R&D and
equipment redesign have been recovered and are no longer part of the estimated cost, are probably better representations of the possible effect of the rule on equipment prices.
                                                                             6-85

-------
Final Regulatory Impact Analysis
6.5.2 Method of Generating Costs for a Specific Piece of Equipment

   To facilitate the effort to duplicate this example analysis for specific pieces of equipment,
this section will briefly describe the necessary steps to create the cost analysis based on the
information in this document.

   The first step required to develop an estimate of our projected cost for control under the Tier
4 program is to define certain characteristics of the engine in the piece of equipment for which a
cost estimate is desired.  Specifically, the following items must be defined:

   displacement of the engine (i.e., the cylinder swept volume) in liters;
•  type of aspiration (i.e., turbocharged or naturally aspirated);
•  number of cylinders;
•  type of combustion system used by the engine (i.e., indirect-injection,  IDI, or direct-
   injection, DI);
•  model year of production; and,
•  the power category of the engine.

   With this information and the data tables elsewhere in this document, it is possible to
estimate the costs of meeting the new standards for any particular piece of equipment.

   As an example, we will estimate the cost of compliance for a 76 hp backhoe in the year 2012.
The first step is to define our engine characteristics, as shown in Table 6.5-6.

                                       Table 6.5-6
             Engine and Equipment Characteristics of an Example Cost  Estimate
76 hp Backhoe Example
Model Year
Displacement (liters)
Cylinder (number)
Aspiration
Combustion System
Power Category
2012
3.9
4
Turbocharged
Direct Injection
75 to 175 hp
reader defined
application specific
application specific
application specific
application specific
regulations define the standards and
the timing of the standards
   For engines produced in the early years of the program, an accounting of the fixed costs
needs to be made. Fixed costs include the engine fixed cost for research and development,
tooling, and certification as well as equipment fixed includes including redesign and manual
costs. These fixed costs are reported in this chapter on a per engine/piece of equipment basis in
each year of the program for which a fixed cost is applied.  The necessary numbers to calculate
the fixed costs can simply be read from these tables.
                                          6-86

-------
                                             Estimated Engine and Equipment Costs
                                       Table 6.5-3
                         Fixed Costs for an Example Cost Estimate
2012 76hp Backhoe Example
Engine R&D
Engine Tooling
Engine Certification
Total Engine Fixed
Total Equipment Fixed
Total Fixed Costs
$50
$16
$14
$80
$109
$189
Table 6.2-6 Engine R&D Costs (per engine)
Table 6.2-8 Engine Tooling Costs (per engine)
Table 6.2-10 Engine Certification Costs (per engine)
Summation (see also Table 6.4-7)
Table 6.4-9 Equipment Fixed Cost per Unit
Summation
   The engine variable costs are related to specific engine technology characteristics in a series
of linear equations described in table 6.2-27'. The table includes all the different variable cost
components for different size ranges of engines meeting applicable emission standards.  It
includes a description of the particular engine categories for which the costs are incurred. The
simplest approach to estimating the variable costs is to repeat the table and then to simply zero
out any components that do not apply for a particular example (see Table 6.5-4 below).
                                          6-87

-------
                                         Table 6.5-4
                   Summary of Cost Equations for Engine Variable Costs
              for a 76hp Backhoe Example (x represents the dependent variable)
Engine Technology
NOx Adsorber System
201 2 76hp Backhoe
CDPF System
201 2 76hp Backhoe
CDPF Regen System -
IDI engines
201 2 76hp Backhoe
CDPF Regen System -
DI engines
201 2 76hp Backhoe
DOC
201 2 76hp Backhoe
CCV System
201 2 76hp Backhoe
Cooled EGR System
201 2 76hp Backhoe
Common Rail Fuel
Injection
(mechanical fuel
system baseline)
201 2 76hp Backhoe
Common Rail Fuel
Injection
(electronic rotary fuel
system baseline)
201 2 76hp Backhoe
Time Frame3
Near term
Long term
2012 is
Near Term
Near term
Long term
2012 is
Near Term
Near term
Long term
2012 is
Near Term
Near term
Long term
2012 is
Near Term
Near term
Long term
2012 is
Near Term
Near term
Long term
2012 is
Near Term
Near term
Long term
2012 is
Near Term
Near term
Long term
2012 is
Near Term
Near term
Long term
2012 is
Near Term
Cost Equation
$103(x)+$183
$83(x)+$160
$103 (3.9)+$183 =
$585
$146(x)+$75
$112(x)+$57
$146(3.9)+$75=
$644
$20(x) + $293
$16(x)+$223
not applicable
$10(x)+$147
$8(x)+$lll
$10(3. 9)+$ 147=
$186
$18(x)+$116
$18(x)+$110
not applicable
$2(x) + $34
$2(x) + $24
$2(3.9)+$34=
$42
$43(x)+$65
$33(x)+$48
not applicable
$78(x) + $636
$58(x) + $484
not applicable
$67(x)+$178
$50(x)+$134
not applicable
Dependent
Variable (x)
Displacement1"
3.9 liters
Displacement
3.9 liters
Displacement
3. 9 liters
Displacement
3.9 liters
Displacement
3.9 liters
Displacement
3.9 liters
Displacement
3.9 liters
# of cylinders/
injectors
3.9 liters
# of cylinders/
injectors
3.9 liters
How Used
>75 hp engines according to
phase-in of NRT4 NOx std.
In 2012 a 76 hp engine in the
NOx phase-in set will require a
NOx adsorber
>25 hp engines according to
NRT4 PM std.
In 2012 all 76hp engines are
projected to require CDPFs
IDI engines adding a CDPF
The example engine has a direct-
injection combustion system, not
indirect-inj ection
DI engines adding a CDPF
The example engine is a DI
engine and has a CDPF
<25 hp engines beginning in
2008;
25-75 hp engines 2008 thru 2012
Example engine rated power is
greater than 75 hp
All turbocharged engines when
they first meet a Tier 4 PM std.
The example engine is
turbocharged
25-50 hp engines beginning in
2013; >750hp engines beginning
in 2011
Example rated power is greater
than 50 hp
25-50 hp DI engines when they
add a CDPF
Example rated power is greater
than 50 hp
50-75 hp DI engines when they
add a CDPF
Example rated power is greater
than 75 ho
a Near term = years 1 and 2; Long term = years 3+. Explanation of near term and long term is in Section 6.1.
b Displacement refers to engine displacement in liters.

-------
                                               Estimated Engine and Equipment Costs
    Summing the applicable variable costs estimated in table 6.5-4 gives a total engine variable
cost for the 76hp Backhoe example of $1457 (Note that this value of $1457 differs from the
value shown in Table 6.4-8 due to that value being based on only 50 percent of engines in this
power range  adding a NOx adsorber in 2012). The equipment variable costs are presented in
table 6.4-10 and are referenced by engine power category.  For the 76hp example here, the
estimated equipment variable costs  are $45.

    Having estimated the engine and equipment fixed and variable costs it is possible to estimate
the total new product costs (excluding operating costs changes) by simply totaling the fixed and
variable costs estimated here. The resulting total is $1691 ($189 + $1457 + $45, note that
rounding may result in slightly different results).  Typically we have presented these total cost
estimates to the nearest ten dollars.

6.5.3 Costs for Specific Examples from the Proposal

    In the proposal, we developed costs and prices for several specific example pieces of
equipment. Here we recreate that analysis using the costs presented above for the final rule.
Table 6.5-5 shows these results. For this table, we have used the same engine and equipment
related inputs (power, displacement, etc.)  as was used in Table  6.5-1 of the draft RIA to facilitate
the comparison.8
     Another important point here is that we have used the same load factor, activity, and fuel consumption inputs, etc..
that were used in the draft RIA to ensure a fair comparison of operating cost differences between the draft analysis and
the final analysis. Note also that the inputs used for the values shown in Table 6.5-5 are for the specific pieces of
equipment and are not the sales weighted inputs used to generate the operating costs shown in Table 6.5-2, this explains
the different results.

                                            6-89

-------
Final Regulatory Impact Analysis
                                        Table 6.5-5
          Near Term and Long Term Costs for Several Example Pieces of Equipment"
         ($2002, for the final emission standards to which the equipment must comply)

Horsepower
Displacement (L)
# of cylinders/injectors
Aspiration
Fuel System
Incremental Engine &
Equipment Cost
Long Term
Near Term
Estimated Equipment
Priceb
Incremental Operating
Costs0
Baseline Operating Costs
(Fuel & Oil only)0
GenSet
9hp
0.4
1
natural
DI
$120
$180
$4,000
-$80
$940
Skid/Steer
Loader
33 hp
1.5
3
natural
DI
$790
$1,160
$20,000
$70
$2,680
Backhoe
76 hp
3.9
4
turbo
DI
$1,200
$1,700
$49,000
$610
$7,960
Dozer
175 hp
10.5
6
turbo
DI
$2,560
$3,770
$238,000
$2,480
$27,080d
Ag
Tractor
250 hp
7.6
6
turbo
DI
$1,970
$3,020
$135,000
$2,110
$23,750
Dozer
503 hp
18
8
turbo
DI
$4,140
$6,320
$618,000
$7,630
$77,850
Off-
Highway
Truck
1000 hp
28
12
turbo
DI
$4,670
$8,610
$840,000
$20,670
$179,530
 a. Near-term costs include both variable costs and fixed costs; long-term costs include only variable costs and
 represent those costs that remain following recovery of all fixed costs.
 b. Updated prices for the final analysis taken from, "Price Database for New Non-road Equipment," memorandum
 from Zuimdie Guerra to docket A-2001-28.55
 c. Present value of lifetime costs.
 d. This value corrects an error that existed in the draft RIA where we incorrectly reported the baseline operating cost
 as $77,850 (the value for the 503 hp dozer).
6.6 Residual Value of Platinum Group Metals

   One element not considered in our cost analysis is the residual value of the platinum group
metals (PGMs) in the aftertreatment devices that may be added to comply with the new engine
standards.  These devices cannot be lawfully removed at the end of an engine's life and reused
on a new engine or piece of equipment due to deterioration and/or agglomeration of the PGMs.
However, virtually all of the PGMs contained in the devices will remain there and can be
removed and recycled back into the open market for use in new aftertreatment devices. This
represents a residual value to these metals much like the residual value to many other parts of a
truck headed for scrappage.  Typically, today, the item of greatest residual value would be the
engine which can be removed from an old vehicle/truck prior to scrappage, rebuilt, and then sold
back into the market.  This same thing can be expected to happen with the PGMs installed in the
                                           6-90

-------
aftertreatment devices.

   From experts in the field,56'57 we learned that there are as many as 50 major used/spent auto
catalyst collection sites in the United States.  Further, roughly 80 percent of spent auto catalysts
are recycled in the US (only 30 percent are recycled currently in Europe, a percentage that will
presumably increase as more PGM containing devices are used in Europe). We also learned that
only one to two percent of platinum is lost during the recovery process and the same is true for
paladium. For rhodium, as much as 10 percent is lost during the recovery process.

   We can estimate the residual value of PGMs being used to comply with the Tier 4 standards
by using the PGM loadings and the aftertreatment device volumes we have estimated will be
used (see section 6.2.2). Doing this results in a 30-year net present value, assuming a three
percent discount rate, of $3 billion (using the NRT4 PGM prices). This is roughly 20 percent of
the $13.6 billion engine variable costs we have estimated. But, according to experts in the field,
we cannot expect all of this value to be returned to the market. To be conservative, we have
assumed that  80 percent of aftertreatment devices would be recycled and that 98 percent of the
platinum in those devices would be recovered and returned to the market while only 90 percent
of the rhodium would be recovered and returned to  the market. Further, we have assumed that
ten percent of the residual value would be kept by the recycler to cover costs associated with
recycling the  material (i.e., energy use, labor, and profit).58 We must also consider the time gap
between installation on a new truck and recovery. For these calculations, we used the average
lifetimes by power category from our NONROAD model and assumed that, at the end of those
lifetimes, 80 percent of devices would be recovered. In this way, we calculate a net present
value of PGMs recovered in the year they first enter the new truck market. We have done this
for each of the 30 years following implementation of the Tier 4 standards giving us a series of
present values of recovered PGMs for each of 30 years.  Note that, when accounting for the
latency period between the new equipment purchase and the ultimate recycling, we have used a
seven percent discount rate rather than three percent.  Had we used a three percent rate, the
savings would have been higher. Table 6.6-1 shows these results along with the total annual
engine variable costs for comparison (see Table 8.2-3).

   The table  shows that the residual value of PGMs could amount to a 30-year net present value
savings  of roughly $1.2 billion, assuming a three percent social discount rate. Note  that, while
we have estimated the residual value at $1.2 billion versus PGM use of $3 billion, this does not
mean that only 40 percent of PGMs are actually returned to the market.  Instead, it means that
the present value of PGMs recovered are 40 percent of the value of those initially used. By our
estimation, nearly 80 percent of platinum will be recovered (98% of 80%) and just over 70
percent of rhodium will be recovered (90% of 80%).  Note also that, to remain conservative in
our cost estimates, we have not used these estimates in any of our cost per ton or our benefit-cost
analyses. We have presented them here only for the information of the reader.
                                          6-91

-------
Final Regulatory Impact Analysis
                                    Table 6.6-1
                      Potential Impact of PGM Recovery on Costs
                             (SMillions of 2002 dollars)
Year
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2031
2032
2033
2034
2035
2036
30YrNPVat3%
30YrNPVat7%
Engine Variable
Costs
(including PGMs)
$ 62
$ 63
$ 61
$ 340
$ 637
$ 798
$ 864
$ 839
$ 852
$ 860
$ 873
$ 887
$ 900
$ 913
$ 927
$ 940
$ 954
$ 967
$ 980
$ 994
$ 1 ,007
$ 1,021
$ 1 ,034
$ 1 ,048
$ 1,061
$ 1 ,074
$ 1 ,088
$ 1,101
$ 1,115
$ 13,562
$ 6,871
PGM Costs
$0
£
$0
£
$ 2
$ 59
$ 113
$ 130
$ 186
$ 193
$ 196
$ 199
$ 202
$ 205
$ 208
$ 211
$ 214
$ 217
$ 220
$ 223
$ 226
$ 229
$ 232
$ 234
$ 237
$ 240
$ 243
$ 246
$ 249
$ 252
$ 255
$ 2,996
$ 1 ,488
PV of PGMs
Recovered
$ 0)
$ 0)
$ 0)
$ (22)
$ (46)
$ (54)
$ (76)
$ (79)
$ (80)
$ (82)
$ (83)
$ (84)
$ (85)
$ (87)
$ (88)
$ (89)
$ (90)
$ (92)
$ (93)
$ (94)
$ (95)
$ (97)
$ (98)
$ (99)
$ (100)
$ (102)
$ (103)
$ (104)
$ (105)
$ (1,231)
$ (611)
                                       6-92

-------
                                           Estimated Engine and Equipment Costs
Chapter 6 References

1.  "Electronic Systems and EGR Costs for Nonroad Engines," Final Report, ICF Consulting,
December, 2002, Public Docket No. A-2001-28, Docket Item II-A-10.

2.  "Estimated Economic Impact of New Emission Standards for Heavy-Duty On-Highway
Engines," Acurex Environmental Corporation Final Report (FR 97-103), March 31, 1997, Public
Docket No. A-1996-40, Docket Item II-A-12.

3.  "Economic Analysis of Vehicle and Engine Changes Made Possible by the Reduction of
Diesel Fuel Sulfur Content, Task 2 - Benefits for Durability and Reduced Maintenance" ICF
Consulting, December 9, 1999, Public Docket No. A-2001-28, Docket Item II-A-75.

4.  "Update of EPA's Motor Vehicle Emission Control Equipment Retail Price Equivalent (RPE)
Calculation Formula," Jack Faucett Associates, Report No. JACKFAU-85-322-3, September
1985, Public Docket No. A-2001-28, Docket Item II-A-74.

5.  "Economic Analysis of Diesel Aftertreatment System Changes Made Possible by Reduction
of Diesel Fuel Sulfur Content," Engine, Fuel, and Emissions Engineering, Incorporated,
December 15, 1999, Public Docket No. A-2001-28, Docket Item II-A-76.

6.  "Learning Curves in Manufacturing," Linda Argote and Dennis Epple, Science, February 23,
1990, Vol. 247, pp. 920-924.

7.  Power Systems Research, OELink Sales Version, 2002.

8.  For the European Union:  Directive of the European Parliament and of the Council amending
Directive 97/68/EC; For Canada: memo to public docket from Todd Sherwood, Public Docket
No. A-2001-28, Docket Item II-B-36.

9.  "Financial Data regarding Geographic Allocation of Nonroad Diesel Equipment and Engine
Company Revenue and Sales," memorandum from William Charmley to Air Docket OAR-2003-
0012, April 7, 2004, EDOCKET OAR-2003-0012-0927.

10. Nonroad Diesel Final Rule, 63 FR 56968, October 23, 1998.

11. "Learning Curves in Manufacturing," Linda Argote and Dennis Epple, Science, February 23,
1990, Vol. 247, pp. 920-924.

12. "Treating Progress Functions As Managerial Opportunity", J.M Dutton and A. Thomas,
Academy of Management Review, Rev. 9, 235, 1984, Public Docket A-2001-28, Docket Item II-
A-73.

13. Nonconformance Penalty Final Rule, 67 FR 51464, August 8,  2002.
                                        6-93

-------
Final Regulatory Impact Analysis
14. "Economic Analysis of Diesel Aftertreatment System Changes Made Possible by Reduction
of Diesel Fuel Sulfur Content," Engine, Fuel, and Emissions Engineering, Incorporated,
December 15, 1999, Public Docket No. A-2001-28, Docket Item II-A-76.

15. "Estimated Economic Impact of New Emission Standards for Heavy-Duty On-Highway
Engines," March 1997, EPA420-R-97-009, Public Docket A-2001-28, Docket Item II-A-136.

16. "Cost Estimates for Heavy-Duty Gasoline Vehicles," Arcadis Geraghty & Miller, September
1998, EPA Air Docket A-2001-28, Docket Item II-A-77.

17. "Economic Analysis of Diesel Aftertreatment System Changes Made Possible by Reduction
of Diesel Fuel Sulfur Content," Engine, Fuel, and Emissions Engineering, Incorporated,
December 15, 1999, Public Docket No. A-2001-28, Docket Item II-A-76.

18. "Economic Analysis of Diesel Aftertreatment System Changes Made Possible by Reduction
of Diesel Fuel Sulfur Content," Engine, Fuel, and Emissions Engineering, Incorporated,
December 15, 1999, Public Docket No. A-2001-28, Docket Item II-A-76.

19. McDonald and Bunker, "Testing of the Toyota Avensis DPNR at U.S. EPA-NVFEL," SAE
2002-01-2877, October 2002.

20. "Cost Estimates for Heavy-Duty Gasoline Vehicles," Arcadis Geraghty & Miller, September
1998, EPA Air Docket A-2001-28, Docket Item II-A-77.

21. U.S. Department of Labor, Bureau of Labor Statistics,  Producer Price Index Home Page at
www.bls.gov/ppi,  Industry: Motor Vehicle Parts and Accessories, Product: Catalytic
Convenors, Series  Id: PCU3714#503.

22. "Economic Analysis of Diesel Aftertreatment System Changes Made Possible by Reduction
of Diesel Fuel Sulfur Content," Engine, Fuel, and Emissions Engineering, Incorporated,
December 15, 1999, Public Docket No. A-2001-28, Docket Item II-A-76.

23. Johnson Matthey Platinum Today, www.platinum.matthey.com/prices .

24. "Economic Analysis of Diesel Aftertreatment System Changes Made Possible by Reduction
of Diesel Fuel Sulfur Content," Engine, Fuel, and Emissions Engineering, Incorporated,
December 15, 1999, Public Docket No. A-2001-28, Docket Item II-A-76.

25. "Economic Analysis of Diesel Aftertreatment System Changes Made Possible by Reduction
of Diesel Fuel Sulfur Content," Engine, Fuel, and Emissions Engineering, Incorporated,
December 15, 1999, Public Docket No. A-2001-28, Docket Item II-A-76.

26. "Economic Analysis of Diesel Aftertreatment System Changes Made Possible by Reduction
of Diesel Fuel Sulfur Content," Engine, Fuel, and Emissions Engineering, Incorporated,
December 15, 1999, Public Docket No. A-2001-28, Docket Item II-A-76.
                                        6-94

-------
                                           Estimated Engine and Equipment Costs
27. "Economic Analysis of Diesel Aftertreatment System Changes Made Possible by Reduction
of Diesel Fuel Sulfur Content," Engine, Fuel, and Emissions Engineering, Incorporated,
December 15, 1999, Public Docket No. A-2001-28, Docket Item II-A-76.

28. "Economic Analysis of Diesel Aftertreatment System Changes Made Possible by Reduction
of Diesel Fuel Sulfur Content," Engine, Fuel, and Emissions Engineering, Incorporated,
December 15, 1999, Public Docket No. A-2001-28, Docket Item II-A-76.

29. U.S. Department of Labor, Bureau of Labor Statistics, Producer Price Index Home Page at
www.bls.gov/ppi, Industry: Motor Vehicle Parts and Accessories, Product: Catalytic
Convenors, Series Id: PCU3714#503.

30. "Economic Analysis of Diesel Aftertreatment System Changes Made Possible by Reduction
of Diesel Fuel Sulfur Content," Engine, Fuel, and Emissions Engineering, Incorporated,
December 15, 1999, Public Docket No. A-2001-28, Docket Item II-A-76.

31. "Economic Analysis of Diesel Aftertreatment System Changes Made Possible by Reduction
of Diesel Fuel Sulfur Content," Engine, Fuel, and Emissions Engineering, Incorporated,
December 15, 1999, Public Docket No. A-2001-28, Docket Item II-A-76.

32. "Economic Analysis of Diesel Aftertreatment System Changes Made Possible by Reduction
of Diesel Fuel Sulfur Content," Engine, Fuel, and Emissions Engineering, Incorporated,
December 15, 1999, Public Docket No. A-2001-28, Docket Item II-A-76.

33. "Economic Analysis of Diesel Aftertreatment System Changes Made Possible by Reduction
of Diesel Fuel Sulfur Content," Engine, Fuel, and Emissions Engineering, Incorporated,
December 15, 1999, Public Docket No. A-2001-28, Docket Item II-A-76.

34. Czerwinski, Jaussi, Wyser, and Mayer, "Particulate Traps for Construction Machines
Properties and Field Experience," SAE 2000-01-1923, June 2000.

35.  "Electronic Systems and EGR Costs for Nonroad Engines," Final Report, ICF Consulting,
December, 2002, Public Docket No. A-2001-28, Docket Item II-A-10.

36.  "Electronic Systems and EGR Costs for Nonroad Engines," Final Report, ICF Consulting,
December, 2002, Public Docket No. A-2001-28, Docket Item II-A-10.

37.  "Electronic Systems and EGR Costs for Nonroad Engines," Final Report, ICF Consulting,
December, 2002, Public Docket No. A-2001-28, Docket Item II-A-10.

38.  "Final Technical Support Document: Nonconformance Penalties for 2004 Highway Heavy
Duty Diesel Engines," EPA420-R-02-021, August 2002.

39.  "Estimate of the Impact of Low Sulfur Fuel on Oil Change Intervals for Nonroad Diesel
Equipment", memo from William Charmley to Public Docket No. A-2001-28, Docket Item II-A-
194.
                                        6-95

-------
Final Regulatory Impact Analysis
40. "Exhaust and Crankcase Emission Factors for Nonroad Engine Modeling:  Compression
Ignition," NR-009b, November 2002, Air Docket A-2001-28, Docket Item II-A-29; and, the
OTAQ web site for the Nonroad Model and supporting documentation at
www. epa. gov/otaq/nonrdmdl. htm

41. "Estimate of the Impact of Low Sulfur Fuel on Oil Change Intervals for Nonroad Diesel
Equipment", memorandum from William Charmley to Public Docket No.  A-2001-28, Docket
Item II-A-194.

42. "Economic Analysis of Vehicle and Engine Changes Made Possible by the Reduction of
Diesel Fuel Sulfur Content; Task 2 Final  Report: Benefits for Durability and Reduced
Maintenance," ICF Consulting, December 9, 1999, Air Docket A-2001-28, Docket Item II-A-75.

43. "Economic Analysis of Vehicle and Engine Changes Made Possible by the Reduction of
Diesel Fuel Sulfur Content; Task 2 Final  Report: Benefits for Durability and Reduced
Maintenance," ICF Consulting, December 9, 1999, Air Docket A-2001-28, Docket Item II-A-75.

44.  Schenk, C., McDonald, J., and Laroo, C. "High-Efficiency NOx and PM Exhaust Emission
Control for Heavy-Duty On-Highway Diesel Engines - Part Two," SAE 2001-01-3619.

45.  LeTavec, C., et al, "Year-Long Evaluation of Trucks and Buses Equipped with Passive
Diesel Particulate Filters," March 20002, SAE 2002-01-0433.

46.  "Economic Analysis of Diesel Aftertreatment System Changes Made  Possible by Reduction
of Diesel Fuel Sulfur Content," Engine, Fuel, and Emissions Engineering, Incorporated,
December 15, 1999, Public Docket No. A-2001-28, Docket Item II-A-76.

47.  Johnson, T., "Diesel Emission Control: 2001 in Review," March 2002, SAE 2002-01-0285.

48.  "Regulatory Impact Analysis: Heavy-Duty Engine and Vehicle Standards and Highway
Diesel Fuel Sulfur Control Requirements," December 2000, EPA420-R-00-026, Docket Item II-
A-01.

49.  Dou, D. and Bailey, O., "Investigation of NOx Adsorber Catalyst Deactivation"
SAE982594.

50.  Herzog, P. et al, NOx Reduction Strategies for DI Diesel Engines, SAE 920470, Society of
Automotive Engineers 1992 (from Figure 1).

51.  Zelenka, P., et al., "Cooled EGR - A Key Technology for Future Efficient HD Diesels",
SAE 980190.

52. Power Systems Research, OELink Sales Version, 2002.

53. Power Systems Research, OELink Sales Version, 2002.
                                        6-96

-------
                                         Estimated Engine and Equipment Costs
54. "Price Database for New Nonroad Equipment," memorandum from Zuimdie Guerra to EPA
Air Docket OAR-2003-0012, EDOCKET OAR-2003-0012-0960.

55.  "Price Database for New Nonroad Equipment," memorandum from Zuimdie Guerra to EPA
Air Docket OAR-2003-0012, EDOCKET OAR-2003-0012-0960.

56.  "Meeting with Johnson Matthey regarding PGM Recycling," memorandum from Todd
Sherwood to Air Docket A-2001-28, Docket Item IV-E-43, EDOCKET OAR-2003-0012-0877,
March 16, 2004.

57. "Telephone Conversation with Jim Roberts of Multimetco, Inc., regarding PGM Recycling,"
memorandum from Todd Sherwood to Air Docket A-2001-28, Docket Item IV-E-39,  EDOCKET
OAR-2003-0012-0869, March 16, 2004.

58. "Telephone Conversation with Jim Roberts of Multimetco, Inc., regarding PGM Recycling,"
memorandum from Todd Sherwood to Air Docket A-2001-28, Docket Item IV-E-39,  EDOCKET
OAR-2003-0012-0869, March 16, 2004.
                                      6-97

-------
CHAPTER 7: Estimated Costs of Low-Sulfur Fuels
   7.1 Production and Consumption of NRLM Diesel Fuel  	7-2
       7.1.1 Overview	7-2
       7.1.2 Distillate Fuel Production and Demand in 2001	7-6
          7.1.2.1 2001 Distillate Demand  	7-6
          7.1.2.2 2001 Distillate Fuel Production  	7-10
       7.1.3 Distillate Fuel Production and Demand in 2014	7-18
          7.1.3.1 Distillate Fuel Demand in 2014	7-19
          7.1.3.2 Future Distillate Fuel Production	7-22
       7.1.4 Sensitivity Cases	7-45
          7.1.4.1 NRLM Regulated to 500 ppm Indefinitely 	7-46
          7.1.4.2 Proposed Rule - 500 ppm NRLM Cap in 2007; 15 ppm Nonroad Fuel Cap in
             2010	7-47
          7.1.4.3 Final NRLM Fuel Program With Nonroad Fuel Demand Derived from EIA
             FOKS and AEO	7-49
       7.1.5 Methodology for Annual Distillate Fuel Demand: 1996 to 2040	7-62
       7.1.6 Annual Distillate Fuel Demand and Sulfur Content	7-67
          7.1.6.1 Sulfur Content 	7-67
          7.1.4.2 Distillate Fuel Demand and Sulfur Content by Year	7-77
   7.2 Refining Costs	7-86
       7.2.1 Methodology	7-86
          7.2.1.1 Overview	7-86
          7.2.1.2 Basic Cost Inputs for Specific Desulfurization Technologies	7-87
          7.2.1.3 Refinery-Specific Inputs  	7-109
          7.2.1.4 Summary of Cost Estimation Factors  	7-148
          7.2.1.5 Projected Use of Advanced Desulfurization Technologies	7-157
       7.2.2 Refining Costs	7-157
          7.2.2.1 15 ppm Highway Diesel Fuel Program	7-158
          7.2.2.2 Costs for Final Two Step Nonroad Program	7-159
          7.2.2.3 Refining Costs for Sensitivity Cases	7-168
          7.2.2.4 Capital Investments by the Refining Industry	7-174
          7.2.2.5 Other Cost Estimates for Desulfurizing Highway Diesel Fuel  	7-177
   7.3 Cost of Lubricity Additives	7-188
   7.4 Cost of Distributing Non-Highway Diesel Fuel  	7-189
       7.4.1 New Production Segregation at Bulk Plants  	7-190
       7.4.2 Reduction in Fuel Volumetric Energy Content  	7-192
       7.4.3 Handling of Distillate Fuel Produced from Pipeline Interface	7-194
       7.4.4 Fuel Marker Costs	7-200
       7.4.5 Distribution and Marker Costs Under Alternative Sulfur Control Options  . . . 7-205
   7.5 Total Cost of Supplying NRLM Fuel Under the Two-Step Program  	7-206
   7.6 Potential Fuel Price Impacts 	7-208

-------
                                               Estimated Costs of Low-Sulfur Fuels
       CHAPTER 7: Estimated Costs of Low-Sulfur Fuels
   This chapter presents the methodology and costs, and discusses the possible price impacts,
for supplying nonroad, locomotive and marine (NRLM) diesel fuel under the final two step
program. It also presents similar information for various sensitivity cases analyzed. Section 7.1
contains our analysis of the volume of NRLM diesel fuel and other distillate fuels which are
affected by this program. This section also presents our estimates of the sulfur levels of NRLM
diesel fuel and other fuels impacted, which is used in our emissions analysis. Section 7.2
discusses our methodology for estimating the refining costs. We present our refining cost
estimates for the final rule program as well as several sensitivity cases. We also compare our
cost estimates to other parties. Section 7.3 contains our estimate of the cost of adding lubricity
additive to NRLM diesel fuel. Section 7.4 presents our analysis of the cost of distributing diesel
fuel under this program. Section 7.5 contains a summary of the refining and distribution cost for
the final rule NRLM program. Section 7.6 discusses the potential price impacts of the final
NRLM program.

   Table 7-1 summarizes the number of refineries we estimate will be affected by the final
NRLM fuel program, as well as the total volume of NRLM fuel affected.

                                      Table 7-1
           Number of Refineries and Refining Costs for the Final NRLM Program

Number of Refineries Producing
500 or 15 ppm NRLM Diesel
Fuel
Production Volume
(Million gallons per year in 2014)
Year of
Program
2007-2010
2010-2012
2012-2014
2014-2020
2007-2010
2010-2012
2012-2014
2014-2020
500 ppm Fuel
All Refineries
36a
26
15
0
13,327
3,792
728
0
Small
Refineries
0
13
13
0
0
393
393
0
1 5 ppm Fuel
All
Refineries
0
32
47
63
0
8,598
12,247
13,030
Small
Refineries
0
2
2
15
0
335
335
728
   Table 2 summarizes the per gallon refining, distribution and lubricity additive costs during
the various phases of the final NRLM fuel program.
                                         7-1

-------
Final Regulatory Support Document
                                       Table 7-2
       Summary of Fuel Costs for NRLM Fuel Control Options (cents per gallon, $2002)
Option
Final Rule
Specification
500 ppm NRLM
5 00 ppm NRLM
5 00 ppm NRLM
1 5 ppm Nonroad
15 ppm NRLM
15 ppm NRLM
Year
2007-10
2010-12
2012-14
2010-12
2012-14
2014+
Refining
Costs
(c/gal)
1.9
2.7
2.9
5.0
5.6
5.8
Distribution &
Additive Costs
(c/gal)
0.2
0.6
0.6
0.8
0.8
1.2
Total
Costs
(c/gal)
2.1
3.3
3.5
5.8
6.4
7.0
   Table 7-3 and 7-4 summarize the potential price impacts of the final NRLM fuel program
during the initial 500 ppm phase (2007-2010) and the final 15 ppm phase (2014 and beyond).
Due to the uncertainty in projecting price impacts from cost estimates, we develop three
potential price impacts to indicate the range of possible outcomes.

                                      Table 7-3
          Range of Possible Total Diesel Fuel Price Increases (cents  per gallon)"

Lower Limit
Mid-Range Estimate
Upper Limit
500 ppm Sulfur Cap: Nonroad, Locomotive and Marine Diesel Fuel (2007-2010)
PADDs 1 and 3
PADD2
PADD4
PADDS
2.9
3.0
3.7
1.2
1.8
2.5
3.5
1.5
4.5
3.8
6.1
1.5
15 ppm Sulfur Cap: NRLM Fuel (fully implemented program: 2014 +)
PADDs 1 and 3
PADD2
PADD4
PADDS
7.7
7.6
8.2
5.1
6.3
7.9
13.0
6.8
9.8
11.2
13.9
7.2
 a  At a wholesale price of approximately $1.00 per gallon, these values also represent the percentage increase in
 diesel fuel price.
7.1 Production and Consumption of NRLM Diesel Fuel

7.1.1 Overview

   This subsection describes how we estimated the distillate fuel production and demand for
land-based nonroad engines, locomotives, and marine vessels that will be affected by the
requirements of this final rule.  This analysis also estimates the volumes of the highway diesel
                                         7-2

-------
                                                  Estimated Costs of Low-Sulfur Fuels
fuel and heating oilA pools which also affect or are affected by the final NRLM fuel program.
Fuel production and demand are estimated for various geographic regions of interest.  We begin
by estimating production and consumption of various distillate fuels in 2001. We then project
these volumes to 2014, which is the year in which we project per gallon costs. We selected
2014, as IRS guidelines allow refinery equipment to be depreciated over 15 years and 2014
represents the mid-point in the depreciation life of new hydrotreaters built for the 2007 500 ppm
NRLM fuel cap. NRLM fuel demand is projected to increase steadily in the future. As the
number of domestic refineries is not projected to increase, the economy of scale will gradually
improve over time.  Selecting 2014 as the year in which to project per gallon fuel costs provides
a reasonable estimate of the average economies of scale which will  exist with the hydrotreaters
constructed in response to the rule.

   These NRLM production and consumption estimates are developed for the final NRLM fuel
program, as well as for a number of alternative scenarios. We then  develop a set of production
and consumption estimates for NRLM fuel for each year from 1996 to 2040, which are used to
estimate annual emission reductions (see Chapter 3) and fuel-related costs (Sections 7.2 through
7.5 below).  Finally, we estimate how the final rule and the various  alternative scenarios affect
the sulfur content of the various types of distillate fuel, which is again used to estimate annual
emission reductions associated with each  of these scenarios.

   It is important early on in this discussion to define distillate fuel and how it is used.
Distillate fuel is often split into three groups according to the range  of temperatures at which the
hydrocarbons comprising the fuel boil (boiling range). No. 1 distillate fuel is the lightest fuel, or
has the lowest boiling range.  Common No. 1 distillate fuels are jet  fuel, No.  1 diesel fuel, and
kerosene (also known as No.  1 fuel oil). No. 2 distillate fuel is somewhat heavier and has a
higher boiling range, though there is significant overlap between No. 1 and No. 2 distillate fuels.
No. 2 distillate fuels are usually excellent diesel fuels. Finally, No.  4 distillate fuel is the
heaviest of the three, having the highest boiling range.B No. 4 distillate fuel is generally a poor
diesel fuel and can only be used in slower speed diesel engines.  This rule does not address the
sulfur content of No. 4 distillate fuel.  Thus, we will not address No. 4 distillate fuels in this
analysis.  All of these distillate fuels boil at higher temperatures than gasoline, though there is
some overlap between the heaviest compounds in gasoline and the lightest compounds in No. 1
distillates.

   The vast majority of the fuel used in NRLM engines falls into the No. 2 distillate fuel
category.  As will be seen below, a very small volume of No. 1  distillate  fuel is used to fuel
   A The term heating oil as used here represents fuel used for stationary source purposes including home heating
industrial boilers, and electrical generation.

   B There is also a No. 6 fuel, but this is usually considered a heavy fuel or heavy oil and not included in
"distillate."

                                           7-3

-------
Final Regulatory Support Document
NRLM engines.0  Also No. 1 distillate fuel is often blended into No. 2 distillate fuels in the
winter in cold climates to avoid fuel gelling. Thus, we will address the impact of this rule on No.
1 distillate fuel in this analysis, though the primary focus will be on No. 2 distillate fuels.

   The American Society of Testing and Materials (ASTM) defines three No. 2 distillate fuels:
1) low sulfur No.  2-D, 2) high sulfur No. 2-D, and 3) No. 2 fuel oil. Low sulfur No. 2-D fuel
must contain 500  ppm sulfur or less, have a minimum cetane number of 40, and have a minimum
cetane index limit of 40 (or a maximum aromatic content of 35 volume percent).  These
specifications match those set by EPA for highway diesel  fuel, so essentially these ASTM limits
are legal  specifications. Per ASTM, both high sulfur No. 2-D  and No. 2 fuel oil (heating oil)
must contain no more than 5000 ppm  sulfur,0 and currently averages about 3000 ppm.  The
ASTM specifications for high sulfur No. 2-D fuel also include a minimum cetane number
specification of 40.  The ASTM  specifications  for high sulfur No. 2-D and No. 2 fuel oil only
have the force of law in those  states which have incorporated the ASTM standards in their state
laws or regulations.  There are no federal standards currently for these two high sulfur fuel.

   We will break down No. 2-D distillate fuel into three fuels, according to the way we regulate
its quality: highway diesel fuel, NRLM diesel fuel, and heating oil. Operators of highway diesel
engines must use low sulfur highway diesel fuel engines, though the low sulfur fuel can be and is
used in other applications. As will be discussed further below, highway diesel fuel must
currently meet a 500 ppm sulfur cap.  Starting in 2006, 80% of highway diesel fuel volume will
have to meet a 15 ppm cap, with 100% having  to do so in  2010.  NRLM diesel fuel is that fuel
used in nonroad, locomotive and marine diesel engines and is the fuel primarily affected by this
rule. Heating oil is all other No. 2 distillate fuel. It includes No. 2 fuel oil used in boilers,
furnaces and turbines. It also  includes No. 2 diesel fuel used in stationary diesel engines (e.g.,
for electricity generation). Heating oil is not covered by the NRLM fuel standards, but is
affected because of limitations in the fuel distribution system.

   We base our estimates of historical distillate fuel demand used in this analysis on EPA's
Nonroad Model (NONROAD) and the Energy Information Administration's (EIA) Fuel Oil  and
Kerosene Sales (FOKS) report for 2001.  NONROAD estimates diesel fuel consumption by the
land-based nonroad  engines based on  the sales, scrappage and use of nonroad engines.   FOKS
contains detailed, comprehensive distillate fuel sales to highway vehicles and ten non-highway
sectors. We use FOKS to estimate the consumption of highway, marine, and locomotive diesel
fuel and heating oil, given the nonroad diesel fuel consumption from NONROAD.

   We base future demand for nonroad diesel  fuel again on estimates from NONROAD. Future
demand for highway diesel fuel and the other non-highway sectors (locomotive, marine and
heating oil) is based on estimates from EIA's Annual Energy Outlook (AEO) for 2002.
   c No. 1 distillate fuels is mostly consumed in jet engines and tends to cost more than No. 2 distillate fuels.
Since diesel engines can burn either fuel, No. 2 distillates are their preferred choice.

   D Some states, particularly those in the Northeast, limit the sulfur content of No. 2 fuel oil to 2000 - 3000 ppm.

                                           7-4

-------
                                                Estimated Costs of Low-Sulfur Fuels
   The methodology used for the final rule differs somewhat from that used in the NPRM.  For
the NPRM, we used different methodologies to estimate distillate fuel demand for the purpose of
estimating emissions and for estimating fuel-related costs. For emissions, we used a
methodology very similar to that being used for this final rule. However, for fuel cost
estimation, we did not use NONROAD to estimate nonroad fuel consumption. We derived all of
our fuel consumption estimates from FOKS and AEO, although we projected future nonroad fuel
consumption with NONROAD. To avoid this inconsistency, we  decided to utilize the same
methodology for both emission and cost estimation purposes.  As discussed in Section 2.3.2.2 of
the Summary and Analysis document for this rule, we decided to use NONROAD to estimate
nonroad fuel consumption for both emission and cost estimation purposes.  In addition, the
analysis for this final rule utilizes more recent information from FOKS 2001 and AEO 2002, as
opposed to FOKS 2000 and AEO 2001, which were used in the analysis for the NPRM.

   We estimate historic production of distillate fuel in these pools by starting with downstream
demand.  We used Information from EIA's Petroleum Supply Annual on the sales of highway
diesel fuel and high sulfur distillate from refinery racks and terminals. The volume of highway
diesel fuel supplied at terminals is compared to that consumed in highway vehicles to estimate
the percentage of highway fuel which is used in other applications. We call highway fuel used in
other applications "spillover." We then adjust the terminal level supply of highway diesel fuel to
represent shifts in the volume of various fuels during distribution, particularly through pipelines.
These shifts are referred to as "downgrades."  The result is an estimate of production needed by
refineries and importers to supply demand in the various sectors.

   The sulfur level of the various distillate  fuels produced at refineries is primarily controlled by
applicable EPA standards.  These of course vary depending on the regulatory scenario being
evaluated. We also consider the impact of the small refiner provisions, which usually allow the
sale of higher sulfur fuel into a particular market than would otherwise be the case.  The
spillover of highway fuel into non-highway sectors also affects the sulfur content of these fuels,
as do the downgrades that occur during distribution.  Our estimate of in-use sulfur levels of the
various distillate fuels begins with in-use survey data and then adjusts these levels for changes in
the sulfur content of fuel being produced, spillover and downgrades  during distribution.

   The two primary regulatory scenarios evaluated are: 1) a reference case, which assumes no
NRLM sulfur standards and 2) the final NRLM fuel program. In addition, we evaluate several
sensitivity cases:

   NRLM control only to 500 ppm in 2007 (no second step to 15 ppm),

   nonroad fuel control to 15 ppm in 2010, but keeping locomotive and marine (L&M) fuel at
   500 ppm indefinitely (the proposal or NPRM case),E and
   E The increment of the final rule program to this regulatory scenario is the basis for our 500 ppm to 15 ppm
locomotive and marine incremental analysis.

                                          7-5

-------
Final Regulatory Support Document
   the final NPRM fuel program with the volume of nonroad diesel fuel derived from FOKS
   and AEO 2003 instead of NONROAD.

7.1.2 Distillate Fuel Production and Demand in 2001

   This section describes our estimates of total production and demand by region for the various
distillate fuels. The primary regions of interest are the different refining districts called PADDs.F
There are five PADDs: 1) the East Coast, 2) the Midwest, 3) the Gulf Coast, 4) the Mountain
states and 5) the West Coast, Alaska and Hawaii. Because the Alaskan and Hawaiian fuel
markets are mostly distinct from the rest of PADD 5 and because California applies distinct
specifications to diesel fuel sold in that state, we split PADD 5 into four pieces: the states of
California,  Hawaii and Alaska and the remainder of PADD 5. We will refer to this remainder of
PADD 5 as PADD 5-O (with "O" denoting "other" than the specific states listed).

   We begin with estimating the demand for each type distillate fuel, highway, NRLM and
heating oil. We then estimate how much highway fuel was supplied at the terminal level to
estimate spillover of highway fuel into the other sectors. Finally, we estimate downgrade of
higher quality fuels to lower quality fuels during distribution to back-calculate the volume of
each fuel produced by refineries.

   7.1.2.1  2001 Distillate Demand

   We obtain our estimate of total distillate demand from EIA's FOKS report for 2001.l This
report presents results of a national statistical survey of approximately 4,700 fuel suppliers,
including refiners and large companies that sell distillate fuels for end use (rather than resale).
The sample design involves classification of fuel suppliers based on sales volume with
subsamples in individual classes optimized to improve sample precision. Distillate fuels
surveyed that are relevant to this analysis include diesel and heating oils in grades No. 1, No. 2
and No. 4.  The survey requests respondents to report estimates of fuel sold for eleven "end
uses" that correspond to broad economic  sectors. These eleven sectors are highway, industrial,
off-highway (construction and other), farm, military, railroad, marine vessel, commercial,
residential, oil  company and electric utility.  Suppliers presumably determine the applicable
sector by the type of entity which purchases the fuel (e.g., farmers buy fuel for farming).  FOKS
is therefore not a direct measure of how fuel  is used, but a measure of who buys fuel. However,
for most of these sectors it should provide a reasonable estimate. The reader is referred to
Section 2.3.2.2 of the Summary and Analysis document for this rule for a more detailed
description of FOKS  and the fuel user surveys which provide an independent assessment of its
accuracy.

   FOKS presents two sets of fuel demand estimates.  The first, labeled unadjusted, includes
adjustments to reflect estimates of highway fuel use from the Federal Highway Administration.
   F The Department of Energy split up the nation into five districts, called Petroleum Allocation for Defense
Districts, or PADDs, during the 1970's. The regions primarily reflect where refineries get their crude oil.

                                          7-6

-------
                                                 Estimated Costs of Low-Sulfur Fuels
The second, labeled adjusted, includes further adjustments to reflect distillate fuel use to
generate electricity and to match total distillate demand to total distillate fuel supply, as
estimated in EIA's Petroleum Supply Annual (PSA).  EIA's PSA reports an aggregation of the
volumes of fuels sold by primary suppliers, which includes refinery racks and terminals.  As the
PSA figures represent recorded sales from all primary suppliers, and not a survey of
representative suppliers, it is a more accurate estimate of total distillate fuel supply than the total
demand estimated in FOKS. Because of this, we use the adjusted FOKS demand estimates here.
Thus, while we refer to total distillate fuel demand as being taken from FOKS, it is just as
accurate to say that it comes from PSA.

   Of the eleven economic sectors evaluated by FOKS, we are interested primarily in three:
highway, railroad and marine vessels. Little fuel used in these sectors involves nonroad
equipment or heating oil. The remaining eight sectors all include significant portions of nonroad
fuel use and heating oil use. Because of this, we use the EPA NONROAD model to estimate
nonroad fuel use and assume that the remainder is heating oil.

   Table 7.1.2-1 shows total distillate fuel demand from the 2001 FOKS report, as well as total
demand for highway, railroad and marine fuel from this same report.0 Nonroad diesel fuel
demand was taken from the draft NONROAD2004 model (see Chapter 3 for a detailed
description of this model). Heating oil demand was set so that the total fuel demand from the
five sectors equaled total fuel demand.

                                       Table 7.1.2-1
                 Total Distillate Demand in 2001 by Region (million gallons)
End Use
Highway
Railroad
Marine
Other
Nonroad
Heating
Oil
Total Demand
Region
1
10,284
506
461
2,935
7,363
21,549
2
10,947
1,051
318
4,174
602
17,092
3
5,743
883
1,153
1,409
1,744
10,932
4
1,570
223
0
597
78
2,468
5-O*
1,901
100
23
631
45
2,700
AK
111
4
67
25
205
412
HI
33
0
20
32
129
214
CA
2,627
183
52
783
(41)
3,604
*  Represents the states of AZ, NV, OR, and WA.

   For this analysis, we made several small modifications to the fuel demand estimates shown in
2001 FOKS. We made one adjustment to the estimate of highway fuel demand. FHWA
   G Since the volume of No. 4 distillate fuel is small compared to total distillate use, we did not attempt exclude
No. 4 distillate use from the 2001 FOKS estimate of total distillate demand. Because of the methodology used, any
incremental volume of No. 4 distillate fuel shows up as heating oil demand in Table 7.1.2-1.
                                           7-7

-------
Final Regulatory Support Document
estimates highway fuel demand based on fuel excise tax receipts. Individuals and businesses that
purchase highway fuel for off-highway use can request a refund of this excise tax on their income
tax forms. FHWA adjusts their estimates for these refund requests.  However, it is possible that
not everyone who uses taxed, highway diesel fuel for non-highway use files for a refund. For
example, many businesses own fleets of both highway and nonroad equipment. Some owners or
operators, particularly rentals, might find it expedient or necessary to purchase at least some of
their nonroad diesel fuel at retail outlets such as gas stations, where high sulfur diesel fuel is
usually not available. It is plausible that some fraction of the fuel attributed by FHWA to
highway use is actually used for non-highway purposes. This fuel would likely be used by
construction and commercial  nonroad equipment users, as they are the most likely to refuel their
nonroad engines at retail fuel  outlets.

   To gain a better understanding of this issue, EPA provided a grant to the Northeast States for
Coordinated Air Use Management (NESC AUM) to conduct a survey of diesel fuel use in
construction equipment in New England.2 The survey was designed to develop methods to
estimate emission inventories for construction equipment.  The study area included two counties,
one in Massachusetts and one in Pennsylvania. Equipment owners in selected sectors were
targeted, including construction, equipment rental, wholesale trade,  and government (local
highway departments).  Surveyors administered a questionnaire requesting information about fuel
purchases and associated tax-credits.  Owners reported quantities and proportions of high-sulfur
(dyed and untaxed) and low-sulfur (undyed and taxed) diesel fuel purchased over the previous
year.  Owners who reported purchases of undyed diesel fuel for use in construction equipment
were also requested to indicate whether they applied for tax credits for which they were eligible
under state or federal law. The survey showed that approximately 20 percent  of all diesel fuel
purchased for use in "construction" was undyed diesel fuel for which the purchaser had not
applied for a tax refund.

   To ensure that this type of adjustment was not already included in the FOKS estimates, we
confirmed with FHWA that they only subtract tax refunds from the total tax receipts from
highway diesel fuel sales.3'4 In other words, they assume that all purchasers of taxed diesel fuel
for non-highway use request a refund. Similarly, we confirmed with EIA that they do not make a
similar type of adjustment.5

   To estimate the volume of nonroad diesel fuel classified as highway fuel demand in FOKS,
we applied the results of the NESCAUM survey to the FOKS  estimates of construction fuel
demand plus a portion of commercial fuel demand.  As discussed in Section 7.1.3. below, fuel
demand in the commercial sector is broken out by the type of distillate purchased. One  of these
fuel types is high sulfur diesel fuel, which we believe is primarily used in nonroad equipment.
We believe that the results of the NESCAUM are equally applicable to these types of nonroad
equipment, as they tend to be used away from the business' primary location (e.g., lawn and
garden equipment). However, because the survey only covered two counties, the results are not
necessarily representative of the entire U.S.  Extrapolating the results to the entire U.S. is
therefore uncertain. Given that we lack any other estimate, we decided to use the results of the
NESCAUM survey with an ad hoc adjustment, where the percentage of unrefunded highway fuel
used is assumed to be 10%, as opposed to the surveyed 20%.

                                           7-8

-------
                                               Estimated Costs of Low-Sulfur Fuels
   Table 7.1.2-2 shows the volume of construction and commercial, high sulfur diesel fuel, and
the portion believed to be made up from unrefunded highway fuel by region. We reduced the
total construction volume by 5% to not base our estimates of unrefunded fuel on that portion
which is estimated to be used as heating oil (see below).  On a nationwide average, this
unrefunded highway fuel represents 0.7% of total highway fuel demand. As will be shown
below, we reduce the volume of highway fuel demand in each region by the volume shown in
Table 7.1.2-2.

                                     Table 7.1.2-2
    Unrefunded Use of Taxed Highway Fuel in Nonroad Equipment in 2001 (million gallons)


Total Construction*
Nonroad Portion (0.95)
Unrefunded Fuel (10%)
Commercial: #2 High Sulfur
Diesel Fuel *
Unrefunded Fuel (10%)
Total Unrefunded Fuel
Region
1
550
523
52
203
20
73
2
602
572
57
155
16
73
o
5
448
425
43
71
7
50
4
124
118
12
8
1
13
5-0
87
83
8
19
2
10
HI
4
3
0.3
2
0.2
1
AK
7
7
0.7
21
2
3
CA
264
251
25
3
0.3
25
*  FOKS 2001

   While we believe that this highway fuel is used in nonroad engines, we did not increase the
nonroad fuel demand shown in Table 7.1.1-1 above. This adjustment is not necessary since the
NONROAD model projects fuel use for the entire in-use nonroad equipment fleet and does not
consider where the fuel is purchased. As will be seen below, the result is that this reduction in
highway fuel demand causes an analogous increase in the demand for heating oil under our
methodology.

   We also made minor adjustments to the FOKS estimates for diesel fuel demand for
locomotive engines and marine vessels.  Based on guidance from EIA staff, 5% of the fuel
purchased by railroads is heating oil, under our definitions described above.6  Thus, we reduced
the railroad fuel demand from FOKS by  5%.  We further reduced the railroad fuel demand by an
additional 1%, which represents  fuel believed to be used in nonroad diesel engines in railyards
and which is already included in the nonroad fuel demand estimates from NONROAD.7 The
FOKS estimates of fuel demand  for marine vessels were multiplied by 90%, to remove the use of
heating oil and No. 4 distillate fuel included in the FOKS estimates. Again, this was based on
guidance from EIA staff.8

   Table 7.1.2-3 shows the FOKS and NONROAD estimates of distillate fuel demand, the
adjustments made and the final estimates.  Only the revised estimate of heating oil demand is
                                         7-9

-------
Final Regulatory Support Document
shown, as this is simply back-calculated from the total demand for the other fuels and total
distillate demand.

                                      Table 7.1.2-3
               Adjusted Distillate Demand by Region in 2001 (million gallons)
End Use
FOKS Highway
Unrefunded fuel (0.7%)
Revised Highway
FOKS Railroad
Revised Railroad
FOKS Marine
Revised Marine
Nonroad
Heating Oil
Total
Region
1
10,284
73
10,211
506
476
461
415
2,935
7,511
21,549
2
10,947
73
10,873
1,051
989
318
286
4,174
769
17,092
o
3
5,743
50
5,694
883
831
1,153
1,037
1,409
1,961
10,932
4
1,570
13
1,557
223
209
0
0
597
105
2,468
5-O
1,901
10
1,890
100
94
23
20
631
64
2,700
AK
111
3
108
4
4
67
60
25
214
412
HI
33
1
32
0
0
20
18
32
132
214
CA
2,627
25
2602
183
172
52
46
783
0
3,604
   7.1.2.2  2001 Distillate Fuel Production

   Refiners do not produce exactly the same volume of fuel which is consumed. This is
especially true for the specific categories of distillate fuel. The largest difference occurs with
highway diesel fuel.  All fuel used in highway diesel engines must meet EPA's 500 ppm sulfur
cap.  Other distillate fuel does not. However, fuel meeting the highway diesel fuel specification
can be used in the other four categories. As is shown below, this occurs to a significant extent.
We refer to this as spillover.  Thus, the production of highway diesel fuel tends to be much larger
than  is actually consumed in highway diesel engines. More importantly for this rule, the highway
fuel used in NRLM engines already meets the sulfur caps of the final NRLM fuel program. Thus,
this spillover fuel faces no new production or distribution costs due to this rule.

   Also, a  certain amount of mixing occurs when fuel is shipped in pipelines, particularly at the
interface between fuel batches. The properties of this interface material are a blend of the
properties of the two distinct fuel batches. Generally, this interface material does not meet the
specification of one of the two fuels and is cut into the batch of the lower quality fuel.  We refer
to the volume of the higher quality fuel that is lost to the lower quality fuel as downgrade.
However, sometimes this interface does not meet the specifications of either fuel and has to be
segregated from both batches and reprocessed. This downgraded material is referred to as
transmix.
                                          7-10

-------
                                                Estimated Costs of Low-Sulfur Fuels
   Downgrade can both increase and decrease the supply of distillate fuel relative to that which
was produced by refineries.  We consider these changes in the supply various distillate fuels
below when estimating the cost of providing NRLM fuel meeting the final NRLM sulfur
standards.

   Spillover

   Spillover is the volume of highway diesel fuel supplied which exceeds highway diesel fuel
demand and is thus used by off-highway users. We estimate spillover volume by subtracting
diesel fuel consumption by highway vehicles from the total supply of low-sulfur, highway fuel.
We already estimated highway fuel consumption by highway engines (see Table 7.1.2-3 above).
We obtain highway fuel supply to each region from EIA's Petroleum Marketing Annual 2001.9 It
should be noted that PMA estimates distillate fuel supply from primary suppliers, which are
primarily refinery racks and terminals. Thus, any downgrades occurring in pipelines have already
occurred. However, fuel sales by transmix processors are included in PMA. Thus, any distillate
fuel recovered from transmix processing is also included in PMA. Table 7.1.2-4 shows the
spillover volumes in each region based on the above information.

                                      Table 7.1.2-4
                      Highway Fuel Spillover in 2001 (million gallons)

Total Supply
Highway Engine Demand
Spillover
1
10,596
10,211
385
2
12,549
10,873
1,676
3
6,532
5,694
838
4
2,067
1,557
510
5-O
2,206
1,890
316
AK
111
108
3
HI
45
32
13
CA
3,568
2,602
966
U.S.
37,674
32,967
4,707
   Information on the use of this spillover of highway fuel in the individual nonroad, locomotive,
marine, and heating oil markets does not exist.  Therefore, we assume that this spillover
represents the same percentage of total demand for each fuel category within a region. Table
7.1.2-5 shows spillover, total non-highway distillate demand, and the percentage of spillover to
non-highway distillate demand by region.

                                      Table 7.1.2-5
     Spillover As Percentage of the Non-Highway Distillate Demand, 2001  (million gallons)

Spillover
Non-Highway
Distillate Demand
Spillover (% of Non-
Highway Demand)
1
385
11,337
3.4
2
1,676
6,218
26.9
3
838
5,238
16.0
4
510
911
55.9
5-O
316
809
38.9
AK
3
303
1.0
HI
13
182
7.1
CA
9
1,001
100
   As can be seen, the degree of spillover varies widely across the U.S.  Spillover is very low in
Alaska and Hawaii, because of the absence of fuel product pipelines.  Spillover is also very low in
                                          7-11

-------
Final Regulatory Support Document
PADD 1, because of its large demand for high sulfur heating oil.  This large demand causes high
sulfur distillate to be available nearly everywhere, particularly in the northern portion of PADD 1.
Thus, there is little reason for highway fuel to be used in non-highway applications. Spillover is
relatively high in PADD 4 due to the fact that several pipelines in the region do not carry high
sulfur distillate. Finally, spillover is very high in California, as that State requires the use of 500
ppm fuel in nonroad engines.

    The final issue is the distribution of this spillover into the four high sulfur distillate markets:
nonroad, locomotive, marine, and heating oil.  Differences do exist in the way that these fuels are
typically shipped, particularly for locomotive and marine fuel.  This could affect the relative
volume of spillover added to that market. However, data are not available which indicate any
difference in the distribution of spillover. Thus, except for the unrefunded use of highway fuel in
the construction and commercial  sectors, we assume that the spillover is distributed into the four
high sulfur distillate markets in proportion to their total demand.  Consistent with the way the
NESCAUM survey was conducted, we assume that the portion of spillover coming from
unrefunded use of highway fuel is all nonroad fuel  demand.

    Downgrade

    When fuel is shipped through pipelines, the batch of one fuel flows immediately next to a
batch of another fuel.  As the fuel flows through the pipeline, the two fuels  start to mix at the
interface of the two batches.  This interface takes on a character of its own and its properties are a
blend of the properties of the two fuels. The mixture is commonly called interface material or
simply interface. Depending on the properties of the two fuels  and the stringency  of the
specifications what each fuel must meet, this interface material can simply be cut in half and
blended into the two batches of fuel.  In this case, there is no loss of volume in either batch.
However, usually one of the two fuels is of higher quality than the other and the interface  is
blended into the lower quality batch. In this case, the lower quality fuel gains volume, while the
higher quality fuel loses volume.  This loss of volume is called  downgrade.

    The loss of higher quality fuel volume through  downgrade means that more of this fuel must
be produced than implied by demand.  Likewise, the gain of lower quality fuel volume through
downgrade means that less of this fuel  must be produced than implied by demand.  The latter is
particularly important after the control of NRLM fuel sulfur content, as heating oil demand (a
sink for high sulfur downgrade) in some of the regions is quite limited. Also, the sulfur content of
downgrade will differ from that of fuels produced at refineries.  Thus, the relative volume of
downgrade being sold in each fuel market will affect the average in-use sulfur content of that fuel
and the emission reductions resulting from  this NRLM rule.

    Figure 7.1-1 shows the order in which petroleum fuels are typically shipped through pipelines
today.10 Jet fuel is often "wrapped" with high  sulfur distillate and highway diesel fuel.  The sides
of the batches of high sulfur distillate and highway diesel fuel not adjacent to jet fuel are often
adjacent to gasoline of some type. The order of fuels can vary from pipeline to pipeline.
However, the specific order will generally not affect the volumes and quality of downgrade
estimated here. According to our methodology, the size of the various interfaces are generally

                                           7-12

-------
                                                  Estimated Costs of Low-Sulfur Fuels
independent of the adjacent fuels and any distillate fuel lost to transmix is recovered by transmix
processors. The only difference might be the percentage of downgraded distillate which is able to
be sold to the 500 ppm highway fuel market versus the high sulfur distillate market. While this
breakdown affects current fuel supply, it is not an issue once diesel fuel must meet a 15 ppm cap.
                  Figure 7.1-1 Pipeline Sequence and Fate of
               the Interface Between Fuel Pipeline Batches in 2001
 500 ppm
1.75% Jet
2.2% Hwy
Gasoline in equal
amounts
    At the interface between these different fuels there is a mixing zone which results in the two
fuels contaminating each other.  There are two different ways this mixed fuel between the two
fuels is dealt with by the pipeline companies. One way that pipeline companies deal with the
interface between the two fuels is to simply downgrade the mixture into the batch of fuel with the
lowest quality. Pipeline companies have informed us that the entire interface zone between jet
fuel and highway diesel fuel and also the interface zone between jet fuel and high sulfur distillate
is simply "cut" into the batches of highway diesel fuel and high sulfur distillate, respectively, by
timing their valve actions.  This can occur because jet fuel would generally comply with the
specifications of the other two pools.H

    The second way to handle this interface occurs when the specifications governing the quality
of each fuel prevents the interface from being blended into  either fuel.  This always occurs
between a batch of gasoline and a batch of any distillate fuel. Even a small amount of gasoline
would cause diesel fuel to exceed its flashpoint limit.  Similarly, a small amount of diesel fuel
would cause gasoline to exceed its endpoint limits.  In this case, the interface is commonly
referred to as transmix. Transmix must be separated from either batch, is usually  stored in a
transmix tank with other types of transmix, and then shipped to a transmix processor. The
    H The sulfur content of jet fuel often exceeds 500 ppm. However, adding a small volume jet fuel to highway
diesel fuel usually will not cause the sulfur content of the highway diesel fuel to exceed 500 ppm.

                                            7-13

-------
Final Regulatory Support Document
physical characteristics of pipeline mixing indicate that the interface would generally contain
roughly even quantities of gasoline and distillate. We assume that this is the case here.

   The transmix processor distills the transmix to produce a reprocessed gasoline and distillate
fuel. However, there is some overlap between the lower temperature boiling components of
distillate, particularly jet fuel and the higher temperature boiling components of gasoline. The
lower temperature boiling components of distillate have a particularly low octane number. If any
significant quantity of distillate is mixed with the gasoline product, the cost of raising the octane
number to back to 87 or higher is economically prohibitive.  Therefore, transmix processors
operate their distillation columns so that roughly one-third of the original gasoline contained in
the transmix leaves with distillate product.

   We are not concerned with the gasoline produced by transmix processors here. However, the
gasoline portion of the original transmix which enters the distillate pool in this fashion affects
both the volume and sulfur content of the distillate fuel pool  and is, thus, relevant to this
discussion.

   The distillate portion of current transmix can consist of highway diesel fuel, jet fuel and high
sulfur distillate, plus the heaviest components of gasoline.  Because most pipelines carry high
sulfur distillate fuel currently and jet fuel often exceeds 500 ppm sulfur, and because most
facilities have only one tank for storing transmix from all interfaces, we assume that the distillate
produced from transmix is usually sold as high sulfur distillate. Thus, per Figure 7.1-1, the
highway diesel fuel portion of transmix is shifted to high sulfur distillate supply.

   The next step in our assessment of downgrade is to estimate its volume.  The jet fuel
downgrade is easiest to estimate because, assuming the shipping order shown in Figure 7.1-1, it is
simply cut into each adjacent pool.  We polled several pipeline companies to obtain an estimate
on the quantity of jet fuel downgraded today. Their estimates of the volume of jet fuel
downgraded during distribution ranged from 1% to 7%.n We assumed that the national average
downgrade percentage was near the mid-point of this range, or3.5%. Per Figure 7.1-1, half of
this volume is shifted to the highway fuel market and half is  shifted to the high sulfur distillate
market. Table 7.1.2-6 shows this shift.
                                           7-14

-------
                                                  Estimated Costs of Low-Sulfur Fuels
                                       Table 7.1.2-6
                       Types of Downgrade and Their Volumes in 2001
Interface
Jet Fuel
Interface
Gasoline -
High Sulfur
Distillate
Interface
Gasoline -
Highway
Diesel Fuel
Interface
Original
Fuel
Jet Fuel
High Sulfur
Distillate
Gasoline
Highway
Diesel
Gasoline
Destination
Highway Diesel Fuel
High Sulfur Distillate
High Sulfur Distillate
High Sulfur Distillate
High Sulfur Distillate
High Sulfur Distillate
Volume
1.75% of jet fuel demand
1.75% of jet fuel demand
Neutral
Equivalent to 0.58% of jet fuel demand
2.2% of highway diesel fuel supply
Equivalent to 0.73% of highway diesel fuel supply
    The other downgrades occur through the creation of transmix and its processing.  Starting
with high sulfur distillate fuel, some of the volume of this fuel is lost to transmix. However,
transmix processors return all of the distillate portion of the original transmix to their distillate
product. As stated above, we assume that all the distillate produced by transmix processors
contains more than 500 ppm sulfur and is sold to the high sulfur distillate market. Thus, the
volume of high sulfur distillate which is lost to transmix is eventually returned to the high sulfur
distillate market by transmix processors. The result is no net loss or gain in the high sulfur
distillate market through its mixture with gasoline. This is shown in Table 7.1.2-6.

    While the high sulfur distillate portion of this transmix returns to the fuel pool from which it
came, the gasoline which abuts high sulfur distillate in the pipeline does not all return to gasoline
supply. The heaviest portion of this gasoline moves from the gasoline market to the high sulfur
distillate market.  We were not able to obtain a direct estimate of the volume of gasoline lost in
this manner or the volume of high sulfur distillate shifted to transmix. Thus, we  estimate this
volume by  comparing it to the volume of jet fuel moved to the high sulfur distillate pool.  As
mentioned  above, the mixing properties of all these fuels are fairly similar. They also have
flowed through the pipeline over the same distance (i.e., all these fuels are major products which
tend to flow the entire length of the pipeline).  Thus, it is reasonable to assume that the interface
on either side of the batch of high sulfur distillate has the same volume. If 1.75% of jet fuel is
lost to high sulfur distillate on one side of the batch, then the  same volume of high sulfur distillate
will be lost to transmix on the other side of the batch. Likewise, the same volume of gasoline will
be lost to this transmix through the interface with high sulfur distillate. The percentages of
gasoline and high sulfur distillate lost will not be the same as the size of the jet fuel, gasoline and
high sulfur distillate batches will likely differ, since their total demands vary widely.  However,
the absolute volumes of jet fuel, gasoline and high sulfur distillate contributing to the interfaces
should be very similar.
                                           7-15

-------
Final Regulatory Support Document
   As mentioned above, two-thirds of the gasoline portion of transmix leaves the transmix
processor as naphtha and returns to the gasoline pool. However, the other one-third leaves as
distillate. As mentioned above, we assume that it does so as high sulfur distillate today.  Thus, a
volume of gasoline equivalent to one-third of 1.75% of jet fuel demand (or 0.58% of jet fuel
demand) is shifted from gasoline to the high sulfur distillate fuel market.  This is shown in Table
7.1.2-6.

   This leaves the downgrade of highway diesel fuel.  In the Final RIA for the 2007 highway
diesel rule, we estimated that a clean cut on one side of highway diesel fuel batches would
downgrade 2.2% of the supply of highway diesel fuel.1 We have applied this estimate in this
analysis, as well.  In Figure 7.1-1, this 2.2% loss occurs via the creation of transmix with
gasoline. We assume that the volume of gasoline contributing to this transmix is the same, 2.2%
of highway diesel fuel supply. All of the highway diesel fuel leaves the transmix processor as
high sulfur distillate. One-third of the gasoline (equivalent to 0.73% of highway diesel fuel
supply) does so, as well.  These downgrades are shown in Table 7.1.2-6.

   The volumes of the various types of downgrade shown in Table 7.1.2-6 fall into two groups.
The first are a function of jet fuel demand, while the second are a function of highway diesel fuel
supply.  To simplify our calculations, we aggregated the volumes of these two types of
downgrades to create just two categories of downgrades, jet-based downgrade and highway fuel-
based downgrade. Jet-based downgrade consists of the jet fuel lost to both the highway and high
sulfur distillate fuel supplies. It also includes the gasoline lost to the high sulfur distillate pool via
interface with high sulfur distillate fuel in the pipeline. In total, the jet-based downgrade
represents 4.08% of jet fuel demand. Of this 4.08%, 1.75% shifts to highway diesel fuel supply,
while 2.33% shifts to high  sulfur distillate supply. Highway fuel-based downgrade consists  of the
highway diesel fuel and gasoline which is shifted to high sulfur distillate supply via the interface
between highway diesel fuel  and gasoline in the pipeline. This downgrade consists of 2.93% of
highway diesel fuel supply.

   The relative volumes of jet fuel demand and highway diesel fuel supply vary across the
various regions of the country being evaluated here. Thus, the relative volumes of the two types
of downgrade will vary, as well. Table 7.1.2-7 shows the demand for jet fuel and highway diesel
fuel, the volume of each type of downgrade and the portions of these downgrades shifted to
highway and high sulfur distillate fuel. Since the States of Alaska and Hawaii have no product
pipelines, we assumed no downgrade occurs there.
   1 When highway diesel fuel must meet a 15 ppm cap standard starting in 2006, we project that the amount of
downgrade will increase to protect the cleaner highway diesel fuel. We discuss this in the next section.

                                          7-16

-------
                                                Estimated Costs of Low-Sulfur Fuels
                                      Table 7.1.2-7
               Downgrade Generation and Disposition in 2001 (Million gallons)

PADD 1
PADD 2
PADD 3
PADD 4
PADD 5-O
AK
HI
CA
et-Based Downgrade
et Fuel Demand (PMA)
downgrade Loss
To Highway Fuel
To High Sulfur Fuel
4,585
187
80
107
3,776
154
66
88
6,095
249
107
142
562
23
10
13
1,580
64
28
37
1,014
0
0
0
325
0
0
0
3,n:
154
6C
8*
-lighway Fuel Based Downgrade
-lighway Fuel Supply
downgrade Loss
Net Highway Fuel Loss*
High Sulfur Fuel Gain
10,596
310
233
310
12,549
368
276
368
6,532
191
144
191
2,067
61
45
61
2,206
65
49
65
111
0
0
0
45
0
0
0
3,56*
lOf
7*
lot
* The difference is due to downgrade from gasoline.
   The final issue is how the new supply of high sulfur distillate is apportioned among the four
uses of high sulfur distillate fuel: nonroad, locomotive, marine, and heating oil. Data are not
available which indicate any difference in the final disposition of high sulfur distillate fuel
produced from transmix compared to that produced by refineries.  Thus, we assume that the
spillover is equally distributed into the four non-highway distillate markets in proportion to their
demand.

   Production

   Distillate fuel production must be sufficient to supply demand, considering changes in supply
during distribution.  Since the net loss in highway fuel produced is 2.2%, highway fuel production
must be 2.2% higher than that indicated in EIA' s PMA for 2001. Likewise, the production of
high sulfur distillate fuel is lower than the estimate of supply from PMA, due to the addition of
some gasoline, jet fuel and highway diesel fuel. The balance of production, gains and losses
during distribution and final supply are shown in Table 7.1.2-8.
                                          7-17

-------
Final Regulatory Support Document
                                       Table 7.1.2-8
                  Distillate Production and Demand in 2001 (million gallons)
Fuel Use
Category
High-
way
Non-
road
Loco-
motive
Marine
Heating
Oil
Fuel Type
Production 500 ppm
Spillover to Non-hwy
Hwy Downgrade
Jet Downgrade
Demand
Production HS
Hwy Spillover
Jet Downgrade
Hwy Downgrade
Demand
Production HS
Hwy Spillover
Jet Downgrade
Hwy Downgrade
Demand
Production HS
Hwy Spillover
Jet Downgrade
Hwy Downgrade
Demand
Production HS
Hwy Spillover
Jet Downgrade
Hwy Downgrade
Demand
PADD
1
10,840
-383
-327
81
10,211
2,672
151
28
83
2,935
445
13
5
14
476
388
11
43
12
415
7,014
207
72
218
7.511
2
12,847
-1,656
-387
69
10,873
2,725
1,130
61
258
4,174
658
255
15
62
989
190
74
4
18
286
511
198
11
48
769
3
6,622
-831
-202
105
5,694
1,064
255
38
53
1,409
651
125
22
32
831
813
156
28
40
1,037
1,537
295
52
76
1.961
4
2,115
-504
-64
10
1,557
215
332
9
41
597
77
114
3
15
209
0
0
0
0
0
39
57
2
7
105
5-O
2,227
-312
-68
43
1,890
289
245
45
53
631
44
36
7
8
94
9
8
1
2
20
30
24
5
5
64
AK
111
-3
0
0
108
22
3
0
0
25
4
0
0
0
4
60
0
0
0
60
214
0
0
0
214
HI
45
-13
0
0
32
29
3
0
0
32
0
0
0
0
0
17
1
0
0
18
123
9
0
0
132
US-
CA
34,806
-3,701
-1,048
309
30,366
7,016
2,118
181
489
9,803
1,878
543
51
131
2,604
1,478
250
37
72
1,838
9,469
791
142
356
10.757
CA
3,468
-830
-95
59
2,602
0
675
61
72
783
0
142
14
17
172
0
38
4
4
46
0
0
0
0
0
US
38,275
-4,532
-1,143
368
32,968
7,015
2,787
242
561
10,586
1,879
685
65
148
2,776
1,477
288
41
77
1,884
9,469
791
142
356
10.757
7.1.3 Distillate Fuel Production and Demand in 2014

   As described in Section 7.2.1, we estimate the cost per gallon of desulfurizating NRLM fuel
using refinery specific production volumes indicative of 2014.  This is the mid-point of the useful
life of hydrotreating equipment built in 2007, per IRS depreciation guidelines. Thus, using
production volumes from 2014 provides a reasonable estimate of the economies of scale of
hydrotreating expected to exist over the life of new equipment built in response to this rule/ As
was the case for 2001, we begin with estimating future demand, and then estimate the fuel
production necessary to satisfy this demand considering spillover and downgrades.
   J In Chapter 8, we project the cost of replacing the hydrotreaters built in 2007. In doing so, we did not increase
the estimated refinery-specific production volumes to represent growth in NRLM fuel demand beyond 2022 (2007
plus the 15 year life of the equipment). This overestimates the cost of replacement equipment to a small extent.
                                           7-18

-------
                                               Estimated Costs of Low-Sulfur Fuels
   7.1.3.1 Distillate Fuel Demand in 2014

   We derive our estimates of growth in highway, locomotive and marine fuel demand from
2001 to 2014 from EIA's AEO for 2003.12 Table 7.1.3-1 shows the projected growth in demand
for these three fuels, as well as projected growth for jet fuel demand.  The fuel demand in each of
these three categories in 2001  (shown in Table 7.1.2-8) were multiplied by the respective growth
factors to estimate fuel demand in 2014. This implicitly assumes that the same growth rate
applies in each region.

                                     Table 7.1.3-1
     Projected Growth in Highway, Locomotive and Marine Fuel Demand: El A 2003 AEO

Demand in 2001 (trillion BTU)
Demand in 2014 (trillion BTU)
Growth Factor to 20 14
Highway
5440
7840
1.44
Locomotive
630
710
1.13
Marine
340
390
1.14
Jet Fuel
3960
2970
1.34
   Nonroad fuel demand in 2014 was estimated using the draft NONROAD2004 model, as was
done for 2001.  Nonroad fuel demand in 2014 is estimated to be 14,379 million gallons per year,
which represents a 36% increase over 2001.

   We projected the growth in heating oil demand from information contained in the 2003 AEO
2003, along with our own estimates of the heating oil portion of each of the economic sectors
tracked in AEO. In its 2003 AEO, EIA projects the demand of petroleum fuels from 2001-2025
based on historical demand and econometric and engineering forecasts. AEO does not provide
forecasts for heating oil demand as we define it here.  Thus, we estimate the heating oil portion
of the fuel demand in each economic sectors tracked in AEO. We then weighted the growth in
the fuel demand in each of the economic sectors by its contribution to total heating oil demand in
2001. Table 7.1.3.2 shows distillate fuel demand in each of the economic sectors tracked by
AEO. (Highway fuel use is not shown, since there is no heating oil use in this category.) The
estimates of demand were taken from the  2001 FOKS report. FOKS breaks down fuel use by fuel
type for several of the sectors. We believe that the use of distillate fuel varies depending on the
type of fuel being consumed (e.g., low sulfur diesel fuel, high sulfur diesel fuel, high sulfur fuel
oil) The FOKS breakdown allows us to apply distinct heating oil percentages to each sector and
fuel type combination.  The information presented in  Table 7.1.3-2 describes the  process we used
to estimate the source of heating oil demand in 2001.
                                         7-19

-------
Final Regulatory Support Document
                                    Table 7.1.3-2
                          Source of Heating Oil Demand: 2001
End Use
Farm
Construction
Other/(Logging)
Industrial
Commercial
Oil Company
Military
Electric Utility
Railroad
Vessel Bunkering
On-Highway
Residential
Total
Fuel Grade
diesel
distillate
distillate
distillate
No. 2 fuel oil
No. 4 distillate
No. 1 distillate
No. 2 low-S diesel
No. 2 high-S diesel
No. 2 fuel oil
No. 4 distillate
No. 1 distillate
No. 2 low-S diesel
No. 2 high-S diesel
distillate
diesel
distillate
distillate
distillate
distillate
diesel
No. 2 fuel oil
No. 1 distillate

Distillate Fuel
FOKS Volume
(1000 gal)
3,351
77
2,086
428
354
44
44
849
1,033
1,546
200
63
1,212
483
820
310
36
1,510
2,952
2,093
33,130
6,151
112
58,971
Percent
Heating Oil
0
100
5
5
100
100
60
0
0
100
100
80
0
0
50
0
100
0
5
10
0
100
100

Heating Oil
Volume
(1000 gal)
0
77
104
21
354
44
26
0
0
1,546
200
50
0
0
410
0
36
1,510
148
209
0
6,151
112
10,998
Percent Heating
Oil Pool
0
0.7
0.9
0.2
3.2
0.4
0.2
0
0
14.1
1.8
0.5
0
0
3.7
0
0.4
13.8
1.3
1.9
0
55.9
1.0
100
   The key figures in Table 7.1.3-2 are the percentages of each economic sector and fuel type
combination which we believe falls into our definition of heating oil. These percentages were
derived using the same methodology which we use in Section 7.1.4 below to derive an estimate of
nonroad fuel demand from FOKS fuel demand estimates. The difference here is that we are not
                                        7-20

-------
                                                Estimated Costs of Low-Sulfur Fuels
focused on nonroad fuel demand, but on heating oil demand. In most of the economic sectors
shown in Table 7.1.3-2, if the fuel is not nonroad fuel, it is heating oil.  The exceptions to this are:
1) locomotive and marine vessel fuel, where the fuel that is not heating oil is locomotive or
marine fuel, respectively, and low sulfur diesel commercial fuel, which is highway fuel which is
not subject to highway fuel excise taxes (e.g., school buses).

    As shown in Table 7.1.3-2, we multiply the total fuel demand for that specific economic
sector and fuel type by its heating oil percentage to estimate the volume of heating oil demanded
in that sector-fuel type combination. We then divide that heating oil demand by total heating oil
demand to derive the percentage of total heating oil demand represented by that sector-fuel type
combination.  The information presented in Table 7.1.3-3 describes the next  step in this process.
Table 7.1.3-3 shows the total distillate fuel demand in 2001 and 2014 from 2003 AEO and the
ratio of these fuel demand volumes.

                                      Table 7.1.3-3
                   Projected Growth in Heating Oil  Demand: 2001 to 2014
Category
Farm
Construction
Logging/Other
Industrial
Commercial
Oil Company
Military
Electric Utility
Railroad
Vessel Bunkering
Residential
Weighted Ave.
2001 Distillate
Demand *
469
238
55.6
1,130
460
6.2
101
170
628
345
910
-
2014 Distillate
Demand *
533
274
59.9
1,270
490
0
124
90
707
394
880
-
Ratio of 20 14 to 2001
Distillate Demand
1.14
1.15
1.08
1.12
1.07
0
1.22
0.70
1.13
1.14
0.97
0.93
Percent of Total
Heating Oil Demand
0.7
0.9
0.2
3.8
16.4
3.7
0.4
13.8
1.3
1.9
56.9

   Trillion BTU from the 2003 AEO.
   We weighted the growth in each sector's distillate fuel demand by that sectors' contribution to
2001 heating oil demand.  For farm, industrial, commercial, residential and military, the
contributions of the various fuel types shown in Table 7.1.3-2 were combined for use in Table
7.1.3-3.  The result is that heating oil demand is projected to shrink by 7% between 2001 and
2014. Thus, we multiplied the heating oil demand in each region shown in Table 7.1.2-8 by 0.93
to estimate heating oil demand in 2014. Table 7.1.3-4 shows the resulting distillate demands
                                          7-21

-------
Final Regulatory Support Document
projected for 2014 for the five fuel categories. Table 7.1.3-4 also shows jet fuel demand in 2014,
which represents a 34% increase over those shown in Table 7'.1.2-7'.

                                     Table 7.1.3-4
                        Distillate Demand in 2014 (million gallons)
End Use
Highway
\Ionroad
Railroad
Vlarine
Seating Oil
Total No. 2
Distillate Demand
let Fuel
Region
1
14,722
3,987
536
475
6,970
26,690
6,143
2
15,676
5,670
1,114
327
714
23,501
5,060
3
8,210
1,914
935
1,187
1,820
14,066
9,313
4
2,245
810
236
0
98
3,389
753
5-O
2,725
857
106
23
59
3,770
2,117
AK
157
34
5
69
199
464
1,359
HI
46
43
0
21
122
232
436
CA
3,752
1,064
194
53
0
5,063
5,054
U.S.
47,533
14,379
3,126
2,155
9,982
77,175
30,235
   7.1.3.2 Future Distillate Fuel Production

   The primary purpose of projecting production of the various types of distillate fuel in 2014 is
to factor in appropriate economies of scale for the investment in new desulfurization equipment to
comply with the NRLM sulfur standards. We use 2014 production volumes to estimate these
costs for all of the steps of the final NRLM fuel program, because 2014 represents the mid-point
of the life of refinery equipment for the purposes of calculating annual depreciation under IRS
guidelines.  The five steps for which production volumes were estimated are:

   1) Reference Case (i.e., no NRLM Program),
   2) Final NRLM fuel Program: 2007-2010,
   3) Final NRLM fuel Program: 2010-2012,
   4) Final NRLM fuel Program: 2012-2014, and
   5) Final NRLM fuel Program: 2014 and beyond

   7.1.3.2.1 Reference Case; no NRLM Fuel Program

   There are two distinct periods which define the reference case which assumes that the NRLM
fuel program was not promulgated.  One is during the period between 2007 and 2010 when the
highway diesel fuel program's temporary compliance option is in effect. During this time,
consistent with the refiners' pre-compliance reports under the highway fuel program, we assume
5% of highway diesel fuel will be produced at 500 ppm.13 The remainder will be 15 ppm fuel.
The second period is after 2010 when the highway diesel fuel program's temporary compliance
                                         7-22

-------
                                                 Estimated Costs of Low-Sulfur Fuels
option expires and all highway diesel fuel must meet a 15 ppm cap. During both of these periods,
NRLM fuel would continue to be high sulfur diesel fuel.

   California has implemented its own sulfur standards for highway and nonroad diesel fuel pool
starting in 2006. Thus, nonroad diesel fuel in California was assumed to already meet the 15 ppm
standard in the reference case.  While California will not be regulating the locomotive and marine
diesel fuel quality as part of its regulation, our analysis shows that the locomotive and marine
diesel fuel demand will be met using spillover and the low sulfur diesel fuel downgrade once the
nonroad pool is regulated to 15 ppm.  Therefore, EPA's NRLM program is not expected to have
any impact on the production or distribution of locomotive and marine diesel fuel in that State.K

   We project the production volume of highway diesel fuel in 2014 using a slightly  different
methodology than we used for 2001 production. For 2001, we started with supply and demand
and calculated spillover. Downgraded volume was then added to estimate total production.  For
2014, we start with highway fuel demand, add the spillover of highway fuel into non-highway
fuel markets based on 2001  estimates, and add the volume of highway fuel which is downgraded
to lower quality fuel.

   The demand for highway diesel fuel was estimated in the previous section. Regarding
spillover, we assume that the same constraints in the distribution system which cause  most
spillover to occur today will continue in the future.  This means that the volume of highway fuel
spilling over into each of the four non-highway fuel markets will grow as each of these markets
grows. Thus, we have increased the  spillover volumes shown in Table 7.1.2-5 for the nonroad,
locomotive, marine and heating oil markets by the 2001 to 2014 growth factors for these fuels
shown in Tables 7.1.3-1 and 7.1.3-3 (and a factor of 1.36 for nonroad fuel). The net effect of this
assumption is that the percentage of demand represented by spillover in each of the four non-
highway fuel markets is the same in 2014 as in 2001. Table 7.1.3-5 shows the demand for
highway fuel, spillover into each of the four non-highway fuel markets, and the resultant supply
of highway fuel needed to provide for this demand and spillover.
   K Our conclusion that California will not be affected by the NRLM program is based on our nationwide analysis
on how fuels are produced and distributed throughout the U.S. focusing on areas outside of California. It is possible
that California fuel production and distribution is different enough that some fuel would in fact be affected by this
rulemaking.

                                           7-23

-------
Final Regulatory Support Document
                                      Table 7.1.3-5
                     Spillover of Highway Fuel in 2014 (million gallons)
End Use
Highway Demand
Region
1
14,722
2
15,676
3
8,210
4
2,245
5-O
2,725
AK
157
HI
46
CA
3,752
Spillover
Nonroad
Railroad
Marine
Heating Oil
Total Spillover
Highway Supply
206
15
13
192
425
15,247
1,535
287
84
184
2,090
17,911
345
141
179
274
939
9,127
451
129
0
53
633
2,900
333
40
9
22
404
3,111
4
0
0
0
4
161
4
0
1
8
13
60
1,054
0
0
0
1,298
4,978
   As mentioned above, the State of California has promulgated regulations requiring that
nonroad fuel meet a 15 ppm cap, as well as highway fuel, in 2006.  We have categorized this 15
ppm nonroad fuel as highway fuel to better distinguish between 15 ppm fuel which would be
produced prior to this NRLM rule and that which will be produced because  of this rule. Because
15 ppm nonroad fuel in California will be produced with or without this rule, we have classified it
as highway fuel in our presentation.  Thus, any production of 15 ppm nonroad fuel shown below
will be due to this rule and not due to California regulations.

   The next step is to estimate the volume of downgrade into and out of the various fuel supply
pools, as was done for 2001. In the Final RIA for the 2007 highway diesel rule, we projected that
the downgrade of 15 ppm highway diesel fuel would increase to 4.4% from  the current estimated
level of 2.2%.  Thus, we assume that 4.4%L of the supply of highway fuel shown in Table 7.1.3-5
will be downgraded to a lower quality distillate.

   The implementation of the 15 ppm highway fuel cap in 2006 could affect sequencing in some
pipelines.  Most pipelines will simply replace their 500 ppm highway fuel with 15 ppm highway
fuel. However, some  pipelines will continue to carry a 500 ppm highway fuel through mid-2010.
In the Final RIA of the highway rule, we projected that roughly 40% of fuel markets would
include a 500 ppm fuel to distribute the roughly 20% of highway fuel which would be at 500
ppm.  However, the highway pre-compliance reports indicate a much lower  percentage of
highway fuel which likely be produced at  500 ppm. Because of this and for simplicity, we
assume that most pipelines would not carry 500 ppm highway fuel  absent the NRLM rule.
However, we believe  that the  sequencing of fuels in pipelines will still likely change from that
   L Due to a miscalculation, the highway diesel fuel downgrade is estimated to be 4.5% instead of 4.4% for all
analyses after 2010.  The overestimated highway downgrade volume overestimates the costs of the program.
                                          7-24

-------
                                                  Estimated Costs of Low-Sulfur Fuels
shown in Figure 7.1.1.  In particular, we believe that pipelines would not wrap 15 ppm highway
fuel with jet fuel and heating oil, but would wrap it with heating oil and gasoline, as shown in
Figure 7.1-2. With the sequence shown in Figure 7.1-1, the interface between jet fuel and 15 ppm
highway fuel could not be cut into either fuel, but would have to be segregated and added to the
heating oil storage tank. With the sequence in Figure 7.1-2, all of the distillate-distillate
interfaces can be cut into heating oil and the only interfaces requiring segregation and processing
are those containing gasoline and distillate, as is currently the case.
   Figure 7.1-2 Pipeline Sequence and Fate of Interface Between Fuel Batches
         in Areas that Carry Heating Oil; Prior to NRLM Rule: 2006+
Jet


NRLM +
Heating Oil


15 ppm
Highway Fuel


Tier 2
Gasoline
0}
1 — 1
            L
Heating Oil
           + 1.75% Batch Swell  +2.2%
             Jet              Hwy
                 Transmix Products
                               Transmix
                                1.75% Jet
                                2.2% Hwy
                                Gasoline in equal amounts
                 Distillate volume = 1.75% Jet + 2.2% Hwy + 1/3 of gasoline in transmix
                 Distillate quality: <500 ppm
    The change in sequencing affects the types of downgrade which will occur. Table 7.1.3-6
shows these downgrades and their volumes.  Overall 3.5% of jet fuel volume is still downgraded
to the distillate market.  In addition, gasoline volume equivalent to 0.58% of jet fuel demand and
0.73% of highway fuel supply will also be downgraded to the distillate market. The volume of
high sulfur distillate supplied should again not be affected. Only the volume of highway fuel
downgraded will increase, from 2.2% to 4.4% of total supply. We assume that the jet fuel and
highway diesel fuel interfaces with high sulfur distillate will be cut directly into the batch of high
sulfur distillate. Therefore, half of the jet fuel downgrade and half of the highway diesel fuel
downgrade will be cut directly into batches of high sulfur distillate. The remaining downgrades
are mixed with gasoline and sent to transmix processors, where distillate fuel is recovered and
sold.  Due to the Tier 2 sulfur standards applicable to gasoline in 2004 and beyond and the 15
ppm highway diesel fuel cap, the sulfur content of distillate produced by transmix processors will
decrease dramatically.  As described in Section 7.7 below, we estimate that the sulfur content of
distillate produced by transmix processors will be well below 500 ppm.  The 500 ppm highway
diesel fuel market should command a price premium over high sulfur distillate fuel during this
timeframe. Therefore, we assume that this distillate will be sold to the 500 ppm highway diesel
fuel market.
                                           7-25

-------
Final Regulatory Support Document
                                     Table 7.1.3-6
         Types of Downgrade and Their Volumes for the Reference Case: 2006-2010
Interface
Jet Fuel- High Sulfur
Distillate Interface
Gasoline - Jet Fuel
Interface
Highway Diesel Fuel-
High Sulfur Distillate
Interface
Gasoline - Highway Diesel
Fuel Interface
Original
Fuel
High Sulfur
Distillate
Jet Fuel
Jet Fuel
Gasoline
High Sulfur
Distillate
Highway
Diesel Fuel
Highway
Diesel
Gasoline
Destination
High Sulfur Distillate
High Sulfur Distillate
500 ppm Highway Fuel
500 ppm Highway Fuel
High Sulfur Distillate
High Sulfur Distillate
500 ppm Highway Fuel
500 ppm Highway Fuel
Volume
Zero
1.75% of jet fuel demand
1.75% of jet fuel demand
Equivalent to 0.58% of jet fuel demand
Zero
2.2% of highway diesel fuel supply
2.2% of highway diesel fuel supply
Equivalent to 0.73% of highway diesel
fuel supply
   We obtained future demand for jet fuel from 2003 AEO. There, EIA projects a 34% increase
in jet fuel demand compared to demand in 2001. We applied this nationwide increase to the 2001
jet fuel demand by region shown in Table 7'.1.2-7'.  The resultant 2014 jet fuel demand by region
is summarized in Table 7.1.3-7.

                                     Table 7.1.3-7
  Downgrade Generation and Disposition for the Reference Case: 2006-2010 (Million gallons)

PADD 1
PADD 2
PADD 3
PADD 4
PADD 5-O
AK
HI
CA
Jet-Based Downgrade
Jet Fuel Demand (PMA)
To High Sulfur Fuel
To 500 ppm Fuel
Total Downgrade
6,144
108
143
251
5,060
89
118
206
8,167
143
190
333
753
13
18
31
2,117
37
49
86
1,359
24
32
55
435
8
10
18
5,054
88
118
206
Highway Fuel Based Downgrade
Highway Fuel Supply
To High Sulfur Fuel
To 500 ppm Fuel
Total Downgrade
15,825
348
464
812
18,487
407
542
948
9,527
210
279
489
2,981
66
87
153
3,254
72
95
167
161
4
5
8
60
1
2
o
6
5,223
115
153
268
   The downgraded jet fuel and highway diesel fuel are cut directly into batches of high sulfur
distillate being carried in the pipeline. Therefore, it is reasonable to assume that this downgrade
                                         7-26

-------
                                                   Estimated Costs of Low-Sulfur Fuels
would be distributed just as the rest of the high sulfur distillate supply. Thus, we allocate this
downgrade to the four high sulfur distillate markets in proportion to the demand for each of these
fuels in each region. The final projections of production, spillover, downgrade and demand for
2006-2010 for the Reference Case which assumes no implementation of this NRLM rule are
shown in Table 7.1.3-8.

                                        Table 7.1.3-8
   Distillate Supply and Demand for the Reference Case: 2006-2010 (million gallons in 2014)M
Fuel Use
Category
High-
way
Non-
road
Loco-
motive
Marine
Heating
Oil
Fuel Type
Production 1 5 ppm
Production 500 ppm
Spillover to Non-hwy
Hwy Downgrade
Jet Downgrade to 500 ppm
1 5 ppm Hwy Downgrade to
500 ppm
Demand 1 5 ppm
Demand 500 ppm
Production HS
Hwy Spillover
Jet Downgrade to 500*
Hwy Downgrade to 500*
Jet Downgrade to HS
Hwy Downgrade to HS
Demand
Production HS
Hwy Spillover
Jet Downgrade to HS
Hwy Downgrade to HS
Demand
Production HS
Hwy Spillover
Jet Downgrade to HS
Hwy Downgrade to HS
Demand
Production HS
Hwy Spillover
Jet Downgrade to HS
Hwy Downgrade HS
Demand
PADD
1
14,363
866
-425
-680
126
453
13,306
1,416
3,626
206
2
6
32
115
3,987
500
14
5
16
536
443
13
4
15
475
6,514
191
57
206
6,970
2
16,648
1,213
-2090
-724
90
452
14,169
1,508
3,726
1,535
9
44
59
297
5,670
755
287
12
60
1,114
222
84
3
18
327
484
184
8
38
714
3
8,616
532
-939
-379
137
235
7,420
790
1,445
345
6
10
40
68
1,914
739
141
20
35
935
938
179
26
44
1,187
1,440
274
39
67
1,820
4
2,658
219
-633
-104
11
62
2,029
216
290
450
2
12
8
47
810
90
128
2
14
236
0
0
0
0
0
37
53
1
6
98
5-0
2,928
200
-404
-126
52
73
2,463
262
408
333
6
9
42
59
857
53
40
5
7
106
12
9
1
2
23
30
22
3
4
59
AK
152
8
-4
0
0
0
149
8
30
4
0
0
0
0
34
5
0
0
0
5
69
0
0
0
69
199
0
0
0
199
HI
56
4
-13
0
0
0
44
2
39
3
0
0
0
0
43
0
0
0
0
0
20
1
0
0
21
114
8
0
0
122
US-
CA
45,436
3,029
-4508
-2012
416
1,276
39,580
4,201
9,565
2,877
25
82
181
586
13,316
2,143
611
45
133
2,932
1,704
287
35
78
2,103
8,819
734
108
321
9,981
CA
4,978
0
-1053
-173
0
0
3,752
0
10
1,054
0
0
0
0
1,064
0
0
144
217
194
0
0
46
59
53
0
0
0
0
0
us
50,377
3,066
-5561
-2185
416
1,276
43,332
4,201
9,575
3,930
25
82
181
586
14,379
2,143
611
189
350
3,126
1,704
287
81
137
2,156
8,819
734
108
321
9,981
*  Highway and jet downgrade to 500 ppm spillover pool.  This is not shown for other PADDs.
    M Due to a miscalculation, the jet fuel downgrade is about 10 percent lower than if calculated as described.
This error results in slightly overestimating the cost and the benefits of the program. This miscalculation occurred in
all the volume analyses prior to 2010.
                                            7-27

-------
Final Regulatory Support Document
   In 2010, the temporary compliance option of the highway program ends.  Therefore, there
would not be any 500 ppm highway fuel, only 15 ppm highway fuel and high sulfur distillate.
The pipeline sequence shown in Figure 7.1-2 applies.  All of the downgrade volumes shown in
Table 7.1.3-6 would still apply.  No downgraded distillate fuel would meet a 15 ppm cap.
Therefore, all the downgraded distillate would be shifted to the high sulfur distillate market.  As
for 2006-2010, we assume that this downgrade is distributed to the four high sulfur distillate
markets in proportion to the demand for each fuel in each region. The projections of production,
spillover, downgrade and demand for 2010 and beyond for the Reference Case which assumes no
implementation of this NRLM rule are shown in Table 7.1.3-9.

                                     Table 7.1.3-9
     Distillate Supply and Demand for the Reference Case: 2010+ (million gallons in 2014)
Fuel Use
Category
High-
way
Non-
road
Loco-
motive
Marine
Heating
Oil
Fuel Type
Production 1 5
Spillover to Non-
Hwy Downgrade
Demand
Production HS
Hwy Spillover
Jet Downgrade
Hwy Downgrade
Demand
Production HS
Hwy Spillover
Jet Downgrade
Hwy Downgrade
Demand
Production HS
Hwy Spillover
Jet Downgrade
Hwy Downgrade
Demand
Production HS
Hwy Spillover
Jet Downgrade
Hwy Downgrade
Demand
PADD
1
15,825
-425
-678
14,722
3,401
206
108
272
3,987
469
15
15
38
536
416
13
13
33
475
6,097
192
194
488
6.970
2
18,487
-2,090
-721
15,676
3,235
1,535
199
702
5,670
647
287
40
140
1,114
190
84
12
41
327
414
184
25
90
714
3
9,527
-939
-378
8,210
1,275
345
133
160
1,914
646
141
69
81
935
820
179
86
103
1,187
1,257
274
131
158
1.820
4
2,981
-633
-103
2,245
221
451
28
111
810
66
129
8
33
236
0
0
0
0
0
27
53
3
14
98
5-0
3,254
-404
-125
2,725
242
333
142
140
857
30
40
18
18
106
7
9
4
4
23
17
22
10
10
59
AK
161
-4
0
157
30
4
0
0
34
5
0
0
0
5
69
0
0
0
69
199
0
0
0
199
HI
60
-13
0
46
39
4
0
0
43
0
0
0
0
0
20
1
0
0
21
114
8
0
0
122
US-
CA
50,294
-4,508
-2,006
43,781
8,443
2,877
610
1,385
13,316
1,863
611
150
310
2,932
1,521
286
114
181
2,103
8,125
734
364
759
9.981
CA
5,223
-1,053
-173
3,752
10
1,054
0
0
1,064
0
0
144
217
194
0
0
46
59
53
0
0
0
0
0
US
55,517
-5,561
-2,178
47,533
8,453
3,930
610
1,385
14,379
1,863
611
294
527
3,126
1,521
286
161
241
2,156
8,125
734
364
759
9.981
    7.7.3.2.2 Final NRLM Fuel Program: 2007-2010
                                         7-28

-------
                                                 Estimated Costs of Low-Sulfur Fuels
   Demand for the various categories of distillate fuel are assumed to not change under the final
NRLM fuel program.  Therefore, the fuel demand estimates shown in Table 7.1.3-5 apply to this
scenario, as well as prior to the NRLM rule. We also assume that spillover will not be affected by
the NRLM rule, because spillover occurs where only one fuel is available and this fuel will still
be 15 ppm highway fuel.  Thus, the production of highway fuel and the spillover of this fuel to
the NRLM and heating oil markets will be the same as shown in Tables 7.1.3-5 and 7.1.3-8.

   With the initiation of the NRLM fuel program in 2007, 500 ppm NRLM fuel will be widely
distributed and available.  Thus, pipeline sequencing will be affected. While most 500 ppm fuel
is likely to be NRLM fuel, the widespread distribution of 500 ppm NRLM fuel will also facilitate
the distribute of 500 ppm  highway fuel. In areas with relatively small heating oil markets, such as
PADDs 2 and 4 and California, we assume that the heating oil volume will be too small to justify
pipelines handling a separate high sulfur distillate fuel for this market. Thus, 500 ppm NRLM
fuel will replace high sulfur distillate in the common carrier distribution systems in these regions.
Generally, this means that most heating oil in these regions will meet a 500 ppm cap.

   Outside of PADDs 2 and 4, we believe that the heating oil  market is  either sufficiently large
or the distribution system is sufficiently flexible to allow the distribution of high sulfur distillate
fuel to this market. The pipelines in PADD 1 are expected to carry heating oil for the large
market there, and PADD 3 pipelines are expected to carry heating oil, in part, to supply the
PADD 1 market. The heating oil market in the Pacific Northwest is not  large. However, this area
has a fairly simple distribution system and much of this heating oil consumption is believed to be
on the coast.  Thus, we believe that it would be feasible for a refiner to produce and distribute
high sulfur distillate fuel to this market, though this distribution will not likely be by pipeline.
The same is true for Hawaii.  Table 7.1.3-10a summarizes these assumptions for the various
regions.

                                     Table 7.1.3-10a
      Production and Distribution of High Sulfur Distillate: Final NRLM Rule: 2007-2010

High Sulfur Distillate in Pipelines
High Sulfur Distillate Produced for
Heating Oil Market
PADDs 1&3
Yes
Yes
PADDs 2 & 4
No
No
PADD 5-O
No
Yes
AKandHI
No pipelines
Yes
CA
No
No
   Figures 7.1-3 depicts pipeline sequencing with 500 ppm NRLM fuel and heating oil both
being carried.  As shown in Table 7.1.3-10, this applies to pipelines in PADDs 1 and 3.
                                          7-29

-------
Final Regulatory Support Document
      Figure 7.1-3 Pipeline Sequence and Fate of the Interface Between Fuel
     Batches in Areas that Carry Heating Oil; After NRLM Rule: 2007 - 2010
      Jet




High Sulfur
Distillate
500 ppm
Hwy&
NRLM


L J J
15 ppm
Highway

Tier 2
Gasoline


1
~r—
\ r
+ 1.75% +2.2% +2.2% V J
Jet 500 Hwy 15 Hwy ^ .
               Transmix Products
Transmix
 1.75% Jet
 2.2% Hwy
 Gasoline in equal amounts
                Distillate volume = 1.75% Jet + 2.2% Hwy + 1/3 of gasoline in transmix
                Distillate quality: < 500 ppm
In this case, 15 ppm highway diesel fuel is downgraded directly to batches of 500 ppm fuel in the
pipeline. A similar volume of 500 ppm fuel will be downgraded to high sulfur heating oil. Thus,
there will be essentially no net loss of 500 ppm fuel from its batch during distribution. The loss of
15 ppm highway fuel is essentially shifted to high sulfur distillate.  The interfaces containing
gasoline and distillate are not affected, relative to that occurring prior to the NRLM rule.  Thus,
the net downgrade of 15 ppm highway diesel fuel, jet fuel and heavy gasoline is the same as that
prior to the NRLM rule during this timeframe. The distillate fuel produced from transmix should
still contain less than 500 ppm sulfur and can be sold to either the highway or NRLM fuel market.
We generally presumed that this fuel would be sold to the highway fuel market, given the higher
prices likely to exist there. However, under the designate and track provisions of the final NRLM
rule, the total volume of highway fuel cannot increase during shipment. Thus, the net loss of 15
ppm highway fuel to the high sulfur distillate market must be greater than the increase in 500 ppm
highway fuel from transmix distillate.  Therefore, we limited the volume of transmix distillate
shifted to the 500 ppm highway fuel  market to the volume of 15 ppm highway fuel lost. Any
remaining 500 ppm fuel  produced from transmix was sent to the 500 ppm NRLM market.  A
detailed description of these downgrades and their volumes is shown in Table 7.1.3-10.
                                           7-30

-------
                                                   Estimated Costs of Low-Sulfur Fuels
                                        Table 7.1.3-10
          Types of Downgrade and Their Volumes Under the NRLM Rule: 2007-2010
   Pipelines Carrying Both 500 ppm NRLM Fuel and High Sulfur Distillate (PADDs 1 and 3)
Interface
Jet Fuel- High Sulfur
Distillate Interface
Gasoline - Jet Fuel
Interface
Highway Diesel Fuel-
500 ppm NRLM Fuel
Interface
500 ppm NRLM Fuel
- High Sulfur
Distillate Interface
Gasoline - Highway
Diesel Fuel Interface
Original Fuel
High Sulfur
Distillate
Jet Fuel
Jet Fuel
Gasoline
Highway Diesel
Fuel
500 ppm NRLM
Fuel
Highway Diesel
Gasoline
Destination
High Sulfur Distillate
High Sulfur Distillate
500 ppm Highway Fuel
500 ppm Highway Fuel
500 ppm NRLM Fuel
High Sulfur Distillate
500 ppm Highway Fuel
500 ppm Highway Fuel
Volume
Zero
1.75% of jet fuel demand
1.75% of jet fuel demand
Equivalent to 0.58% of jet fuel
demand
2.2% of highway diesel fuel supply
2.2% of highway diesel fuel supply
2.2% of highway diesel fuel supply
Equivalent to 0.73% of highway diesel
fuel supply
   Figure 7.1-4 depicts pipeline sequencing in systems that no longer carry high sulfur heating
oil. This applies to pipelines in PADDs 2, 4 and 5.
    Figure 7.1-4 Pipeline Sequence and Fate of the Interface Between Batches
     in Areas that do not Carry Heating Oil; After NRLM Rule: 2007 - 2010

Jet







i
j 500 ppm Hwy
& NRLM
|





| 15 PPm !
i Highway I
! !
!
Tier 2
j Gasoline
|








L J ^ f

^
^

J
+ 1 75% +97% 'r ' •
-r i ./j/o TZ.././O Iransinix
Jet Hwy
                 Transmix Products
                                                   1.75% Jet
                                                   2.2% Hwy
                                                   Gasoline in equal amounts
                 Distillate volume = 1.75% Jet + 2.2% Hwy + 1/3 of gasoline in transmix
                 Distillate quality:  < 500 ppm
                                             7-31

-------
Final Regulatory Support Document
   The absence of high sulfur distillate in the pipeline affects the types of downgrade occurring.
Both downgraded 15 ppm highway diesel fuel and jet fuel are cut directly into batches of 500
ppm fuel in the pipeline. The interfaces containing gasoline and distillate are not affected by the
NRLM rule during this timeframe. As discussed in Section 7.1.6, the sulfur level of the distillate
produced by transmix operators is estimated to be less than 500 ppm.

   We made different assumptions regarding the disposition of this downgrade in the four
applicable regions due to varying circumstances existing in each one. Because of the small size
of the heating oil market in PADDs 2 and 4 (see Table 7.1.3-8), we assume that refiners will not
produce high sulfur distillate fuel for the heating oil market. Thus,  in these areas, we assume that
this downgraded distillate will preferentially fulfill remaining heating oil demand. This might
entail some additional  distribution costs to reach all heating oil users, but no sulfur content testing
would be required. If the volume of downgrade exceeded heating oil demand in these areas, we
assumed that the downgrade would then be used in the 500 ppm highway fuel market, up to the
volume of 15 ppm highway fuel lost during distribution(due to designate and track limitations).
Any remaining downgrade distillate was assumed to be used as 500 ppm NRLM fuel, in
proportion to each region's demand for nonroad, locomotive and marine fuel.

   In California, we also assumed that refiners would not produce  high sulfur distillate fuel for
the heating oil  market. However, California's regulations require that all highway and nonroad
fuel meet a 15  ppm cap in this timeframe. Also, we project essentially no demand for heating oil
in California. Thus, all downgrade distillate was assumed to be used in the L&M markets, in
proportion to the demand for each fuel.

   Finally, in PADD 5-O, we assumed that refiners could produce  high sulfur distillate for the
heating oil market, but that this would not be shipped inland in pipelines. Therefore, we assumed
that the downgrade distillate would not be used to fulfill heating oil demand, but would be used as
500 ppm highway fuel up to the point allowed by the designate and track procedures. The
remainder would then be used as 500 ppm NRLM fuel, in proportion to the region's demand for
nonroad,  locomotive and marine fuel.  Table 7.1.3-11 summarizes these priorities of downgrade
use in PADDs  2, 4, and 5 from 2007 - 2010 uncer the fuel rule provisions.

                                      Table  7.1.3-11
           Use of Distillate Downgrade by Region: Final NRLM Rule: 2007 to 2010

1st Priority
2nd Priority
3rd Priority
PADD 2
HO
500 ppm Highway *
500 ppm NRLM
PADD 4
HO
500 ppm Highway *
500 ppm NRLM
PADD 5-O
500 ppm Highway *
500 ppm NRLM
-
CA
L&M
-
-
 : Volume limited by loss of 15 ppm highway fuel
                                          7-32

-------
                                                  Estimated Costs of Low-Sulfur Fuels
    Table 7.1.3-12 shows the sources of downgrades and their volumes.

                                       Table 7.1.3-12
          Types of Downgrade and Their Volumes Under the NRLM Rule: 2007-2010
          Pipelines Not Carrying High Sulfur Distillate (PADDs 2, 4, 5-O, California)

Jet Fuel- 500 ppm
Diesel Fuel
Gasoline - Jet Fuel
Interface
1 5 ppm Highway
Diesel Fuel- 500 ppm
Diesel Fuel Interface
Gasoline - Highway
Diesel Fuel Interface
Original Fuel
Jet Fuel
Jet Fuel
Gasoline
Highway
Diesel Fuel
Highway
Diesel
Gasoline
Quality of Downgrade *
500 ppm Diesel Fuel
500 ppm Diesel Fuel
500 ppm Diesel Fuel
500 ppm Diesel Fuel
500 ppm Diesel Fuel
500 ppm Diesel Fuel
Volume
1.75% of jet fuel demand
1.75% of jet fuel demand
Equivalent to 0.58% of jet fuel
demand
2.2% of highway diesel fuel supply
2.2% of highway diesel fuel supply
Equivalent to 0.73% of highway diesel
fuel supply
* Destination of the new 500 ppm diesel fuel varies by region.
    One last effect of the NRLM rule during the 2007-2010 timeframe is the provision for small
refiners to be able to sell high sulfur distillate fuel to the NRLM market.  If a small refiner
chooses to produce 500 ppm NRLM fuel, then they can sell credits to other refiners, which allows
them to produce and market high sulfur NRLM fuel.  In either case, the volume of fuel potentially
affected by this provision is the production of high sulfur distillate fuel by small refiners. The
production of both highway fuel and high sulfur distillate by small refiners is addressed in Section
7.2.1.  Since so much of the fuel produced in PADD 3 is distributed to PADD 1, we spread the
volume of PADD 3 small refiner fuel over the two PADDs in proportion to the demand for
NRLM fuel in the two PADDs.N Within each PADD we assume that the high sulfur, small refiner
NRLM fuel is blended into the nonroad, locomotive and marine markets in proportion to the
demand in each market. The volume of small refiner fuel is summarized in Table 7.1.3-13.
   N The final NRLM rule includes an Northeast/Mid-Atlantic Area within which no high sulfur NRLM fuel can
be sold. This area covers the most of the Northeast and Middle Atlantic states. Thus, it might be difficult for the
levels of small refiner fuel assumed here to be sold in PADD 1 under these provisions. If this were the case, this
small refiner fuel would likely stay in PADD 3. The net result would be that the sulfur content of NRLM fuel in
PADD 1 would decrease and that in PADD 3 would increase. The net nationwide impact would be negligible.
                                           7-33

-------
Final Regulatory Support Document
                                   Table 7.1.3-13
                  Small Refiner NRLM Fuel: 2007-2010 (million gallons)
PADDl
420
PADD2
140
PADD3
291
PADD4
0
PADD 5-O
60
AK
104
HI
0
CA
0
   The final projections of production, spillover, downgrade and demand under the final NRLM
fuel program from 2007-2010 are shown in Table 7.1.3-14.
                                       7-34

-------
                                                     Estimated Costs of Low-Sulfur Fuels
                                         Table 7.1.3-14
         Distillate Supply and Demand: Final Rule: 2007-2010 (million gallons in 2014)c
Fuel Use
Category
High-
way
Non-
road
Loco
motive
Marine
Heating
Oil
Fuel Type
Production 15 ppm
Production 500 ppm
Spillover to Non-Hwy
Hwy Dwngr 1 5 ppm
Jet Downgrade
Hwy Downgrade
Demand 1 5 ppm
Demand 500 ppm
Production 500 ppm
Small Refiner Fuel
Hwy Spillover
Jet Downgrade
Hwy Downgrade
Reproc. Downgrade
Demand
Production 500 ppm
Small Refiner Fuel
Hwy Spillover
Jet Downgrade
Hwy Downgrade
Demand
Production 500 ppm
Small Refiner Fuel
Hwy Spillover
Jet Downgrade
Hwy Downgrade
Demand
Production HS
Hwy Spillover
Jet Downgrade
Hwy Downgrade
Demand
PADD
1
14,363
866
-425
-678
130
466
13,284
1,438
3,448
333
206
0
0
0
3,987
476
46
15
0
0
536
421
41
13
0
0
475
6,329
192
98
351
6,970
2
16,648
1,213
-2,090
-714
107
542
13,986
1,690
4,025
111
1,535
0
0
0
5,670
805
22
287
0
0
1,114
236
7
84
0
0
327
0
184
88
442
714
3
8,616
532
-939
-375
139
239
7,357
853
1,402
135
345
11
19
0
1,914
710
69
141
6
10
935
901
87
179
7
13
1,187
1,210
274
124
212
1,820
4
2,658
219
-633
-101
15
85
1,973
271
329
0
451
5
26
0
810
98
0
129
1
8
236
0
0
0
0
0
0
0
53
7
38
98
5-0
2,928
200
-404
-124
52
73
2,427
299
330
52
333
59
83
0
857
41
7
40
7
10
106
9
1
9
2
2
23
37
22
0
0
59
AK
152
8
-4
0
0
0
148
8
0
30
4
0
0
0
34
0
5
0
0
0
5
0
69
0
0
0
69
199
0
0
0
199
HI
56
4
-13
0
0
0
44
3
39
0
4
0
0
0
43
0
0
0
0
0
0
20
0
1
0
0
21
115
8
0
0
122
US-
CA
45,436
3,029
-4,508
-1,991
437
1,378
39,219
4,562
9,573
661
2,877
75
129
0
13,316
2,130
148
611
15
28
2,932
1,588
205
286
9
15
2,103
7,888
734
316
1,043
9,981
CA
4,760
0
-835
-173
0
0
3,752
0
10
0
835
0
0
219
1,064
0
0
0
141
213
194
0
0
0
46
59
53
0
0
0
0
0
us
50,196
3,029
-5,343
-2,164
437
1,378
42,971
4,562
9,584
661
3,712
75
129
219
14,379
2,130
148
612
159
245
3,126
1,588
205
286
55
74
2,156
7,888
734
316
1,043
9,981
    0 Due to a miscalculation, the jet fuel downgrade is about 10 percent lower than if calculated as described.  This
error results in slightly overestimating the costs and the benefits of the program. This miscalculation occurred in all
the volume analyses prior to 2010.
                                              7-35

-------
Final Regulatory Support Document
    7.7.3.2.3 Final Rule Program - 2010 to 2012

    Beginning in mid-2010, two regulatory requirements change: 1) the temporary compliance
option under the highway fuel program ends and all highway fuel must meet a 15 ppm cap and 2)
nonroad fuel must meet a 15 ppm cap (L&M fuel continues to meet a 500 ppm cap).  However,
downgraded 500 ppm fuel produced during shipment of 15 ppm highway diesel fuel and jet fuel
(or produced by small refiners or with small refiner credits) can continue to be sold to the NRLM
fuel markets outside of the Northeast/Mid-Atlantic Area. Within the Northeast/Mid-Atlantic
Area, downgraded 500 ppm fuel produced during shipment of 15 ppm fuel and jet fuel can only
be sold to the L&M fuel market.

    As was the case from 2007-2010, the demand for each distillate fuel  and the spillover of
highway fuel into these markets are assumed to remain unchanged  from  those occurring prior to
the NRLM rule (see Table 7.1.3-5).  With the application of the 15 ppm  cap on nonroad fuel in
2010, 500 ppm fuel is not likely to be widely distributed through pipelines.  Thus, pipeline
sequencing will again be affected. All pipelines will  continue to carry 15 ppm fuel, now for both
the highway and NRLM markets.  Pipelines serving PADD 1 will continue to carry high sulfur
distillate for the heating oil market. However, due to the small size of the heating oil markets
elsewhere (or the lack of pipelines, as in Alaska and Hawaii), we do not  expect that pipelines
other than those serving PADD 1 will carry high sulfur distillate. While some pipelines are likely
to carry some 500 ppm L&M or small refiner fuel,  this is likely to be in proprietary shipments and
not as a fungible product. Thus, in assessing pipeline sequencing, we assume that no 500 ppm
fuel will be regularly present.

    Figure 7.1-5 shows the pipeline sequence for the pipelines in PADDs 1 and 3  which are
expected to carry high  sulfur heating oil in the 2010-2012 timeframe (applies to the period 2012 -
2014 period as well).
                                          7-36

-------
                                                Estimated Costs of Low-Sulfur Fuels
    Figure 7.1-5 Pipeline Sequence and Fate of Interface Between Fuel Batches
         in Areas that Carry Heating Oil; After NRLM Rule: 2010-2012
Jet
L
+ 1

H
75

*•
°/c
Heating Oil
Heating Oil
Batch Swell

•4
+ .

^^
2.2°
15 ppm
Highway and
NRLM Fuel
L
/o



^
Tier 2
Gasoline
\ /

7


ts
i 	 !
J
             Jet
                             Hwy
                                                 1.75% Jet
                                                 2.2% Hwy
                                                 Gasoline in equal amounts
                 Transmix Products
                 Distillate volume = 1.75% Jet + 2.2% Hwy + 1/3 of gasoline in transmix
                 Distillate quality: <500 ppm
   The primary difference between the sequencing in these pipelines in 2010-2012 and 2007-
2010 is the elimination of 500 ppm fuel. However, as discussed in Section 7.1.3.2.2, there was no
net gain or loss in the size of the 500 ppm batch, as it gained fuel from the adjacent batch of 15
ppm fuel and lost the same volume of 500 ppm fuel to the adjacent batch of high sulfur heating
oil. Now, in the absence of the 500 ppm batch, the loss of 15 ppm fuel is cut directly to the
heating oil batch in 2010-2012. The quality of the distillate produced from transmix is also the
same as in 2007-2010. Thus, the volumes and quality of distillate downgrades remain unchanged
from 2007-2010.

   The destination of these downgrades changes, however, due to the elimination of the 500 ppm
highway fuel market. The downgrades of jet fuel and 15 ppm fuel which are cut directly into the
heating oil batch still go directly to the heating oil market. The 500 ppm downgrade material
produced from transmix now is assumed to be used in only the NRLM markets, in proportion to
the demand for nonroad, locomotive and marine fuel in PADD 3. In most of PADD 1, the
Northeast/Mid-Atlantic Area provisions of the final rule prohibit the use of 500 ppm fuel in the
nonroad market.  As the volume of downgrade produced from transmix in PADD 1 was
significantly less than L&M fuel demand, we assumed that all of the distillate produced from
transmix in PADD  1 was used in the L&M fuel market from 2010-2012.

   It should be noted that we continue to assume that 4.4% of highway diesel fuel supply will be
downgraded to protect the quality of 15 ppm diesel fuel. We do not apply the 4.4% downgrade to
the new volume of 15 ppm NRLM diesel fuel supply, because the new 15 ppm NRLM fuel is
assumed to simply increase the size of the existing batches of 15 ppm highway diesel fuel and not
increase the number of interfaces created.
                                          7-37

-------
Final Regulatory Support Document
   Figure 7.1-6 shows the pipeline sequence for the pipelines in PADDs 2, 4 and 5 which are not
expected to carry high sulfur heating oil in the 2010-2012 timeframe (applies to the period 2012 -
2014 period as well).
    Figure 7.1-6 Pipeline Sequence and Fate of Interface Between Fuel Batches
      in Areas that Do Not Carry Heating Oil; After NRLM Rule: 2010-2012
Jet


15 ppm
Highway and
NRLM Fuel


Tier 2
Gasoline


Jet
  Segregated Interface
Volume = 1.75% Jet + 2.2% Hwy
Quality: <500 ppm
                                                 r
                                         Transmix
                                         1.75% Jet
                                         2.2% Hwy
                                         Gasoline in equal amounts
              Transmix Products
              Distillate volume = 1.75% Jet + 2.2% Hwy + 1/3 of gasoline in transmix
              Distillate quality: <500ppm
    The primary difference between the sequencing in these pipelines in 2010-2012 and 2007-
2010 is again the elimination of 500 ppm fuel. Now, in the absence of the 500 ppm batch, the
interface between the batch of jet fuel and the batch of 15 ppm fuel can no longer be cut into
either fuel. The jet fuel specifications will not allow the addition of No. 2 distillate material due
its higher aromatic levels and higher boiling points. The 15 ppm cap will not allow the blending
of jet fuel with its much higher sulfur levels. Thus, this interface will have to be segregated from
both adjacent batches and stored separately at the terminal. We do not expect that this jet-
highway fuel interface  will be mixed with other transmix which contains some gasoline.
Transmix processors simply separate gasoline from distillate material via distillation. Adding a
mixture of jet fuel and  highway fuel to a transmix distillation column will just cause all of this
material to flow to the  distillate product.  No separation will occur. Thus, there is no benefit to
offset the cost of shipping this distillate transmix to the transmix processor and distilling it.
Instead we expect that  the terminal will store this interface in a separate tank and sell it directly to
a market which can use 500 ppm fuel.  In the 2010-2012 timeframe, this is either the NRLM fuel
market or the heating oil market. As assumed for 2007-2010 in Section 7.1.3.2.2, in PADDs 2
and 4 from 2010-2012, we assume that this 500 ppm interface will be sold first to the heating oil
market and then to the  NRLM markets, in proportion to demand. In California, it will  be sold to
the L&M market. In PADD 5 outside of California, it will be sold to the NRLM markets, in
proportion to demand.
                                            7-38

-------
                                                Estimated Costs of Low-Sulfur Fuels
   The volume of the downgrade from jet fuel and 15 ppm highway fuel to this 500 ppm
interface does not change from 2007-2010, as there was no net change in the size of the 500 ppm
batch in 2007-2010. The quality of the distillate produced from transmix is also the same as in
2007-2010. Thus, the volumes and quality of distillate downgrades remain unchanged from those
in 2007-2010.  Table 7.1.3-15 summarizes the destination of downgrade from 2010 to 2012.

                                     Table 7.1.3-15
                Blending of Downgrade Under the NRLM Rule: 2010 to 2012

1st Priority
2nd Priority
PADD1
HO & L&M
-
PADD2
HO
NRLM
PADD3
HO & NRLM
-
PADD4
HO
NRLM
PADD 5-O
NRLM
-
CA
L&M
-
   Finally, small refiners can produce and sell 500 ppm fuel to the NRLM markets during this
timeframe.  We assume that this fuel is generally not distributed in pipelines, so it does not affect
the product shipment sequences shown in Figures 7.1-5 and 7.1-6. We expect that the volume of
this 500 ppm small refiner fuel will decrease somewhat relative to that in 2007-2010.  This occurs
because we do not believe that a small refiner would invest to produce 500 ppm NRLM fuel for
four years unless they also planned to produce 15 ppm NRLM fuel after 2014.  Therefore, we
assumed that only those small refiners which our cost analysis shows as competitive with other
refiners in producing 15 ppm diesel fuel would produce 500 ppm NRLM fuel in the 2010-2014
timeframe.  We assume that the 500 ppm small refiner fuel which is exempted from the 15 ppm
nonroad sulfur standard is blended into the nonroad pool. As in 2007-2010, we combined small
refiner fuel  production in PADDs 1 and 3 and then apportioned it to the two PADDs based on the
relative demands for NRLM fuel in each PADD.P The volume of 500 ppm small  refiner fuel
expected to be exempted in each region is summarized in Table 7.1.3-16.

                                     Table 7.1.3-16
        Small Refiner Fuel Exempted by Region: 2010 - 2012 (million gallons in 2014)
PADD 1
261
PADD 2
140
PADDS
165
PADD 4
4
PADD 5-O
60
AK
30
HI
0
CA
0
   The final projections of production, spillover, downgrade and demand for 2010-2012 under
this final NRLM rule are shown in Table 7.1.3-17.
   p Given the low likelihood that small refiner fuel would be shipped through pipelines, it would have been more
realistic to assume that small refiner fuel produced in PADD 3 would be consumed in that region. This has no
impact on the nationwide emission reductions projected here. However, a greater volume of small refiner fuel would
have been slightly higher emissions of sulfur dioxide and sulfate PM in PADD 3 and slightly lower emissions in
PADD1.
                                          7-39

-------
Final Regulatory Support Document
                                     Table 7.1.3-17
        Distillate Supply and Demand: Final Rule: 2010-2012 (million gallons in 2014)
Fuel Use
Category
High-
way
Non-
road
Loco-
motive
Marine
Heating
Oil
Fuel Type
Production 1 5 ppm
Spillover to Non-hwy
Hwy Downgrade
Demand
Production 1 5 ppm
Small Refiner Fuel
Hwy Spillover
Jet Downgrade
Hwy Downgrade
Reproc. Downgrade
Demand
Production 500 ppm
Hwy Spillover
Jet Downgrade
Hwy Downgrade
Demand
Production 500 ppm
Hwy Spillover
Jet Downgrade
Hwy Downgrade
Demand
Production HS
Hwy Spillover
Jet Downgrade
Hwv Downgrade
PADD
1
15,825
-425
-678
14,722
3,498
283
206
0
0
0
3,987
195
15
76
251
536
173
13
67
222
475
6,313
192
108
357
2
18,487
-2,090
-721
15,676
3,477
139
1,535
92
427
0
5,670
723
287
18
85
1,114
212
84
5
25
327
0
436
94
436
3
9,527
-939
-378
8,210
1,215
136
345
85
133
0
1,914
684
141
43
67
935
868
179
54
85
1,187
1,193
215
137
215
4
2,981
-633
-103
2,245
245
5
451
18
93
0
810
74
129
5
28
236
0
0
0
0
0
0
53
7
37
5-0
3,254
-404
-125
2,725
200
60
333
115
149
0
857
33
40
14
19
106
7
9
3
4
23
37
22
0
0
AK
161
-4
0
157
0
30
4
0
0
0
34
5
0
0
0
5
69
0
0
0
69
199
0
0
0
HI
60
-13
0
46
39
0
4
0
0
0
43
0
0
0
0
0
20
1
0
0
21
114
8
0
0
US-
CA
50,294
-4,508
-2,006
43,781
8,674
654
2,877
310
801
0
13,316
1,714
611
157
450
2,932
1,349
286
130
337
2,103
7,856
734
347
1,045
CA
4,760
-835
-173
3,752
10
0
835
0
0
219
1,064
0
0
144
217
194
0
0
46
59
53
0
0
0
0
us
55,056
-5,343
-2,178
47,533
8,684
654
3,712
310
801
219
14,379
1,714
611
301
667
3,126
1,349
286
176
396
2,156
7,856
734
347
1,045
    7.1.3.2.4 Final Rule Program - 2012 to 2014

    Beginning in mid-2012, the sulfur cap applicable to L&M fuel changes from 500 ppm to 15
ppm.  Also, 500 ppm fuel produced during shipment of 15 ppm fuel (and by small refiners or
using small refiner credits) can continue to be sold to the NRLM fuel markets outside of the
Northeast/Mid-Atlantic Area. However, within the Northeast/Mid-Atlantic Area, downgraded
distillate or small refiner fuel containing more than 15 ppm sulfur can only be sold as heating oil.

    As was the case for 2007-2010 and 2010-2012, the demand for each distillate fuel and the
spillover of highway fuel into these markets are assumed to remain unchanged from those
occurring in the Reference Case (see Table 7.1.3-5). Since we assumed that 500 ppm L&M fuel
would not be widely distributed as a fungible fuel from 2010-2012, the pipeline sequencing
described in Figures 7.1-5 and 7.1-6 continue to apply. Thus, the types and volumes of
downgrade generated in 2010-2012 will continue in 2012-2014.
                                         7-40

-------
                                               Estimated Costs of Low-Sulfur Fuels
   The destination of these downgrades stays the same outside of the Northeast/Mid-Atlantic
Area, as downgraded distillate can continue to be sold to the NRLM market through 2014 (and to
the L&M fuel market thereafter). Within the Northeast/Mid-Atlantic Area, however, downgraded
distillate can no longer be sold to the L&M fuel market.  Thus, starting in mid-2012, the
downgraded distillate generated in the Northeast/Mid-Atlantic Area shifts from the L&M market
to the heating oil market, where it displaces high sulfur distillate. This also causes the volume of
L&M fuel which must be produced to the 15 ppm cap to be larger than that needed under the 500
ppm cap. The small refiner fuel exempted and blended into the 15 ppm sulfur NRLM diesel fuel
pool remains the same as in 2010-2012 except for Alaska.  The volume of small refiner fuel
eligible for exemptions in Alaska is limited by the volume of the 15 ppm market. The additional
production of 15 ppm fuel to satisfy the locomotive and marine market in 2012 in Alaska
increases the volume of small refiner fuel exempted there to the total production of NRLM diesel
fuel. The volume of small refiner fuel exempted is summarized in Table 7.1.3-18.

                                     Table 7.1.3-18
         Small Refiner Fuel Exempted by Region: 2012 - 2014 (million gallons in 2014)
PADDl
261
PADD2
140
PADD3
165
PADD4
4
PADD 5-O
60
AK
104
HI
0
CA
0
   The final projections of production, spillover, downgrade and demand for 2012-2014 under
this final NRLM rule are shown in Table 7.1.3-19.
                                         7-41

-------
Final Regulatory Support Document
                                     Table 7.1.3-19
        Distillate Supply and Demand: Final Rule: 2012-2014 (million gallons in 2014)
Fuel Use
Category
High-
way
Non-
road
Loco
motive
Marine
Heating
Oil
Fuel Type
Production 1 5 ppm
Spillover to Non-hw
Hwy Downgrade
Demand
Production 1 5 ppm
Small Refiner Fuel
Hwy Spillover
Jet Downgrade
Hwy Downgrade
Reproc. Downgrade
Demand
Production 1 5 ppm
Small Refiner Fuel
Hwy Spillover
Jet Downgrade
Hwy Downgrade
Demand
Production 1 5 ppm
Small Refiner Fuel
Hwy Spillover
Jet Downgrade
Hwy Downgrade
Demand
Production HS
Hwy Spillover
Jet Downgrade
Hwy Downgrade
Demand
PADD
1
15,825
-425
-678
14,722
3,574
207
206
0
0
0
3,987
493
29
15
0
0
536
437
25
13
0
0
475
5,697
192
252
830
6.970
2
18,487
-2,090
-721
15,676
3,506
111
1,535
92
427
0
5,670
701
22
287
18
85
1,114
205
7
84
6
26
327
0
184
94
436
714
3
9,527
-939
-378
8,210
1,278
74
345
85
133
0
1,914
647
37
141
43
67
935
820
48
179
54
85
1,187
1,193
274
137
215
1.820
4
2,981
-633
-103
2,245
246
3
451
18
93
0
810
73
1
129
5
28
236
0
0
0
0
0
0
0
53
7
37
98
5-0
3,254
-404
-125
2,725
209
52
333
115
149
0
857
26
7
40
14
19
106
7
3
9
3
4
23
37
22
0
0
59
AK
161
-4
0
157
0
30
4
0
0
0
34
0
5
0
0
0
5
0
69
0
0
0
69
199
0
0
0
199
HI
60
-13
0
46
39
0
4
0
0
0
43
0
0
0
0
0
0
20
0
1
0
0
21
114
8
0
0
122
US-
CA
50,294
-4,508
-2,006
43,781
8,851
477
2,877
310
801
0
13,316
1,931
100
611
82
203
2,932
1,489
150
286
63
116
2,103
7,240
734
490
1,518
9.981
CA
4,760
-835
-173
3,752
10
0
835
0
0
219
1,064
0
0
0
144
217
194
0
0
0
46
59
53
0
0
0
0
0
us
55,054
-5,343
-2,178
47,533
8,861
477
3,712
310
801
219
14,379
1,931
100
611
226
421
3,126
1,489
150
286
109
175
2,156
7,240
734
490
1,518
9.981
    7.1.3.2.5 Final Rule Program - 2014 and Beyond

    The primary changes occurring in 2014 are: 1) the end of the small refiner provisions and 2)
the prohibition on the use of any 500 ppm fuel in the nonroad fuel market. These changes have
no effect on fuel demand in any of the markets of interest here. Spillover of highway fuel into the
other markets is also assumed to be unaffected, with one exception, as discussed below. As
pipelines still  carry the same fuels, the volume of each fuel downgraded is also unaffected.
                                         7-42

-------
                                                Estimated Costs of Low-Sulfur Fuels
   Only the use of 500 ppm downgrade changes, as this fuel can no longer be sold into the
nonroad fuel market.  Therefore, we assumed that it would be used in either the L&M fuel market
or the heating oil market according to the same relative priorities described in Table 7.1.3-15.  In
a few cases, the volume of downgrade exceeds the demand for all L&M fuel and heating oil in a
region, considering the historical level of highway fuel spillover. In those cases, we reduced the
volume of spillover of highway fuel into these markets until  demand for non-spillover fuel
equaled that of the available downgrade.  If the volume of available downgrade exceeded total
demand for L&M fuel and heating oil in a region (i.e., zero spillover), we assume that the excess
downgrade fuel will be returned to a refinery and be reprocessed into 15 ppm fuel.  The
projections of production, spillover, downgrade and demand for 2014 and beyond under this
NRLM rule are shown in Table 7.1.3-20.
                                         7-43

-------
Final Regulatory Support Document
                                     Table 7.1.3-20
     Distillate Supply and Demand: Final Rule: 2014 and Beyond (million gallons in 2014)
Fuel Use
Category
High-
way
Non-
road
Loco
motive
Marine
Heating
Oil
Fuel Type
Production 15 ppm
Spillover to Non-
Hwy Downgrade
Demand
Production 1 5 ppm
Hwy Spillover
Jet Downgrade
Hwy Downgrade
Reprocessed
Downgrade
Demand
Production 15 ppm
Hwy Spillover
Jet Downgrade
Hwy Downgrade
Demand
Production 15 ppm
Hwy Spillover
Jet Downgrade
Hwy Downgrade
Demand
Production HS
Hwy Spillover
Jet Downgrade
Hwy Downgrade
Demand
PADD
1
15,825
-425
-678
14,722
3,781
206
0
0
0
3,987
522
15
1
0
536
462
13
0
0
475
5,697
192
252
830
6,970
2
18,487
-2,090
-721
15,676
4,136
1,535
0
0
0
5,670
142
287
122
563
1,114
243
84
0
0
327
0
184
94
436
714
3
9,527
-939
-378
8,210
1,568
345
0
0
0
1,914
443
141
137
215
935
894
179
45
70
1,187
1,193
274
137
215
1,820
4
2,981
-633
-103
2,245
321
490
0
0
0
810
0
90
24
122
236
0
0
0
0
0
0
53
7
37
98
5-0
3,254
-404
-125
2,725
336
404
0
0
116
857
0
0
46
60
106
0
0
61
78
23
0
0
26
33
59
AK
161
-4
0
157
30
4
0
0
0
34
5
0
0
0
5
69
0
0
0
69
199
0
0
0
199
HI
60
-13
0
46
39
4
0
0
0
43
0
0
0
0
0
20
1
0
0
21
114
8
0
0
122
US-
CA
50,294
-4,508
-2,006
43,781
10,211
2,986
0
0
116
13,316
1,111
532
328
960
2,932
1,687
Til
105
149
2,103
7,202
712
516
1,552
9,981
CA
4,760
-835
-173
3,752
10
835
0
0
219
1,064
0
0
144
217
194
0
0
46
59
53
0
0
0
0
0
us
55,056
-5,343
-2,178
47,533
10,221
3,821
0
0
335
14,379
1,111
532
472
1,177
3,126
1,687
111
151
208
2,156
7,202
712
516
1,552
9,981
7.1.4 Sensitivity Cases

   Distillate fuel production and demand were estimated for three sensitivity cases. The first
sensitivity case represents an indefinite 500 ppm cap on NRLM fuel that takes effect in 2007 (i.e.,
no subsequent 15 ppm cap).  The second sensitivity case analyzes the proposed rule, which would
not require locomotive and marine diesel fuel be desulfurized to 15 ppm.  The last sensitivity case
                                          7-44

-------
                                               Estimated Costs of Low-Sulfur Fuels
analyzes the final rule, but bases the demand for nonroad fuel on information from EIA reports
rather than EPA's draft NONROAD2004 model.

   7.1.4.1 NRLM Regulated to 500 ppm Indefinitely

   To support the legal justification of the 500 ppm cap on NRLM fuel in 2007, we evaluate the
costs and benefits of this standard in the absence of a subsequent 15 ppm cap on NRLM fuel.
Here, we estimate the production and demand for the various distillate fuels in 2014 under this
indefinite 500 ppm cap on NRLM fuel.

   During the period from 2007 to 2010, distillate fuel production and demand  under this
indefinite 500 ppm NRLM fuel cap are assumed to be the same as under the FRM (see Table
7.1.3-14). After 2010, the only differences are the end of the small  refiner provisions for
producing high sulfur NRLM fuel and the end of the temporary compliance option under the
highway fuel program. These two changes are assumed to not affect the demand for the various
distillate fuels, nor the spillover of highway fuel into the NRLM fuel and heating oil markets.

   The types and volumes of distillate downgrade is not affected, since 500 ppm NRLM fuel will
still be carried in all pipelines. However, the disposition of this downgraded distillate is affected
slightly, since 500 ppm downgraded distillate can no longer be sold into the 500 ppm highway
market.  The disposition of downgraded distillate as summarized in Tables 7.1.3-10 through
7.1.3-12 still apply except for the removal of 500 ppm highway fuel as an option for use of this
downgraded distillate. The final projections of production, spillover, downgrade and demand for
2010 and beyond under this NRLM rule are shown in Table 7.1.4-1.
                                         7-45

-------
Final Regulatory Support Document
                                    Table 7.1.4-1
       Distillate Fuel Supply and Demand in 2010 and Beyond (million gallons in 2014)
                             NRLM at 500 ppm Indefinitely
Fuel Use
Category
High-
way
Non-
road
Loco-
motive
Marine
Heating
Oil
Fuel Type
Production 1 5 ppm
Spillover to Non-
Hwy Downgrade
Demand
Production 500 ppm
Hwy Spillover
Jet Downgrade
Hwy Downgrade
Reproc. Downgrade
Demand
Production 500 ppm
Hwy Spillover
Jet Downgrade
Hwy Downgrade
Demand
Production 500 ppm
Hwy Spillover
Jet Downgrade
Hwy Downgrade
Demand
Production HS
Hwy Spillover
Jet Downgrade
Hwy Downgrade
Demand
PADD
1
15,825
-425
-678
14,722
3,293
206
114
375
0
3,987
454
15
16
52
536
402
13
14
46
475
6,313
192
108
357
6.970
2
18,487
-2,090
-721
15,676
3,617
1,535
92
427
0
5,670
723
287
18
85
1,114
211
84
6
26
327
0
184
94
436
714
3
9,527
-939
-378
8,210
1,351
345
84
133
0
1,914
685
141
43
67
935
869
179
54
85
1,187
1,193
274
137
215
1.820
4
2,981
-633
-103
2,245
249
451
18
93
0
810
73
129
5
28
236
0
0
0
0
0
0
53
7
37
98
5-0
3,254
-404
-125
2,725
261
333
115
149
0
857
33
40
14
19
106
7
9
3
4
23
37
22
0
0
59
AK
161
-4
0
157
30
4
0
0
0
34
5
0
0
0
5
69
0
0
0
69
199
0
0
0
199
HI
60
-13
0
46
39
4
0
0
0
43
0
0
0
0
0
20
1
0
0
21
114
8
0
0
122
US-
CA
50,294
-4,508
-2,006
43,781
8,839
2,877
424
1,177
0
13,316
1,973
611
98
255
2,932
1,578
286
77
161
2,103
7,856
734
347
1,045
9.981
CA
4,760
-835
-173
3,752
10
835
0
0
219
1,064
0
0
144
217
194
0
53
46
59
53
0
0
0
0
0
us
55,056
-5,343
-2,178
47,533
8,849
3,712
424
1,177
219
14,379
1,973
611
242
472
3,126
1,578
339
123
221
2,156
7,856
734
347
1,045
9.981
   7.1.4.2 Proposed Rule - 500 ppm NRLM Cap in 2007; 15 ppm Nonroad Fuel Cap in 2010

   This second sensitivity case evaluates the NRLM fuel program proposed in the NPRM. This
case is the same as that proposed, except that the Northeast/Mid-Atlantic Area provisions were
added not allowing small refiner fuel and downgrade to be used in the 15 ppm nonroad diesel fuel
pool in most of PADD 1 after 2010. Thus, from 2007 to 2012, the program is the same as the
final NRLM fuel program. After 2012, the difference is that L&M fuel remains at 500 ppm and
that the Northeast/Mid-Atlantic Area restrictions would apply to only the nonroad pool in PADD
1, not the NRLM pool as is the case for the final NRLM program.  Since there are no differences
between this case and the final NRLM program during the period from 2007 to 2010 the distillate
production and demand estimates shown in Table 7.1.3-14 are assumed to apply here,  as well.
                                        7-46

-------
                                               Estimated Costs of Low-Sulfur Fuels
   From 2010 to 2012, there are no differences in the regulatory requirements of the proposed
and final NRLM fuel programs.  Thus, distillate fuel demand, spillover of highway fuel to non-
highway markets, and the types and volume of downgrade are the same under both programs.
The small refiner fuel volume exempted from the 15 ppm sulfur standard and is blended into the
nonroad diesel fuel pool.  The small refiner fuel volume is the same as that summarized in Table
7.1.3-16. Nothing changes in 2012 under the proposed NRLM program.  Thus, the production,
downgrade, spillover and demand volumes are the same over the entire period from 2010 to 2014.
The final projections of production,  spillover, downgrade and demand for 2010 to 2014 under this
proposed rule sensitivity case are shown in Table 7.1.4-2.

                                     Table 7.1.4-2
          Distillate Fuel Supply and Demand in 2010 - 2014 (million gallons in 2014)
                        15 ppm Nonroad Cap, 500 ppm L&M Cap
Fuel Use
Category
High-
way
Non-
road
Loco-
motive
Marine
Heating
Oil
Fuel Type
Production 1 5 ppm
Spillover to Non-hwy
Hwy Downgrade
Demand
Production 1 5 ppm
Small Refiner Fuel
Hwy Spillover
Jet Downgrade
Hwy Downgrade
Reproc. Downgrade
Demand
Production 500 ppm
Hwy Spillover
Jet Downgrade
Hwy Downgrade
Demand
Production 500 ppm
Hwy Spillover
Jet Downgrade
Hwy Downgrade
Demand
Production HS
Hwy Spillover
Jet Downgrade
Hwv Downarade
PADD
1
15,825
-425
-678
14,722
3,498
283
206
0
0
0
3,987
195
15
76
251
536
173
13
67
222
475
6,313
192
108
357
2
18,487
-2,090
-721
15,676
3,477
139
1,535
92
427
0
5,670
723
287
18
85
1,114
212
84
5
25
327
0
436
94
436
3
9,527
-939
-378
8,210
1,215
136
345
85
133
0
1,914
684
141
43
67
935
868
179
54
85
1,187
1,193
215
137
215
4
2,981
-633
-103
2,245
245
5
451
18
93
0
810
74
129
5
28
236
0
0
0
0
0
0
53
7
37
5-O
3,254
-404
-125
2,725
200
60
333
115
149
0
857
33
40
14
19
106
7
9
3
4
23
37
22
0
0
AK
161
-4
0
157
0
30
4
0
0
0
34
5
0
0
0
5
69
0
0
0
69
199
0
0
0
HI
60
-13
0
46
39
0
4
0
0
0
43
0
0
0
0
0
20
1
0
0
21
114
8
0
0
US-
CA
50,294
-4,508
-2,006
43,781
8,674
654
2,877
310
801
0
13,316
1,714
611
157
450
2,932
1,349
286
130
337
2,103
7,856
734
347
1.045
CA
4,760
-835
-173
3,752
10
0
835
0
0
219
1,064
0
0
144
217
194
0
0
46
59
53
0
0
0
0
US
55,056
-5,343
-2,178
47,533
8,684
654
3,712
310
801
219
14,379
1,714
611
301
667
3,126
1,349
286
176
396
2,156
7,856
734
347
1.045
   After 2014, the small refiner provisions end and downgraded distillate can no longer be sold
to the nonroad fuel market.  Downgrade can only be used in the L&M and heating oil markets.
                                         7-47

-------
Final Regulatory Support Document
The final projections of production, spillover, downgrade and demand for 2014 and beyond for
the proposed rule are shown in Table 7.1.4-3.

                                    Table 7.1.4-3
       Distillate Fuel Supply and Demand in 2014 and Beyond (million gallons in 2014)
                        15 ppm Nonroad Cap, 500 ppm L&M Cap

High-
way
Non-
road
Loco-
motive
Marine
Heating
Oil
Fuel Type
Production 1 5 ppm
Spillover to Non-hwy
Hwy Downgrade
Demand
Production 1 5 ppm
Hwy Spillover
Jet Downgrade
Hwy Downgrade
Reprocessed
Downgrade
Demand
Production 500 ppm
Hwy Spillover
Jet Downgrade
Hwy Downgrade
Demand
Production 500 ppm
Hwy Spillover
Jet Downgrade
Hwy Downgrade
Demand
Production HS
Hwy Spillover
Jet Downgrade
Hwy Downgrade
Demand

1
15,825
-425
-678
14,722
3,781
206
0
0
0
3,987
195
15
76
251
536
172
13
67
222
475
6,313
192
108
357
6,970
2
18,487
-2,090
-721
15,676
4,136
1,535
0
0
0
5,670
142
287
122
563
1,114
243
84
0
0
327
0
184
94
436
714
3
9,527
-939
-378
8,210
1,568
345
0
0
0
1,914
443
141
137
215
935
894
179
45
70
1,187
1,193
274
137
215
1,820
4
2,981
-633
-103
2,245
323
488
0
0
0
810
0
90
24
122
236
0
0
0
0
0
0
53
7
37
98
5-0
3,254
-404
-125
2,725
338
404
0
0
116
857
0
0
46
60
106
0
0
61
78
23
0
0
26
33
59
AK
161
-4
0
157
30
4
0
0
0
34
5
0
0
0
5
69
0
0
0
69
199
0
0
0
199
HI
60
-13
0
46
39
4
0
0
0
43
0
0
0
0
0
20
1
0
0
21
114
8
0
0
122
US-
CA
50,294
-4,508
-2,006
43,781
10,215
2,985
0
0
116
13,316
816
1,106
399
1,183
2,932
1,398
Til
173
371
2,103
7,819
712
373
1,079
9,981
CA
4,760
-835
-173
3,752
10
835
0
0
219
1,064
0
0
144
217
194
0
0
46
59
53
0
0
0
0
0
US
55,056
-5,343
-2,178
47,533
10,225
3,820
0
0
335
14,379
816
1,106
543
1,401
3,126
1,398
111
219
430
2,156
7,819
712
373
1,079
9,981
   7.1.4.3 Final NRLM Fuel Program With Nonroad Fuel Demand Derived from EIA
   FOKS and AEO

   This sensitivity case evaluates the final NRLM fuel program assuming a reduced level of
nonroad fuel demand.  As discussed in Section 2.4.5 of the Summary and Analysis document for
this rule, a number of commenters claimed that EPA's NONROAD model overestimates nonroad
fuel demand. To ensure that uncertainties in the level of nonroad fuel demand do not affect the
decisions being made in this NRLM rule, we evaluate the cost, emission reductions and cost
effectiveness of the final NRLM fuel program using an estimate of nonroad fuel demand derived
                                        7-48

-------
                                               Estimated Costs of Low-Sulfur Fuels
from EIA's FOKS and AEO reports. Thus, the first step in this sensitivity analysis is to derive
this lower nonroad fuel demand. Then, we will discuss how this affects spillover, downgrade and
production of the various distillate fuels.

   We based nonroad fuel demand for the purpose of estimating fuel costs in the NPRM on the
information contained in EIA's FOKS and AEO reports. The methodology used here is
essentially the same as that used in the NPRM. The primary difference is the use of more recent
EIA FOKS and AEO reports.  In the NPRM, we used the 2000 FOKS and 2002 AEO reports.
Here, we use the 2001 FOKS and 2003 AEO reports. We start with our derivation of nonroad
fuel demand in 2001 using 2001 FOKS and then adjust this estimate for growth using 2003 AEO.

    7.1.4.3.1 Nonroad Fuel Demand in 2001 Derived from EIA FOKS

   This section describes our methodology for deriving nonroad fuel demand from information
collected and projections made by EIA. For a more detailed description of the EIA FOKS
information collection process and how estimates of nonroad fuel can be derived from it, the
reader is referred to the draft RIA for this rule.  As described in Section 7.1.2, EIA's FOKS
estimates distillate demand in eleven economic sectors. FOKS also breaks down the distillate
demand for several of these sectors according to the physical type of distillate used.  Table 7.1.4-4
presents the "adjusted" estimated of distillate fuel demand for PADD 1 from the 2001 FOKS
report.
                                         7-49

-------
Final Regulatory Support Document
                                    Table 7.1.4-4
                Nonroad Fuel Demand, PADD 1 Estimates from 2001 FOKS
End Use
Farm
Construction
Other/(Logging)
Industrial
Commercial
Oil Company
Military
Electric Utility
Railroad
Vessel Bunkering
On-Highway
Residential
Total
Fuel Grade
diesel
distillate
distillate
distillate
No. 2 fuel oil
No. 4 distillate
No. 1 distillate
No. 2 low-S diesel
No. 2 high-S diesel
No. 2 fuel oil
No. 4 distillate
No. 1 distillate
No. 2 low-S diesel
No. 2 high-S diesel
distillate
diesel
distillate
distillate
distillate
distillate
diesel
No. 2 fuel oil
No. 1 distillate

Distillate*
(Mgal)
447
41
550
149
226
40
1
118
374
1,369
200
2
450
203
21
45
28
564
506
461
10,284
5,464
5
21,548
Diesel
(%)
100
0
95
95
0
0
40
100
100
0
0
40
100
100
50
100
0
100
95
90
100
0
0
-
Diesel
(M gal)
447
0
523
142
0
0
0.4
118
374
0
0
0.8
450
203
10.5
45
0
564
481
415
10,284
0
0
14,058
Nonroad
(%)
100
0
100
100
0
0
100
100
100
0
0
50
0
100
100
85
0
0
1.0
0
0.7
0
0

Nonroad
(Mgal)
447
0
523
142
0
0
0.4
118
374
0
0
0.4
0
203
11
38
0
0
5
0
73
0
0
1,934
   The key step in our methodology is the estimation of the portion of each sector's fuel demand
that is used in nonroad engines.  These percentages are summarized in Table 7.1.4-4.  We
describe these estimates below.
                                        7-50

-------
                                                 Estimated Costs of Low-Sulfur Fuels
   Farm. FOKS estimates fuel demand in this sector for two fuel grades: "diesel fuel" and
"distillate." We assume that 100 percent of the diesel fuel represents nonroad use, and 100
percent of the distillate represents uses other than in nonroad engines, such as heating and crop
drying.

   Construction/Other Off-Highway (Logging).  For the construction and logging/other-non-
highway end uses, we assume that 95 percent of the total distillate sold is diesel fuel, and that 100
percent of the diesel fuel is used in nonroad engines.

   Industrial. FOKS breaks down distillate sales in this sector into five individual fuel grades:
No. 1 distillate, low sulfur No. 2 diesel, high sulfur No. 2 diesel fuel, high sulfur No. 2 fuel oil
and No. 4 distillate. No. 4 distillate is not covered by the NRLM rule and is rarely used in
nonroad engines, if at all.  Therefore, we exclude all sales of No. 4 distillate from our estimate of
nonroad fuel use. Since sales of No. 2 diesel fuel and No. 2 fuel oil  are categorized separately,
we assume that no No. 2 fuel oil is used in diesel engines. Thus, no No. 2 fuel oil sales are
assumed to fall into nonroad fuel demand.  Conversely, we assume that all No. 2 diesel fuel, low-
sulfur and high-sulfur, is used in diesel engines and that all of this diesel fuel represents nonroad
use.  As will be seen below, the low sulfur diesel fuel in the commercial sector is most often used
in highway vehicles owned by "commercial" entities not subject to highway  excise taxes. We are
not aware of any "industrial" entities which are not subject to the excise tax.  Thus, should an
industrial entity use this low sulfur diesel fuel  in a highway vehicle that it owns, this use would be
included in the FOKS estimate of highway diesel fuel sales, since the latter is based on excise tax
receipts.  Therefore, it is reasonable to assume that the  low  sulfur diesel fuel is not used in
highway vehicles.  The industrial  sector does not include either locomotives or marine vessels.
Thus, the non-highway diesel engines must be either nonroad engines or stationary diesel engines
likely used for power generation.  We assume that the latter use is negligible.  For the remaining
category, No. 1 distillate, diesel and fuel oil are not distinguished. After consulting with EIA
staff, we estimate that 40 percent of No.  1 distillate sales represent diesel fuel, that 100 percent of
this diesel represents nonroad use, and that the remainder represents No. 1 fuel oil used in other
applications, such as space heating.

   Commercial. As with the industrial end use, distillate sales in this sector are reported by fuel
grade.  As in the industrial sector, we assume that none of the No. 2  fuel oil, and No. 4 fuel
represents nonroad diesel fuel. However, in the commercial sector, we assume that all low sulfur
diesel fuel sold is used in highway vehicles. This sector includes school-bus and government
(local, state  and federal) fleets. Fuel used by these fleets are exempt from the federal excise tax,
as is fuel for nonroad use.  Thus, we assume that none of the low-sulfur No. 2 diesel fuel sold to
this sector is used in nonroad engines. As in the industrial sector, we assume that 100 percent of
the high-sulfur No. 2 diesel fuel sold is used in nonroad engines.  Also as in the industrial sector,
after consultation with EIA staff, we estimate that 40 percent of the No. 1 distillate  sold is diesel
fuel. However, due to the presence of public fleet fuel use in this sector, we estimate that only 50
percent of this diesel fuel is used in nonroad engines.
                                           7-51

-------
Final Regulatory Support Document
   Oil Company. Sales to this sector include fuel purchased for drilling and refinery operations.
We assume that 50 percent of the reported distillate is diesel fuel, and that all of this diesel fuel is
used in nonroad equipment. We assume that the remainder represents other uses such as
underground injection under pressure to fracture rock.

   Military. Fuel sales to the military are reported as being either diesel fuel or distillate. We
assume that 85 percent of diesel fuel sales is used in 'non-tactical' nonroad equipment, and that
none of the distillate sales represents nonroad use. We assume that 15% of the diesel fuel is not
used in nonroad engines because the NONROAD model does not attempt to represent fuel use or
emissions from 'tactical' military equipment,  such as tanks and personnel carriers because they
are not covered by EPA emission standards.

   Railroad.  We believe that the vast majority of fuel sales to railroads is used by locomotives.
Based on guidance from a major railroad, we assume that a small fraction (1%) of reported fuel
sales is used in nonroad equipment operated by railroads.

   Electric Utility, Vessel Bunkering and Residential, We assume that all of the fuel sold to these
sectors falls into  our definition of marine fuel or heating oil and that none of it is used in nonroad
engines..

   The EIA FOKS report presents fuel sales by sector for each region of interest here. Thus, we
applied the diesel fuel and nonroad percentages shown in Table 7.1.4-4 to the fuel sales in each
sector and region to estimate nonroad fuel demand. The results are summarized in Table 7.1.4-5.
                                          7-52

-------
                                               Estimated Costs of Low-Sulfur Fuels
         2001
                       Table 7.1.4-5
Nonroad Fuel Consumption Derived From EIA FOKS (million gallons)
End Use
Farm
Construction
Other/(Logging)
Industrial
Commercial
Oil Company
Military
Electric Utility
Railroad
Subtotal
Highway (Retail
Purchases)
Total
Fuel Grade
diesel
distillate
distillate
distillate
No. 2 fuel oil
No. 4 distillate
No. 1 distillate
No. 2 low-S diesel
No. 2 high-S diesel
No. 2 fuel oil
No. 4 distillate
No. 1 distillate
No. 2 low-S diesel
No. 2 high-S diesel
distillate
diesel
distillate
distillate
distillate

diesel

Region
1
447
0
523
142
0
0
0.5
118
374
0
0
0.5
0
203
11
38
0
0
5
1,862
73
1,934
2
1,764
0
572
66
0
0
8
210
355
0
0
7
0
155
26
15
0
0
10
3,188
73
3,261
3
627
0
425
136
0
0
1
196
204
0
0
0.3
0
71
344
105
0
0
8
2,119
50
2,169
4
155
0
118
21
0
0
4
175
15
0
0
2
0
8
10
4
0
0
2
514
13
527
5-O
90
0
83
23
0
0
0.2
101
66
0
0
0.4
0
19
1.5
50
0
0
1
436
10
446
AK
0
0
7
3
0
0
4
2
13
0
0
2
0
21
14
5
0
0
0.04
69
3
72
HI
7
0
3
0
0
0
0
2
0.6
0
0
0
0
3
0
22
0
0
0
38
1
39
CA
281
0
251
17
0
0
0
44
5
0
0
0
0
3
4
24
0
0
2
611
25
636
   Table 7.1.4-5 shows that, according to the above methodology, the farm, construction,
commercial, and industrial categories are the largest consumers of nonroad diesel fuel. Nonroad
fuel use on farms is concentrated in PADD 2 (the Midwest), while nonroad fuel demand in the
other sectors is spread out more evenly across the nation.

   We replaced the year 2001 nonroad fuel demand estimates shown in Table 7.1.2-3 from
EPA's NONROAD model with those shown in the last line of Table 7.1.4-5.  We recalculated the
heating oil demand in each region so that the total fuel demand in the five categories matched the
total distillate demand shown. Table 7.1.4-6 shows the revised estimates of fuel demand by
region for each of the five usage categories.
                                         7-53

-------
Final Regulatory Support Document
                                     Table 7.1.4-6
          2001 Distillate Fuel Demand as Derived From EIA FOKS (million gallons)

EPA Use Category
Highway Fuel
Nonroad Fuel
Locomotive Fuel
Marine Fuel
Heating Oil
Total Demand
Region
1
10,211
1,934
476
415
8,512
21,549
2
10,873
3,261
989
286
1,682
17,092
3
5,694
2,169
831
1,037
1,202
10,932
4
1,557
527
209
0
175
2,468
5-O
1,890
446
94
20
249
2,700
AK
108
72
4
60
167
412
HI
32
38
0
18
125
214
CA
2,602
637
172
46
146
3,604
   The volume of spillover of highway fuel into the four non-highway fuel categories is the same
as that shown in Table 7.1.2-5. We considered the volume of unrefunded fuel for this case as
well.  Since we are basing nonroad fuel demand in this sensitivity case on information contained
in FOKS, we adjust both the highway fuel demand and the nonroad fuel demand for unrefunded
use of highway fuel in nonroad equipment. The volume of unrefunded fuel is the same as that
used for the final rule case, shown in Table 7.1.2-2.  The types and volume percentages of
downgrade of highway fuel, jet fuel and gasoline are the same as those shown in Table 7.1.2-6.
However, we do not show a complete breakdown of production, spillover, downgrade and
demand for each usage category and region for 2001 (analogous to that shown in Table 7.1.2-8),
since these figures are not used directly in the estimates of either costs, nor emission reductions in
this sensitivity analysis.

   7.1.4.3.2 Nonroad Fuel Demand in 2014 Derived from EIA AEO 2003

   We developed an estimate of nonroad fuel  demand in 2014 from EIA's AEO 2003 report. We
began with a detailed set of distillate fuel consumption estimates for the various economic sectors
presented in AEO 2003.  AEO 2003 presents distillate fuel consumption estimates at roughly
three levels of detail, as shown in Table 7.1.4-7 below.
                                         7-54

-------
                                                Estimated Costs of Low-Sulfur Fuels
                                      Table 7.1.4-7
                   Distillate Fuel Consumption Demand within AEO 2003
First Level
Total
Second Level
Transportation
Residential
Commercial
Industrial
Electricity Generation
Third Level
Highway
Rail
Marine
Military
Residential
Commercial
Farm
Oil Company
Construction
Other *
Electricity Generation
Nonroad Fuel Percentage
0.7%
1%
0%
76%
0%
14%
98%
50%
95%
82%
0%
* Not explicitly shown in AEO 2003. Backcalculated from total "Industrial" fuel use.
   At the third level of detail from AEO 2003, we utilized distillate fuel consumption estimates
from AEO to estimate future nonroad demand. The one exception was the "other" industrial
sector. This estimate was obtained by subtracting the demand in the farm, construction and oil
company sectors from that in the total industrial sector. We converted all these estimates of fuel
consumption from AEO from quadrillion BTU per year to gallons per year using EIA's
conversion factor of 138,700 BTU/gal.  When available, we estimated the nonroad percentage of
each sector's total distillate fuel consumption using the same methodology which we used with
the FOKS estimates above. These estimates are available for all the sectors except commercial,
"other" industrial, farm, and military. The estimates of the nonroad portion of total distillate
demand for these four sectors depended on the type of distillate fuel consumed, such as low sulfur
diesel fuel, kerosene, etc.  AEO 2003 does not provide projections broken down by the type of
distillate fuel, only total distillate. In these cases, we used the nonroad diesel fuel fractions found
                                          7-55

-------
Final Regulatory Support Document
from the analysis of the 2002 FOKS.Q  All of these nonroad fuel percentages are shown in Table
7.1.4-8.

   Table 7.1.4-8 presents total distillate demand by sector for 2002 and projected total distillate
demand for 2014 from AEO 2003, the percentage of each fuel demand that is assumed to be
nonroad, and the resulting 2014 nonroad fuel demand by sector.

                                      Table 7.1.4-8
       2002 and 2014 Nonroad Diesel Fuel Demand: 2003 AEO (million gallons per year)
Category
Year
Commercial
Other Industrial
Highway
Oil Company
Farm
Railroad
Military
Construction
Total
Total Distillate Demand
2002
3244
2653
32,242
43
3403
3669
800
1687
—
2014
3533
3331
48,839
0
3843
4196
894
1983
—
Nonroad Diesel (%)*
2002 & 20 14
14%
82%
0.7%
50%
98%
1%
76%
95%
—
Nonroad Diesel Fuel Demand
2002
458
2164
221
22
3320
35
607
1603
8428
2014
498
2717
257
0
3749
40
678
1884
9823
 : Derived by applying EPA estimates of nonroad fuel use to FOKS 2002 fuel sales.
   As shown in Table 7.1.4-8, from information contained in both FOKS 2002 and AEO 2003,
total nonroad fuel demand in 2014 is projected to be 9.82 billion gallons per year.  This represents
a 17% increase over the 8.43 billion gallons demand estimated for 2002, or 1.37% per year linear
growth from a 2002 base.  The growth rates embedded in AEO 2003 vary slightly from year to
year and decade to decade.  However, as the purpose of this analysis is simply to evaluate the
sensitivity of the cost effectiveness of the NRLM rule to uncertainty in nonroad fuel consumption,
we have applied this 1.37% growth rate from 2001 through the final year of analysis, 2040. We
based the growth rate off of fuel consumption in 2002, rather than 2001, because FOKS 2002
shows a significant drop in distillate fuel consumption in 2002.  The AEO 2003 estimates reflect
this decrease in 2002 and projects relatively steady growth starting from 2002.  Thus, reflecting
   Q The projection of nonroad fuel demand using the NONROAD model was already complete and subsequent
analyses of emission benefits, monetized benefits and economic impacts were underway when FOKS 2002 was
issued in late November 2003. Therefore, it was not possible to utilize FOKS 2002 for the primary estimates
presented in this Final RIA. However, it was possible to utilize this more recent information for this sensitivity
analysis.
                                           7-56

-------
                                               Estimated Costs of Low-Sulfur Fuels
this drop in nonroad diesel fuel consumption in 2002 and steady growth thereafter better reflects
the AEO 2003 projections.  Projecting growth from 2001 would have reduced the annual growth
rate considerably, over-predicting fuel consumption prior to 2014 and under-predicting fuel
consumption after 2014.

   We used the same 2001-2014 growth ratios for the other four fuel use categories as shown in
Tables 7.1.3-1 and 7.1.3-3.  These growth ratios were applied to the demand volumes in Table
7.1.4-7 to estimate fuel demand in 2014.  We increased the 2001 nonroad fuel consumption of
9.084 billion gallons (shown in Table 7.1.4-7) by 8.14%, which is the total increase between the
2014 fuel demand of 9.823 billion gallons shown in Table 7.1.4-8 and 2001 nonroad fuel demand.
These volumes are summarized in Table 7.1.4-9.

                                     Table 7.1.4-9
      2014 Distillate Fuel Demand based on AEO 2003 and FOKS 2002 (million gallons)

EPA Use Category
Highway Fuel
Nonroad Fuel
Locomotive Fuel
Marine Fuel
Heating Oil
Region
1
14,738
2,104
536
475
7,898
2
15,693
3,603
1114
327
1,561
3
8,221
2,394
935
1187
1,115
4
2,248
581
236
0
162
5-0
2,728
492
106
23
231
AK
157
78
5
69
155
HI
47
43
0
21
116
CA
3,758
691
194
53
136
    The volume of spillover of highway fuel into the four non-highway fuel categories is the same
as that shown in Table 7.1.3-5. The types and volume percentages of downgrade of highway fuel,
jet fuel and gasoline are the same as those shown in Table 7.1.3-6. Jet fuel demand is the same as
shown in Table 7.1.3-7. We also used the same methodology to assign downgrade to the various
distillate markets. Finally, the volume of NRLM fuel produced by small refiners is the same as
that shown in Table 7.1.3-16.

    We do not show a complete breakdown of production, spillover, downgrade and demand for
each usage category and region for 2010-2014 or 2014 and beyond in a Reference Case (which
assumes no implementation of this nonroad rule).  This is not necessary because we used a
different methodology to estimate the emission reductions for this case than for the final rule case
which did not require the estimation of reference case sulfur levels. Tables 7.1.4-10 through
7.4.1-13 present the estimates of distillate demand and production for the four time periods
relevant to this nonroad rule: 2007-2010, 2010-2012, 2012-2014, and 2014 and beyond,
respectively.
                                         7-57

-------
Final Regulatory Support Document
                                    Table 7.1.4-10
        Distillate Supply and Demand: Final Rule: 2007-2010 (million gallons in 2014)
                Nonroad Fuel Demand Derived from EIA FOKS and AEO R
Fuel Use
Categor
y
High-
way
Non-
road
Loco-
motive
Marine
Heating
Oil
Fuel Type
Production 1 5 ppm
Prod 500 ppm
Spillover
Hwy Downgrade 15
Jet Downgrade
Hwy Downgrade
Demand 1 5 ppm
Demand 500 ppm
Production 500
ppm
Small Refiner Fuel
Hwy Spillover
Jet Downgrade
Hwy Downgrade
Reproc. Downgrade
Demand
Production 500
ppm
Small Refiner Fuel
Hwy Spillover
Jet Downgrade
Hwy Downgrade
Demand
Production 500
ppm
Small Refiner Fuel
Hwy Spillover
Jet Downgrade
Hwy Downgrade
Demand
Production HS
Hwy Spillover
Jet Downgrade
PADD
1
14,347
860
-388
-679
129
465
13,303
1,433
1,825
211
143
0
0
0
2,178
468
54
15
0
0
536
414
48
13
0
0
475
7,233
217
98
2
16,382
1822
-1798
-717
106
534
14,048
1,642
2,606
100
1,025
0
0
0
3,730
797
31
287
0
0
1,114
234
9
84
0
0
327
28
402
187
3
8,589
540
-910
-375
139
239
7,358
861
1,807
212
423
14
23
0
2,479
698
82
141
5
9
935
886
104
179
6
11
1,187
612
168
124
4
2,601
199
-553
-101
15
83
1,987
261
261
3
335
0
2
0
601
105
1
129
0
1
236
0
0
0
0
0
0
0
89
11
5-0
2,882
181
-336
-125
51
71
2,441
286
139
48
200
51
72
0
510
29
10
40
11
15
106
6
2
9
2
3
23
144
87
0
AK
152
8
-3
0
0
0
149
8
28
49
3
0
0
0
81
2
3
0
0
0
5
25
44
0
0
0
69
155
0
0
HI
56
4
-13
0
0
0
44
3
41
0
4
0
0
0
44
0
0
0
0
0
0
20
0
1
0
0
21
109
8
0
US-
CA
45,030
3595
-4001
-1,997
440
1,392
39,328
4,494
6,706
623
2,132
65
97
0
9,624
2,098
181
611
16
25
2,932
1,585
207
286
9
15
2,103
8,280
971
419
CA
4,547
0
-622
-173
0
0
3,752
0
7
0
614
0
0
95
715
0
0
0
85
110
194
0
0
0
64
83
53
0
8
56
US
49,577
3595
-4623
-2,170
440
1,392
43,080
4,494
6,712
623
2,746
65
97
95
10,339
2,098
181
611
102
135
3,126
1,585
207
286
74
98
2,156
8,953
980
475
   R The jet and highway-based downgrade volumes shown in this table were over-estimated by 10% and 2%,
respectively.
                                        7-58

-------
                                     Estimated Costs of Low-Sulfur Fuels

Hwy Downgrade
Demand
351
7,898
944
1,561
212
1,115
63
162
0
231
0
155
0
116
1,569
11,239
72
136
l,64l|
11,37s!
                           Table 7.1.4-11
Distillate Supply and Demand: Final Rule: 2010-2012 (million gallons in
         Nonroad Fuel Demand Derived from EIA FOKS and AEO
2014)
Fuel Use
Category
High-
way
Non-
road
Loco-
motive
Marine
Heating
Oil
Fuel Type
Production 1 5 ppm
Spillover
Hwy Downgrade
Demand
Production 1 5 ppm
Small Refiner fuel
Hwy Spillover
Jet Downgrade
Hwy Downgrade
Proc. Downgrade
Demand
Production 1 5 ppm
Hwy Spillover
Jet Downgrade
Hwy Downgrade
Demand
Production 1 5 ppm
Hwy Spillover
Jet Downgrade
Hwy Downgrade
Demand
Production HS
Hwy Spillover
Jet Downgrade
Hwy Downgrade
Demand
PADD
1
15,801
-388
-678
14,735
1,835
283
145
0
0
0
2,263
195
15
76
250
536
173
13
67
222
475
7,217
217
108
356
7,898
2
18,210
-1,798
-722
15,690
2,630
139
1,047
0
0
0
3,816
821
287
1
5
1,114
241
84
0
1
327
0
402
206
953
1,561
3
9,507
-910
-378
8,219
1,970
136
431
0
0
0
2,537
589
141
80
126
935
747
179
102
160
1,187
595
168
137
215
1,115
4
2,903
-553
-103
2,247
265
5
344
0
0
0
616
0
126
18
92
236
0
0
0
0
0
0
89
12
62
162
5-0
3,189
-336
-126
2,727
182
60
280
0
0
0
522
0
14
40
52
106
0
3
9
11
23
0
44
81
105
231
AK
161
-3
0
157
51
30
3
0
0
0
84
5
0
0
0
5
69
0
0
0
69
155
0
0
0
155
HI
59
-13
0
47
41
0
4
0
0
0
45
0
0
0
0
0
20
1
0
0
21
108
8
0
0
116
US-
CA
49,831
-4,001
-2,008
43,822
6,974
654
2,256
0
0
0
9,884
1,610
582
215
525
2,932
1,250
280
178
394
2,103
8,076
928
544
1,691
11,239
CA
4,552
-622
-173
3,757
7
0
614
0
0
96
715
0
0
85
110
194
0
0
65
84
53
0
8
56
72
136
US
54,383
-4,623
-2,180
47,579
6,981
654
2,870
0
0
96
10,599
1,610
582
300
635
3,126
1,250
280
244
479
2,156
8,076
936
601
1,764
11,375
                               7-59

-------
Final Regulatory Support Document
                                  Table 7.1.4-12
       Distillate Supply and Demand: Final Rule: 2012-2014 (million gallons in
                Nonroad Fuel Demand Derived from EIA FOKS and AEO
2014)
Fuel Use
Category
High-
way
Non-
road
Loco-
motive
Marine
Heating
Oil
Fuel Type
Production 1 5 ppm
Spillover
Hwy Downgrade
Demand
Production 1 5 ppm
Small Refiner Fuel
Hwy Spillover
Jet Downgrade
Hwy Downgrade
Proc. Downgrade
Demand
Production 1 5 ppm
Small Refiner Fuel
Hwy Spillover
Jet Downgrade
Hwy Downgrade
Demand
Production 1 5 ppm
Small Refiner Fuel
Hwy Spillover
Jet Downgrade
Hwy Downgrade
Demand
Production HS
Hwy Spillover
Jet Downgrade
Hwy Downgrade
Demand
PADD
1
15,801
-388
-678
14,735
1,903
143
143
0
0
0
2,178
487
34
15
0
0
536
432
30
13
0
0
475
6,602
217
251
828
7,898
2
18,210
-1798
-722
15,690
2,554
100
1,025
9
42
0
3,730
781
31
287
3
13
1,114
229
9
84
1
4
327
65
402
194
899
1,561
3
9,507
-910
-378
8,219
1,690
118
423
97
152
0
2,479
653
46
141
38
60
935
828
58
179
47
74
1,187
595
168
137
215
1,115
4
2,903
-553
-103
2,247
182
3
335
13
68
0
601
73
1
129
5
28
236
0
0
0
0
0
0
4
89
11
58
162
5-O
3,189
-336
-126
2,727
25
48
200
103
133
0
510
5
10
40
22
29
106
1
2
9
5
6
23
144
87
0
0
231
AK
161
-3
0
157
24
53
3
0
0
0
81
1
3
0
0
0
5
22
47
0
0
0
69
155
0
0
0
155
HI
59
-13
0
47
41
0
4
0
0
0
44
0
0
0
0
0
0
20
0
1
0
0
21
108
8
0
0
116
US-
CA
49,831
-4001
-2,008
43,822
6,419
455
2,132
222
395
0
9,624
2,001
125
611
69
129
2,932
1,532
147
286
53
84
2,103
7,674
971
593
2,001
11,239
CA
4,552
-622
-173
3,757
7
0
614
0
0
95
715
0
0
0
85
109
194
-95
0
0
65
84
53
0
8
56
72
136
US
54,383
-4623
-2,180
47,579
6,425
455
2,746
222
395
95
9,622
2,001
125
611
178
322
3,126
1,597
147
286
137
137
2,156
7,674
979
665
2,073
11,375
                                      7-60

-------
                                               Estimated Costs of Low-Sulfur Fuels
                                    Table 7.1.4-13
     Distillate Supply and Demand: Final Rule: 2014 and Beyond (million gallons in 2014)
                 Nonroad Fuel Demand Derived from EIA FOKS and AEO
Fuel Use
Category
High-
way
Non-
road
Loco-
motive
Marine
Heating
Oil
Fuel Type
Production 1 5 ppm
Spillover
Hwy Downgrade
Demand
Production 1 5 ppm
Hwy Spillover
Jet Downgrade
Hwy Downgrade
Reproc. Downgrade
Demand
Production 1 5 ppm
Hwy Spillover
Jet Downgrade
Hwy Downgrade
Demand
Production 1 5 ppm
Hwy Spillover
Jet Downgrade
Hwy Downgrade
Demand
Production HS
Hwy Spillover
Jet Downgrade
Hwy Downgrade
Demand
PADD
1
15,801
-388
-678
14,735
2,036
143
0
0
0
2,178
522
15
0
0
536
462
13
0
0
475
6,602
217
251
828
7,898
2
18,210
-1,798
-722
15,690
2,706
1,025
0
0
0
3,730
755
287
13
59
1,114
243
84
0
0
327
66
402
194
898
1,561
3
9,507
-910
-378
8,219
2,056
423
0
0
0
2,479
443
141
136
215
935
894
179
45
70
1,187
595
168
137
215
1,115
4
2,903
-553
-103
2,247
260
335
0
0
0
601
0
129
18
95
236
0
0
0
0
0
4
89
11
58
162
5-O
3,189
-336
-126
2,727
229
200
0
0
0
510
0
0
46
60
106
0
0
10
13
23
8
87
74
95
231
AK
161
-3
0
157
77
3
0
0
0
81
5
0
0
0
5
69
0
0
0
69
155
0
0
0
155
HI
59
-13
0
47
41
4
0
0
0
44
0
0
0
0
0
20
1
0
0
21
108
8
0
0
116
US-
CA
49,831
-4,001
-2,008
43,822
7,404
2,132
0
0
0
9,624
1,723
516
214
429
2,932
1,688
111
55
83
2,103
7,538
971
667
2,095
11,239
CA
4,552
-622
-173
3,757
7
614
0
0
96
715
0
0
85
110
194
0
0
65
84
53
0
134
56
72
136
US
54,383
-4623
-2,180
47,579
7,411
2,746
0
0
96
10,339
1,723
516
298
539
3,126
1,688
111
120
167
2,156
7,538
1,106
723
2,167
11,375
   The primary difference resulting from estimating nonroad fuel demand using FOKS and AEO
is that nonroad demand is lower (and therefore, heating oil demand is larger) in PADDs 2, 4, and
5.  This eliminates the need to reprocess any downgraded fuel after 2014 when this fuel can only
be used in the L&M fuel and heating oil markets.

7.1.5 Methodology for Annual Distillate Fuel Demand: 1996 to 2040

   The environmental impact and cost-effectiveness analyses presented in this Final RIA require
estimates of fuel demand from 1996 through 2040. This section presents the methodology used to
develop these estimates. The actual levels of fuel demand are presented in Section 7.1.6 along
with the sulfur contents of the various fuels on an annual basis.
   In this section, we develop a set of year-over-year (compound) growth rates from 1996-2040
for the four non-highway fuel categories.  We did not address highway fuel demand, as this is not
                                         7-61

-------
Final Regulatory Support Document
affected by this NRLM rule. For nonroad, locomotive and marine fuels, we obtained annual
estimates of fuel demand for as much of this time period as was available. We then calculated
year-over-year growth rates over the period of time that the data were available.  Finally, we
extrapolated or interpolated these growth rates to cover any years for which specific fuel demand
projections were not available.

   We obtained our estimates of annual fuel demand by nonroad engines from EPA's
NONROAD emission model. These estimates of fuel demand and the resulting annual growth
rates are shown in Table 7.1.5-1. As can be seen, NONROAD projects a linear increase in fuel
consumption over time.  This results in a slightly decreasing year-over-year growth rate over
time.
                                         7-62

-------
                              Estimated Costs of Low-Sulfur Fuels
                    Table
Annual Growth In the Demand
7.1.5-1
of Nonroad and Locomotive Fuel
Year
1996
1997
1998
1999
2000
2001
2002
2003
2004
2005
2006
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2031
2032
2033
2034
2035
2036
2037
2038
2039
2040
Nonroad Fuel Demand
(million gallons)
9,158
9,450
9,742
10,024
10,319
10,613
10,906
11,200
11,493
11,787
12,078
12,370
12,661
12,952
13,244
13,537
13,830
14,123
14,416
14,709
14,999
15,289
15,579
15,869
16,159
16,449
16,739
17,029
17,319
17,609
17,897
18,185
18,473
18,761
19,049
19,337
19,625
19,912
20,201
20,489
20,777
21,065
21,353
21,641
21 928
Annual Growth Rate

1.032
1.031
1.029
1.030
1.028
1.028
1.027
1.026
1.026
1.025
1.024
1.024
1.023
1.023
1.022
1.022
1.021
1.021
1.020
1.020
1.020
1.019
1.019
1.018
1.018
1.018
1.017
1.017
1.017
1.016
1.016
1.016
1.016
1.015
1.015
1.015
1.015
1.015
1.014
1.014
1.014
1.014
1.014
1 013
Locomotive Fuel Demand
(trillion btu)




609.2
628.4
610.2
617.0
621.4
626.1
638.9
650.2
657.4
666.3
676.9
689.7
696.6
702.1
707.6
713.5
721.1
727.7
733.1
740.3
745.4
749.2
755.9
762.6
769.2
776.6
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
(million gallons)
3072



2692








































Annual
Growth Rate

0.969
0.968
0.967
0.966
1.032
0.971
1.011
1.007
1.008
1.020
1.018
1.011
1.014
1.016
1.019
1.010
1.008
1.007
1.008
1.011
1.009
1.007
1.010
1.007
1.005
1.009
1.009
1.009
1.010
1.008
1.008
1.008
1.008
1.008
1.008
1.008
1.008
1.008
1.008
1.007
1.007
1.007
1.007
1 007
                        7-63

-------

-------
                                              Estimated Costs of Low-Sulfur Fuels
   Locomotive diesel fuel growth rates for the period from 1996 to 2000 were estimated from
historic estimates of fuel consumption taken from the 1996 and 2000 FOKS reports. We assume
that locomotive diesel fuel demand decreased linearly between 1996 and 2000. We assume a
constant linear growth rate for this time period, as this seemed most consistent with EIA's
projection of growth in locomotive fuel demand in the post-2000 time period.  For the period after
2000, we use the annual demand for locomotive diesel fuel projected by EIA in the AEO 2003 to
calculate year-over-year growth rates from 2000 to 2025 (the last projection year in AEO 2003).
Beyond 2025,  we assume that locomotive fuel demand grows linearly at the average rate of
growth between 2021 and 2025.  The FOKS and AEO estimates of fuel demand and the year-
over-year growth rates for locomotive diesel fuel are summarized in Table 7.1.5-1.

   According to EIA FOKS reports, the demand for marine diesel fuel decreased slightly
between 1996  and 2001. We estimated annual demand for marine  diesel fuel for 1997-2000 by
assuming a constant compound growth rate between 1996 and 2001.  (Constant compound growth
is more consistent with EIA's projection of growth in marine fuel demand in the post-2000 time
period than constant linear growth.) For the period after 2000, we  use the annual demand for
marine diesel fuel projected by EIA in the AEO 2003 to calculate a year-over-year growth rates
2000 to 2025 (the last projection year in AEO 2003). Beyond 2025, we assume that marine fuel
demand grows at a constant compound growth rate between 2001 and 2025, which was 1.3%.
The FOKS and AEO estimates of fuel demand and the year-over-year growth rates for marine
diesel fuel are summarized in Table 7.1.5-2.
                                        7-65

-------
                 Table 7.1.5-2
Annual Growth in the Demand for Marine Diesel Fuel
Year

1996
1997
1998
1999
2000
2001
2002
2003
2004
2005
2006
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2031
2032
2033
2034
2035
2036
2037
2038
2039
2040
Marine Fuel Consumption
AEO2003 (trillion BTU)
-
-
-
-
-
344.6
338.4
342.6
346.1
348.4
356.5
361.7
366.7
371.1
375.7
381.2
386.1
389.6
394.3
398.7
402.5
407.0
413.1
420.1
425.0
430.2
437.2
442.1
448.0
453.2
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
FOKS 2001 (million gallons)
1960
-
-
-
-
1884
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
Annual Growth Rate

0.992
0.992
0.992
0.992
0.992
0.982
1.012
1.010
1.007
1.023
1.015
1.014
1.012
1.012
1.015
1.013
1.009
1.012
1.011
1.010
1.011
1.015
1.017
1.012
1.012
1.016
1.011
1.013
1.012
1.013
1.013
1.013
1.013
1.013
1.013
1.013
1.013
1.013
1.013
1.013
1.013
1.013
1.013
1.013

-------
                                                Estimated Costs of Low-Sulfur Fuels
   We applied a simpler approach to estimating the growth in the demand for heating oil for a
number of reasons. One, this rule does not regulate the sulfur content of heating oil. Two, EIA
does not present estimates of heating oil demand, as it is defined here. Three, heating oil demand
between 2001 and 2014 is very close to zero. Thus, the effect of differing assumptions regarding
the shape of this growth, such as linear versus compound, have a negligible effect on any
extrapolated growth.

   As shown in Table 7.1.3-3, heating oil demand declined by  7% from 2001 to 2014. We
assumed that this decline was occurring at a constant compound rate, which we calculated to be -
0.006% for this time period. We assumed that this decline would continue through 2040.

7.1.6 Annual Distillate Fuel Demand and Sulfur Content

   In this section we estimate the sulfur content of the various  types of distillate fuel prior to this
rule and how they are affected by the NRLM rule.  We then present year-by-year estimates of
both distillate fuel demand and sulfur content for the purpose of estimating the environmental
benefits of this rule.

   7.1.6.1 Sulfur Content

   The sulfur content of high sulfur distillate before and after this NRLM rule is used in two
ways in this regulatory impact analysis: 1) to estimate the reductions in emissions of sulfur
dioxide and sulfate PM, and 2) to estimate the cost of desulfurizing this fuel to meet 500 and 15
ppm caps. In this section we estimate  the current sulfur content of the four non-highway distillate
fuels by region.  We then estimate how these sulfur contents change during the various phases of
the final NRLM fuel program. Finally, we estimate the sulfur content of these fuels for two
sensitivity cases: 1) a long-term 500 ppm sulfur NRLM program and 2) the proposed NRLM fuel
program (15 ppm nonroad fuel and 500 ppm L&M fuel in 2010).

   We estimate the current sulfur content of high sulfur distillate from diesel fuel survey data
collected by TRW Petroleum Technologies (TRW) at its facility in Bartlesville, Oklahoma.  This
facility was formerly known as the National Institute for Petroleum and Energy Research
(NIPER)).  Surveys performed for 1999 through 2002 were published by TRW.  Surveys prior to
1999 were published by the NIPER. We evaluated their survey data from 1996 through 2002.  As
the methodology of conducting the surveys and the presentation of the data have not changed
over this time period, we will simply refer to these surveys as TRW surveys.

   No comments were received on our methodology for estimating the sulfur content of high
sulfur distillate for the NPRM. However, we have made three changes to that analysis which we
believe improve the estimate. The first is to include the 2002 survey data, which is now available.
The second is to include sample data which were assigned a production volume by TRW. The
third is to adjust the sample data for the addition of downgraded jet fuel, highway diesel fuel and
heavy gasoline during distribution.
                                          7-67

-------
Final Regulatory Support Document
   TRW collects sulfur data voluntarily provided by domestic refiners, including a refiner
located in the Virgin Islands.  These refiners analyze the sulfur content of their diesel fuel
production and submit the results to TRW. TRW states that the survey results reflect the average
quality of distillate fuel produced at refineries for use in each geographical  area.  However, TRW
also states that the data may not be representative of the full range of sulfur content of these fuels
at their point of use. This appears to be due to either TRW or refiners reporting the average
quality of their high sulfur diesel fuel versus a set of individual samples, in addition to the effect
of convenience sampling.

   TRW presents survey results for five geographic regions containing 16  districts. According to
TRW, these areas are based on fuel distribution systems, refinery locations, centers of population,
temperature zones, and arteries of commerce. A map of the regions and districts is shown in
Figure 7.1-6 below. Each sample is assigned to both a region and to one or more districts.  We
primarily use the TRW district assignments, as they provide a more precise indication of where
the fuel was eventually sold. A map of the Petroleum Administration Defense Districts (PADDs)
is shown for comparison in Figure 7.1-7.  Since all of our estimates for distillate  production and
demand were developed by PADD (with PADD 5 split up further), we assigned each TRW
district to one or more PADDs as described in Table 7.1.6-1.

                    Figure 7.1-7 TRW Fuel Survey Regions and Districts

                  ROCKY MOUNTAIN
                      REGION
                               •\SOUTH DAKOTA
                IDAHO f
                       WYOMING
                                           7-68

-------
                                               Estimated Costs of Low-Sulfur Fuels
           Figure 7.1-8. Petroleum Administration for Defense Districts (PADDs)
           .
                                     Table 7.1.6-1
                   Assignments of TRW Regions and Districts to PADDs
Region

Eastern

Southern

Central



Rocky Mountain




Western


TRW District
A
B
C
D
E
F
G
H
I

J
K
L
M
N
O
P
Assigned PADD
1
1
1,2
1,3
2
2
2
4
4

3
4
5
5
5
5
5
   TRW provides a rough indication of the annual volume of fuel represented by each sulfur
measurement by assigning each data point one of four numbers.  Table 7.1.6-2 presents the
numbering system used by TRW and the range of diesel fuel production represented by each
                                         7-69

-------
Final Regulatory Support Document
numeral assignment.  In order to weight the sulfur measurements by volume, we assigned an
average volume to each range.  These averages are also shown in Table 7.1.6-2.

                                      Table 7.1.6-2
                        Production Volumes of Fuel Sulfur Samples
TRW Sample Quantity Number
1
2
3
4
Fuel Volume (Barrels Per Year)
TRW: Range
Over 1,500,000
500,000 to 1,500,000
50,000 to 500,000
Under 50,000
EPA: Assumed Average Volume
1,500,000
1,000,000
275,000
50,000
   Within each region, the TRW reports generally list the sulfur samples by their Sample
Quantity Number, starting with 1 and moving to 2, 3, and 4. Thus, the sulfur data representing
the largest fuel batches are listed first and those representing the smallest fuel batches are listed
last.  However, some sulfur data points in the TRW reports do not have a Sample Quantity
Number. These data points always appear at either top of the list or the bottom of the list. When
the data missing a Sample Quantity Number appeared at the top of the list, we assigned that data a
production volume of 2 million barrels per year. When the data appeared at the bottom of the list,
we assigned it a volume of 25,000 barrels per year. In the analysis performed for the NPRM, we
excluded this data from the analysis.

   The survey reports often list the same sample number under more than one region. Each of
these listings shows the districts in both regions. For example, Sample 45 may be listed in both
the Eastern and Central Regions. Both listing show C2 and E2, indicating that 0.5-1.0 million
barrels of fuel were shipped that year to Districts C and E.  Since both districts are listed under
both regions, we assumed that this was in fact only one data point and that 0.5-1 million barrels
were shipped to District C in the Eastern Region and that 0.5-1 million barrels were shipped to
District E  in the Central Region, not twice this volume.

   In this case, the numeral 2 was assigned to each district, so we assumed that 0.5-1 million
barrels of fuel were provided to each district.  In some cases, two or more districts are listed with
only a single numeral following the district letter (i.e., C, E 2). In this case, we assumed that the
total volume of fuel produced was 0.5-1 million barrels and that this volume was split between
the two districts. TRW indicates that the district receiving the most fuel was listed first, etc.
However,  lacking any quantitative information about the relative volumes of fuel supplied to each
district, we simply assumed that each district received the same proportion.

   TRW segregates their reporting of fuel quality by fuel type, namely No. 1 diesel fuel, No. 2
highway diesel fuel and No. 2 off-highway diesel fuel. We focused solely  on the data for No. 2
off-highway diesel fuel.  However, we assumed that off-highway diesel fuel with a sulfur content
                                          7-70

-------
                                               Estimated Costs of Low-Sulfur Fuels
of less than 500 ppm was highway diesel fuel "spillover." These data were excluded from this
analysis since we account for the lower sulfur content of spillover fuel separately below.

   After applying the PADD assignments shown in Table 7.1.6-1, we volume weighted the sulfur
data in each PADD using the average volumes shown in Table 7.1.6-2 in order to derive a PADD
average  sulfur content for each calendar year.  These PADD averages are shown in Table 7.1.6-3.
                                         7-71

-------
Final Regulatory Support Document
                                     Table 7.1.6-3
                         Sulfur Content of High Sulfur Diesel Fuel
PADD
1
2
3
4
5
U.S.
Year Volume (bbls/vear) Sulfur (pprn) PADD Average
1996
1997
1998
1999
2000
2001
2002
1996
1997
1998
1999
2000
2001
2002
1996
1997
1998
1999
2000
2001
2002
1996
1997
1998
1999
2000
2001
2002
1996
1997
1998
1999
2000
2001
2002
1996
1997
1998
1999
2000
2001
2002
7,170,833
13,250,000
5,887,500
4,137,500
10,525,000
4,437,500
2,662,500
4,158,333
5,100,000
2,775,000
2,912,500
10,412,500
5,212,500
1,000,000
2,420,833
4,500,000
2,387,500
3,000,000
3,387,500
1,775,000
2,387,500
275,000
275,000
275,000
275,000
275,000
275,000
275,000
2,050,000
3,550,000
1,550,000
1,550,000
2,175,000 *
2,175,000 *
2,175,000 *
16,075,000
26,675,000
12,875,000
11,875,000
26,775,000
14,375,000
8.500.000
3,482
2,601
2,418
3,257
2,691
3,061
4,343
3,497
3,008
2,241
1,717
2,939
3,854
1,620
4,539
3,945
5,004
4,177
4,361
4,298
4,359
4,100
1,000
3,400
2,000
2,600
2,340
2,400
3,076
2,268
3,077
2,065
2,566 *
2,566 *
2,566 *
3,623
2,710
2,669
2,818
2,886
3,440
3.510
2,925
2,973
3,776
2,549
2,566
3,030
       * No data reported. Estimated from the average from 1996-1999.
                                         7-72

-------
                                                 Estimated Costs of Low-Sulfur Fuels
    We next calculated a national average sulfur content for each year. This was done by
weighting the PADD average sulfur contents in each year by the volume of fuel represented by all
the samples in that PADD. No data were reported for the Western Region for 2000, 2001 and
2002.  Thus, we substituted the 1996-1999 average production volume and sulfur content for
these missing years when calculating the national average for 1999-2002.  These national
averages are also shown in Table 7.1.6-3.  It should be noted that these national average sulfur
contents were not used in either the emissions nor cost analysis.  The emission and cost analyses
used the PADD average sulfur contents.  However, we present them here for illustrative purposes
and to simply the evaluation of the presence of any temporal trends in the sulfur content of high
sulfur diesel fuel.

    We examined the annual average sulfur contents for possible trends. However, as indicated
by the national averages shown in Table 7.1.6-3, the sulfur content of high sulfur diesel fuel
seems to vary randomly.  Therefore, we average the data once more across calendar years, again
using the fuel volumes represented by all the samples from each year. As shown in Table 7.1.6-3,
this overall average sulfur content is 3030 ppm.

    While the TRW reports indicate that the sulfur data was supplied by refiners, we assume that
these sulfur levels are actually those existing at the point-of-use (i.e. retail). Thus, this average
sulfur content of 3030 ppm is used in Chapter 3 to project emissions of sulfur dioxide and sulfate
PM from the burning of NRLM fuel and heating oil.  Because of the absence of a trend in the
1996-2002 data, we assume that these sulfur contents will not change in the future, absent NRLM
fuel standards.

    In order to project desulfurization costs, however, an estimate of the current sulfur content of
NRLM fuel at the refinery is needed. As discussed in Sections 7.1.2 and 7.1.3, small volumes of
jet fuel, highway diesel fuel and heavy gasoline become mixed with high sulfur distillate during
pipeline shipment. These other fuels generally contain less sulfur than high sulfur diesel fuel, so
the sulfur content of high sulfur diesel fuel actually decreases during shipment.  In order to better
estimate desulfurization costs, we estimated the sulfur content of high sulfur diesel fuel prior to
this mixing during shipment.

    The volumes of high sulfur distillate produced at refineries and the volume of material
downgraded to high sulfur distillate is estimated in Sections 7.1.2 and 7.1.3 (see, for example,
Tables 7.1.2-8 and 7.1.3-8). Here, we estimate the sulfur content of these various materials so
that the combination matches the PADD average sulfur contents shown in Table 7.1.6-3.

    Table 7.1.2-6 shows the types of downgrades and their volumes and destinations. This table
shows that 1.75% of jet fuel demand, 2.2% of highway  diesel fuel  production, and a volume of
heavy gasoline equivalent to 0.58% of jet fuel demand and 0.73%  of highway diesel fuel
production is shifted to high sulfur distillate during pipeline shipment. We estimate that jet fuel
averages 550 ppm sulfur.14 From the Final RIA for the  highway diesel rule, highway diesel fuel
averages 340 ppm sulfur.  The sulfur level of today's gasoline, before the Tier 2 rule has been
implemented, averages about 300 ppm.  The vast majority of this sulfur is contained in the

                                          7-73

-------
Final Regulatory Support Document
naphtha produced in the fluidized catalytic cracker (FCC naphtha).  The sulfur content of FCC
naphtha increases significantly with distillation temperature.  Therefore, we estimate that the
heaviest one-third of gasoline distilled into transmix contains essentially all the sulfur in the
whole gasoline.  Thus, we estimate the sulfur level of the heaviest one-third of gasoline to be
about 900 ppm.

    As described in Section 7.1.2, to simplify the analysis of downgrade distillate volume, we
combined the jet fuel downgrade with the portion of the heavy gasoline downgrade which was
dependent on jet fuel demand.  Of this jet-based downgrade, jet fuel represents 75%
(1.757(1.75+0.58)) and heavy gasoline represents 25% (0.587(1.75+0.58)).  Weighting the sulfur
content of jet fuel and heavy gasoline by these percentages produces an average sulfur content of
638 ppm.

    Likewise, we combined the highway diesel fuel downgrade with the portion of the heavy
gasoline downgrade which was dependent on highway diesel fuel production.  Of this highway-
based downgrade, highway diesel fuel represents 75% (2.27(2.2+0.73)) and heavy gasoline
represents 25% (0.737(2.2+0.73)). Weighting the sulfur content of jet fuel and heavy gasoline by
these percentages produces an average sulfur content of 480 ppm.s

    Table  7.1.6-4 presents the levels of high sulfur distillate production and demand, as well as
the volumes of downgraded material which are added to this fuel  during distribution.  All of these
figures were taken directly from Table 7.1.2-8. Table 7.1.6-4 also shows the sulfur content of
high sulfur diesel fuel at retail (from Table 7.1.6-3) and of the two types of downgrade, as
discussed  above. We  determined the sulfur content of high sulfur distillate at the refinery which,
when combined  with the volumes and sulfur content of the two types of downgrade, matched  the
sulfur content from the TRW surveys. The sulfur content of high sulfur distillate at the refinery
gate in each PADD are shown in Table 7.1.6-4.  Because there are no product pipelines in Alaska
and Hawaii, we assume that there is no downgrade in these areas. Also, because we assumed
100% spillover into the high sulfur distillate  market in California, there is no high sulfur distillate
in California pipelines to receive this downgrade.  Distillate downgrade is assumed to be used
directly as L&M fuel. Thus, we assume that the sulfur content of 2,570 ppm for high suflur
distillate in PADD 5 applies at both retail and the refinery in Alaska, Hawaii, and California.
    s The distillate sulfur contents presented at the end of this section for 1996-2006 assume that jet-based
downgrade contains 700 ppm rather than 638 ppm and that highway-based downgrade contains 560 ppm rather than
480 ppm.  These errors have a very small effect on the final sulfur content of high sulfur distillate fuels during these
years. As the NRLM fuel program has no effect during these years, neither the costs nor benefits associated with
this rule are affected.

                                           7-74

-------
                                                Estimated Costs of Low-Sulfur Fuels
                                      Table 7.1.6-4
                Sulfur Content of High Sulfur Diesel Fuel at Refineries in 2001

PADD1
PADD2
PADD3
PADD4
PADD 5-O
AK, HI, CA
High Sulfur Distillate Fuel Volume
Demand
Jet-Based Downgrade
Highway -Based Downgrade
Refinery Production
10,955
95
327
10,533
4,562
80
387
4,095
4,407
123
202
4,082
408
12
64
332
497
51
68
378
486
0
0
486
High Sulfur Distillate Sulfur Content (ppm)
At Retail
Jet-Based Downgrade
Highway -Based Downgrade
Sulfur level of HS Dist Pool at
Refineries
2,930
638
480
3,041
2,970
638
480
3,295
3,780
638
480
4,059
2,550
638
480
3,102
2,570
638
480
3,280
2,570
638
480
2,570
   As can be seen, downgrade occurring in pipelines decreases the sulfur content of high sulfur
distillate by as little as 111 ppm in PADD 1 and as much as 710 in PADD 5-O. The difference is
due to the very small volume of downgrade relative to the demand for high sulfur distillate in
PADD 1, with the opposite being true in PADD 5-O.

   After completion of this analysis, we discovered that the TRW data represented sulfur levels
at the refinery and not downstream.  Thus, the TRW sulfur levels should have been used to
estimate desulfurization costs in Section 7.2.2 and the adjustments shown in Table 7.1.6-4 should
have been used to estimate lower sulfur levels downstream. The result  of this error is an
overestimation of the baseline sulfur content of high sulfur distillate by roughly 150 ppm on
average. Given the limited data set and the resulting year-to-year variation, the resulting estimate
is still well within the range of possible actual sulfur levels. This 150 ppm difference, if real,
results in an overestimation of the cost to produce 500 ppm NRLM fuel of roughly 0.02 cent per
gallon (i.e., roughly 1%) and an overestimation of the sulfur dioxide and sulfate PM emission
reductions due to the 500 ppm NRLM fuel cap of roughly 4-5%.

   The next step in this analysis is to project the sulfur content of the various distillate fuels
during the various phases of the final NRLM fuel program, as well as under the two sensitivity
cases. We assume that the sulfur content of NRLM fuel produced under 15 and 500 ppm caps
will be the same as those we estimate for highway diesel fuel produced under the same standards.
Thus, we assume that NRLM fuel produced to meet  a 500 ppm cap will contain 340 ppm sulfur.
We assume that NRLM fuel produced to meet a 15 ppm cap will contain 7 ppm sulfur at the
refinery. However, as discussed in the Final RIA for the highway diesel rule, we assume that this
fuel will contain 11 ppm at the time of final sale. This increase of 4 ppm is due to very small
                                          7-75

-------
Final Regulatory Support Document
volumes of higher sulfur fuel being incorporated into batches of 15 ppm diesel fuel during
shipment.  This volume is by necessity very small compared to the volume of pipeline interface.
Thus, this 4 ppm increase in 15 ppm fuel during shipment does not affect our estimation of the
creation and disposition of downgrade created in the pipeline during shipment.

    As just mentioned, highway fuel in the pipeline will contain between 7 and 11 ppm sulfur.
We assume that the highway fuel contributing to interface contains 11 ppm sulfur.  We assume
that the sulfur content of jet fuel will remain 550 ppm in the future. Under the Tier 2 standards,
gasoline will average 30 ppm sulfur. With this degree of sulfur control, essentially all the sulfur
in gasoline will be in the heavy portion of FCC naphtha. Thus, we apply the same factor of 3
discussed above and estimate that the heaviest one-third of gasoline will contain 90 ppm sulfur.

    Prior to the NRLM rule, the volume of jet-based downgrade stays the same as that shown in
Table 7.1.6-4 (compare the jet-based downgrade in Table 7.1.2-6 (2001) to that in Table 7.1.3-6
(2014 prior to the NRLM rule)).  Only the sulfur levels change.  A 75%/25% weighting of the
sulfur content of jet fuel (550 ppm) and heavy gasoline (90 ppm) produces an average sulfur
content of 435  ppm.

    As indicated in Table 7.1.3-6, the volume  of highway-based downgrade increases
significantly with the onset of the 15 ppm highway program, due to the need to make more
protective interface cuts to maintain the quality of this fuel.  As described in Table 7.1.3-6, 2.2%
of highway diesel fuel supply will be cut directly into high sulfur distillate fuel. We assume that
this highway fuel contains 11 ppm sulfur. Also, 2.2% of highway fuel supply plus a volume of
heavy gasoline equivalent to 0.73% of highway fuel supply will be processed as transmix and
added to the 500 ppm highway fuel supply. This downgrade will have an average sulfur content
of 31 ppm (25% of 90 ppm plus 75% of 11 ppm).T

    Under the NRLM fuel program, after 2007, some pipelines are projected to continue carrying
heating oil, while others are expected to drop this fuel. For those pipelines still carrying heating
oil (PADDs 1 and 3), the sulfur content of jet-based downgrade will continue to be 435 ppm, as
described above. The sulfur content of the highway-based downgrade to high sulfur distillate and
500 ppm diesel fuel will continue to be 11 ppm and 31 ppm, respectively, as described above.u
    T The distillate sulfur contents presented at the end of this section assume that jet-based downgrade in this time
period contains 400 ppm rather than 435 ppm and that highway-based downgrade contains 35 ppm rather than 31
ppm. The net effect of these partially offsetting errors on the final sulfur content of high sulfur distillate fuels in the
base case is very minor.

    u TRW also surveys the quality of distillate fuel oil. These surveys which we received after completion of this
analysis, show national average sulfur levels of roughly 2200 ppm, versus 3000 ppm for high sulfur diesel fuel.
However, it is not clear how much distillate actually burned in heating oil uses is defined as heating oil at the
refinery and how much is defined as diesel fuel. Thus, we chose not to use the heating oil survey results here.
However, given that at least a portion of the heating oil market must meet state sulfur caps of 2000-4000 ppm,
extrapolation of the diesel fuel survey results to heating oil probably over-estimates the sulfur content to some
degree.  Given that the sulfurous emission reductions from heating oil are only ancillary to the benefits of this rule,
this likely  small degree of overestimation is not critical. However, the heating oil related benefits are a large portion

                                            7-76

-------
                                                  Estimated Costs of Low-Sulfur Fuels
    For those pipelines not carrying heating oil, the nature of the downgrade and its disposition
changes, as shown in Table 7.1.3-12. For these pipelines (all PADDs except 1 and 3), all of the
jet-based downgrade is combined, as is the highway-based downgrade.  The total jet-based
downgrade consists of 3.5% of jet fuel demand and a volume of heavy gasoline equivalent to
0.58% of jet fuel demand. This is a 6:1 ratio of jet fuel to gasoline. With jet fuel at 550 ppm and
heavy gasoline at 90 ppm, the average sulfur content of the jet-based downgrade is 485 ppm.
Similarly, the total highway-based downgrade consists of 4.4% of highway fuel supply and a
volume of heavy gasoline equivalent to 0.73% of highway fuel supply.  This is a 6:1 ratio of
highway fuel to gasoline.  With highway fuel at 11 ppm and heavy gasoline at 90 ppm, the
average sulfur content of the highway-based downgrade is 22 ppm.v  While the disposition of
this downgrade changes during the various phases of the NRLM fuel program, the sulfur content
of these two types of downgrade remain the same.

    7.1.4.2 Distillate Fuel Demand and  Sulfur Content by Year

    We present the final estimates of distillate fuel demand and sulfur content for each year from
1996-2040 in this section. We develop these estimates by combining:

    1) The sulfur contents developed in Section 7.1.4.1 with
    2) The sources of each distillate fuel's supply in 2014 developed in Sections 7.1.2 (Reference
    Case), 7.1.3 (after implementation of the final NRLM fuel program), and 7.1.4 (sensitivity
    cases), and
    3) The growth in distillate fuel demand developed in Section 7.1.5.

    We did this for the entire U.S. (50-state) and for 48 states (the U.S. minus the states of Alaska
and Hawaii). The results are  summarized in Tables 7.1.6-5  to 7.1.6-12. In all cases, we assume
that a new sulfur standard becomes effective on June 1. Therefore, the average sulfur levels in
any transition year is a 5:7 weighting of the previous year's sulfur level and the following year's
sulfur level.
of the incremental benefits of associated with the 15 ppm cap for L&M fuel.  Thus, we address the possibility of a
lower sulfur content for heating oil in Section 8.3, where we evaluate the incremental cost effectiveness of the 15
ppm cap for L&M fuel.

    v The distillate sulfur contents presented at the end of this section assume that jet-based downgrade in this time
period contains 470 ppm rather than 485 ppm and that highway-based downgrade contains 25 ppm rather than 22
ppm. The net effect of these partially offsetting errors on the final sulfur content of high sulfur distillate fuels in the
base case is minor.

                                            7-77

-------
Table 7. 1 .6-5 Annual Distillate Fuel Demand and Sulfur Content for the Reference Case;
U.S. minus AK and HI (million gallons and ppm)

Year
1996
1997
1998
1999
2000
2001
2002
2003
2004
2005
2006
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2031
2032
2033
2034
2035
2036
2037
2038
2039
2040
Nonroad
Demand
9,087
9,376
9,665
9,945
10,238
10,530
10,821
11,112
11,403
11,694
11,983
12,272
12,562
12,851
13,140
13,430
13,721
14,012
14,302
14,593
14,881
15,169
15,456
15,744
16,032
16,319
16,607
16,895
17,183
17,470
17,756
18,042
18,328
18,613
18,899
19,185
19,470
19,756
20,042
20,328
20,613
20,899
21,185
21,470
21.756
Sulfur
2,283
2,283
2,283
2,283
2,283
2,283
2,283
2,283
2,283
2,283
2,243
2,214
2,214
2,214
2,159
2,120
2,120
2,120
2,120
2,120
2,120
2,120
2,120
2,120
2,120
2,120
2,120
2,120
2,120
2,120
2,120
2,120
2,120
2,120
2,120
2,120
2,120
2,120
2,120
2,120
2,120
2,120
2,120
2,120
2.120
Locomotive
Demand
3,065
2,971
2,876
2,782
2,687
2,772
2,692
2,722
2,741
2,762
2,818
2,868
2,900
2,939
2,986
3,043
3,073
3,097
3,121
3,148
3,181
3,210
3,234
3,266
3,288
3,305
3,335
3,364
3,393
3,426
3,453
3,481
3,508
3,536
3,564
3,591
3,619
3,646
3,674
3,701
3,729
3,756
3,784
3,811
3.839
Sulfur
2,454
2,454
2,454
2,454
2,454
2,454
2,454
2,454
2,454
2,454
2,437
2,424
2,424
2,424
2,254
2,133
2,133
2,133
2,133
2,133
2,133
2,133
2,133
2,133
2,133
2,133
2,133
2,133
2,133
2,133
2,133
2,133
2,133
2,133
2,133
2,133
2,133
2,133
2,133
2,133
2,133
2,133
2,133
2,133
2.133
Marine
Demand
1,878
1,863
1,849
1,834
1,820
1,805
1,773
1,795
1,813
1,825
1,868
1,895
1,921
1,944
1,968
1,997
2,023
2,041
2,066
2,089
2,109
2,132
2,164
2,201
2,226
2,254
2,290
2,316
2,347
2,374
2,405
2,436
2,467
2,499
2,532
2,564
2,598
2,631
2,665
2,700
2,735
2,770
2,806
2,842
2.879
Sulfur
2,918
2,918
2,918
2,918
2,918
2,918
2,918
2,918
2,918
2,918
2,904
2,893
2,893
2,893
2,712
2,583
2,583
2,583
2,583
2,583
2,583
2,583
2,583
2,583
2,583
2,583
2,583
2,583
2,583
2,583
2,583
2,583
2,583
2,583
2,583
2,583
2,583
2,583
2,583
2,583
2,583
2,583
2,583
2,583
2.583
L&M
Demand
4,943
4,834
4,725
4,616
4,507
4,577
4,465
4,517
4,554
4,587
4,686
4,763
4,821
4,883
4,954
5,039
5,096
5,138
5,187
5,236
5,290
5,342
5,398
5,466
5,515
5,559
5,625
5,680
5,740
5,800
5,858
5,917
5,976
6,035
6,095
6,155
6,216
6,277
6,339
6,401
6,463
6,526
6,590
6,653
6.718
Sulfur
2,641
2,641
2,641
2,641
2,641
2,637
2,638
2,638
2,639
2,639
2,623
2,611
2,611
2,611
2,436
2,312
2,312
2,312
2,312
2,313
2,313
2,313
2,314
2,314
2,315
2,316
2,316
2,317
2,317
2,317
2,318
2,319
2,319
2,320
2,320
2,321
2,321
2,322
2,322
2,323
2,324
2,324
2,325
2,325
2.326
Heating Oil
Demand
10,715
10,654
10,593
10,532
10,471
10,411
10,352
10,292
10,233
10,174
10,116
10,058
10,000
9,943
9,886
9,829
9,772
9,716
9,661
9,605
9,550
9,495
9,441
9,386
9,333
9,279
9,226
9,173
9,120
9,068
9,016
8,964
8,913
8,861
8,811
8,760
8,710
8,660
8,610
8,561
8,511
8,463
8,414
8,366
8,318
Sulfur
2,871
2,871
2,871
2,871
2,871
2,871
2,871
2,871
2,871
2,871
2,860
2,853
2,853
2,853
2,722
2,628
2,628
2,628
2,628
2,628
2,628
2,628
2,628
2,628
2,628
2,628
2,628
2,628
2,628
2,628
2,628
2,628
2,628
2,628
2,628
2,628
2,628
2,628
2,628
2,624
2,628
2,628
2,628
2,628
2.628
Table 7.1.6-6 Annual Distillate Fuel Demand and Sulfur Content: Final NRLM Rule:
U.S. minus AK and HI (million gallons and ppm)

Nonroad
Locomotive
Marine
L&M
Heating Oil

-------
Year
1996
1997
1998
1999
2000
2001
2002
2003
2004
2005
2006
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2031
2032
2033
2034
2035
2036
2037
2038
2039
2040
Demand
9,087
9,376
9,665
9,945
10,238
10,530
10,821
11,112
11,403
11,694
11,983
12,272
12,562
12,851
13,140
13,430
13,721
14,012
14,302
14,593
14,881
15,169
15,456
15,744
16,032
16,319
16,607
16,895
17,183
17,470
17,756
18,042
18,328
18,613
18,899
19,185
19,470
19,756
20,042
20,328
20,613
20,899
21,185
21,470
21.756
Sulfur
2,283
2,283
2,283
2,283
2,283
2,283
2,283
2,283
2,283
2,283
2,243
1,127
330
330
155
30
30
19
11
11
11
11
11
11
11
11
11
11
11
11
11
11
11
11
11
11
11
11
11
11
11
11
11
11
11
Demand
3,065
2,971
2,876
2,782
2,687
2,772
2,692
2,722
2,741
2,762
2,818
2,868
2,900
2,939
2,986
3,043
3,073
3,097
3,121
3,148
3,181
3,210
3,234
3,266
3,288
3,305
3,335
3,364
3,393
3,426
3,453
3,481
3,508
3,536
3,564
3,591
3,619
3,646
3,674
3,701
3,729
3,756
3,784
3,811
3.839
Sulfur
2,454
2,454
2,454
2,454
2,454
2,454
2,454
2,454
2,454
2,454
2,435
1,225
361
361
177
45
45
45
61
72
72
72
72
72
72
72
72
72
72
72
72
72
72
72
72
72
72
72
72
72
72
72
72
72
72
Demand
1,878
1,863
1,849
1,834
1,820
1,805
1,773
1,795
1,813
1,825
1,868
1,895
1,921
1,944
1,968
1,997
2,023
2,041
2,066
2,089
2,109
2,132
2,164
2,201
2,226
2,254
2,290
2,316
2,347
2,374
2,405
2,436
2,467
2,499
2,532
2,564
2,598
2,631
2,665
2,700
2,735
2,770
2,806
2,842
2.879
Sulfur
2,918
2,918
2,918
2,918
2,918
2,918
2,918
2,918
2,918
2,918
2,902
1,469
445
445
208
39
39
39
33
28
28
28
28
28
28
28
28
28
28
28
28
28
28
28
28
28
28
28
28
28
28
28
28
28
28
Demand
4,943
4,834
4,725
4,616
4,507
4,577
4,465
4,517
4,554
4,587
4,686
4,763
4,821
4,883
4,954
5,039
5,096
5,138
5,187
5,236
5,290
5,342
5,398
5,466
5,515
5,559
5,625
5,680
5,740
5,800
5,858
5,917
5,976
6,035
6,095
6,155
6,216
6,277
6,339
6,401
6,463
6,526
6,590
6,653
6.718
Sulfur
2,641
2,641
2,641
2,641
2,641
2,637
2,638
2,638
2,639
2,639
2,621
1,321
394
394
189
43
43
43
49
54
54
54
54
54
54
54
54
54
54
54
54
54
54
54
54
54
54
54
54
54
54
54
54
54
54
Demand
10,715
10,654
10,593
10,532
10,471
10,411
10,352
10,292
10,233
10,174
10,116
10,058
10,000
9,943
9,886
9,829
9,772
9,716
9,661
9,605
9,550
9,495
9,441
9,386
9,333
9,279
9,226
9,173
9,120
9,068
9,016
8,964
8,913
8,861
8,811
8,760
8,710
8,660
8,610
8,561
8,511
8,463
8,414
8,366
8.318
Sulfur
2,871
2,871
2,871
2,871
2,871
2,871
2,871
2,871
2,871
2,871
2,860
2,667
2,530
2,530
2,424
2,349
2,349
2,349
2,336
2,327
2,327
2,327
2,327
2,327
2,327
2,327
2,327
2,327
2,327
2,327
2,327
2,327
2,327
2,327
2,327
2,327
2,327
2,327
2,327
2,327
2,327
2,327
2,327
2,327
2.327
Table 7.1.6-7 Annual Distillate Fuel Demand and Sulfur Content: NRLM to 500 ppm in 2007, no
15 ppm Step; U.S. minus AK and HI (million gallons and ppm)

Year
1996
Nonroad
Demand
9,087
Sulfur
2,283
Locomotive
Demand
3,065
Sulfur
2,454
Marine
Demand
1,878
Sulfur
2,918
L&M
Demand
4,943
Sulfur
2,641
Heating Oil
Demand
10,715
Sulfur
2,871

-------
1997
1998
1999
2000
2001
2002
2003
2004
2005
2006
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2031
2032
2033
2034
2035
2036
2037
2038
2039
2040
9,376
9,665
9,945
10,238
10,530
10,821
11,112
11,403
11,694
11,983
12,272
12,562
12,851
13,140
13,430
13,721
14,012
14,302
14,593
14,881
15,169
15,456
15,744
16,032
16,319
16,607
16,895
17,183
17,470
17,756
18,042
18,328
18,613
18,899
19,185
19,470
19,756
20,042
20,328
20,613
20,899
21,185
21,470
21.756
2,283
2,283
2,283
2,283
2,283
2,283
2,283
2,283
2,283
2,242
1,126
330
330
276
237
237
237
237
237
237
237
237
237
237
237
237
237
237
237
237
237
237
237
237
237
237
237
237
237
237
237
237
237
237
2,971
2,876
2,782
2,687
2,772
2,692
2,722
2,741
2,762
2,818
2,868
2,900
2,939
2,986
3,043
3,073
3,097
3,121
3,148
3,181
3,210
3,234
3,266
3,288
3,305
3,335
3,364
3,393
3,426
3,453
3,481
3,508
3,536
3,564
3,591
3,619
3,646
3,674
3,701
3,729
3,756
3,784
3,811
3.839
2,454
2,454
2,454
2,454
2,454
2,454
2,454
2,454
2,454
2,435
1,225
361
361
293
245
245
245
245
245
245
245
245
245
245
245
245
245
245
245
245
245
245
245
245
245
245
245
245
245
245
245
245
245
245
1,863
1,849
1,834
1,820
1,805
1,773
1,795
1,813
1,825
1,868
1,895
1,921
1,944
1,968
1,997
2,023
2,041
2,066
2,089
2,109
2,132
2,164
2,201
2,226
2,254
2,290
2,316
2,347
2,374
2,405
2,436
2,467
2,499
2,532
2,564
2,598
2,631
2,665
2,700
2,735
2,770
2,806
2,842
2.879
2,918
2,918
2,918
2,918
2,918
2,918
2,918
2,918
2,918
2,902
1,469
445
445
348
280
280
280
280
280
280
280
280
280
280
280
280
280
280
280
280
280
280
280
280
280
280
280
280
280
280
280
280
280
280
4,834
4,725
4,616
4,507
4,577
4,465
4,517
4,554
4,587
4,686
4,763
4,821
4,883
4,954
5,039
5,096
5,138
5,187
5,236
5,290
5,342
5,398
5,466
5,515
5,559
5,625
5,680
5,740
5,800
5,858
5,917
5,976
6,035
6,095
6,155
6,216
6,277
6,339
6,401
6,463
6,526
6,590
6,653
6.718
2,641
2,641
2,641
2,641
2,637
2,638
2,638
2,639
2,639
2,621
1,323
394
394
315
259
259
259
259
259
259
259
259
259
259
259
259
259
259
259
259
259
259
259
259
259
259
259
259
259
259
260
260
260
260
10,654
10,593
10,532
10,471
10,411
10,352
10,292
10,233
10,174
10,116
10,058
10,000
9,943
9,886
9,829
9,772
9,716
9,661
9,605
9,550
9,495
9,441
9,386
9,333
9,279
9,226
9,173
9,120
9,068
9,016
8,964
8,913
8,861
8,811
8,760
8,710
8,660
8,610
8,561
8,511
8,463
8,414
8,366
8.318
2,871
2,871
2,871
2,871
2,871
2,871
2,871
2,871
2,871
2,860
2,667
2,530
2,530
2,526
2,523
2,523
2,523
2,523
2,523
2,523
2,523
2,523
2,523
2,523
2,523
2,523
2,523
2,523
2,523
2,523
2,523
2,523
2,523
2,523
2,523
2,523
2,523
2,523
2,523
2,523
2,523
2,523
2,523
2.523

-------
Table 7. 1.6-8 Proposed Rule Program: NRLM to 500 ppm in 2007,
Nonroad Only to 15 ppm in 2010; U.S. minus AK and HI (million gallons and ppm)

Year
1996
1997
1998
1999
2000
2001
2002
2003
2004
2005
2006
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2031
2032
2033
2034
2035
2036
2037
2038
2039
2040
Nonroad
Demand
9,087
9,376
9,665
9,945
10,238
10,530
10,821
11,112
11,403
11,694
11,983
12,272
12,562
12,851
13,140
13,430
13,721
14,012
14,302
14,593
14,881
15,169
15,456
15,744
16,032
16,319
16,607
16,895
17,183
17,470
17,756
18,042
18,328
18,613
18,899
19,185
19,470
19,756
20,042
20,328
20,613
20,899
21,185
21,470
21.756
Sulfur
2,283
2,283
2,283
2,283
2,283
2,283
2,283
2,283
2,283
2,283
2,242
1,127
330
330
152
25
25
25
17
11
11
11
11
11
11
11
11
11
11
11
11
11
11
11
11
11
11
11
11
11
11
11
11
11
11
Locomotive
Demand
3,065
2,971
2,876
2,782
2,687
2,772
2,692
2,722
2,741
2,762
2,818
2,868
2,900
2,939
2,986
3,043
3,073
3,097
3,121
3,148
3,181
3,210
3,234
3,266
3,288
3,305
3,335
3,364
3,393
3,426
3,453
3,481
3,508
3,536
3,564
3,591
3,619
3,646
3,674
3,701
3,729
3,756
3,784
3,811
3.839
Sulfur
2,454
2,454
2,454
2,454
2,454
2,454
2,454
2,454
2,454
2,454
2,437
1,226
361
361
293
245
245
245
200
168
168
168
168
168
168
168
168
168
168
168
168
168
168
168
168
168
168
168
168
168
168
168
168
168
168
Marine
Demand
1,878
1,863
1,849
1,834
1,820
1,805
1,773
1,795
1,813
1,825
1,868
1,895
1,921
1,944
1,968
1,997
2,023
2,041
2,066
2,089
2,109
2,132
2,164
2,201
2,226
2,254
2,290
2,316
2,347
2,374
2,405
2,436
2,467
2,499
2,532
2,564
2,598
2,631
2,665
2,700
2,735
2,770
2,806
2,842
2.879
Sulfur
2,918
2,918
2,918
2,918
2,918
2,918
2,918
2,918
2,918
2,918
2,904
1,469
445
445
343
270
270
270
259
252
252
252
252
252
252
252
252
252
252
252
252
252
252
252
252
252
252
252
252
252
252
252
252
252
252
L&M
Demand
4,943
4,834
4,725
4,616
4,507
4,577
4,465
4,517
4,554
4,587
4,686
4,763
4,821
4,883
4,954
5,039
5,096
5,138
5,187
5,236
5,290
5,342
5,398
5,466
5,515
5,559
5,625
5,680
5,740
5,800
5,858
5,917
5,976
6,035
6,095
6,155
6,216
6,277
6,339
6,401
6,463
6,526
6,590
6,653
6.718
Sulfur
2,641
2,641
2,641
2,641
2,641
2,637
2,638
2,638
2,639
2,639
2,623
1,323
394
394
313
255
255
255
224
202
202
202
202
202
202
202
202
202
202
203
203
203
203
203
203
203
203
203
203
204
204
204
204
204
204
Heating Oil
Demand
10,715
10,654
10,593
10,532
10,471
10,411
10,352
10,292
10,233
10,174
10,116
10,058
10,000
9,943
9,886
9,829
9,772
9,716
9,661
9,605
9,550
9,495
9,441
9,386
9,333
9,279
9,226
9,173
9,120
9,068
9,016
8,964
8,913
8,861
8,811
8,760
8,710
8,660
8,610
8,561
8,511
8,463
8,414
8,366
8.318
Sulfur
2,871
2,871
2,871
2,871
2,871
2,871
2,871
2,871
2,871
2,871
2,860
2,667
2,530
2,530
2,526
2,523
2,523
2,516
2,512
2,512
2,512
2,512
2,512
2,512
2,512
2,512
2,512
2,512
2,512
2,512
2,512
2,512
2,512
2,512
2,512
2,512
2,512
2,512
2,512
2,512
2,512
2,512
2,512
2,512
2.512

-------
Table 7.1.6-9 Annual Distillate Fuel Demand and Sulfur Content for the Reference Case;
U.S. (million gallons and ppm)

Year
1996
1997
1998
1999
2000
2001
2002
2003
2004
2005
2006
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2031
2032
2033
2034
2035
2036
2037
2038
2039
2040
Nonroad
Demand
9,136
9,426
9,717
9,999
10,293
10,586
10,879
11,172
11,465
11,757
12,048
12,339
12,629
12,920
13,210
13,503
13,795
14,087
14,379
14,672
14,961
15,250
15,539
15,829
16,118
16,407
16,986
17,275
17,564
17,852
18,139
18,426
18,714
19,001
19,575
19,288
19,575
19,863
20,150
20,437
20,724
21,012
21,299
21,586
21.873
Sulfur
2,284
2,284
2,284
2,284
2,284
2,284
2,284
2,284
2,284
2,284
2,244
2,214
2,214
2,214
2,160
2,121
2,121
2,121
2,121
2,121
2,121
2,121
2,121
2,121
2,121
2,121
2,121
2,121
2,121
2,121
2,121
2,121
2,121
2,121
2,121
2,121
2,121
2,121
2,121
2,121
2,121
2,121
2,121
2,121
2.121
Locomotive
Demand
3,072
2,977
2,882
2,787
2,691
2,776
2,696
2,726
2,745
2,766
2,823
2,873
2,904
2,944
2,990
3,047
3,077
3,102
3,126
3,152
3,186
3,215
3,239
3,271
3,293
3,310
3,339
3,369
3,398
3,431
3,458
3,486
3,514
3,541
3,569
3,596
3,624
3,651
3,679
3,707
3,734
3,762
3,789
3,817
3.844
Sulfur
2,455
2,455
2,455
2,455
2,455
2,455
2,455
2,455
2,455
2,455
2,437
2,424
2,424
2,424
2,255
2,134
2,134
2,134
2,134
2,134
2,134
2,134
2,134
2,134
2,134
2,134
2,134
2,134
2,134
2,134
2,134
2,134
2,134
2,134
2,134
2,134
2,134
2,134
2,134
2,134
2,134
2,134
2,134
2,134
2.134
Marine
Demand
1,960
1,945
1,929
1,914
1,899
1,884
1,850
1,873
1,892
1,905
1,949
1,977
2,005
2,029
2,054
2,084
2,111
2,130
2,156
2,180
2,200
2,225
2,258
2,297
2,323
2,352
2,390
2,417
2,449
2,478
2,510
2,542
2,575
2,608
2,642
2,676
2,711
2,746
2,781
2,817
2,854
2,891
2,928
2,966
3.004
Sulfur
2,902
2,902
2,902
2,902
2,902
2,902
2,902
2,902
2,902
2,902
2,888
2,878
2,878
2,878
2,705
2,581
2,581
2,581
2,581
2,581
2,581
2,581
2,581
2,581
2,581
2,581
2,581
2,581
2,581
2,581
2,581
2,581
2,581
2,581
2,581
2,581
2,581
2,581
2,581
2,581
2,581
2,581
2,581
2,581
2.581
L&M
Demand
5,032
4,922
4,811
4,701
4,590
4,660
4,546
4,599
4,637
4,671
4,772
4,850
4,909
4,972
5,044
5,131
5,188
5,232
5,282
5,332
5,386
5,440
5,497
5,567
5,617
5,662
5,730
5,786
5,847
5,909
5,968
6,028
6,089
6,150
6,211
6,273
6,335
6,497
6,460
6,524
6,588
6,652
6,717
6,783
6,849
Sulfur
2,640
2,640
2,640
2,640
2,640
2,635
2,637
2,637
2,637
2,637
2,621
2,609
2,609
2,609
2,438
2,316
2,316
2,316
2,316
2,317
2,317
2,317
2,318
2,318
2,319
2,320
2,320
2,321
2,321
2,321
2,322
2,322
2,323
2,324
2,324
2,325
2,325
2,326
2,326
2,327
2,328
2,328
2,329
2,329
2.330
Heating Oil
Demand
11,071
11,088
10,945
10,882
10,819
10,757
10,695
10,634
10,573
10,512
10,452
10,392
10,332
10,273
10,214
10,155
10,097
10,039
9,982
9,924
9,867
9,811
9,754
9,698
9,643
9,587
9,532
9,478
9,423
9,369
9,315
9,262
9,209
9,156
9,103
9,051
8,999
8,947
8,896
8,845
8,794
8,744
8,694
8,644
8,594
Sulfur
2,859
2,859
2,859
2,859
2,859
2,859
2,859
2,859
2,859
2,859
2,849
2,842
2,842
2,842
2,712
2,624
2,624
2,624
2,624
2,624
2,624
2,624
2,624
2,624
2,624
2,624
2,624
2,624
2,624
2,624
2,624
2,624
2,624
2,624
2,624
2,624
2,624
2,624
2,624
2,624
2,624
2,624
2,624
2,624
2.624

-------
Table 7.1.6-10 Annual Distillate Fuel Demand and Sulfur Content: Final NRLM Rule:
U.S. (million gallons and ppm)

Year
1996
1997
1998
1999
2000
2001
2002
2003
2004
2005
2006
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2031
2032
2033
2034
2035
2036
2037
2038
2039
2040
Nonroad
Demand
9,136
9,426
9,717
9,999
10,293
10,586
10,879
11,172
11,465
11,757
12,048
12,339
12,629
12,920
13,210
13,503
13,795
14,087
14,379
14,672
14,961
15,250
15,539
15,829
16,118
16,407
16,697
16,986
17,275
17,564
17,852
18,139
18,426
18,714
19,001
19,288
19,575
19,863
20,150
20,437
20,724
21,012
21,299
21,586
21.873
Sulfur
2,284
2,284
2,284
2,284
2,284
2,284
2,284
2,284
2,284
2,284
2,242
1,130
335
335
157
30
30
30
19
11
11
11
11
11
11
11
11
11
11
11
11
11
11
11
11
11
11
11
11
11
11
11
11
11
11
Locomotive
Demand
3,072
2,977
2,882
2,787
2,691
2,776
2,696
2,726
2,745
2,766
2,823
2,873
2,904
2,944
2,990
3,047
3,077
3,102
3,126
3,152
3,186
3,215
3,239
3,271
3,293
3,310
3,339
3,369
3,398
3,431
3,458
3,486
3,514
3,541
3,569
3,596
3,624
3,651
3,679
3,707
3,734
3,762
3,789
3,817
3.844
Sulfur
2,455
2,455
2,455
2,455
2,455
2,455
2,455
2,455
2,455
2,455
2,435
1,228
364
364
178
46
46
46
61
71
71
71
71
71
71
71
71
71
71
71
71
71
71
71
71
71
71
71
71
71
71
71
71
71
71
Marine
Demand
1,960
1,945
1,929
1,914
1,899
1,884
1,850
1,873
1,892
1,905
1,949
1,977
2,005
2,029
2,054
2,084
2,111
2,130
2,156
2,180
2,200
2,225
2,258
2,297
2,323
2,352
2,390
2,417
2,449
2,478
2,510
2,542
2,575
2,608
2,642
2,676
2,711
2,746
2,781
2,817
2,854
2,891
2,928
2,966
3.004
Sulfur
2,902
2,902
2,902
2,902
2,902
2,902
2,902
2,902
2,902
2,902
2,886
1,500
512
512
242
49
49
49
36
27
27
27
27
27
27
27
27
27
27
27
27
27
27
27
27
27
27
27
27
27
27
27
27
27
27
L&M
Demand
5,032
4,922
4,811
4,701
4,590
4,660
4,546
4,599
4,637
4,671
4,772
4,850
4,909
4,972
5,044
5,131
5,188
5,232
5,282
5,332
5,386
5,440
5,497
5,567
5,617
5,662
5,730
5,786
5,847
5,909
5,968
6,028
6,089
6,150
6,211
6,273
6,335
6,497
6,460
6,524
6,588
6,652
6,717
6,783
6.849
Sulfur
2,640
2,640
2,640
2,640
2,640
2,635
2,637
2,637
2,637
2,637
2,620
1,340
425
425
204
47
47
47
51
53
53
53
53
53
53
53
53
53
53
53
53
53
53
53
53
53
53
53
52
52
52
52
52
52
52
Heating Oil
Demand
11,071
11,088
10,945
10,882
10,819
10,757
10,695
10,634
10,573
10,512
10,452
10,392
10,332
10,273
10,214
10,155
10,097
10,039
9,982
9,924
9,867
9,811
9,754
9,698
9,643
9,587
9,532
9,478
9,423
9,369
9,315
9,262
9,209
9,156
9,103
9,051
8,999
8,947
8,896
8,845
8,794
8,744
8,694
8,644
8.594
Sulfur
2,859
2,859
2,859
2,859
2,859
2,859
2,859
2,859
2,859
2,859
2,849
2,662
2,529
2,529
2,420
2,343
2,343
2,343
2,337
2,333
2,333
2,333
2,333
2,333
2,333
2,333
2,333
2,333
2,333
2,333
2,333
2,333
2,333
2,333
2,333
2,333
2,333
2,333
2,333
2,333
2,333
2,333
2,333
2,333
2.333

-------
Table 7. 1.6-11 Annual Distillate Fuel Demand and Sulfur Content: NRLM to 500 ppm in 2007, no
15 ppm Step; U.S. (million gallons and ppm)

Year
1996
1997
1998
1999
2000
2001
2002
2003
2004
2005
2006
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2031
2032
2033
2034
2035
2036
2037
2038
2039
2040
Nonroad
Demand
9,136
9,426
9,717
9,999
10,293
10,586
10,879
11,172
11,465
11,757
12,048
12,339
12,629
12,920
13,210
13,503
13,795
14,087
14,379
14,672
14,961
15,250
15,539
15,829
16,118
16,407
16,697
16,986
17,275
17,564
17,852
18,139
18,426
18,714
19,001
19,288
19,575
19,863
20,150
20,437
20,724
21,012
21,299
21,586
21.873
Sulfur
2,284
2,284
2,284
2,284
2,284
2,284
2,284
2,284
2,284
2,284
2,242
1,130
335
335
278
237
237
237
237
237
237
237
237
237
237
237
237
237
237
237
237
237
237
237
237
237
237
237
237
237
237
237
237
237
237
Locomotive
Demand
3,072
2,977
2,882
2,787
2,691
2,776
2,696
2,726
2,745
2,766
2,823
2,873
2,904
2,944
2,990
3,047
3,077
3,102
3,126
3,152
3,186
3,215
3,239
3,271
3,293
3,310
3,339
3,369
3,398
3,431
3,458
3,486
3,514
3,541
3,569
3,596
3,624
3,651
3,679
3,707
3,734
3,762
3,789
3,817
3.844
Sulfur
2,455
2,455
2,455
2,455
2,455
2,455
2,455
2,455
2,455
2,455
2,435
1,227
364
364
295
245
245
245
245
245
245
245
245
245
245
245
245
245
245
245
245
245
245
245
245
245
245
245
245
245
245
245
245
245
245
Marine
Demand
1,960
1,945
1,929
1,914
1,899
1,884
1,850
1,873
1,892
1,905
1,949
1,977
2,005
2,029
2,054
2,084
2,111
2,130
2,156
2,180
2,200
2,225
2,258
2,297
2,323
2,352
2,390
2,417
2,449
2,478
2,510
2,542
2,575
2,608
2,642
2,676
2,711
2,746
2,781
2,817
2,854
2,891
2,928
2,966
3.004
Sulfur
2,902
2,902
2,902
2,902
2,902
2,902
2,902
2,902
2,906
2,906
2,886
1,502
512
512
378
282
282
282
282
282
282
282
282
282
282
282
282
282
282
282
282
282
282
282
282
282
282
282
282
282
282
282
282
282
282
L&M
Demand
5,032
4,922
4,811
4,701
4,590
4,660
4,546
4,599
4,637
4,671
4,772
4,850
4,909
4,972
5,044
5,131
5,188
5,232
5,282
5,332
5,386
5,440
5,497
5,567
5,617
5,662
5,730
5,786
5,847
5,909
5,968
6,028
6,089
6,150
6,211
6,273
6,335
6,497
6,460
6,524
6,588
6,652
6,717
6,783
6.849
Sulfur
2,640
2,640
2,640
2,640
2,640
2,635
2,637
2,637
2,637
2,637
2,620
1,340
425
425
329
260
260
260
260
260
260
260
260
260
260
260
260
260
260
260
260
261
261
261
261
261
261
261
261
261
261
261
261
261
261
Heating Oil
Demand
11,071
11,088
10,945
10,882
10,819
10,757
10,695
10,634
10,573
10,512
10,452
10,392
10,332
10,273
10,214
10,155
10,097
10,039
9,982
9,924
9,867
9,811
9,754
9,698
9,643
9,587
9,532
9,478
9,423
9,369
9,315
9,262
9,209
9,156
9,103
9,051
8,999
8,947
8,896
8,845
8,794
8,744
8,694
8,644
8.594
Sulfur
2,859
2,859
2,859
2,859
2,859
2,859
2,859
2,859
2,859
2,859
2,849
2,662
2,529
2,529
2,525
2,522
2,522
2,522
2,522
2,522
2,522
2,522
2,522
2,522
2,522
2,522
2,522
2,522
2,522
2,522
2,522
2,522
2,522
2,522
2,522
2,522
2,522
2,522
2,522
2,522
2,522
2,522
2,522
2,522
2.522

-------
Table 7.1.6-12 Annual Distillate Fuel Demand and Sulfur Content: Proposed Rule Program: 500
ppm NRLM ppm in 2007, 15 ppm Nonroad Only in 2010; U.S. (million gallons and ppm)

Year
1996
1997
1998
1999
2000
2001
2002
2003
2004
2005
2006
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2031
2032
2033
2034
2035
2036
2037
2038
2039
2040
Nonroad
Demand
9,136
9,426
9,717
9,999
10,293
10,586
10,879
11,172
11,465
11,757
12,048
12,339
12,629
12,920
13,210
13,503
13,795
14,087
14,379
14,672
14,961
15,250
15,539
15,829
16,118
16,407
16,697
16,986
17,275
17,564
17,852
18,139
18,426
18,714
19,001
19,288
19,575
19,863
20,150
20,437
20,724
21,012
21,299
21,586
21.873
Sulfur
2,284
2,284
2,284
2,284
2,284
2,284
2,284
2,284
2,284
2,284
2,242
1,130
335
335
163
40
40
40
23
11
11
11
11
11
11
11
11
11
11
11
11
11
11
11
11
11
11
11
11
11
11
11
11
11
11
Locomotive
Demand
3,072
2,977
2,882
2,787
2,691
2,776
2,696
2,726
2,745
2,766
2,823
2,873
2,904
2,944
2,990
3,047
3,077
3,102
3,126
3,152
3,186
3,215
3,239
3,271
3,293
3,310
3,339
3,369
3,398
3,431
3,458
3,486
3,514
3,541
3,569
3,596
3,624
3,651
3,679
3,707
3,734
3,762
3,789
3,817
3.844
Sulfur
2,455
2,455
2,455
2,455
2,455
2,455
2,455
2,455
2,455
2,455
2,435
1,228
364
364
295
245
245
245
200
169
169
169
169
169
169
169
169
169
169
169
169
169
169
169
169
169
169
169
169
169
169
169
169
169
169
Marine
Demand
1,960
1,945
1,929
1,914
1,899
1,884
1,850
1,873
1,892
1,905
1,949
1,977
2,005
2,029
2,054
2,084
2,111
2,130
2,156
2,180
2,200
2,225
2,258
2,297
2,323
2,352
2,390
2,417
2,449
2,478
2,510
2,542
2,575
2,608
2,642
2,676
2,711
2,746
2,781
2,817
2,854
2,891
2,928
2,966
3.004
Sulfur
2,902
2,902
2,902
2,902
2,902
2,902
2,902
2,902
2,902
2,902
2,888
1,502
512
512
373
273
273
273
255
242
242
242
242
242
242
242
242
242
242
242
242
242
242
242
242
242
242
242
242
242
242
242
242
242
242
L&M
Demand
5,032
4,922
4,811
4,701
4,590
4,660
4,546
4,599
4,637
4,671
4,772
4,850
4,909
4,972
5,044
5,131
5,188
5,232
5,282
5,332
5,386
5,440
5,497
5,567
5,617
5,662
5,730
5,786
5,847
5,909
5,968
6,028
6,089
6,150
6,211
6,273
6,335
6,497
6,460
6,524
6,588
6,652
6,717
6,783
6.849
Sulfur
2,640
2,640
2,640
2,640
2,640
2,635
2,637
2,637
2,637
2,637
2,621
1,340
425
425
326
256
256
256
223
199
199
199
199
199
199
199
199
199
199
199
199
199
200
200
200
200
200
200
200
200
200
200
201
201
201
Heating Oil
Demand
11,071
11,088
10,945
10,882
10,819
10,757
10,695
10,634
10,573
10,512
10,452
10,392
10,332
10,273
10,214
10,155
10,097
10,039
9,982
9,924
9,867
9,811
9,754
9,698
9,643
9,587
9,532
9,478
9,423
9,369
9,315
9,262
9,209
9,156
9,103
9,051
8,999
8,947
8,896
8,845
8,794
8,744
8,694
8,644
8.594
Sulfur
2,859
2,859
2,859
2,859
2,859
2,859
2,859
2,859
2,859
2,859
2,849
2,662
2,529
2,529
2,525
2,522
2,522
2,522
2,516
2,511
2,511
2,511
2,511
2,511
2,511
2,511
2,511
2,511
2,511
2,511
2,511
2,511
2,511
2,511
2,511
2,511
2,511
2,511
2,511
2,511
2,511
2,511
2,511
2,511
2.511

-------
Final Regulatory Support Document
7.2 Refining Costs

   The most significant cost involved in providing diesel fuel meeting more stringent sulfur
standards is the cost of removing the sulfur at the refinery. In this section, we describe the
methodology used and present the estimated costs for refiners to:
•  comply with the 2007 Nonroad, Locomotive, and Marine (NRLM) 500 ppm diesel fuel sulfur
   standards and the 15 ppm nonroad diesel fuel standard in 2010 and the 15 ppm L&M
   standard in 2012,
•  comply with other NRLM diesel fuel sulfur sensitivity cases considered, and
   comply with the 2006 sulfur standards already adopted for highway diesel fuel (an update of
   a previous cost analysis).
   Finally, we compare our estimated costs with those developed by Mathpro (for the Engine
Manufacturers Association) and Baker and O'Brien (for the American Petroleum Institute).

7.2.1 Methodology

   7.2.1.1  Overview

   This section describes the methodology used to estimate the refining cost of reducing diesel
fuel sulfur content. Costs are estimated based on two distinct desulfurization technologies:
conventional hydrotreating and the Process Dynamics  IsoTherming process. Conventional
hydrotreating cost estimates were based on information from two vendors, while the cost
estimates for the more advanced process was made from information provided by the respective
vendor. For both technologies, costs are estimated for each U.S. refinery currently producing
distillate fuel. Conventional hydrotreating technology was projected to be used to desulfurize
distillate to meet a 500 ppm sulfur cap. A mix comprised of advanced desulfurization
technology with some conventional hydrotreating technology was projected to be used to meet
the 15 ppm sulfur cap.  This mix of technology varied  depending on the timing of the 15 ppm
sulfur standard. To meet the 500 ppm and 15 ppm  sulfur standards, refiners are expected to
desulfurize to 340 ppm and 7 ppm, respectively.

   Refining costs were developed for revamping existing hydrotreaters that produce low-sulfur
diesel fuel, as well as new, grass roots desulfurization units.  The lower revamped costs were
primarily used when streams or parts of streams were already desulfurized (i.e., highway), while
the grassroots costs applied normally for untreated  streams (mostly nonroad). In both cases,
costs were developed for our refinery cost model and used to estimate the desulfurization cost
for each refinery in the United States producing distillate fuel in 2001. These refinery-specific
costs consider the volume of distillate fuel produced, the composition of this distillate fuel, and
the location of the refinery (e.g., Gulf Coast, Rocky Mountain region, etc.). The estimated
composition of each refinery's distillate included the fraction of hydrotreated and
nonhydrotreated straight-run distillate, light cycle oil (LCO), other cracked stocks (coker,
visbreaker, thermal cracked) and hydrocracked distillate, and the cost to desulfurize  each of
those stocks.  The cost information provided by the various vendors was used to develop the
desulfurization cost for each blendstock; however, when lacking, engineering judgment was used
to develop the needed specific cost estimate.  The average desulfurization cost for each refinery

                                          7-86

-------
                                                 Estimated Costs of Low-Sulfur Fuels
was based on the volume-weighted average of desulfurizing each of those blendstocks. The
production volumes used were those indicative of 2014, a midyear of the estimated 15 year
project life of the year 2007 capital investments by the refining industry.

   7.2.1.2 Basic Cost Inputs for Specific Desulfurization Technologies

   To obtain a comprehensive basis for estimating the cost of desufurizing diesel fuel, over the
past few years we have held meetings with a large number of vendors of desulfurization
technologies. These firms include: Criterion Catalyst, UOP, Akzo Nobel, Haldor Topsoe, and
Process Dynamics. We have also met with numerous refiners of diesel fuel considering the use
of these technologies and reviewed the literature on this subject.  The information and estimates
described below represent the culmination of these efforts.  See Chapter 5 of the RIA for a more
complete discussion of conventional hydrotreating and Process Dynamics Isotherming, as well
as other desulfurization technologies evaluated in the course of this rulemaking.

   The information used in our refinery cost model for estimating the cost of meeting 500 and
15 ppm sulfur caps using conventional  hydrotreating is presented first.  The cost methodology
for conventional hydrotreating was developed for the HD2007 rulemaking for highway diesel
fuel. Only the final process-design parameters are presented here. For a complete description of
the methodology used to develop the cost estimates for conventional hydrotreating, consult
Chapter 5 of the HD2007 Regulatory Impact Analysis.15  The few variations from the HD2007
methodology are described below.

   Next we present the methodology and resulting cost information used for developing  the
refinery costs for the Process Dynamics IsoTherming processs. In this case, we begin by
presenting the estimates of the process-design parameters provided by the developers of this
process.  These projections are then evaluated to produce sets of process-design parameters that
can be used to estimate the cost of meeting 500 ppm and 15 ppm NRLM diesel fuel standards for
each domestic refiner. The resulting refining cost projections are presented and discussed in
Section 7.2.2.

   7.2.1.2.1  Conventional Desulfurization Technology

   The cost of desulfurizing diesel fuel includes the capital  cost related to designing and
constructing the desulfurization unit, as well as the cost of operating the unit. We were able to
obtain fairly complete sets of such process-design parameters from two out of the five or six
licensors of conventional desulfurization technologies16'17'18.  These designs addressed the
production of 15 ppm diesel fuel by retrofitting existing hydrotreaters originally designed to
produce 500 ppm diesel fuel, as well as building new, grass roots units. These two sets of
process-design parameters were also used to estimate the cost of hydrotreating high-sulfur diesel
fuel down to 500 ppm.

   In addition to the information obtained from these two vendors, we reviewed similar
information  submitted to the National Petroleum Council (NPC) by Akzo Nobel, Criterion,
Haldor Topsoe, UOP and IFF for its study of diesel fuel desulfurization costs and discussed them

                                          7-87

-------
Final Regulatory Support Document
with the vendors.19 These submissions were generally not as comprehensive as those provided
by the two vendors mentioned above.  In all cases, these submissions corroborated the costs from
the two vendors.

   All the vendors identified operating pressures sufficient to produce fuel meeting a 15 ppm
sulfur cap under 900 psi. Most of the vendors projected that 650 psi is sufficient, while others
indicated that pressures well below 1000 psi are sufficient. A contractor for API indicated that
they believe a 850 psi unit is enough to meet a 15 ppm cap, though lower-pressure units would
not be sufficient. We therefore based our estimate of capital cost on two different vendor
submissions based on units operating at 650 and 900 psi.

   Based on the information obtained from the two vendors of conventional hydrotreating
technologies, as well as that obtained from Process Dynamics, we project that refiners will use
conventional hydrotreating to produce NRLM diesel fuel meeting the 500 ppm  standard in 2007.
This unit would include heat exchangers, a fired pre-heater, a reactor, a hydrogen compressor
and a make up compressor, and both high-pressure and low-pressure strippers.  The refinery
would also need a source of new hydrogen, an amine scrubber and a sulfur plant.  Most refineries
already have sources of hydrogen, an amine scrubber and a sulfur plant.  However, considering
the hydrogen demand for complying with Tier 2 sulfur standards for gasoline and the 15 ppm
cap on highway diesel sulfur, no residual refinery production hydrogen is expected to exist. We
therefore project that any new hydrogen demand will likely be produced from the addition of a
new steam reforming hydrogen plant using natural gas as the feedstock, either on-site or by a
third party.  Likewise,  a refinery's amine scrubber and sulfur plant would need  modest
expansion.

   Producing diesel fuel meeting a 15 ppm standard generally requires much greater reactor
volume and a larger hydrogen capacity, both in terms of compressor capacity and ability to
introduce this hydrogen into the reactor, than are required to meet a 500 ppm cap.  Since the  15
ppm sulfur cap for nonroad diesel fuel follows the 500 ppm NRLM sulfur cap by  only three
years and L&M by 5 years, we project that refiners will design any new hydrotreaters built for
2007 to be easily retrofitted with additional equipment, such as a second reactor, a hydrogen
compressor, a recycle scrubber, an inter-stage stripper and other associated process hardware.
The technical approach described by each vendor to achieve a 15 ppm sulfur cap (average level
of 7-8 ppm) is summarized in Table 7.2.1-1.

-------
                                                  Estimated Costs of Low-Sulfur Fuels
                                       Table 7.2.1-1
             Modifications Necessary to Reduce 500 ppm Sulfur Levels to 15 ppm
    Diesel Fuel
   Sulfur Level
               Vendor A
             Vendor B
    7-8 ppm
   (15 ppm cap)
Change to a more active catalyst
Install recycle gas scrubber
Modify compressor
Install a second reactor, high pressure (900 psi)
Use existing hot oil separator for inter-stage
   stripper	
Change to a more active catalyst
Install a recycle gas scrubber
Install a second reactor (650 psi)
Install a color reactor
Install an interstage stripper
   It is important to note that back when the highway rulemaking was being promulgated, the
vendors of conventional hydrotreating technology believed that a high pressure interstage
stripper was needed for each hydrotreating unit to meet the 15 ppm sulfur cap standard, and
included the costs for such a unit in their cost estimates. However, since that time the vendors
are no longer recommending that the 15 ppm hydrotreaters include such a stage in the
desulfurization process thus negating the need for the associated piece of capital.  Our costs
estimates are nevertheless still based on the vendor capital cost estimates which include the
interstage stripper. Thus, the capital costs on which this rulemaking is based are,  with respect to
this single factor, somewhat conservative compared to the costs which refiners would likely
incur to comply with the 15 ppm sulfur standard.

   The vendors assumed that the existing highway desulfurization unit in place could be utilized
(revamped) to comply with the 15 ppm sulfur standards. This includes hydrotreater sub-units
necessary for desulfurization.  Revamping the highway unit saves on both capital  and operating
costs for a two-stage revamp compared with whole new grassroots unit.  These sub-units include
heat exchangers, a heater,  a reactor filled with catalyst, two or more vessels used for separating
hydrogen and any light ends produced by cracking during the desulfurization process, a
compressor, and sometimes a hydrogen recycle gas scrubber.  The desulfurization subunits listed
here are discussed in detail in Chapter 5.

   To estimate the cost of meeting the NRLM diesel fuel sulfur standards, it was necessary to
evaluate three situations refiners may face:  (1) producing NRLM diesel fuel meeting a 15 ppm
cap from diesel fuel already being hydrotreated to meet a 500 ppm cap (i.e., a highway revamp),
(2) producing NRLM diesel fuel meeting a 15 ppm cap from high-sulfur distillate (i.e., grass
roots 15 ppm hydrotreater), and (3) producing 15 ppm NRLM diesel fuel meeting a 500 ppm cap
by replacing the existing hydrotreater with a grass roots 15 ppm hydrotreater. Sets of process-
design parameters for the first  two of these desulfurization configurations were developed for the
HD2007 rule and summarized  in the Regulatory Impact Analysis.20 As discussed above,  only
the results of the previous derivations are presented below. The third configuration was not
addressed for the highway diesel fuel rule, as highway diesel fuel was already meeting a 500
ppm cap.  The section that develops the process-design parameters for this third configuration
includes a short description of the methodology used in its development, as it is very similar to
those used to develop the first  two sets of process-design parameters.
                                           7-89

-------
Final Regulatory Support Document
   One straightforward adjustment was made to all the capital costs developed for the HD2007
rule. The capital costs developed for that rule were in terms of 1999 dollars. These costs were
updated to represent 2002 dollars by increasing them by 2.5 percent to reflect inflation in
construction costs occurring between 1999 and 2002.21

       7.2.1.2.1.1 Revamping to Process 500 ppm Diesel Fuel to Meet a 15 ppm Cap

   The process-design projections developed in this section apply to a revamp of an existing
desulfurization unit with additional  hardware to enable the combined older and new unit to meet
a 15 ppm sulfur cap.  The portion of these projections that apply to operating costs are also
relevant if a refiner decides to replace an existing diesel fuel desulfurization unit with a new
grassroots unit.  In this case, the entire capital cost of the grass roots unit is incurred.  However,
the incremental operating costs would be those of the new grass roots unit, less those of the
existing hydrotreater (which are developed in this section).

   The process-design parameters shown below were taken directly from those shown in the
HD2007 Regulatory Impact Analysis, with two adjustments.  The first adjustment relates to the
amount of desulfurization required from the current low sulfur diesel pool, while the second
adjustment relates to the amount of fuel gas consumed in the process.

   Diesel fuel complying with the current 500 ppm sulfur standard typically contains 340 ppm
sulfur.  We expect refiners complying with the 500 ppm NRLM diesel fuel sulfur cap also to
desulfurize down to roughly 340 ppm sulfur. Thus, in revamping an existing 500 ppm
hydrotreater to comply with a 15 ppm cap, refiners will have to desulfurize from about 340 ppm
down to 7 ppm. This is analogous to what we assumed in the analysis for the HD2007 rule.
After the highway diesel fuel rule was finalized, however, it became evident that the vendor
projections assumed a starting sulfur level of 500 ppm and not 340 ppm.  Thus, the vendor
projections assumed more desulfurization would be needed than is the case here.  Based on a
curve of hydrogen consumption versus initial and final sulfur level developed in the Regulatory
Impact Analysis supporting the proposed HD2007 program, reducing the initial sulfur level from
500 ppm to 340 ppm reduces hydrogen consumption by 3.5 percent.22  We assumed that all cost-
related parameters (capital cost,w catalyst cost, yield losses, and utilities) will be reduced by the
same 3.5 percent.

   For the second adjustment, the fuel gas rates were adjusted to account for the heat produced
by the saturation of the aromatic compounds that occurs during desulfurization. In the Draft RIA
for the NPRM, we presumed  that the highly aromatic blendstocks, which are LCO and coker,
would consume more fuel gas than straight run distillate, which has much less aromatics.
However, because the aromatic compounds are exothermic in the hydrotreating reactor,  they
actually contribute some heat which lowers the heat load compared to straight run distillate.
Furthermore, when updating the fuel gas consumption values, we found and corrected an error in
   w Capital costs are also affected, as a higher starting sulfur level requires a larger reactor to provide a greater
residence time to remove the sulfur and a larger compressor for the greater volume of hydrogen which must be fed to
the reactor.

                                          7-90

-------
                                                Estimated Costs of Low-Sulfur Fuels
our interpretation of fuel gas consumption information from one of the two vendors which
provided us with the unit operations information for their diesel fuel desulfurization technology.
The error was that we had interpreted that vendor's information to read as thousands of British
thermal units (BTUs) per day instead of millions of BTUs per day.

    Some of the information from one of the two vendors (which was referred to as Vendor A in
the 2007 Highway Final Rule) was used to estimate the relative heat demand for the two mixed
distillate streams.  The heat demand information was presented as million BTU per hour a
25,000 bbl/day grassroots unit producing 15 ppm diesel. We converted this estimate to BTU/bbl
and summarized the values in Table 7.2.1-2.

                                     Table 7.2.1-2
                 Fuel Gas Demand for a 15 ppm Grassroots Unit (BTU/bbl)
67% cracked stocks, 33% SR
20% cracked stocks, 80% SR
1100
1480
   The above table shows a 380 btu/bbl difference in heat consumption between the two feeds
for a grassroots unit. Based on this information, we were able to estimate that cracked stocks
require only 56 percent of the heat input of straight run stocks.  The fuel gas consumption
estimate for the cracked stocks (LCO and coker light gas oil) is 920 btu/bbl while the fuel gas
consumption for straight run gas oil is 1640 btu/bbl.  Since this is the heat consumption for only
Vendor A, it was necessary to merge the fuel gas consumption information from Vendor B.
Vendor B reported fuel gas consumption of 16,000 btu/bbl.  This value is much higher probably
because it incorporates the fuel gas used to generate steam for pumping. Because both vendors
were providing cost estimates on the same feeds (69 percent straight run 31 percent cracked
stocks) to achieve the same desulfurization target, it is likely that both were assuming similar
levels of aromatics saturation, thus we assume that both vendors would estimate a similar
absolute difference in heat consumption between the  different blendstocks. To estimate an
average heat consumption representing the heat consumption estimates from both vendors, we
averaged the average heat for the two vendors  (assuming an average of 1320 btu/bbl  for Vendor
A) resulting in an average heat consumption of 8660  btu/bbl. Assuming that the heat consumed
by each blendstock maintains the same differential as that calculated based on Vendor A's
information alone, the heat  consumed is 8880 btu/bbl for straight run and 8160 for cracked
stocks which maintains the  same 720 btu/bbl difference from above.

   Since we need to estimate the incremental fuel gas demand for a unit treating diesel fuel
meeting a 500 ppm cap standard to comply with a 15 ppm cap standard for this section, the fuel
consumption information from Vendors A and B was evaluated for this sulfur reduction
increment.  Both vendors show essentially zero fuel gas consumption for this interval, yet
aromatics are  still being saturated  similar to about half the increment of going from untreated to
15 ppm sulfur. Thus, half the difference in fuel gas consumed for cracked stocks and straight run
was assumed for this interval with a typical blend of diesel fuel (69 percent straight run and 31
                                         7-91

-------
Final Regulatory Support Document
percent cracked stocks) having a zero net fuel gas consumption.  Thus, cracked stocks are
estimated to require -250 btu/bbl of fuel gas and straight run is estimated to require 110 btu/bbl
of fuel gas for a difference of 360 scf/bbl or half of that for a grassroots unit.

   Table 7.2.1-3 presents the process-design parameters for desulfurizing 500 ppm sulfur diesel
fuel to meet a 15 ppm standard.

                                      Table 7.2.1-3
    Process Projections for Revamping an Existing Diesel Fuel Hydrotreater Desulfurizing
                 Diesel Fuel  Blendstocks from 500 ppm Cap to 15 ppm Cap

Capacity (BPSD)
Capital Cost (ISBL) (Smillion)
Liquid Hour Space Velocity (hr"1)
Hydrogen Consumption (scf/bbl)
Electricity (kW-hr/bbl)
HP Steam (Ib/bbl)
Fuel Gas (BTU/bbl)
Catalyst Cost (S/BPSD)
Yield Loss (wt%)
Diesel
Naphtha
LPG
Fuel Gas
Straight-Run
25,000
16
1.25
96
0.4
-
110
0.2
1.0
-0.7
-0.04
-0.04
Other Cracked Stocks
25,000
19
0.7
230
0.7
-
-250
0.4
1.9
-1.3
-0.07
-0.11
Light Cycle Oil
25,000
22
0.6
375
0.8
-
-250
0.5
2.1
-1.4
-0.08
-0.13
       7.2.1.2.1.2 Process-Design Projections for a Grassroots Unit Producing 15 ppm Fuel

   The process-design parameters presented in this section were taken directly from those
derived in the HD2007 Regulatory Impact Analysis. These costs apply primarily to refineries
currently producing only, or predominantly, high-sulfur diesel fuel.  In addition, the capital cost
portion of these costs apply to a refinery replacing an existing hydrotreater with a grassroots unit
instead of revamping their existing hydrotreater. In this case, these refiners would incur the
capital costs outlined here, but their operating costs would be based on a revamp, as described
above. Most refineries currently producing high-sulfur distillate fuel also produce some
highway diesel fuel.  In this case, we project costs reflecting those of a revamp and a grass roots
unit. The methodology for this merging of the two costs is described in Section 7.2.1.5 below.

   Table 7.2.1-4 presents the process-design parameters for desulfurizing high-sulfur distillate
fuel to meet a 15 ppm standard in a grassroots unit.
                                           7-92

-------
                                                Estimated Costs of Low-Sulfur Fuels
                                      Table 7.2.1-4
          Process Projections for Installing a New Grassroots Unit for Desulfurizing
              Untreated Distillate Fuel Blendstocks to Meet a 15 ppm Standard

Capacity BPSD
(bbl/day)
Capital Cost (ISBL)
(MM$)
Liquid Hour Space Velocity
(Hi"1)
Hydrogen Consumption
(SCF/bbl)
Electricity
(KwH/bbl)
HP Steam
(Lb/bbl)
Fuel Gas
(BTU/bbl)
Catalyst Cost
($/BPSD)
Yield Loss (%)
Diesel
Naphtha
LPG
Fuel Gas
Straight-Run
25,000
32
0.8
240
0.6
-
8880
0.3
1.5
-1.1
-0.06
-0.06
Other Cracked Stocks
25,000
38
0.5
850
1.1
-
8160
0.6
2.9
-2.0
-0.11
-0.17
Light Cycle Oil
25,000
43
0.4
1100
1.2
-
8160
0.8
3.3
-2.3
-0.12
-0.20
   Unlike processing highway diesel fuel, which is assumed to contain 340 ppm sulfur, the
sulfur content of high-sulfur distillate fuel can vary dramatically from refinery to refinery and
region to region. To account for varying starting sulfur levels, an adjustment in hydrogen
consumption. The basis for the amount of sulfur needing to be removed is that the starting feed,
comprised of 69 percent straight-run, 23 percent LCO and 8 percent cracked stocks, contains
9000 ppm sulfur (0.9 weight percent). However, as described below in Section 7.2.1.3, the
average concentration of sulfur in the overall distillate pool, and especially the untreated part of
the pool, varies by PADD. After estimating this sulfur level, we adjusted the hydrogen
consumption for this varying sulfur level. (According to Vendor B, removing sulfur from diesel
fuel consumes 125 scf/bbl for each weight percent of sulfur removed.23) We did not adjust the
hydrogen consumption for the other qualities, mono- and poly-aromatics and olefins, but
assumed that the hydrogen consumption from saturating olefins and aromatics, or from breaking
aromatic rings would depend more on whether the feedstock had been previously hydrotreated or
not, and less on whether the starting sulfur level was 5000 or 8000 ppm. Since sulfur removal
consumes less than half the hydrogen of desulfurizing from untreated 9000 ppm sulfur
                                          7-93

-------
Final Regulatory Support Document
feedstocks to 15 ppm,x the adjustment is always less than 50 percent.  The adjustment is applied
as an adjustment ratio to each untreated blendstock type for a refinery with a distillate
hydrotreater. The adjustment ranged from 0.80 for PADD 5, which has an estimated untreated
distillate sulfur level of 3010 ppm, to  1.0 for PADD 3, which has an estimated untreated
distillate sulfur level of 9,350 ppm. No adjustment was necessary for the already hydrotreated
part of the distillate pool since this subpool is always assumed to contain 340 ppm sulfur.

    For refineries without a distillate hydrotreater, our adjustment to account for differing
starting  sulfur levels assumes that they currently blend only unhydrotreated blendstocks into the
distillate that comprises the high-sulfur pool.  Thus, we are making our adjustments based on a
lower starting sulfur level. Our adjustment for these refineries ranged from 0.79 for PADD 4,
which has an estimated untreated sulfur level of 2550 ppm, to 0.83 for PADD 3, which  has a
starting  sulfur level of 3780 ppm.  The various hydrogen consumption adjustment values are
summarized in Table 7.2.1-5.

                                      Table 7.2.1-5
                Hydrogen Consumption Adjustment Factors: Grassroots Units

Refinery with Distillate HT
No Distillate HT
PADD 1
0.84
0.80
PADD 2
0.89
0.80
PADD 3
1.0
0.83
PADD 4
0.81
0.79
PADD 5
0.80
0.79
       7.2.1.2.1.3 Desulfurizing High-Sulfur Distillate Fuel to a 500 ppm Cap

   Finally, we needed to provide inputs for our cost model for desulfurizing untreated, high-
sulfur distillate to meet a 500 ppm sulfur standard, which is the first step of our two-step
program. These inputs are estimated by simply subtracting the inputs for the revamped unit for
desulfurizing 500 ppm diesel fuel down to 15 ppm from the inputs for a grassroots unit for
desulfurizing untreated diesel fuel down to 15 ppm.  The untreated to 500 ppm inputs for our
refinery cost model are summarized in Table 7.2.1-6.
   x Much of the hydrogen consumption is due to the saturation of olefms, or partial saturation of aromatics.

                                          7-94

-------
                                                Estimated Costs of Low-Sulfur Fuels
                                      Table 7.2.1-6
               Process Projections for Installing a New Unit for Desulfurizing
            Untreated Diesel Fuel Blendstocks to Meet a 500 ppm Sulfur Standard

Capacity BPSD
(bbl/day)
Capital Cost (ISBL)
(MM$)
Liquid Hour Space Velocity
(Hi"1)
Hydrogen Consumption
(SCF/bbl)
Electricity
(KwH/bbl)
HP Steam
(Lb/bbl)
Fuel Gas
(BTU/bbl)
Catalyst Cost
($/BPSD)
Yield Loss (%)
Diesel
Naphtha
LPG
Fuel Gas
Straight-Run
25,000
15
2.4
144
0.2
-
8770
0.1
0.5
-0.4
-0.02
-0.02
Coker Distillate
25,000
18
1.9
620
0.4
-
8410
0.2
1.1
-0.7
-0.04
-0.06
Light Cycle Oil
25,000
21
1.3
725
0.4
-
8410
0.3
1.2
-0.8
-0.04
-0.07
   Again, a hydrogen consumption adjustment was made for starting sulfur levels that differ
from 9000 ppm.  In this case, the hydrogen adjustment ended up being larger than the grassroots
desulfurization unit as the adjustment to the hydrogen consumption for going from untreated to
500 ppm comprises a larger percentage of the total hydrogen consumption. This adjustment is
for a refinery with a distillate hydrotreater.  The adjustment is applied as an adjustment ratio to
each unhydrotreated blendstock type and it ranged from 0.69 for PADD 5, which has an
estimated untreated distillate sulfur level of 3010 ppm, to 1.0 for PADD 3, which has an
estimated untreated distillate sulfur level of 9,350 ppm. No adjustment was necessary for the
already hydrotreated part of the distillate pool since this subpool is always assumed to contain
340 ppm sulfur.

   For refineries without a distillate hydrotreater, our analysis does not assume that they
currently hydrotreat any of the distillate that comprises the high-sulfur pool. Thus, we estimate a
somewhat lower starting sulfur level. Our adjustment for these refineries ranged from 0.67 for
PADD 4, which has an estimated untreated sulfur level of 2550 ppm, to 0.73 for PADD 3, which
                                          7-95

-------
Final Regulatory Support Document
has a starting sulfur level of 3780 ppm.  The various hydrogen consumption adjustment values
are summarized in Table 7'.2.1-7'.

                                      Table 7.2.1-7
             Hydrogen Consumption Adjustment Factors: High Sulfur to 500 ppm

Refinery with Distillate HT
No Distillate HT
PADD 1
0.75
0.69
PADD 2
0.83
0.69
PADD 3
1.0
0.73
PADD 4
0.70
0.67
PADD 5
0.69
0.67
       7.2.1.2.1.4 Hydrocrackate Processing and Tankage Costs

   We believe refineries with hydrocrackers will have to invest some capital and incur some
operating costs to ensure that recombination reactions at the exit of the second stage of their
hydrocracker do not cause the diesel fuel being produced by their hydrocracker to exceed the
standard. The hydrocracker is a very severe hydrotreating unit capable of hydrotreating its
product from thousands of ppm sulfur to nearly zero ppm sulfur; however, hydrogen sulfide
recombination reactions that occur at the end of the cracking stage, and fluctuations in unit
operations, such as temperature and catalyst life, can result in the hydrocracker diesel product
having up to 30 ppm sulfur in its product stream.24 25  Thus, refiners may need to install a
finishing reactor for the diesel stream produced by the hydrocracker. According to vendors, this
finishing reactor is a low-temperature, low-pressure hydrotreater that can desulfurize the simple
sulfur compounds formed in the cracking stage of the hydrocracker.

   Additionally, since the 15 ppm diesel sulfur standard is very stringent, we take into account
tankage that will likely be needed. We believe refiners could store high-sulfur batches of
highway diesel fuel or nonroad diesel fuel during a shutdown of the diesel fuel hydrotreater.
Diesel fuel production would cease in the short term, but the rest of the refinery could  remain
operative.  To account for this, we provided for the cost of installing a tank that would store ten
days of 15 ppm sulfur diesel production, sufficient for a ten-day emergency turnaround, which is
typical for the industry; the estimated cost for a 270,000  barrel  storage tank is $3 million.26 The
cost of the land needed for this tank is assumed to be negligible relative to the cost of the tank.
This amount of storage should be adequate for most unanticipated turnarounds. We presumed
that each refinery will need to add such  storage, though for some refineries, off-spec diesel fuel
could also be sold  as high-sulfur heating oil or fuel oil.

   The cost inputs for the storage tank and the finishing reactor are summarized in Table 7.2.1-
                                          7-96

-------
                                                  Estimated Costs of Low-Sulfur Fuels
                                       Table 7.2.1-8
                       Process Operations Information for Additional
                       Units used in the Desulfurization Cost Analysis

Capacity
Capital Cost
(MM$)
Electricity
(KwH/bbl
HP Steam
(Lb/bbl)
Fuel Gas
(BTU/bbl)
Cooling Water
(Gal/bbl)
Operating Cost
($/bbl)
Diesel
Storage Tank
50,000 bbls
0.75
—
—
—
—
none3
Distillate Hydrocracker
Post Treat Reactor
25,000 (bbl/day)
5.727
0.98
4.2
18
5
see above
     No operating costs are estimated directly; however both the ISBL to OSBL factor and the capital contingency
       factor used for desulfurization processes is used for the tankage as well, which we believe to be excessive
       for storage tanks so it is presumed to cover the operating cost.
   Refiners will also likely invest in a diesel fuel sulfur analyzer.28 A sulfur analyzer at the
refinery provides nearly real-time information regarding the sulfur levels of important streams in
the refinery and facilitate operational modifications to prevent excursions above the sulfur cap.
Based on information from a manufacturer of such an analyzer, the analyzer costs about $50,000,
with an additional $5,000 estimated for installation.29 Compared with the capital and operating
cost of desulfurizing diesel fuel, the cost for this instrumentation is far below 1 percent of the
total cost of this program.  Because the cost is so small, the cost of an analyzer was assumed
covered as a cost contingency described in Section 7.2.1.4.1.

   7.2.1.2.2  Process Dynamics IsoTherming

   Process Dynamics has licensed a technology called IsoTherming, which is designed to
desulfurize both highway and non-highway distillate fuel.  At our request, Process Dynamics
provided basic design parameters that can be used to project the cost of using their process to
meet tighter sulfur caps,30 which is summarized in the process information table. Subsequently,
EPA spoke to a Linde engineer responsible for implementing the IsoTherming unit at the Giant
refinery.31  The hydrogen and utility consumption information obtained earlier from Process
Dynamics was adjusted based on these comments, as described in the text further below.
                                           7-97

-------
Final Regulatory Support Document
   Specifically, Process Dynamics provided design parameters for a revamp of an existing
highway desulfurization unit to meet a 15 ppm standard. The revamp involves putting an
IsoTherming unit upstream of the existing highway diesel fuel hydrotreater. Thus, when
applying the Process Dynamics unit in our cost estimates for meeting the 15 ppm standard, the
new Process Dynamics unit itself is  assumed to be used as a first stage. As described in more
detail in Chapter 5 of the RIA, this configuration takes the most advantage of the inherent
benefits of the Process Dynamics IsoTherming desulfurization process.

   Process Dynamics provided to EPA process information for the IsoTherming process based
on three revamp situations. In the first revamp design, the feedstock consisted of 60 percent
straight-run and 40 percent LCO. The unhydrotreated sulfur level was just under 2000 ppm and
both the existing hydrotreater and the IsoTherming unit operated at 600 psi. In the second
design, the feedstock consisted of 60 percent straight-run, 30 percent LCO and 10 percent light-
coker gas oil with an unhydrotreated sulfur level of 9950 ppm. The existing hydrotreater and the
IsoTherming unit operated at 950 psi. In the third design, the feedstock was the same as in the
second, but the IsoTherming unit was designed to operate at 1500 psi, while the conventional
hydrotreating unit operated at 950 psi.

   We largely based our cost projections for the IsoTherming process on the second design.
The unhydrotreated sulfur level of more than 9000 ppm is more typical for most refiners than
2000 ppm.  The 950 psi design pressure for the IsoTherming unit was also thought to preferable
to 1500 psi for most refiners. The higher-pressure unit reduces capital and catalyst costs, but
higher hydrogen consumption  offsets much of the cost savings. The higher-pressure reactors and
compressors also have a longer delivery time and there would likely be fewer  fabricators.  Thus,
given that the savings associated with the higher pressure unit were small, we  decided to focus
on the 950 psi design.

   The information provided by Process Dynamics for the 950 psi IsoTherming desulfurization
unit is summarized in Table 7.2.1-9. The operation and product quality of the IsoTherming unit
is shown separatly from those for the existing conventional hydrotreater.  Again,  prior to the
revamp, the conventional hydrotreater would have processed this feedstock down to roughly 340
ppm sulfur.
                                         7-98

-------
                                                Estimated Costs of Low-Sulfur Fuels
                                      Table 7.2.1-9
                         Process Dynamics IsoTherming Revamp
                  Design Parameters to Produce 10 ppm Sulfur Diesel Fuel

LCD vol %
Straight-Run vol %
Light-Coker Gas Oil vol%
Sulfur ppm
Nitrogen
API gravity (degrees)
Cetane Index
H2 Consumption (scf/bbl)
Relative H2 Consumption
LHSV (hr1)
Relative Catalyst Volume
Reactor Delta T
H2 Partial Pressure
Electricity (kW)
Natural Gas (mmbtu/hr)
Steam (Ib/hr)
Feed Quality
30
60
10
9950
340
33.98
44.5









IsoTherming Unit and its
Product Quality



850
38
34.42
48.5
320
75
15/15
45
15
950
Conventional Hydrotreater and
Final Product Quality



10
2
35.84
50.8
100
25
3
100
15
950
1525
0
0
       7.2.1.2.2.1 Hydrotreating High-Sulfur Distillate Fuel to 15 ppm

   The design parameters provided by Process Dynamics involve the revamp of an existing
conventional hydrotreater currently producing highway diesel fuel (i.e., less than 500 ppm
sulfur) to produce diesel fuel with a sulfur level well below 15 ppm. Before addressing this
situation, however, we will use the Process Dynamics revamp design to project the costs of an
IsoTherming unit that processes unhydrotreated distillate fuel (e.g., 3400-10,000 ppm sulfur)
down to 7-8 ppm sulfur. This type of unit was not projected to be used under the two-step fuel
program. However, we considered such a sulfur reduction step for alternative programs, for
which costs are also estimated later in this chapter.

   Also, as was done for conventional hydrotreating, we develop cost estimates for applying the
IsoTherming process to three individual blendstocks—straight-run, LCO and light-coker gas
oil—to be able to project desulfurization costs for individual refineries whose diesel fuel
compositions vary dramatically.
                                          7-99

-------
Final Regulatory Support Document
   We have broken down the derivation of the cost of a stand-alone IsoTherming unit capable of
producing 15 ppm diesel fuel into four parts:  hydrogen consumption, utilities and yield losses,
catalyst cost and capital cost.

   Hydrogen Consumption: In this section, we estimate the hydrogen consumption to process
individual refinery streams from their uncontrolled levels down to 7-8 ppm sulfur. Process
Dynamics provided hydrogen consumption estimates for desulfurizing a mixed feedstock of 60
percent straight-run, 30 percent LCO and 10 percent coker distillate, but not for specific refinery
streams.  Additionally, Process Dynamics provided information for a hybrid desulfurization unit
comprised of a Process Dynamics IsoTherming unit revamping a conventional highway
hydrotreater. For the proposed rule, we used the hydrogen consumption values provided by
Process Dynamics to estimate the hydrogen consumption for the IsoTherming unit for the
individual diesel fuel blendstocks which we model. This information resulted in a hydrogen
consumption which was somewhat lower than that of conventional hydrotreating.  After the
proposal, we asked the Linde engineers to provide their most recent estimate of the hydrogen
consumption values for the IsoTherming process based on the in-use data from their commercial
demonstration unit.  The resulting hydrogen consumption estimates for the IsoTherming process
are similar to that of conventional hydrotreating.  Consequently, for the final rule analysis we set
the hydrogen consumption of the Process Dynamics IsoTherming process to be the same as
conventional hydrotreating. The resulting hydrogen consumptions were 1100 scf/bbl  for LCO,
850 scf/bbl for other cracked stocks, and 240 scf/bbl for  straight-run.

   Consistent with the methodology used for conventional hydrotreating,  we developed
adjustments to each blendstock hydrogen consumption values to reflect differing unhydrotreated
sulfur levels. We assumed that the hydrogen consumption for IsoTherming process varied in the
same proportions as those for conventional hydrotreating because the treated feed sulfur levels
were about the same. Thus, the same hydrogen adjustment factors were used as for conventional
hydrotreating, and they can be found in Table 7.2.1-5 and Table7.2.1-7.

   Utilities and Yield  Losses: We next established the IsoTherming utility inputs for individual
blendstocks. The Process Dynamics IsoTherming process saves a substantial amount of heat
input by conserving the heat of reaction that occurs in the IsoTherming reactors. This conserved
energy is used to heat the feedstock to the unit. This differs from conventional hydrotreating that
normally rejects much  of this energy to avoid coking the catalyst. According to Process
Dynamics, this allows the IsoTherming process to operate with negligible  external heat input. In
the highway hydrotreater revamp, which is the source of the information provided by  Process
Dynamics, the  existing heater for the highway hydrotreater was hardly needed after the
IsoTherming process was added. However, there is still  the need for a small  heater to heat up the
feedstock during unit startup. This affects capital costs.  However, when averaged over
production between start-ups (generally at least two years), the little amount of fuel used during
start-up is negligible. Thus, we estimate no need  for either fuel or steam with the IsoTherming
process.

   As shown in Table 7.2.1-9, Process Dynamics estimated electricity demand to be  1525
kilowatts per 20,000 bbl/day unit in their early estimate of the demands for their unit.  However,

                                        7-100

-------
                                                Estimated Costs of Low-Sulfur Fuels
since the commercial demonstration unit has been operating, Process Dynamics has collected
information on the actual electrical consumption of the IsoTherming unit. Process Dynamics
engineers estimate that the electrical consumption is about that same as conventional
hydrotreating.  Thus, for desulfurizing untreated diesel fuel down to 15 ppm, we set the
electricity demand as the same as conventional hydrotreating. Thus, we estimate electricity
demand at 0.6, 1.1 and 1.2 kW-hr/bbl for straight-run, light-coker gas oil, and LCO, respectively.

   This is a decline in electricity consumption compared to the values which Process Dynamics
reported in  their original document. That the IsoTherming unit would consume the same (or
potentially  less) electricity as conventional hydrotreating is reasonable considering that no
recycle compressor is needed with this technology because large excesses of hydrogen are not
fed to the IsoTherming reactor. Recycle compressors are a large electricity consumer. This
electricity savings is somewhat offset because of the increased liquid pumping demands required
to recycle the diesel fuel through the reactors. While some savings are likely, Process Dynamics
suggested we assume that the electricity costs are about the same as conventional hydrotreating.

   Process Dynamics did not estimate the specific yield losses for the IsoTherming process. On
our request for further information, Process Dynamics indicated that their process causes slightly
less than half of the yield loss of conventional hydrotreating.  Thus, the yield loss of the  Process
Dynamics unit was projected to be 50 percent that of conventional hydrotreating, which  is
proportional to the relative catalyst volume.  The resulting projected yield losses are shown in
Table 7.2.1-10 below:

                                     Table 7.2.1-10
         Estimated Yield Loss for a Process Dynamics IsoTherming Grassroots Unit
Fuel Type
Diesel Fuel
Naphtha
LPG
Fuel Gas
Straight Run
0.75
-0.55
-0.03
-0.03
Light Coker Gas Oil
1.45
-1.00
-0.055
-0.085
Light Cycle Oil
1.65
-1.15
-0.06
-0.10
   Catalyst Costs: The catalyst cost for the Process Dynamics process was estimated based on
the relative catalyst volume compared with conventional hydrotreating. As shown in
Table 7.2.1-9, Process Dynamics indicated that the catalyst volume for the new IsoTherming
reactors contained only 45 percent of the volume of the new conventional hydrotreating reactors
that Process Dynamics projects would be needed to revamp the existing hydrotreater to produce
10 ppm fuel. We assumed that this same relationship holds for a stand-alone IsoTherming unit.
Thus, we multiplied the catalyst costs for conventionally hydrotreating specific blendstocks
(shown in Table 7.2.1-4) by 45 percent.  The resulting IsoTherming catalyst costs were 0.14,
0.27 and 0.36 $/BPSD for straight-run, light-coker gas  oil and LCO, respectively.
                                         7-101

-------
Final Regulatory Support Document
   Capital Costs: The last aspect of the IsoTherming process to be determined on a per-
blendstock basis is its capital cost.  Process Dynamics's initial submission of process-design
parameters did not include an estimate of the capital cost.  We developed our own estimate from
the process equipment included, compared with those involved in conventional hydrotreating.
As indicated in Table 7.2.1-9, the catalyst volume of the two IsoTherming reactors unit
(combined LHSV of 7.5) is roughly 8 times smaller than that of a conventional hydrotreating
revamp (LHSV of 0.9 per LHSVs for individual blendstocks from Table 7.2.1-4). Also, because
the IsoTherming reactors use a much higher flowrate and is a totally liquid process (no need for
both gas and liquid in the reactor), it eliminates the need for an expensive distributor.  As
mentioned above, the feed pre-heater can be much smaller and less durable, since it is required
only for startup. Finally, the IsoTherming process does not require an amine scrubber to scrub
the H2S from the recycle hydrogen stream.

   Based on these differences, we estimated that the total capital cost of a stand-alone
IsoTherming unit is two-thirds that for a conventional hydrotreater.  Thus, the  capital costs for a
25,000 bbl per day conventional hydrotreater were reduced by one-third. The  resulting
IsoTherming capital costs for a 25,000 BPSD unit were $21, $25, and $29 million for treating
straight-run, light-coker gas oil and LCO, respectively. The estimated overall  capital cost for the
specific feed composition shown in Table 7.2.1-9 is $900 per BPSD  for the IsoTherming unit,
versus $1400 per BPSD for a conventional hydrotreater. More recently, Linde indicated that the
capital cost will be roughly $800 per barrel for a 25,000 bbl per day unit.32 For this analysis,  we
consequently retained the two-thirds factor relative to conventional hydrotreating ($900 per
BPSD).

   Summary of Process-Design Parameters: Table 7.2.1-11 summarizes the design parameters
used for using the Process Dynamics IsoTherming process to desulfurize untreated distillate fuel
to 10 ppm.
                                         7-102

-------
                                               Estimated Costs of Low-Sulfur Fuels
                                    Table 7.2.1-11
                    Process Parameters for a Stand-Alone IsoTherming
       25,000 BPSD Unit to Produce 10 ppm Sulfur Fuel from Untreated Distillate Fuel

Capital Cost ($MM)
Hydrogen Demand (scf/bbl)
Electricity Demand (kwh/bbl)
Fuel Gas Demand (btu/bbl)
Catalyst Cost ($/bpsd)
Yield Loss (wt%): Diesel
Naphtha
LPG
Fuel Gas
Straight-Run (SR)
21
240
0.6
220
0.15
0.75
-0.55
-0.03
-0.03
Other Cracked Stocks
25
850
1.1
-500
0.29
1.45
-1.00
-0.055
-0.085
Light Cycle Oil (LCD)
29
1100
1.2
-500
0.44
1.65
-1.15
-0.06
-0.10
       7.2.1.2.2.2 Desulfurizing 500 ppm Sulfur Diesel Fuel to Meet a 15 ppm Sulfur Cap

   The derivation of process design parameters for a IsoTherming unit revamp of a conventional
hydrotreater is much more straightforward than that of a stand-alone IsoTherming unit, as the
design parameters provided by Process Dynamics in Table 7.2.1-9 were for a revamp. The
revamp would occur by placing the new Process Dynamics IsoTherming unit as a first stage
(uncontrolled  to under 500 ppm), before the existing highway highway, thus converting the
highway hydrotreater to treating diesel fuel from several hundred ppm to under 15 ppm.  Similar
to how we characterized the cost inputs above, we have broken down the derivation of the cost
of a stand-alone IsoTherming unit capable of producing 15 ppm diesel fuel into four parts:
hydrogen consumption, utilities and yield losses, catalyst cost and capital cost.

   Hydrogen Consumption: Determining the incremental hydrogen consumption of a Process
Dynamics IsoTherming revamp of a conventional hydrotreater requires that the existing
hydrogen consumption of the existing conventional hydrotreater be accounted for.  As described
above, we now estimate that the hydrogen consumption of the Process Dynamics unit to be the
same as the conventional hydrotreating unit for the same service. Thus, there would be no
change in hydrogen consumption when the Process Dynamics unit replaces the conventional
hydrotreating  unit for treating diesel fuel from uncontrolled levels down to 500 ppm sulfur. The
conventional hydrotreater's new role would be to desulfurize 500 ppm sulfur down to 15 ppm
sulfur. The new service of the conventional hydrotreater will define the hydrogen consumption
for this Process Dynamics IsoTherming revamp of the conventional hydrotreater unit. The
hydrogen consumption of a conventional hydrotreater for treating 500 ppm diesel fuel down to
15 ppm is  contained in Table 7.2.1-6 above, which is 96, 230 and 375 standard cubic feet per
minute of hydrogen for straight run, coker,  and LCO, respectively.
                                        7-103

-------
Final Regulatory Support Document
   Utilities and Yield Losses: The electricity consumption for a Process Dynamics IsoTherming
revamp of a conventional hydrotreater follows the same logic as that for hydrogen.  Again the
Process Dynamics unit is assumed to have the same electrical demand as the conventional
hydrotreater for desulfurizing untreated diesel fuel down to 500 ppm.  Thus, the incremental
electricity demand for this revamp is the electrical demand for the conventional hydrotreater in
its new 500 ppm to 15 ppm service. The electric demand of a conventional hydrotreater for
treating 500 ppm diesel fuel down to 15 ppm is contained in Table 7.2.1-6 above, which is 0.4,
0.7 and 0.8 kilowatt hours per barrel for straight run, coker, and LCO, respectively.

   Estimating fuel gas consumption for a Process Dynamics revamp of a conventional
hydrotreater is more complex because the Process Dynamics unit's fuel gas consumption is not
the same as a conventional hydrotreater for desulfurizing undesulfurized diesel fuel down to 500
ppm.  This calculation is best shown in Table 7.2.1-12. The table shows the addition of the
Process Dynamics unit for desulfurizing each undesulfurized blendstock to 500 ppm, the
subtraction of the conventional hydrotreater for the same increment of sulfur control for each
blendstock, the addition of the conventional hydrotreater now treating 500 ppm diesel fuel down
to 15 ppm for each blendstock, and the net change in fuel gas consumption.

                                     Table 7.2..1-12
      Estimate of Fuel Gas Consumption of an IsoTherming Revamp; 500 ppm to 15 ppm

IsoTherming Unit: High Sulfur
to 500 ppm (added)
Conv. HT: High Sulfur to 500
ppm (subtracted)
Conv. HT 500 ppm to 15 ppm
(added)
Net Fuel Gas Consumption
Straight Run
110
8770
110
-8550
Coker
-250
8410
-250
-8910
LCO
-250
8410
-250
-8910
   As mentioned above, Process Dynamics did not provide estimates of yield losses for the
IsoTherming process. Using engineering judgement based on the relative exposure to the
catalyst (the Process Dynamics unit only uses 45 percent of the catalyst as a conventional
hydrotreater), we estimated that a stand-alone IsoTherming unit would reduce yield losses by 45
percent compared to a stand-alone convention hydrotreater. We applied this factor to the
conventional hydrotreater yield loss to estimate the Process Dynamics yield loss. Table 7.2.1-6
shows that the yield loss for straight run feed is 1.0 percent for a conventional hydrotreating
revamp (500 ppm to 15  ppm) and Table 7.2.1-4 shows a 1.5 percent loss for a grass roots
conventional hydrotreater (uncontrolled to 15 ppm). Thus, the original highway fuel
hydrotreater (uncontrolled to 500 ppm) has a yield loss of 0.5 percent for straight run, consistent
with that shown in Table 7.2.1-3.
                                         7-104

-------
                                                Estimated Costs of Low-Sulfur Fuels
   If the IsoTherming revamp reduces the yield loss by 45 percent, its yield loss for straight run
is 55 percent of 1.5 percent, or 0.82 percent.  Subtracting out the 0.5 percent loss of the original
highway hydrotreater means that the IsoTherming revamp had an incremental yield loss of 0.32
percent, or 32 percent of the 1.0 percent yield loss projected for the conventional hydrotreating
revamp. Thus, we projected that all of the yield losses shown in Table 7.2.1-13 for a
conventional hydrotreating revamp would be only 32 percent as large for an IsoTherming
revamp.

                                     Table 7.2.1-13
             Estimated Yield Loss for a Process Dynamics  IsoTherming Revamp
Fuel Type
Diesel Fuel
Naphtha
LPG
Fuel Gas
Straight Run
0.32
-0.22
-0.01
-0.01
Light Coker Gas Oil
0.61
-0.42
-0.02
-0.035
Light Cycle Oil
0.70
-0.48
-0.03
-0.04
   Catalyst Costs: Consistent with the relative catalyst cost for a stand-alone IsoTherming unit,
we project that the catalyst cost for an IsoTherming revamp would be 45 percent of that for a
conventional hydrotreating revamp.

   Capital Costs: Consistent with the relative capital cost for a stand-alone IsoTherming unit,
we project that the capital cost for an IsoTherming revamp would be 45 percent of that for a
conventional hydrotreating revamp.

   Summary of Process Design Parameters:  The inputs into our cost model for treating already
treated non-highway diesel fuel by the individual refinery streams which is presumed to be 340
ppm is summarized in Table 7.2.1-14.
                                         7-105

-------
Final Regulatory Support Document
                                     Table 7.2.1-14
                      Process Projections for an IsoTherming Revamp
               of a Conventional Hydrotreater to Meet a 15 ppm Cap Standard

Capital Cost ($MM)
Unit Size (bbl/stream Day)
Hydrogen Demand (scf/bbl)
Electricity Demand (kwh/bbl)
Fuel Gas Demand (btu/bbl)
Catalyst Cost ($/bpsd)
Yield Loss (wt%)
Diesel
Naphtha
LPG
Fuel Gas
Straight Run (SR)
10.6
25,000
96
0.4
-8550
0.09
0.25
-0.18
-0.01
-0.01
Other Cracked Stocks
12.5
25,000
230
0.7
-8910
0.18
0.48
-0.33
-0.02
-0.03
Light Cycle Oil (LCD)
14.5
25,000
375
0.8
-8910
0.23
0.55
-0.38
-0.02
-0.03
    7.2.1.2.3 Characterization of Vendor Cost Estimates

    Applicability to Specific Refineries:  The information provided by the vendors is based on
typical diesel fuels or diesel fuel blendstocks. However, in reality, diesel fuel (especially LCO,
and to a lesser degree other cracked stocks) varies in desulfurization difficulty based on the
amount of sterically hindered compounds present in the fuel, which is determined by the
endpoint of diesel fuel, and also by the type of crude oil being refined and other unit processes.
The vendors provided cost information based on diesel fuels with T-90 distillation points which
varied from 605 °F to 630 °F, which would roughly correspond to distillation endpoints of 655 °F
to 680 °F. These endpoints can be interpreted to mean that the diesel fuel would, as explained in
Chapter V above, contain sterically hindered compounds.  Other diesel fuels or diesel fuel
blendstocks, such as a straight run diesel fuel with a lower end boiling point,  are lighter and
would not contain sterically hindered compounds. However, a summer time diesel fuel survey
for  1997 shows that the endpoint of highway diesel fuel varies from 600 °F to 700  °F, thus the
lighter diesel fuels would contain no sterically hindered compounds, and the heavier diesel fuels
would contain more.33 Our analysis attempts to capture the cost for each refinery to produce
highway diesel fuel which meets the 15ppm cap sulfur standard,  however, we do not have
specific information for how the highway diesel endpoints vary from refinery to refinery, or from
season to season.  Similarly, we do not have information on what type of crude oil is being
processed by each refinery as the quality of crude oil being processed by a refinery affects the
desulfurization difficulty of the various diesel fuel blendstocks. Diesel fuel processed by a
particular refiner can  either be easier or more difficult to treat than what we estimate depending
on how their diesel fuel endpoint compares to the average endpoint of the industry, and
depending on the crude oil used. For a nationwide analysis, we believe it is appropriate to base
our cost analysis for each refinery on what we estimate would be typical or average qualities for
                                         7-106

-------
                                                 Estimated Costs of Low-Sulfur Fuels
each diesel fuel blendstock. Some estimates of individual refinery costs will be high, others will
be low, but be representative on average.

   Accuracy of Vendor Estimates:  We have heard from refiners in the past that the vendor
costs are optimistic and need to be adjusted higher to better assess the costs.  While the vendors
costs may be optimistic, we believe that there are a multitude of reasons why the cost estimates
could be optimistic and adjusting these estimates isn't necessary.

   First, in specific situations, capital costs can be lower than what the vendors project for a
generic refinery. Many refiners own used reactors, compressors, and other vessels which can be
employed in a new or revamped diesel hydrotreating unit. We do not know to what extent that
additional hydrotreating capacity can be met by employing used vessels, however, we believe
that at least a portion of the capital costs can be offset by used equipment. Additionally, the
vendors of conventional hydrotreating which provided cost estimate information for our analysis
based their capital costs on the inclusion of an interstage stripper to strip out the hydrogen
sulfide between the first and second reactor stages (see Chapter 5  of the RIA).  However,
vendors today are saying that interstage strippers are not necessary.  Thus, the capital costs upon
which our conventional hydrotreating costs are based are conservative,  which offsets optimism
on the part of the vendors.

   There are also operational changes which refiners can make to reduce the difficulty and the
cost of desulfurizing highway diesel fuel. Based on  the information which we received from
vendors and as made apparent in our cost analysis which follows, refiners with LCO in their
diesel fuel would need to hydrotreat their highway diesel pool more severely resulting in a
higher cost to meet the cap standard. We believe that these refiners could potentially avoid some
or much of this higher cost by pursuing two specific options.  The first option which we believe
these refiners would consider would be to shift LCO to heating oil which does not face such
stringent sulfur control. The more lenient sulfur limits which regulate heating oil provide room
for blending in substantial amounts of LCO. The refineries which could take advantage of
shifting LCO to the heating oil pool are those in the Northeast and on the Gulf Coast which have
access to the large heating oil market in the Northeast. If refiners could not shift all the LCO to
the heating oil pool because of market limitations, refiners could distill  its LCO into light and
heavy fractions  and only shift the heavy fraction to the heating oil pool. Essentially all of the
sterically hindered compounds distill above 630°F, so if refiners undercut their LCO to omit
these compounds, they would cut out about 30 percent of their LCO.  We expect that refiners
could shift the same volume of non-LCO distillate from these other distillate pools to the NRLM
pool to maintain current production volumes of all fuels. The T-90 maximum established by
ASTM may limit the amount of LCO, and especially heavy LCO, which can be moved from
NRLM diesel fuel into the heating oil pool. Another option, of course,  would be to move this
dirty distillate fraction into number 4 or number 6 marine bunker fuel. For those refineries
which could trade the heavy portion of LCO with other blendstocks in the heating oil pool from
their own refinery or other refineries, we presume that those refiners could make the separations
cheaply by using a splitting column for separating the undercut LCO from the uncracked heavy
gasoil in the FCC bottoms.
                                         7-107

-------
Final Regulatory Support Document
   Another option for refineries which are faced with treating LCO in its nonroad diesel fuel
would be to sell off or trade their heavy LCO to refineries with a distillate hydrocracker. This is
a viable option only for those refineries which are located close to another refinery with a
distillate hydrocracker.  The refinery with the distillate hydrocracker would upgrade the
purchased LCO into gasoline or high quality diesel fuel. To allow this option, there must be a
way to transfer the heavy LCO from the refinery with the unwanted LCO to the refinery with the
hydrocracker, such as a pipeline or some form of water transport. We asked a refinery
consultant to review this option. The refinery consultant corroborated the idea, but commented
that the trading of blendstocks between refineries is a complicated business matter which is not
practiced much outside the Gulf Coast, and that the refineries with hydrocrackers that would buy
up and process this low quality LCO may have to modify their distillate hydrocrackers.34 The
modification which may be needed would be due to the more exothermic reaction temperature of
treating LCO which could require refiners to install additional quenching in those hydrocrackers.
Additionally, LCO can demand 60 to 80 percent more hydrogen for processing than straight run
material. The refiners which could potentially take advantage of selling or trading their LCO to
these other refineries are mostly located in the Gulf Coast where a significant number of
refineries have hydrocrackers and such trading of blendstocks is common. However, there are
other refineries outside of the Gulf Coast which could take advantage of their very close location
to another refinery with a distillate hydrocracker. Examples for these refining areas where a
hydrocracker could be shared include the Billings, Montana area and Ferndale, Washington.

   As we summarized in Chapter  5, catalysts are improving and expected to continue to
improve. Our costs are based on vendor submissions and incorporate the most advanced new
catalysts available at that time.  However, there are several new lines of catalysts available now
which are more active than the  previous lines of catalysts upon which our costs are based.  As
catalysts continue to improve, the cost of desulfurizing diesel fuel will continue to decrease.

   In summary, while some contend that the vendor  cost estimates are optimistically low, there
are a number of reasons why we believe the cost of desulfurizing diesel fuel to meet the 15 ppm
cap standard may be even lower than estimated. Vendors are expected to continue to improve
their desulfurization technology such as the activity of their catalysts. Also, refiners have
several cost cutting options at their disposal, such as using existing spare equipment, to lower
their capital costs which is not considered here.  Also, refiners may be able  to resort to either of
two operational options to reduce the amount of LCO in their highway diesel fuel.

   We are aware that there are potentially other capital and operating costs in the refinery which
would contribute the projected  cost of desulfurizing diesel  fuel  beyond that provided to us by the
vendors.  For example, refiners may need to expand their amine plant or their sulfur plant to
enable the processing of the sulfur compounds removed from diesel fuel. Then the small amount
of additional sulfur compounds treated would incur additional operating costs. Thus, as
described below, we adjusted the projected capital and operating costs upward to account for
these other potential costs which we have not accounted for explicitly.
                                          7-108

-------
                                                  Estimated Costs of Low-Sulfur Fuels
    7.2.1.3 Refinery-Specific Inputs

    There are a number of reasons why we estimated refining costs on a refinery-specific basis.
First, it provides more precise and realistic estimates of desulfurization costs, as some
differences between individual refineries can be represented (e.g., distillate fuel composition,
production volumes, etc.). These costs are approximate, as we do not have precise data on the
distillate composition for all U.S. refineries.  While we do know historic distillate production
levels, we do not know how these will change in the future. Still, the distribution of costs across
refineries facillitated by the factors developed in this section will provide much more insight into
how desulfurization costs can vary between refineries. The alternative would be to estimate
desulfurization costs for the average U.S. refinery and assume that this cost applied to all
refineries. Given the wide range in refinery capacities and their relative production of highway
diesel fuel and high sulfur distillate, the national  average approach would be overly simplistic.

    Second,  a refinery specific approach to costs allows us to better represent the potential
interactions  between the 15 ppm cap for highway diesel fuel and  the NRLM sulfur caps
associated with this rule. We recently received refiners' plans regarding their compliance with
the 15  ppm highway diesel fuel sulfur cap. Being projections, these plans are subject to change.
However, these projections allow us to reasonably estimate the ways in which refiners might
take advantage of efforts to comply with the highway fuel standards in complying with the
NRLM standards.

    Third, the refinery specific costs can be combined into a distribution of costs for the entire
refining industry. This distribution of costs allows us to better estimate the number of refineries
likely to be affected by this rule. It also provides insight into the  range of costs likely to be
experienced by refineries, particularly the difference in costs between those facing the lowest
costs and those facing the highest costs. This will also provide greater insight into how NRLM
diesel fuel prices might be affected by this rule, as well as refiners' ability to recover capital
costs.
    Fourth, the development of refinery specific costs allows us to better estimate how small
refiners might be affected by this rule, in particular how their costs differ from their larger
competitors.

    Of the many factors which affect desulfurization costs, there are  four which vary
significantly from refinery to refinery and which we have estimated  quantitatively:

    1)  the composition of its no. 2 distillate pool (e.g., the percentages of LCO and other
       cracked stocks),
    2)  the percentage of its no. 2 distillate which is already being hydrotreated,
    3)  the volume of no. 2 distillate
    4)  which specific refineries are most likely to produce lower sulfur NRLM fuel.

    The following four subsections discuss how we developed refinery-specific factors for each
of these four factors.
    7.2.1.3.1 Composition of Distillate Fuel by Refinery

                                          7-109

-------
Final Regulatory Support Document
   In section 7.2.1.2, we developed desulfurization costs as a function of the blend stocks
comprising the diesel fuel being processed, as well as other factors. In this section, we describe
how we estimated each refinery's distillate blendstock diesel composition.

   Refiners do not publish blendstock composition data, nor do they submit it to regulators as
part of any regulatory requirements. The only available information is an industry survey
conducted in 1996, which published compositional data for all the surveyed refiners within a
PADD. Thus, we developed a methodology to estimate each refinery's diesel fuel composition
from the aggregated data available from 1996.  We then revised these compositions to reflect
changes in the capacities of those types of equipment which produce distillate blendstock which
have occurred since that time. Finally, we applied one further change to the compositional data
which we believe will occur as a result of the 15 ppm highway fuel cap.

   The only available data on the composition of diesel blend stocks is from a survey conducted
by API and NPRA in 1996.  This survey was sent to all domestic refiners and the responses
covered 79 percent of the total distillate produced by domestic refineries in 1996.  The
blendstock composition of highway diesel fuel and No. 2 high sulfur distillate fuel were
surveyed separately.  The blendstock composition of the combined pool can also be estimated by
volume weighting the compositions of the two distillate pools.

   Table 7.2.1-15 summarizes the survey results for highway diesel fuel, high sulfur distillate
fuel and the combined distillate pool for refiners outside of California. California refiners were
excluded due to the unique specifications which California distillate must meet, namely low
aromatics and high cetane limits. Also, due to the fact that California has already passed
regulations requiring 15 ppm nonroad fuel, this NRLM rule will have a small impact on
California refiners.  The survey also included whether or not the particular blendstock was
hydrotreated. This hydrotreating information will be used in the next section which addresses
the hydrotreated fraction of each refinery's distillate.  According to the cost estimation
methodology described above, desulfurization costs depend on blendstock composition and
overall hydrotreated fraction, but not on the specific blendstocks which are hydrotreated.
Therefore, we do not consider whether the particular blendstock has been hydrotreated here.

                                     Table 7.2.1-15
    Distillate Composition (Excluding California Refiners): 1996 API/NPRA Survey (vol%)

Straight Run
LCD
Other Cracked Stocks
Hydrocrackate
Highway Diesel Fuel
64%
23%
9%
4%
High Sulfur Distillate
63%
22%
5%
10%
All No. 2 Distillate
64%
22%
8%
6%
   As can be seen, the composition of national average highway fuel and high sulfur distillate
are quite similar. This led us to assume, for the purpose of this analysis, that each refinery sent
                                         7-110

-------
                                                 Estimated Costs of Low-Sulfur Fuels
the same fraction of LCO and other cracked stocks to its highway fuel and high sulfur distillate
pools.  This same information was used as the basis for our cost projections presented in the
NPRM for this rule.

    The next step in this analysis was to determine how each refinery's distillate pool might
differ in composition.  For example, some refineries do not have an FCC unit.  Thus, their
distillate would contain no LCO. Others do not have cokers, hydrocrackers, etc.  Thus, we
allocated the volume of each blendstock in the national distillate pool to each refinery in
proportion to the capacity of its equipment which produces each blendstock. As described in
Section 5.1, LCO is produced in FCC units, hydrocrackate is produced by hydrocrackers and
other cracked stocks are primarily produced by cokers, as well as other thermal cracking units.

    While general rules of thumb are available which estimate the volume of distillate produced
in each of these units, in most cases, we have sufficient information available to estimate, on a
national average basis, these conversion factors.  EIA's Petroleum Supply Annual for 1996 states
that domestic refiners produced a total of 3.06 million barrels per day of No. 2  distillate in 1996.
By multiplying this volume by the percentages of LCO, other cracked stocks, and hydrocrackate
in all No. 2 distillate from Table 7.2.1-15 above, we can estimate the total volume of each of
these blendstocks which was produced in 1996. EIA also publishes the capacity of each
refinery's processing units.  By summing these up, we can estimate the total FCC, coker and
thermal cracking and hydrocracker units existing in domestic refineries in 1996.

    The situation with cokers and other thermal crackers is somewhat more complex, as the
conversion of feedstock into distillate does not tend to be the same in these units.  Thus, their
capacities cannot simply be summed and assumed to have the same conversion rate. One
industry consultant estimated that delayed cokers tend to convert 30 percent of their feedstock
into distillate, while fluidized cokers, visbreakers, and other thermal  crackers are less efficient in
this regard, converting only 15 percent.  Thus, we assumed that the conversion rate for other
thermal crackers was half that of cokers. Practically, we effected this assumption by discounting
the capacity of other thermal crackers by a factor of two before adding them to coking capacity.

    Prior to making this comparison, however, one more adjustment must be made.  Refiners
outside of California with hydrocrackers typically feed LCO and other cracked stocks to their
hydrocracker.  Straight run distillate might also be fed to a hydrocracker which produces
gasoline blendstock. However, we believe that after 2006, the 15 ppm highway diesel fuel cap
will encourage refiners to shift as much LCO and other cracked stocks as possible to their
hydrocrackers.  Thus, for refineries with hydrocrackers and FCC units, we assumed  that any
LCO produced would be sent to the hydrocracker, up to the  capacity of the hydrocracker.Y
Similarly, for refiners with hydrocrackers and cokers or other thermal crackers, we assumed that
any other cracked stocks produced would be sent to the hydrocracker, up to the capacity of the
   Y This assumes that both the FCC unit and the hydrocracker operate at the same percent of capacity, which is
reasonable.

                                          7-111

-------
Final Regulatory Support Document
hydrocracker minus any LCO sent to the hydrocracker. Table 7.2.1-16 summarizes this
information.

                                     Table 7.2.1-16
                       Conversion of Heavy Oils to Distillate in 1996

Total U.S. Refining
Capacity (BPD)
Total Distillate Blendstock
Produced (BPD)
Percentage of Capacity
Converted to Blendstock
FCC Units (LCO)
Total
After Shift to
Hydrocrackers
4,936,940
2,951,287
1,053,610
643,043
—
22%
Coking and other thermal crackers * (Other cracked stocks)
Total
After Shift to
Hydrocrackers
Hydrocracker
(hydrocrackate)
2,664,400
1,771,505
927,390
400,193
256,728
177,265
—
15%
19%
  100% of coker capacity plus 50% of the capacity of other thermal crackers
   By taking the ratio of the volume of distillate blendstock produced to the total capacity of the
type of equipment which produces it, we can estimate the percentage of this capacity which is
converted into each type of blendstocks. These percentage are also shown in Table 7.2.1-16. It
should be noted that these figures are likely lower than the conversions which would be actually
seen during unit operation. The conversions shown in Table 7.2.1-16 are based on rated unit
capacity and actual distillate production. Units typically operate at less than capacity over the
course of a year.  This utilization percentage does not need to be explicitly considered here as the
unit capacity for each refinery and that for the nation as a whole are both on a nameplate rating
basis. Use of a capacity utilization rate would  simply adjust both figures and cancel out within
the methodology.

   Since we know the capacity of the various unit in each refinery in 1996, we could estimate
the volume of each blendstock produced by each U.S. refinery in  1996 by multiplying these
capacities by the above conversion factors. However, many refineries have increased the
capacities of various units since 1996.  As we are using these blendstock compositions to project
desulfurization costs in 2007 and beyond, it would be desirable to reflect the impact of these
changes in capacity in our analysis. The latest data are from 2002. Thus, we multiplied each
refinery's 2002 unit capacities (per EIA) by the above conversion factors to estimate the volume
of each blendstock produced by each refinery in this year.

   This is a marked improvement from the NPRM analysis. In the NPRM, we used refinery
unit capacities existing in the year 2000 (as estimated in the Oil and Gas Journal). These 2000
                                         7-112

-------
                                                 Estimated Costs of Low-Sulfur Fuels
capacities were combined with the 1996 API/NPRA survey results and distillate production data
from 2000 to develop an analogous set of conversion factors. The use of 1996 unit capacities to
develop the conversion factors is more consistent with the survey results. The use of 2002 unit
capacities incorporates two additional years of changes in refinery configurations into the
analysis.

   We also decided to use unit capacities as estimated by EIA in lieu of those published by the
Oil and Gas Journal. Reviewing both sets of unit capacities, particularly that for hydrotreating
capacity used in Section 7.2.1.3.2 below, we found greater consistency between the production
volumes of various distillate fuels, as well as between the capacities of the various units, with the
EIA estimates than with those published by the Oil and Gas Journal.  Therefore, we decided to
use the EIA estimates for this final NRLM rule analysis.  Also, in the NPRM, the use of distillate
compositions from 1996 and unit capacities from 2000 was inconsistent to some degree and the
above methodology eliminates this problem.

   In addition, the use of 2002 unit capacities provides an automatic adjustment for changes in
refinery configurations from 1996 to 2002.  In the NPRM, our methodology basically assumed
that the overall distillate composition in 1996 continued unchanged into the future.  One of the
comments we received on the NPRM cost estimates was  that we had under-estimated
desulfurization costs by assuming that the 1996 distillate composition was not changing over
time.  The commenters pointed out that the average crude oil being processed in domestic
refineries was getting heavier (lower API gravity) and more  sour (higher sulfur) over time,
which would negatively affect distillate composition from the point of view of desulfurization.
They suggested that we should adjust our mix of blendstocks and the amount of sulfur needing to
be removed to account for this trend.

   We reviewed the quality of the U.S. crude oil slate between 1996 and 2002 and indeed found
that the API gravity of average crude oil had decreased by 2.3 percent from 31.1 to 30.4. (The
sulfur content of crude oil also increased, but this will be considered in Section 7.2.1.3.2 below
when we estimate the percentage of NRLM fuel which is hydrotreated prior to this rule.)
Heavier crude oils tend to produce heavier feedstocks to the FCC, coker and hydrocrackers,
which can affect the conversion  of these feedstocks into distillate.  The yield of LCO from an
FCC unit tends to vary inversely with conversion,2 with higher volumes of LCO produced at
lower conversion rates. Heavier crude oils generally produce a heavier FCC feed stock which
lowers FCC conversion. This would tend to increase the production of LCO from FCC units.
The same would be generally true for cokers and other thermal cracking units.

    However, since 1996 refiners have made several process changes which tend to increase
FCC conversion.  Since 1996, FCC feed hydrotreating capacity has increased by 24 percent,
while FCC capacity only increased by 6 percent.35 FCC feed hydrotreating reduces the density
(increases the API gravity) of the FCC feedstock, which increases conversions and decreases
   ZFCC conversion is defined as the volume percent of FCC feed throughput that is converted to products lighter
than LCO and clarified oil/slurry oil, ((FCC feed - LCO product-slurry oil product)/ FCC feed )* 100, per volume
basis.

                                         7-113

-------
Final Regulatory Support Document
LCO yields in the FCC unit.  Also, hydrocracking capacity has increased by 20 percent. Since
these units can process poor quality LCO, this mitigates the effect of heavier crude oils.
According to several FCC technology licensors, refiners are also using more active FCC
catalysts and have added or upgraded their FCC process technologies since 1996. These changes
should also increase FCC conversions and decreases LCO yields. Thus, changes have occurred
since 1996 which both increase and decrease the production of LCO from FCC units. It is not
possible to quantitatively estimate the impact of each of these changes, nor the net change in
LCO yield. In general, we believe that the impact of heavier crude oil is smaller than the impact
of newer FCC technology and increased FCC hydrotreating capacity. Thus, the inability to
quantitatively account for these changes should not lead to an under-estimation of
desulfurization costs. However, due to the compensating nature of these changes, we believe
that the overall change in the quantity and quality of LCO and other cracked stocks being
produced today is small and would not significantly affect desulfurization costs.

   Also, the processing of heavier crude oil has led the U.S. refining industry to increase
capacity of cokers and hydrocrackers relative to crude oil processing capacity.  As mentioned
above, our methodology  automatically adjusted distillate composition for this trend. Thus, we
believe that our current methodology reflects current crude oil quality as much as possible using
available information.  While our methodology does not account for future changes in crude oil
quality, the changes seen below between 1996 and 2002 are quite small and indicate that changes
likely in the future would also be very small.

   Table 7.2.1-17 shows how updating these estimates from 1996 to 2002 affected national
average distillate composition outside of California.

                                     Table 7.2.1-17
            National Average Distillate Composition Excluding California (Vol%)

Straight Run
LCO
Other Cracked Stocks
Hydrocrackate
1996
65%
21%
8%
6%
2002
62%
21%
10%
7%
   We made one last adjustment to distillate composition to reflect a shift we believe will occur
when the 15 ppm sulfur cap begins to apply to highway diesel fuel in 2006.  As shown in Table
7.2.1-17 above, the API/NPRA survey found that the hydrocrackate fraction of high sulfur
distillate was much greater than that in highway diesel fuel.  The reason for this is not obvious,
as the low sulfur level of hydrocrackate would presumably been valuable in producing 500 ppm
highway fuel.  It may be that most highway fuel has be hydrotreated regardless of the percentage
of hydrocrackate added, and the use of hydrocrackate in high sulfur distillate allows a significant
portion of this  fuel to avoid hydrotreating. In any event, the primary properties which differ
                                         7-114

-------
                                                 Estimated Costs of Low-Sulfur Fuels
between highway diesel fuel and high sulfur distillate are sulfur content and cetane number and
refiners can use a wide range of blendstock compositions to meet these specification.

   When the 15 ppm cap starts to apply to highway diesel fuel, however, the economic incentive
to blend hydrocrackate into highway diesel fuel will increase dramatically.  Thus, we believe that
refiners will shift hydrocrackate from high sulfur distillate to highway diesel fuel. However,
most high sulfur distillate is either NRLM diesel fuel or sold as either NRLM fuel or heating oil.
Thus, it must have a minimum cetane number of 40.  Therefore, we did not believe that it would
be feasible  for a refiner to shift unhydrotreated LCO or other cracked stocks from highway diesel
fuel to high sulfur distillate.  Therefore, we assumed that refiners would only shift hydrotreated
blendstocks to compensate for the hydrocrackate shift.  We assumed that the composition of this
shift would reflect the refinery's average distillate composition (i.e., percentage of straight run,
LCO and other cracked stocks). We assumed that a refiner would shift all of their hydrocrackate
to highway diesel fuel as long as there was sufficient hydrotreated material to shift from highway
fuel to high sulfur distillate.  (The hydrotreated fraction of each refinery's distillate is discussed
in the next  section.) For all except five refineries, all of the hydrocrackate was shifted to
highway fuel.  Three refiners lacked sufficient volume of hydrotreated blendstocks for all their
hydrocrackate to be shifted.  Two refiners produced less highway diesel fuel than their estimated
production  of hydrocrackate.  Overall, the hydrocrackate portion of highway diesel fuel
increased to 8.9 percent, while that for high sulfur distillate decreased to 1.6 percent.

   The final compositions of highway and high sulfur distillate after implementation of the 15
ppm sulfur  cap on highway fuel, but prior to this NRLM rule are shown below in Table 7.2.1-18.
These national averages were calculated by 1) applying the above conversion factors to each
refinery's unit capacities to estimate the volume of each blendstock being produced by that
refinery, 2) spreading the volume of each blendstock to the refinery's highway diesel fuel and
high sulfur  distillate fuel pools in proportion to the refinery's production of each of the two fuels
pool (as estimated in Section 7.2.3.3 below), 3) shifting hydrocrackate to highway fuel in return
for other hydtrotreated blendstocks, as  discussed above, 4) summing the volumes of each
blendstock  type in each fuel pool across all refineries and 5) dividing these blendstock volumes
by the total production of highway and high sulfur fuel, respectively. We used each refinery's
projected distillate composition to estimate its cost of meeting the 500 and 15 ppm NRLM sulfur
caps, not the national average composition.

                                      Table 7.2.1-18
    Distillate Composition: After Implementation of the  15 ppm Highway Fuel Sulfur Cap*

Straight Run
LCO
Other Cracked Stocks
Hydrocrackate
Highway Diesel Fuel
61%
20%
10%
9%
High Sulfur Distillate
66%
23%
9%
2%
All No. 2 Distillate
62%
21%
10%
7%
* excludes California.
                                          7-115

-------
Final Regulatory Support Document
   In order to provide an indication of the range of distillate compositions which we projected
using this methodology, we developed distributions of the percentages of LCO and other cracked
stocks in various refiners distillate. These are shown in Table 7.2.1-19 below.

                                      Table 7.2.1-19
   Distribution of LCO and Other Cracked Stocks in High Sulfur Distillate Prior to the NRLM
                   Rule (U.S. Refineries Producing High Sulfur Distillate)

Percentage of LCO and Other Cracked Stocks in the Distillate Pool
0%
<10%
<20%
<25%
<30%
<40%
<50%
<80%
100%
LCO
Number of Refineries
Cumulative % of High
Sulfur Distillate
Volume
47
35
48
36
53
45
60
49
76
71
92
87
96
94
99
98
101
100
Other Cracked Stocks
Number of Refineries
Cumulative % of High
Sulfur Distillate
Volume
71
53
73
61
79
66
87
85
92
88
97
90
101
100
101
100
101
100
   As shown above, in 2002, high sulfur distillate fuel produced by U.S. refineries contains
between zero to over 80 percent LCO. Forty-seven U.S. refineries, which produce about 35
percent of the high sulfur distillate in the U.S., blend no LCO into their distillate.  The high
sulfur distillate from the remaining 54 refineries averages about 33 percent LCO by volume.  On
average, high sulfur distillate contains 21.1 percent LCO in 2002 versus 21.3 percent in 1996.
This reflects the fact that FCC unit capacity grew slightly less between 1996 and 2002 than total
domestic distillate production volume.

   Similarly, we  estimate that about half of the high sulfur distillate fuel in the U.S, which is
produced by 71 refineries, does not contain  any other cracked stocks from cokers,  visbreakers
and thermal crackers.  Of the refineries which produce other cracked stocks, their distillate fuel
contains an average of 20.0 percent of other cracked stocks in 2002. On average, the estimated
percentage of other cracked stocks being blended into high sulfur distillate increased slightly
from 9.2 percent in 1996 to 9.4 percent in 2002.  Thus, coking capacity increased slightly faster
than total distillate production.

   7.2.1.3.2 Sulfur Content and Hydrotreated Fraction of High Sulfur Distillate

   Like distillate composition, per the cost  methodology developed above, the sulfur content
and hydrotreated fraction of high sulfur distillate affects the cost of desulfurization.  There are
two effects.  One relates to the amount of hydrogen consumed in hydrotreating.  The other
relates to the capital cost of a hydrotreater.

   Regarding hydrogen consumption, in addition to removing sulfur,  hydrotreating  also
saturates olefins and most poly-nuclear aromatics. These latter effects occur almost regardless of
                                          7-116

-------
                                                 Estimated Costs of Low-Sulfur Fuels
the degree of sulfur reduction.  Thus, distillate which is being hydrotreated today has already had
its olefins and poly-nuclear aromatics removed. Thus, subsequent hydrotreating of already
hydrotreated blendstocks to reduce sulfur further in response to this NRLM rule does not
consume hydrogen related to olefin or poly-nuclear aromatic saturation. The other effect relates
to the capital  investment needed to meet the 500 ppm NRLM cap in 2007. Material that is
already being hydrotreated to 500 ppm or less need not be treated at all during the first step of
the NRLM fuel program.

   As mentioned in Section  7.2.1.2.1.2, we were not able to incorporate the change in hydrogen
consumption  due to olefin and poly-nuclear aromatic saturation associated with changing
degrees of current hydrotreating.  Differences in total hydrogen consumption between various
refineries should only be a few tenths of a penny per gallon.  Thus, the use of an average level of
olefin and poly-nuclear aromatic  saturation lessened the refinery-specific nature of our estimates
to a slight degree.

   Regarding capital costs, we were able to incorporate differences in expected capital
investment needed to desulfurize unhydrotreated and hydrotreated blendstocks to meet the 2007
500 ppm NRLM cap. This improved our ability to predict overall desulfurization costs, the
number of refineries affected by the NRLM rule and how small refiners might be differentially
impacted by the rule.

   In addition to whether a blendstock has been previously hydrotreated or not, the starting
sulfur content also affects the volume of hydrogen  needed to reduce sulfur to meet a 500 ppm
cap.  In the NPRM, we  started with the 1996 API/NPRA fuel quality survey to obtain estimates
of the portion of highway and high sulfur distillate which receives at least some hydrotreating.
We then used in-use fuel survey data to estimate the sulfur level of high sulfur distillate
produced in 1996.  Assuming that the sulfur content of the hydrotreated portion of this fuel was
the same as that for highway diesel fuel (340 ppm), we then back-calculated the sulfur content of
the non-hydrotreated portion of high sulfur distillate,  so that the blend matched the in-use sulfur
level of finished high sulfur distillate. We then assumed that these 1996 estimates also applied
to current and future high sulfur distillate prior to the NRLM rule.

   We received comment on the NPRM that the sulfur content of crude oil had been increasing
since the 1996 API/NPRA survey was conducted.  The commenters argued that this would
increase the sulfur content of high sulfur distillate and increase desulfurization costs.  Therefore,
we have expanded the methodology used in the NPRM analysis to estimate both the sulfur
content and hydrotreated fraction of high sulfur distillate.

   We first reviewed data on the sulfur content of crude oils processed by U.S. refineries and
found that sulfur content had indeed increased.  We have incorporated this increase in crude oil
sulfur content into the estimates developed in this section. However, as described in Section 7.1
above, there is no evidence so suggest that the sulfur content of high sulfur distillate has
increased since 1996. Thus,  it is likely that a greater percentage of the volume of high sulfur
distillate blendstocks are being hydrotreating than was the case in 1996. We have incorporated a
change in the hydrotreated fraction from 1996 into this analysis, as well. Finally, we also

                                          7-117

-------
Final Regulatory Support Document
reviewed the hydrotreating and hydrocracking capacities of U.S. refineries in 1996 and 2002, as
well as the relative production of highway diesel fuel and high sulfur distillate to confirm that
sufficient hydrotreating capacity exists to hydrotreat a greater fraction of high sulfur distillate
blendstocks.

    Table 7.2.1-20 presents many of the primary inputs for our analysis. These estimates are
intended to represent high sulfur distillate produced in the year 2002, but without consideration
of an increase in crude oil sulfur content.  Due to the significant differences in hydrotreating
percentages seen across PADDs, we incorporated these PADD-specific estimates as much as
possible.

                                      Table 7.2.1-20
                           Quality of High Sulfur Distillate from
     Non-California Refineries: "2002" Prior to Consideration of Increased Crude Oil Sulfur


PADD
1
2
3
4
5
High Sulfur Distillate Pool
Sulfur content (ppm)
% Hydrotreated *
2925
27
2973
31
3776
44
2549
17
2566
2
High Sulfur Distillate Produced by Refineries with Hydrotreaters
% of high sulfur distillate pool
% Hydrotreated
Sulfur content of portion not
hydrotreated (ppm)
81
33
4214
70
45
5081
95
46
6739
40
43
4237
48
4
2646
 : Assumed to be the same as in 1996 API/NPRA survey.
    The sulfur content of the high sulfur distillate pool in each PADD were taken from Table 7.1-
40 in Section 7.1 above. A direct estimate of the portion of the 2002 distillate pool which is
hydrotreated is not available. Therefore, we assumed that this figure has not changed since the
API/NPRA survey.  This necessitates the consideration of increased sulfur content between 1996
and 2002, which is addressed below.  As can be seen, a significant percentage of high sulfur
distillate received some hydrotreating in 1996, despite the fact that the final sulfur level is 2000
ppm or more.  This is likely necessary to improve the stability of untreated LCO, as well as meet
applicable cetane and sulfur specifications with blend stocks which can exceed 10,000 ppm
sulfur and have a cetane number of less than 15 prior to hydrotreating.  The PADD with the
highest percentage of hydrotreated high sulfur distillate is PADD 3, while the lowest is PADD 5
(outside of California). Within PADD 5, Alaska's refineries are believed to have the lowest
hydrotreated percentage (zero),  since none of the Alaskan refineries have distillate hydrotreaters.

    The hydrotreated blendstocks sent to the high sulfur distillate pool are assumed to be part of
a larger pool of hydrotreated blendstocks also used to produce highway diesel fuel. We believe
that this is reasonable because many refiners likely only have a single hydrotreater and they are
simply blending more hydrotreated material into their highway diesel fuel than into their high
                                          7-118

-------
                                                  Estimated Costs of Low-Sulfur Fuels
sulfur distillate.  In this case, we assume that all of the hydrotreated material contains 340 ppm
sulfur, the current average sulfur level for highway diesel fuel.  Some larger refiners likely have
two or more hydrotreaters which could be treating highway diesel fuel blendstocks and high
sulfur distillate blendstocks differently. However, in this case, we have no way of estimating the
sulfur levels of either the hydrotreated or non-hydrotreated portions of the high sulfur distillate.
Thus, we assumed that the 340 ppm sulfur content applied to all hydrotreated blendstocks.
Overall, this assumption has little effect on the estimation of NRLM desulfurization costs. As
will be seen below, we have estimates of both the hydrotreated fraction of high sulfur distillate
and of its final sulfur level. If the sulfur level of hydrotreated blendstocks going to the high
sulfur distillate pool contain more than 340 ppm sulfur, the the sulfur content of the non-
hydrotreated portion of the pool much contain less sulfur than estimated below. The total
amount of sulfur requiring removal is the same in either case.

   Some refiners do not have a distillate hydrotreater. Therefore, the percentage  of their high
sulfur distillate which is hydrotreated is zero.  In order for the entire high sulfur distillate pool to
be hydrotreated to the degree  shown in Table 7.2.1-17, the portion of distillate produced by
refiners with distillate hydrotreaters must be higher.  In order to estimate these percentages, we
reviewed EIA data for both distillate production and distillate hydrotreating capacity. The
former data are confidential and were received directly from EIA. The latter came from their
2002 Petroleum Supply Annual. For each  PADD, we determined the percentage of all high
sulfur distillate produced by refiners with distillate hydrotreaters.  These figures are shown in
Table 7.2.1-20 above. We calculated the percentage of the high sulfur distillate pool produced
by refineries with hydrotreaters by dividing the hydrotreated percentage for the entire pool by
the percentage of distillate produced by refineries with hydrotreaters.  These higher hydrotreated
percentages are shown on the second to the last line of Table 7.2.1-20.

   As discussed above, we assume that the sulfur content of the hydrotreated portion of high
sulfur distillate is the same as that of highway diesel fuel, or 340 ppm.  As discussed in Chapter
5, the sulfur content of hydrocrackate is very low,  less than 50 ppm. Knowing the final sulfur
level and the percentage of hydrotreated blendstock in high sulfur distillate from Table 7.2.1-20
above (which includes hydrocrackate) and the percentage of hydrocrackate from Table 7.2.1-18,
we can back-calculate the sulfur content of the unhydrotreated blendstocks comprising the rest of
the high sulfur distillate pool. These sulfur levels are also shown in Table 7.2.1-20.

   The final step is to incorporate the effect of an increase in crude oil sulfur content. Table
7.2.1-21 shows the average sulfur content of crude oil processed in each PADD in both 1996 and
2002.  As can be seen, crude oil became more sour in all but PADD 1.
                                          7-119

-------
Final Regulatory Support Document
                                       Table 7.2.1-21
             Sulfur Content of Crude Oil Processed by U.S. Refineries (weight %)
PADD
1
2
3
4
5 (Non-California)
Overall
1996
0.94
1.08
1.22
1.31
1.14
1.15
2002
0.86
1.31
1.65
1.40
1.22
1.41
Percent Change
-8.5
21.3
35.3
6.9
7.0
22.6
 : Annual crude properties from EIA's Petroleum Supply Annual 1996 and 2002
   We next used published information to estimate how changes in crude oil sulfur content
would impact the sulfur level of unhydrotreated distillate blendstocks.AA Table 7.2.1-22 depicts
estimated sulfur contents for straight run distillate for a variety of crude oils containing both 1.15
and 1.41 weight percent sulfur.

                                       Table 7.2.1-22
                    Straight Run Middle Distillate Sulfur Content (ppm) *
Crude Oil
Sulfur Content

1.15wt%
1.41 wt%
Change in
Distillate
Sulfur
Sweet U.S.
Crude Oil

4400
5400
22.7%

West Texas
Crude Oil

6400
7800
21.9%

California
Crude Oil

7800
9800
25.6%

Middle East
Crude Oil

4500
5300
17.7%

Venezuelan
Crude Oil

3500
4400
25.7%

Average of
All Crude
Oils
5330
6540
22.7%

 : Middle distillate assumed to have mid-boiling point of 500 F.
    As can be seen, the 22.6 percent increase in crude oil sulfur content is estimated to increase
the sulfur content of straight run distillate by 17.7-25.7 percent, with an average increase of 22.7
percent.  Thus, on average, the sulfur content of straight run distillate increases to essentially the
same degree as that of the crude oil. Therefore,  it is reasonable to assume that the increases in
crude oil sulfur content shown in Table 7.2.1-21 above increased the sulfur content of straight
run distillate proportionally. In addition, we assume that the sulfur content of the other
blendstocks, namely LCO and other cracked stocks, also increased to the same  degree.

    As discussed in Section 7.1 above, the average sulfur content of high sulfur distillate does
not appear to have changed substantially since 1996.  A significant portion of this distillate is
      Petroleum Refining Fourth Edition, Gary Handewerk, 2001, pages 41 to 45.

                                           7-120

-------
                                                  Estimated Costs of Low-Sulfur Fuels
produced by refineries without distillate hydrotreating, where an increase in crude oil sulfur
would by necessity have been reflected in their distillate production.  This implies that the
increases in crude oil sulfur content occurred primarily at refineries with distillate hydrotreating
capacity. To account for this, we adjusted the changes in crude oil sulfur shown for the
percentage of high sulfur distillate produced by refiners with hydrotreaters. For example, crude
oil sulfur in PADD 2 increased by 21.3 percent.  Of all the distillate produced in PADD 2, 70
percent was produced by refineries with distillate hydrotreaters.  Therefore, if the crude oil sulfur
at the refineries producing the other 30 percent of high sulfur distillate did not change, the crude
oil sulfur at refineries with hydrotreaters increased by 30 percent (21.3/0.7). The results for all
five PADDs are shown in Table 7.2.1-23 below.

                                       Table 7.2.1-23
      Quality of High Sulfur Distillate from Non-California Refineries: 2002 and Beyond

PADD
1
2
3
4
5
High Sulfur Distillate Pool
Sulfur content (ppm)
% Hydrotreated
2925
20
2973
41
3776
58
2549
21
2566
83
High Sulfur Distillate Produced by Refineries with Hydrotreaters
Increase in crude oil sulfur content
% of high sulfur distillate pool
% Hydrotreated
Sulfur content of portion not
hydrotreated (ppm)
-11%
81
25
3771
30%
70
58
6623
37%
95
61
9248
17%
40
52
4964
15%
48
17
3034
    The next step was to increase the sulfur content of the unhydrotreated distillate at refineries
with hydrotreaters by the same percentage that crude oil sulfur increased. For example, in
PADD 2, the sulfur content of 5081 ppm was increased by 30 percent to yield a final non-
hydrotreated distillate sulfur content of 6623 ppm. The sulfur content of the 2002 high sulfur
distillate is the same as that shown in Table 7.2.1-23 and the sulfur content of the hydrotreated
distillate is 340 ppm.  Therefore, the percentage of high sulfur distillate at these refineries which
is hydrotreated can be calculated. For example, in PADD 2, a mix of 42 percent hydrotreated
distillate at 340 ppm and 58 percent unhydrotreated distillate at 6623 produces a pool of high
sulfur distillate at 2973 ppm.  Finally, given the percent of all high sulfur distillate being
produced by refineries with hydrotreaters (for PADD 2, 70 percent), the portion of the entire
high sulfur distillate pool which is hydrotreated can be calculated.  For example, for PADD 2,
the portion of the entire high sulfur distillate pool which is hydrotreated is 41 percent, the
product of the the percent of all high sulfur distillate being produced by refineries with
hydrotreaters (70 percent) and the hydrotreated percentage of high sulfur distillate at those
refineries with hydrotreaters (58 percent). These figures  are summarized in Table 7.2.1-23
above.
                                           7-121

-------
Final Regulatory Support Document
   High sulfur distillate produced by refineries without hydrotreaters is assumed to have sulfur
contents equal to the average high sulfur distillate produced in that PADD.  High sulfur distillate
produced by refineries with hydrotreaters is a mix of unhydrotreated blendstocks at the sulfur
levels shown in Table 7.2.1-23 and hydrotreated blendstock containing 340 ppm sulfur.  The
average sulfur content of this distillate is also the average sulfur content of the high sulfur
distillate produced in that PADD. We assume that these hydrotreated percentages and sulfur
contents remain constant beyond 2002.

   A comparison of the hydrotreated portion of all high sulfur distillate in 1996 (Table 7.2.1-20)
and 2002 (Table 7.2.1-23) shows that except in PADD 1, we are projecting that a significant
increase in the degree of hydrotreating has occurred. This implies that refiners built new
hydrotreaters or expanded existing hydrotreaters during this time period. We desired to confirm
that this in fact occured. The first step in this confirmation was to estimate the increased
capacity of distillate hydrotreating. The second step was to show that this increase was sufficient
to provide for the increased production of highway  diesel fuel, as  well as the increase in the
hydrotreated percentage of high sulfur distillate.

   Table 7.2.1-24  presents hydrotreating and hydrocracking capacity at U.S. refineries located
outside of California in 1996 and 2002, according to EIA's Petroleum Supply Annual reports
from these two years (assuming an annual average utilization rate of 90 percent). Both processes
produce distillate blendstocks which likely meet the 500 ppm highway fuel cap  and which have
had their olefins and some aromatics removed,  reducing the cost of further hydrotreating. As
described above, hydrocrackers are assumed to convert roughly 21 percent of their feed to
distillate.

                                      Table 7.2.1-24
   Effective Non-California Distillate Hydrotreating and Hydrocracker Capacity 1996 to 2002

1996 Capacity
2002 Capacity
Increase in capacity
Increase in low sulfur distillate
Distillate Hydrotreating
3,108,285
3,380,323
272,038
272,038
Hydrocrackers
834,651
1,003,050
168,399
35,869*
  90 percent of rated capacity. Hydrocrackers assumed to convert 21 percent of feedstock to distillate.
   As can be seen, the total capacities of both processes increased substantially. In total, these
capacity expansions increased the production capacity of low sulfur distillate by 307,900 barrels
per day.

   Table 7.2.1-25 shows the distillate fuel production in 1996 and 2002, again from EIA's
Petroleum Supply Annual reports. We show the production of jet fuel and kerosene, since much
                                          7-122

-------
                                                 Estimated Costs of Low-Sulfur Fuels
of the volume of these No. 1 distillate fuels is also hydrotreated and the above distillate
hydrotreating capacities do not distinguish between No. 1 and No. 2 distillates.

                                      Table 7.2.1-25
                        Non-California Distillate Production (BPD)

1996
2002
Increase
Jet Fuel and Kerosene *
1,577,000
1,571,000
-6,000
Highway Diesel Fuel
1,842,797
2,298,507
455,710
High Sulfur Distillate
1,213,490
964,184
-249,307
* Jet fuel includes production from California refineries.
    As can be seen, the production of jet fuel and kerosene was essentially constant in 1996 and
2002.  Thus, we assume that no additional hydrotreating capacity was used in the production of
jet fuel and kerosene in 2002 versus 1996.  It is possible that the increased sulfur content of
crude oil occurring over this 6 year period caused refiners to increase a greater percentage of the
No. 1 distillate blendstocks used to produce these two fuels. However, no data are available to
estimate this effect. Since the sulfur standards for these No.l distillate fuels are not stringent, the
overall change in hydrotreating should be small.

    As also shown in Table 7.2.1-25, the production of highway diesel fuel increased by nearly
25 percent, while the production of high sulfur distillate decreased by 20 percent. As described
above, the hydrotreated fraction of highway fuel was 83.8 percent in 1996. Thus, the production
of 455,710 barrels per day more highway diesel fuel likely utilized 382,000 (455,710 * 0.838)
barrels per day of effective hydrotreating or hydrocracking  capacity. However, as discussed
below, crude oil sulfur levels increased between 1996 and 2002 by nearly 20 percent. Thus, to
be conservative, we will also consider the possibility that 100 percent of this additional
production of highway diesel fuel was hydrotreated.  Thus,  we estimate that the production of
455,710 barrels per day more highway diesel fuel might have utilized as much as 455,710 barrels
per day of effective hydrotreating or hydrocracking capacity.  Combining these two estimates to
produce a range, the additional production of highway diesel fuel utilized 74,100-147,810 more
barrels per day of effective hydrotreating and hydrocracking capacity than the 307,000 barrels
per day of effective capacity which was added between 1996 and 2002.

    Regarding the production of high sulfur distillate, two factors changed, volume  and
percentage which was hydrotreated. In 1996, 1.213 million BPD of high sulfur distillate was
produced, 34 percent of which was hydrotreated.  In 2002, 0.964 million BPD of high sulfur
distillate was produced, 41  percent of which was hydrotreated. This implies  a net reduction of
hydrotreated volume of 20,300 BPD. This provides some but not  all of the hydrotreating
capacity needed to produce the additional highway fuel. The shortfall ranges from  53,800-
127,510 barrels per day of effective hydrotreating capacity.
                                          7-123

-------
Final Regulatory Support Document
   We believe that this remaining hydrotreating capacity needed to produce the additional
highway diesel fuel likely came from an increase in the utilization of hydrotreating capacity
between 1996  and 2002. The API/NPRA survey showed that only 78 percent of the total rated
hydrotreating capacity was utilized in 1996. We believe that full utilization can be closer to 90
percent. (Crude oil utilization rates today are over 95 percent.) A 12 percent increase in the
utilization rate of hydrotreating capacity in  1996 would be 373,000 barrel per day.  This far
exceeds the 53,800-127,510 barrel per day shortfall estimated above.  Thus, we conclude that the
increase in overall hydrotreating percentage of high sulfur distillate are reasonable.

   7.2.1.3.3 Refinery Specific Distillate Production Volumes

   In the NPRM, we projected refinery's volumes of no. 2 distillate fuel in two steps. First, we
obtained each refinery's production of no. 2 distillate fuel in 2000 from EIA. (This data is
considered confidential and is based on information which refiners are required to submit to EIA
periodically.)  These production volumes include a breakdown of how much fuel was certified to
meet the 500 ppm highway fuel sulfur cap and how much fuel was not so certified. Second,
these year 2000 production volumes were increased to represent 2008 production using EIA
projections from their 2002 AEO report.  We applied separate growth rates for highway diesel
fuel and high sulfur distillate. We assumed that refineries would not change their relative
production of highway diesel fuel and high  sulfur distillate except as reflected in the distinct
national average growth projections for the  two fuels.

   For the final rule, we have made a number of changes to improve this portion of our cost
analysis. First, since the NPRM analysis was conducted, we received refiners' projection of the
volume  of 15 and 500 ppm highway diesel fuel which they plan to produce in 2006-2010. In
some cases, these volumes differ  significantly from their historic production of highway diesel
fuel. Thus, we have incorporated these projections into our projection of refineries' relative
production of highway diesel fuel and high  sulfur distillate prior to the implementation of this
rule. Second, we have shifted our base year for historic production volumes from 2000 to 2002
to reflect more recent data available from EIA.  Third, we have shifted the future year for which
we project desulfurization costs from 2008 to 2014. Fourth, and finally, we are using EIA
projections of distillate production growth from their 2003 AEO report36, instead of their 2002
AEO report. The methodology for estimating refinery specific production volumes of highway
diesel fuel and high sulfur distillate is described in more detail below, as well as the results of
this analysis.

   As described above, the first step was to estimate each refinery's historic production volumes
of highway diesel fuel and high sulfur distillate.  Except for using more recent 2002 data from
EIA, versus 2000 in the NPRM, this step was identical to that performed in the NPRM analysis.

   The second step increased these 2002 production volumes of highway and high sulfur
distillate fuel to represent growth through 2014.  We chose 2014, because it represents the mid-
point of the life of the desulfurization equipment build in response to this rule (per IRS rules, this
equipment has a  15 year life). We obtained EIA's projected growth factors for domestic
production of these two fuels over this time period, which were consistent with those underlying

                                          7-124

-------
                                                 Estimated Costs of Low-Sulfur Fuels
their 2003 AEO projections. EIA projects that highway fuel production will increase 42.1
percent over this time period, while production of high sulfur distillate will only increase 8.1
percent. Each refinery's 2002 production volumes of these two fuels werw increases by these
percentages to represent their likely production in 2014.  The sum of the production volumes for
the two fuels was taken to be each refinery's total distillate production in 2014.  It should be
noted that the combination of these two growth rates results in a greater increase in the
production of distillate fuel from domestic refineries than indicated by the growth in crude oil
consumption by these refineries (typically assumed to be the driver of increased fuel production).
This difference occurs because EIA projects that domestic refiners will increasingly process
heavy oils in addition to virgin crude oils. This step was analogous to that performed in the
NPRM, with the exception that growth was projected to 2014 instead  of 2008.  The historic and
future production volumes by PADD are shown in  Table 7.2.1-26.

                                      Table 7.2.1-26
                    U.S. Distillate Fuel Production: AEO 2003 (BPSD) *

PADDl
PADD 2
PADD 3
PADD 4
PADD 5
Total
2002
Highway
Fuel
239,375
647,170
1,245,605
129,397
396,475
2,658,022
High Sulfur
Distillate
223,063
159,688
520,142
29,973
95,775
1,028,641
Total
Distillate
462,438
806,858
1,765,747
159,370
492,250
3,686,663
2014
Highway
Fuel
337,936
913,637
1,758,473
182,676
559,720
3,752,442
High Sulfur
Distillate
241,161
172,644
562,345
32,404
103,546
1,112,100
Total
Distillate
579,098
1,086,281
2,320,818
215,080
663,266
4,864,542
* Growth from AEO 2003 Table 17. Includes U.S. Virgin Island refineries.
    The third step differed from the NPRM analysis in that we utilized refiners' confidential
projections of how they planned to produce highway diesel fuel in 2006-2010 under the
upcoming 2007 highway diesel fuel program. Under this program, refiners must submit their
projected production volumes of 15 and 500 ppm diesel fuel to EPA every year starting in 2003
(called a pre-compliance report).  EPA would then publish aggregated results to help refiners
optimize their compliance plans and better ensure sufficient supply of highway diesel fuel under
the rule. Shell oil's refinery in Bakersfield, California and Carribean Petroleum's refinery in
Puerto Rico were removed from the analysis due to recent  shutdowns  or plans to shut down.

    The highway diesel fuel program begins to take effect in June 2006. Some refiners
submitted 2006 production volumes on an annualized basis, while others submitted volumes for
just the  seven months affected by the program.  To avoid these differences, we focused on
refiners' projections for 2007, the first full calendar year affected by the program. We assumed
these projections, made by refiners, represented the best estimate of future production levels of
                                          7-125

-------
Final Regulatory Support Document
highway diesel fuel on a refinery-specific basis.  While refiners projected their production
volumes for highway diesel fuel, they did not have to submit their plans for producing high
sulfur distillate. Therefore, we estimated their production of high sulfur distillate subtracting
their production of highway diesel fuel from our estimate of the refinery'  total production of No.
2 distillate from step two above.

   The fourth and final step was to put refiner's projected 2007 highway diesel fuel production
volumes on the same basis as these 2014 total distillate volumes in order to back-calculate a high
sulfur distillate volume. To do this, we assumed that the refiners' highway pre-compliance
reports represented the absolute volumes which they planned to produce in 2007 including any
increases in total distillate production which might occur due to refinery debottlenecking, new or
expanded heavy oil processing capacity,  etc. Using information supplied in a number of these
reports, it appeared that some refiners simply estimated their 2007 production volumes by
applying some fraction to their historical 2002 production volumes. However, it is possible that
other refiners did include such planned capacity increases.  Overall, our methodology could
under-estimate highway fuel production in 2007 to some degree, but we believe that the degree
of this under-estimation should be small. We then increased these 2007 highway fuel production
volumes by EIA's projected increase in total domestic highway diesel fuel production between
2007 and 2014, which is 14.5 percent

   We then compared the total projected production of highway diesel fuel in 2007 in each
PADD to the projected demand for highway diesel fuel developed in section 7.1 above. Again,
in both cases, the volumes are representative of those expected for 2014.  The highway diesel
fuel sulfur standards are those representative of 2007 prior to this NRLM rule. Production and
demand for PADDs 1 and 3 were combined, due to the large volume of fuel which PADD 3
refiners ship to PADD 1. The results are shown in Table 7'.2.1-27'.
                                         7-126

-------
                                                Estimated Costs of Low-Sulfur Fuels
                                     Table 7.2.1-27
           Projected Production of Highway Fuel in 2007 (Thousand BPD in 2014)

Required Highway Fuel Production *
Projected Production: 15 ppm Highway Fuel
Projected Production: 500 ppm Highway Fuel
Projected Production: All Highway Fuel
Shortfall
PADD's 1 & 3
1,588.3
1,878.0
62.5
1940.5
-352.2
PADD2
1,162.4
914.8
49.5
964.3
198.1
PADD4
187.5
148.4
4.1
152.5
35.0
PADD5
530.9
468.2
20.3
488.5
42.4
Additional Production of Highway Fuel
Current highway fuel refiners with excess 500 ppm
capacity
1 5 ppm highway fuel produced from high sulfur distillate
Final 1 5 ppm Highway Fuel Production
Final 500 ppm Highway Fuel Production
Final Total Highway Fuel Production
0
0
1,723.9
62.5
1,786.4
0
0
914.8
49.5
964.3
0
41.8(4)**
190.2
4.1
194.3
2.2(1)
40.5 (4)
508.7
22.5
531.2
* Demand from highway vehicles, spillover of highway fuel to other markets plus highway fuel lost during distribution.
** Number of refineries producing this fuel is shown in parenthesis.
   As can be seen, projected 2007 production of highway diesel fuel in PADDs 1 and 3
significantly exceeds projected demand, while the opposite is true in PADDs 2, 4 and 5.  PADD
3 refiners currently supply much of PADD 2's diesel fuel consumption.  A comparison of current
shipments from PADD 3 to PADD 2 shows that these shipments far exceed the 198,000 barrel
per day shortfall projected for PADD 2. Therefore, we assumed that PADD 3 refineries would
balance demand for highway fuel in PADD 2. However, PADD 3  currently supplies little or no
fuel to PADDs 4 and 5.  Therefore, we assumed that additional refineries would have to produce
highway diesel fuel in 2007 to satisfy demand.  A comparison of 2002 production of highway
diesel fuel and refiners' projected production in 2007 revealed one refinery in PADD 5 which
had excess capacity to produce 500 ppm diesel fuel using its current hydrotreater.  Therefore, we
assumed that this refinery would likely produce 500 ppm highway diesel fuel in 2007 by
purchasing credits from other refiners. We projected that the remaining shortfalls would be
made up by refiners constructing new desulfurization capacity to process high  sulfur distillate to
15 ppm. We assumed that these refineries would go straight to 15  ppm for two reasons.  First, as
long as they  were investing to produce highway diesel fuel, they would likely design their
equipment to meet the 15 ppm cap, which would affect all highway fuel in 2010.  Second,
whether or not these refiners invested to produce 500 ppm highway diesel fuel in 2006 and
revamped this equipment in 2010 to produce 15 ppm highway diesel fuel has no effect on the
cost of other refiners producing NRLM fuel under this NRLM fuel rule.  It was simpler to
assume these refiners invested in one step rather than two.
                                         7-127

-------
Final Regulatory Support Document
   This left an excess highway fuel production of 154,100 barrels per day in PADDs 1 and 3
beyond that necessary to meet the shortfall in PADD 2. We assumed that refiners would adjust
their plans to produce 15 ppm highway diesel fuel in 2007 based on the results of the refiners'
pre-compliance reports.  Therefore, we assumed that this excess production would not in fact
occur.  To represent this on a refinery specific basis, we assumed that the refiners estimated to
have the highest cost  of producing 15 ppm fuel in PADDs 1 and 3 would decide not to produce
this fuel until the 154,100 barrel per day excess was eliminated. We also assumed that this
excess production capacity would be available to produce 500 ppm NRLM fuel in 2007 with
only incremental operation costs, no capital cost. This would be the case for excess 15 ppm fuel
capacity deriving from a revamp of an existing hydrotreater.  However, it would not be the case
for grass roots 15 ppm fuel capacity which never was built.  Thus, this assumption might have
led to a slight underestimation of the cost of 500 ppm NRLM fuel from 2007-2010. We believe
that the degree of this underestimation is small.

   Having developed refinery-specific projections of both total and highway distillate
production, we assumed that the difference was high sulfur distillate. The resulting total
production volumes for 2007 (projected to year 2014) by PADD and for the nation are shown in
Table 7.2.1-28.

                                     Table 7.2.1-28
          "2007" Refiner's Production of Distillate Fuels (Thousand BPD in 2014) *
PADD
1&3
2
4
5
Total
Highway Fuel
1,786
964
194
531
3,476
High Sulfur Distillate
1,116
122
21
132
1,391
Total Distillate
2,903
1,086
215
663
4,867
* Growth from AEO 2003 Table 17. Includes U.S. Virgin Island refineries.
   We repeated this analysis using refiners' projections of their production of highway diesel
fuel in 2010.  One limitation in doing so is that the refiners' pre-compliance reports for 2010
only apply to the first half of 2010 when they can still use banked credits to produce some 500
ppm highway fuel. We are more interested here in the last half of 2010, when all highway fuel
must meet a 15 ppm cap and NRLM fuel will also have to meet a 15 ppm cap under the final
NRLM program. To accommodate this difference, we assumed that refiners would simply
continue producing 15 ppm fuel at the same rate as they did in the first half of 2010. We also
assumed that refiners would convert production of 500 ppm highway fuel to high sulfur distillate
starting  on June 1, 2010 absent the NRLM fuel standards contained in this rule.
                                         7-128

-------
                                                Estimated Costs of Low-Sulfur Fuels
   As was done for the 2007 projections, we then increased these 2010 highway fuel production
volumes by EIA's projected increase in total domestic highway diesel fuel production between
2010 and 2014, which is 11.0 percent.  The results are shown in Table 7.2.1-29 below.

                                     Table 7.2.1-29
    Projected Production and Demand for Highway Fuel in 2010 (Thousand BPD in 2014)

Required Highway Fuel Production *
Projected 15 ppm Highway Fuel Production
Shortfall
Additional Production of 1 5 ppm Highway Fuel
Produced from high sulfur distillate
Final Production of 1 5 ppm Highway Fuel
PADD's 1 & 3
1,651.9
2008.3
-356.4


1942.4
PADD2
1,205.3
959.5
245.8


914.8
PADD4
194.2
153.7
40.6

41.8(4)**
195.5
PADD5
567.2
474.1
93.2

93.2 (7)
567.3
* Demand from highway vehicles, spillover of highway fuel to other markets plus highway fuel lost during distribution.
** Number of refineries producing this fuel is shown in parenthesis.
   As for 2007, the projected volume of highway diesel fuel in 2010 by PADD 1 and 3 refiners
exceeds projected demand (plus downgrades in the distribution system), while those of the other
PADDs are less than projected demand. In PADDs 4 and 5, we again assumed that additional
refineries would produce 15 ppm highway diesel fuel from their high sulfur distillate.  The
number of PADD 4 refiners was the same as in 2007. In PADD 5, seven additional refineries
were assumed to  produce 15 ppm highway diesel fuel, three more than in 2007.

   PADD 2's shortfall was again assumed to be supplied from PADD 3. Again, we assumed
that a number of PADD 1 and 3 refiners would decide not to produce 15 ppm highway fuel so
that these PADD's production would match demand, after supplanting PADD 2's supply. In
doing this, we also assumed that one PADD 2 refinery would decide not to produce 15 ppm
highway fuel due its much higher desulfurization costs compared to other PADD 2 refineries and
PADD 3 refineries able to supply that area via pipeline transport.

   Having the refinery-specific projections of both total and highway distillate production, we
assumed that the  difference was high sulfur  distillate. The resulting total production volumes for
2010 (grown to year 2014) by PADD and for the nation are shown in Table 7.2.1-30 below.
                                         7-129

-------
Final Regulatory Support Document
                                     Table 7.2.1-30
      "2010" Refiner's Production of Distillate Fuels Projected (Thousand BPD in 2014)

PADD's 1&3
PADD2
PADD4
PADD5
Total
Highway Fuel
1,942
915
196
567
3,620
High Sulfur Distillate
960
172
20
96
1,247
Total Distillate
2,903
1,086
215
663
4,867
* Growth from AEO 2003 Table 17. Includes U.S. Virgin Island refineries.
   Note that we made no changes in the production volumes of distillate fuel to account for any
reduction in wintertime blending of kerosene that might occur as a result of the 15 ppm highway
or NRLM sulfur caps. Kerosene added to 15 ppm diesel fuel must itself meet a 15 ppm sulfur.
Sometimes, kerosene is added at the refinery and the winterized diesel fuel is sold or shipped
directly from the refinery. At other times, the kerosene blending is done at the terminal,
downstream of the refinery.  The former approach may mean adding kerosene to more diesel fuel
than actually requires it. The latter approach requires that a distinct 15 ppm kerosene grade be
produced and distributed. Much of this 15 ppm kerosene might be used in applications not
requiring 15 ppm sulfur content. Adding pour point depressant is an alternative to blending
kerosene.  This can be done very flexibly at the terminals in areas facing very cold weather.
Thus, we expect that the use of pour point depressants will increase and the terminal blending of
kerosene will decrease. For kerosene blended into winter diesel fuel, the kerosene can simply be
added to the distillate being  fed to the hydrotreater and desulfurized along with the rest of the 15
ppm diesel fuel pool.
   In summary, the primary purpose of developing these future production volumes is to
reasonably project the economies of scale of the desulfurization equipment being constructed in
response to the NRLM fuel program, including the interaction of this program with the 2007
highway fuel program.  Larger capacity equipment costs more than smaller equipment in total,
but is less expensive on a per gallon basis. Operating costs are not affected, as these are
proportional to volume.  In the NPRM we projected production volumes for calendar year 2008,
as this was the first full year that the NRLM sulfur caps were effective. However, we now
believe that 2014 is more reasonable, because the assumed life of desulfurization equipment is
15 years and 2014 marks the mid-point of the life of equipment built in 2007.

   7.2.1.3.4 Selection of Refineries Producing 500 and 15 ppm NRLM Fuel

   We used two basic criteria to select those refineries most likely to produce 500 and 15 ppm
NRLM fuel under this NRLM rule.  The first criterion was refineries' ability to avoid producing
                                         7-130

-------
                                                 Estimated Costs of Low-Sulfur Fuels
lower sulfur NRLM fuel (i.e., continue producing high sulfur heating oil). The second criterion
was the estimated cost of compliance. We assumed that those refineries facing the lowest
desulfurization costs in a given region would be the most likely to invest.  A key factor in
estimating desulfurization costs on a refinery specific basis is whether the refinery: 1) would be
able to produce 500 or 15 ppm NRLM fuel with its existing hydrotreater, 2) would be able to
revamp an existing hydrotreater to produce NRLM fuel, or 3) would have to build a grass roots
hydrotreater to produce NRLM fuel. These three factors are described below.

       7.2.1.3.4.1 Geographic and Logistic Limitations Affecting the Production of Heating Oil

   It goes without saying that refiners have to be able to market the fuels which they produce.
That is the nature of business. This includes the No. 2 distillate that they produce. Most No. 2
distillate volume comes directly from the crude oil itself. It is not feasible, or economical, to
convert all this distillate fuel to other products.  Thus, under this NRLM rule, refiners basically
have three choices for this distillate; produce 15 ppm highway diesel fuel, produce 500 and 15
ppm NRLM fuel (depending on the time period)  or produce high sulfur heating oil.  Producing
high sulfur heating oil should require no change in current refinery configurations, as all of the
No. 2 distillate produced today essentially meets heating oil specifications.

   However, as alluded to above, refiners must be able to deliver their fuel to the geographical
market where it is consumed. The market for high sulfur distillate will decrease by 50 percent
upon the implementation of this NRLM rule.  Over two-thirds of all  high sulfur distillate use
after 2010 will be concentrated in the Northeast.  Thus, PADD 1 refineries should have no
difficulty in  selling high-sulfur distillate to this market if they desired. Likewise, PADD 3
refineries which are connected to one of the two large  pipelines running from the Gulf Coast to
the Northeast (Plantation and Colonial)  or which have access to ocean transport should also be
able to market high sulfur distillate.  In addition,  selected markets in PADD 5,  such as Hawaii,
also have significant heating oil demand, so some PADD 5 refineries were also assumed to have
the flexibility to continue producing high-sulfur distillate if they desired.

   As discussed in Section 7.1 above, however, the heating oil markets in PADDs 2 and 4 will
be very small after the NRLM rule takes effect. Thus, we believe that it is unlikely that pipelines
in these PADDs will continue to carry heating oil as a  fungible product. Therefore, we do not
believe that refineries located in PADDs 2 and 4  will have the option of choosing to avoid
complying with the NRLM fuel program by producing high sulfur distillate. To the degree that
they are not  already producing 15 ppm highway diesel fuel, they will have to take steps to
produce 500 ppm and 15 ppm NRLM fuel. The same is true for refineries located in PADDs 3
and 5 which do not have access to a large local market for heating oil or which are not connected
to efficient transport to the Northeast. The final NRLM rule does not require that these refineries
produce NRLM fuel, per se.  We simply believe that this is a reasonable assumption for cost-
estimation purposes.

   We reviewed the geographical location of each domestic refinery and those of pipelines
serving the Northeast and identified those falling into the two groups described above. The
number of refineries projected to have no choice  but to produce NRLM diesel fuel is shown in

                                         7-131

-------
Final Regulatory Support Document
Table 7.2.1-31 along with the total number of refineries projected to produce high-sulfur
distillate fuel after implementation of the 2007 highway diesel rule. These projections consider
the small refiner provisions included in the NRLM final rule. These provisions reduce the
number of refineries projected to have to produce 500 ppm NRLM fuel in 2007, as small refiners
are assumed to be able to sell high sulfur diesel fuel to the NRLM market.

                                     Table 7.2.1-31
              Number of Refineries Lacking the Option to Produce Heating Oil

PADD1
PADD2
PADD3
PADD4
PADD5
Prior to NRLM Rule Implementation considering Fully Implemented Highway Diesel Program
Refineries Producing Some High-Sulfur
Distillate Fuel
13
17
37
8
17
Starting June 1 , 2007 (Considers Small Refiner Provisions)
Must produce 500 NRLM fuel
Refineries Producing Some High-Sulfur
Distillate Fuel
0
13
14
3
4
33
7
1
0
17
Starting June 1, 2010 (Considers Small Refiner Provisions)
Must produce 1 5 Nonroad fuel
Must produce 500 NRLM fuel
Refineries Producing Some High-Sulfur
Distillate Fuel
0
1
12
6
11
0
0
9
28
3
5
0
0
5
12
Starting June 1, 2012 (Considers Small Refiner Provisions)
Must produce 1 5 NRLM fuel
Must produce 500 NRLM fuel
Refineries Producing Some High-Sulfur
Distillate Fuel
0
1
12
14
3
0
4
5
28
7
1
0
0
5
12
   We repeated this analysis for 2010. The number of refineries producing some high sulfur
distillate fuel in 2010 is less than in 2007, as additional refineries produce either 15 or 500 ppm
NRLM fuel. The number of refineries projected to have to produce NRLM fuel in 2010 due to
distribution system constraints increases over that in 2007 due to the expiration of the small
refiner provisions.  While we project that the vast majority of 15 ppm nonroad fuel will be
produced by those refineries facing the lowest desulfurization costs, we project that a few
refineries will have to invest to produce 15 ppm nonroad fuel because of limited ability to
distribute higher sulfur fuel to the L&M and heating oil markets.  These refineries produce a
large volume of 500 ppm NRLM fuel in 2007  and are not directly connected to a pipeline or
navigable waterway.  Given the volume of fuel involved, we decided that shipping all of it via
rail was also not economically feasible long term. The number of these constrained refineries is
                                         7-132

-------
                                                Estimated Costs of Low-Sulfur Fuels
much fewer than those which we project will be unable to distribute all of their distillate fuel to
the heating oil market and thus had to produce make 500 ppm NRLM fuel in 2007.

   In 2012, the number of refineries that must produce NRLM fuel is the same as 2010.
However in 2012, the non-small refineries that we project have to produce 500 ppm L&M fuel in
2010 invest further to produce 15 ppm L&M fuel.

   In 2014, the only change is the expiration of the small refiner provisions. The small
refineries producing 500 ppm nonroad fuel in 2012 invest to produce 15 ppm NRLM fuel.  The
refinery estimates for years 2007-2012 are shown in Table 7.2.1-31.

   Table 7.2.1-32 shows how the NRLM fuel volume produced by these refineries compares
with the total  required NRLM fuel production volume during the 2007-2010 period. This table
starts with the total demand for NRLM fuel, as well as the volume of highway fuel used in the
NRLM fuel markets as developed in Section 7.1. Table 7.2.1-32 also shows the volume of high
sulfur distillate projected for small refiners which are able to sell high sulfur diesel fuel to the
NRLM market during this period. Subtracting the volumes of highway spillover and small
refiner fuel from total demand results in the net volume of 500 ppm NRLM fuel which needs to
be produced in response to this NRLM rule. The 500 ppm fuel volumes from refineries having
to produce this fuel are then shown, along with any remaining volume. It should be noted that
we have excluded demand for NRLM fuel in California from Table 7.2.1-32 and the analogous
tables for 2010, 2012 and 2014. Nonroad fuel sold in California is already required to meet a 15
ppm cap in this timeframe per State regulation. L&M fuel demand in California is totally
satisfied by spillover of highway fuel and downgrade.  Thus, we project no on-purpose
production of L&M fuel for use in California. However, distillate production from two
California refineries which current produce high sulfur distillate fuel is considered in satisfying
NRLM fuel demand in PADD 5.

                                     Table 7.2.1-32
       500 ppm NRLM Fuel Production: 2007-2010 (million gallons per year in 2014) *

Total NRLM Fuel Demand
Highway Fuel Spillover
Fuel Produced Under Small
Refiner Provisions
NRLM Requiring Desulfurization
Refineries Having to Produce 500
ppm NRLM Fuel
Remaining Production of 500
ppm NRLM Diesel Fuel
PADDs 1 & 3
9,034
898
671
7,465
281
7,184
PADD 2
7,111
1,906
139
5,066
2,549
2,517
PADD 4
1,046
580
5
461
303
158
PADDS
1,159
381
165
613
0
613
U.S.
18,350
3,765
980
13,605
3,133
10,472
* Excludes NRLM fuel demand in California.
                                        7-133

-------
Final Regulatory Support Document
   As can be seen, more than enough 500 ppm fuel will be produced in PADDs 2 and 4 by
refineries having to produce this fuel.  This is a direct result of assuming that no refinery in
either of these PADDs will be able to market all of their current high sulfur distillate fuel solely
as heating oil.  Significant volumes of 500 ppm NRLM fuel will still have to be produced by
PADD 1, 3  and 5 refineries. As discussed above, we assume that the refineries facing the lowest
desulfurization costs in each PADD will choose to invest to produce any remaining fuel demand
in that PADD.

   It should be noted that we evaluated small refiners' ability to distribute their production
volume of high-sulfur NRLM diesel fuel, even if they do not have access to a common carrier
pipelines carrying this fuel.  Starting with the total demand for NRLM diesel fuel in each PADD
in 2014 from Section 7.1 above, we divided this demand by the square mileage of each PADD to
estimate NRLM diesel fuel demand per square mile.  We then determined the area over which
each small refiner would have to distribute its high-sulfur NRLM fuel to maintain its current
high sulfur  distillate production level.  In all cases, assuming a circular shaped area, the radius of
the circle was 100 miles or less. As this is easily within trucking distance, we concluded that it
was reasonable to assume that all small refiners can continue selling all their high-sulfur
distillate fuel as either high-sulfur distillate fuel or heating oil, and delay producing any 500 ppm
NRLM diesel fuel until at least 2010.

   Table 7.2.1-33 presents the same breakdown of nonroad fuel supply for the period 2010-
2012, with the implementation of the 15 ppm cap. Just over 20% of nonroad fuel demand is
satisfied by highway spillover and just under 10% by distribution downgrade. Small refiner 500
ppm fuel supplies roughly 5% of the market, with the remainder being new 15 ppm fuel
production. Less than 10% of the new 15 ppm nonroad fuel production is by refineries having
no economic choice but to do so, the vast majority of 15 ppm nonroad fuel is produced by
refineries with the lowest cost of production. The volume of 15 ppm nonroad fuel that has to be
produced by refineries with no other economic choice is significantly than was the case for 500
ppm NRLM fuel in 2007. This  occurs, because the L&M market is much larger than the heating
oil market in PADDs 2, 4 and 5 and most refineries can ship their fuel via pipeline or waterway
to the L&M market.
                                         7-134

-------
                                               Estimated Costs of Low-Sulfur Fuels
                                    Table 7.2.1-33
       15 ppm Nonroad Fuel Production: 2010-2012 (million gallons per year in 2014) *

Total Nonroad Fuel Demand
Highway Spillover
Distribution Downgrade
small Refiner Volume (500 ppm nonroad fuel)
slew Production of 1 5 ppm Nonroad Fuel
Refineries Having to Produce 15 ppm Nonroad
"uel
Remaining Production of 1 5 ppm Nonroad Fuel
PADDs 1 & 3
5901
551
217
419
4,714
0
4,714
PADD2
5,670
1,535
519
139
3,477
631
2,846
PADD4
810
451
111
5
243
157
86
PADDS
934
341
264
165
164
0
164
U.S.
13,315
2,878
1,111
728
8,598
728
7,810
 : Excludes NRLM fuel demand in California.
   Table 7.2.1-34 presents the same breakdown of L&M fuel supply for the period 2010-2012.
Just under 20% of nonroad fuel demand is satisfied by highway spillover and another 20% by
distribution downgrade. We project that small refiner 500 ppm fuel will be used in the nonroad
fuel market, where it has an economic advantage. Distribution of this fuel should be
economically feasible, given the small volumes involved and the ubiquitous nature of the
nonroad fuel market. Thus, no L&M fuel is supplied by small refiners during this time frame.
Thus, roughly 60% of 500 ppm L&M fuel is being produced for the L&M market. Nearly 80%
of this 500 ppm L&M fuel production is by refineries which are unable to economically
distribute heating oil, so they have to produce a lower sulfur fuel. In PADDs 2 and 4, the
volume of 500 ppm fuel produced by refineries with no other economic choice is greater than the
remaining demand for L&M fuel. We assumed that the excess production of 500  ppm fuel
refineries in the eastern and southern regions of PADD 2 could be satisfy L&M demand in
PADDs  1 and 3, respectively.  This still leaves a significant volume of 500 ppm L&M fuel
needing  to be produced by refineries in PADDs 1 and 3. We assumed that excess 500 ppm fuel
in PADD 4  would be used in the heating oil market. As usual, we assumed that refineries with
the lowest desulfurization costs in PADDs 1,3 and 5 would invest to produce the remaining 500
ppm fuel demand.
                                        7-135

-------
Final Regulatory Support Document
                                    Table 7.2.1-34
       500 ppm NRLM Fuel Production: 2010-2012 (million gallons per year in 2014) *

Total L&M Fuel Demand
Highway Fuel Spillover
Distribution Downgrade
NRLM Requiring Desulfurization
Refineries Having to Produce 500
ppm L&M Fuel
Remaining Production of 500
ppm NRLM Diesel Fuel
500 ppm Nonroad Fuel Produced
by Small Refiners
Total New 500 ppm Production
PADDs 1 & 3
3,133
347
866
1,920
281
1,639
419
2,058
PADD2
1,441
371
134
936
1,918
(982)
139
(843)
PADD4
236
129
33
74
153
(79)
5
(74)
PADDS
224
50
40
134
0
134
165
299
U.S.
5,034
897
1,073
3,064
2,352
712
728
1,440
* Excludes NRLM fuel demand in California.
   Table 7.2.1-35 presents the same breakdown of 15 ppm NRLM fuel volumes for the period
2012-2014 when the L&M standard goes to 15 ppm.

                                    Table 7.2.1-35
        15 ppm NRLM Fuel Production: 2012-2014 (million gallons per year in 2014) *

Total NRLM Fuel Demand
Tighway Spillover
Distribution Downgrade
"uel Produced Under Small Refiner Provisions
-"reduction of 1 5 ppm NRLM Fuel
Refineries Having to Produce 15 ppm NRLM Fuel
Remaining Production of 1 5 ppm NRLM Fuel
PADDs 1 & 3
9,034
898
467
419
7,250
281
6,969
PADD2
7,111
1,906
685
139
4,381
2,549
1,832
PADD4
1,046
579
147
5
316
310
6
PADDS
1,159
390
304
165
300
0
300
U.S.
18,350
3,773
1,603
728
12,247
3,140
9,107
* Excludes NRLM fuel demand in California.
   Finally, Table 7.2.1-36 presents the same breakdown of 15 ppm NRLM fuel volumes for the
2014 and beyond. The required production volumes of 15 ppm NRLM fuel in 2014 are larger
than those in 2012, as the small refiner provisions expire and downgraded 15 ppm fuel can no
longer be sold to the nonroad fuel market.
                                       7-136

-------
                                               Estimated Costs of Low-Sulfur Fuels
                                    Table 7.2.1-36
    15 ppm Nonroad Fuel Production: 2014 and Beyond (million gallons per year in 2014) *

Total NRLM Fuel Demand
Highway Spillover
Downgraded "500 ppm" NRLM Fuel
"uel Produced Under Small Refiner Provisions
slew Volume of 1 5 ppm Nonroad Fuel
Refineries Having to Produce 15 ppm NRLM Fuel
Remaining Production of 1 5 ppm NRLM Fuel
PADDs 1 & 3
9,034
898
467
0
7,668
701
6,967
PADD2
7,111
1,906
685
0
4,520
2,688
1,832
PADD4
1,046
579
146
0
321
315
6
PADDS
1,159
390
246
0
523
165
358
U.S.
18,350
3,773
1,544
0
13,032
3,869
9,163
 : Excludes NRLM fuel demand in California.
   Sensitivity Case: Long-Term 500 ppm NRLM cap. Table 7.2.1-37 presents an analogous
set of 500 ppm NRLM production volumes for 2010 assuming that no 15 ppm NRLM fuel cap
was implemented. (This situation is analyzed to allow the long-term analysis of the 500 ppm
NRLM diesel fuel cap independent of the 15 ppm nonroad diesel fuel cap).  The primary
difference between these volumes and those for 2007 above is the absence of the small-refiner
volume and fuel to the NRLM pool from distribution downgrade.

                                    Table 7.2.1-37
    500 ppm NRLM Fuel Production: 2010 and beyond* (million gallons per year in 2014)

MRLM Diesel Fuel Demand
Distribution Downgrade
Highway Spillover
3ase High-Sulfur NRLM Demand
"uel Produced Under Small Refiner Provisions
Volume Having to Produce 500 ppm NRLM Fuel
Remaining Demand for 500 ppm NRLM Diesel Fuel
PADDs 1 & 3
9,034
1,084
898
7,052
0
701
6,351
PADD2
7,111
685
1,906
4,520
0
2,688
1,832
PADD4
1,046
147
579
320
0
315
5
PADDS
1,159
304
390
465
0
165
300
U.S.
18,350
2,220
3,773
12,357
0
3,869
8,488
' After all small refiner provisions have expired.
   Sensitivity Case: 15 ppm Nonroad and 500 ppm L&M Fuel

   This case examines the proposed fuel control program, which is identical to that being
promulgated, except that locomotive and marine fuel remains at 500 ppm indefinitely. The only
difference in the geographical constraints assumed to exist is that PADD 2 refiners were allowed
                                        7-137

-------
Final Regulatory Support Document
to continue producing 500 ppm locomotive and marine fuel in 2010 and beyond.  The result was
that some 15 ppm nonroad fuel being consumed in PADD 2 is being produced in PADD 3.  This
shipment of 15 ppm fuel from PADD 3 to PADD 2 occurs under the final NRLM fuel program,
as well.

       7.2.1.3.4.2 Low Sulfur NRLM Fuel Via Existing. Revamped or Grass Roots Hydrotreater

   This section presents the methodology that we used to determine what actions refiners would
likely take to produce 500 and 15 ppm NRLM diesel fuel during the implementation of the
NRLM diesel fuel program.  The timing of the various steps in both the highway and NRLM fuel
programs are summarized in Table 7.2.1-38.

                                     Table 7.2.1-38
                   Sequence of Sulfur Caps for Highway and NRLM Fuel


June 1,2006 -May 3 1,2007
June 1,2007- May 3 1,20 10
June 1,2010 -May 3 1,2012
June 1,2012 -May 31,2014
June 1, 2014 and beyond
Highway Fuel
80vol% 15 ppm
20 vol% 500 ppm
80vol% 15 ppm
20 vol% 500 ppm
15 ppm
15 ppm
15 ppm
Non-Small Refiners
Nonroad Fuel
High Sulfur
500 ppm
15 ppm
15 ppm
15 ppm
L&M Fuel
High Sulfur
500 ppm
500 ppm
15 ppm
15 ppm
Small Refiners
High Sulfur
High Sulfur
500 ppm
500 ppm
15 ppm
   In Section 7.2.1.3.3, we describe how we coupled refiners' projected highway fuel volumes
with historic total distillate production fuel volumes and EIA future growth rates for highway
and high sulfur distillate fuels to project each refinery's production of highway and high sulfur
distillate fuel prior to this NRLM fuel program.  The issue in this section is the steps which
refiners have to take to produce 15 and 500 ppm NRLM fuel beyond this baseline to comply
with the NRLM standards.  The primary question answered in this section is whether they will
be able to revamp an existing hydrotreater, or must build a new hydrotreater. For 15 ppm
highway fuel, we basically assumed, as we did in the Final RIA for the 2007 highway fuel
program, that 80 percent of 15 ppm highway fuel volume would be produced using revamped
hydrotreaters. The remaining 20 percent would be produced with new, grass-roots units.  The
remainder of this section develops analogous projections for the production of 500 ppm and 15
ppm NRLM fuel during the various steps of the NRLM fuel program.

   To facilitate this discussion, we divided refineries which are projected to produce some high
sulfur distillate after 2010 into three categories:
                                        7-138

-------
                                                  Estimated Costs of Low-Sulfur Fuels
    1) "Highway" refineries: refineries which produce 95 percent or more of their total distillate
    production as 15 ppm highway diesel fuel;BB
    2) "High Sulfur" refineries: refineries which produce 90 percent or more of their total
    distillate production as high sulfur distillate;
    3) "Mix" refineries: refineries which produce some high sulfur distillate and which do not
    fall into categories one or two above.

    Table 7.2.1-39 presents the percentages of high-sulfur distillate fuel production that falls in
the categories described above. The number of refineries in each category is further broken
down as to whether or not it currently has a distillate hydrotreater. This latter aspect is relevant
to desulfurization costs as discussed in Section 7.2.1.3.2 above.

                                      Table 7.2.1-39
                   Distribution of High-Sulfur Distillate Production fo-



Number of
Refineries
Percent of
Nonroad Fuel
High-Sulfur
Refineries
W/Dist
HT
10

31

No Dist
HT
25

15

Mixed Refineries
Producing 1 5 ppm
Highway Fuel in 2006
W/Dist
HT
37

38

No Dist
HT
11

14

Mixed Refineries
Producing 1 5 ppm
Highway in 20 10
W/Dist
HT
1

1

No Dist
HT
0

0

Highway Refineries
W/Dist
HT
7

1

No Dist
HT
1

0

  " W/Dist HT" means refineries currently having a distillate hydrotreater
 "No Dist HT means refineries that do not currently have a distillate hydrotreater
    The next three sub-sections address how we project that each of these groups of refineries
could produce either 500 or 15 ppm NRLM fuel. The final sub-section summarizes the results.

    Highway Refineries: This category primarily includes refineries which are projected to
produce 95 percent or more of their the No. 2 distillate fuel in 2010 to the 15 ppm highway
standard prior to this NRLM rule. Refineries producing 100 percent highway fuel have no
distillate fuel left from which to produce 500 or 15 ppm NRLM fuel. Thus, with one exception,
they are ignored in this analysis.  The exception is that the refiners' pre-compliance reports
showed an excess supply of 15 ppm highway fuel in PADDs 1 and 3. Production of NRLM fuel
by highway refineries presumed to supply this excess is addressed slightly differently below.

    Refineries in this category produce a very small amount of high-sulfur distillate fuel
compared with their volume of highway diesel  fuel.  This small volume of high-sulfur distillate
fuel is likely either off-specification diesel fuel or opportunistic sales to the non-highway diesel
   BB We also included a few refineries which project producing 15 ppm highway fuel in 2010,
but whose highway fuel is not needed to fulfill highway fuel demand in 2010.
                                          7-139

-------
Final Regulatory Support Document
fuel market because of advantageous prices, market relationships, etc. Thus, we assumed that
the refinery could incorporate this high-sulfur distillate into its highway hydrotreater design.
The incremental capital cost assigned to the NRLM diesel fuel program was assumed to be the
difference between the capital cost associated with a grass-roots hydrotreater sized to process all
the refinery's distillate fuel and that for a grass-roots hydrotreater sized to treat just the highway
diesel fuel volume.  Thus, this approach assumed that the incremental cost of this small increase
in capacity could occur at a high degree of economy of scale, but would also encompass the full
cost of hydrotreating from uncontrolled levels to 7 ppm. We did this because it seems
reasonable to assume that a refinery producing so much highway fuel would design its 15 ppm
hydrotreater in such a way that it could be modified to process all the refinery's distillate.  This
is particularly true given the public attention given to the need for 15 ppm nonroad diesel fuel
over the past few years.

   This approach is applied to both the production of 500 and 15  ppm NRLM fuel. While
incorporating the production of 500 ppm NRLM fuel into a 15 ppm highway fuel hydrotreater is
not necessarily straightforward, the net effect of our assumption here is that roughly half the
capital cost to produce 15 ppm NRLM fuel at these refineries is required to produce 500 ppm
NRLM fuel.  This seems reasonable. Also, this assumption only affects capital costs, not
operating costs, as the  latter are only a function of the distillate composition and refinery
location (i.e., PADD).

   As described in Section 7.2.1.3.3 above, the highway pre-compliance reports showed that an
excess of 15 ppm fuel  capacity was likely in PADD 3 in 2007. Thus, we assumed that this
capacity could supply  500 ppm NRLM to PADDs 1, 2 and 3 through 2010 at a relatively low
cost. To approximate these "low" costs we assumed that 500 ppm NRLM fuel could be
produced by these hydrotreaters at the national average cost of the remainder of the 500 ppm
NRLM fuel.

   Figure 7.2-6 presents a flowchart of this process for highway refineries.
                                         7-140

-------
                                              Estimated Costs of Low-Sulfur Fuels
                                     Figure 7.2-6
                  "Hgjrvvay" Refineries MRLMf^drotteater Mdifications
              500 ppm
               NRLM
               in2007
               15 ppm
               NRLM
               in2010
                        Number in box equals nurrber of refineries.
   Mix Refineries: Mix refineries produce substantial volumes of both highway and high
sulfur distillate fuels prior to the NRLM rule. Because of the substantial volumes of both fuels
being produced, we assumed that the 15 ppm hydrotreater being used to produce highway fuel
could not be revamped to incorporate production of 500 or 15 ppm NRLM fuel.  Thus, with one
exception, we assumed that the production of 500 ppm NRLM fuel by mix refineries would
require would require a grass roots hydrotreater. The later production of 15 ppm NRLM fuel
was assumed to be a revamp of this 500 ppm hydrotreater, given that the 500 ppm unit was
designed knowing that the nonroad and L&M caps would soon be 15 ppm.  Thus, with two

                                       7-141

-------
Final Regulatory Support Document
exceptions, there are no presumed synergies between the highway and NRLM fuel programs for
these refineries.

   One exception to this assumption involved the way certain refineries are expected to produce
their 15 ppm highway fuel. As described above, we project that 80 percent of 15 ppm highway
fuel can be produced via a revamp of the existing highway fuel hydrotreater.  The remaining 20
percent of highway fuel volume will be produced with a new grass roots hydrotreater.  In these
latter cases, the current highway hydrotreater will be available to produce 500 ppm NRLM fuel
at no capital cost.

   We did not attempt to identify the specific refineries which were likely to build a new grass
roots hydrotreater for 15 ppm highway fuel production.  This decision depends on many factors,
most of which involve proprietary data. Thus, we assumed that 20 percent of the highway fuel
from highway refiners and 20 percent of the highway fuel from mix refiners was being produced
with a new grass roots unit.  We assumed that 20 percent of the high sulfur distillate production
from mix refiners could be produced with these hydrotreaters  at no capital cost. Then in 2010
and 2012, new grass roots units would be required to produce 15  ppm nonroad and 15 ppm L&M
fuel, as was assumed for the other mix refineries.

   The other exception was a single refinery which projected that they would not begin
producing 15  ppm highway diesel until 2010. In this case, there would be sufficient leadtime for
these refineries to combine their plans to produce 15 ppm highway fuel with those to produce  15
ppm NRLM fuel.cc This provides an opportunity for economy of scale by combining both
highway and NRLM fuel volumes in a single process unit, as well as affording an opportunity
for the use of advanced  desulfurization technology.

   Figure 7.2-7 presents a flowchart of this process for mix refineries.
   cc The calculation of incremental capital costs in this situation is not straightforward.  We
provided an example calculation below to better explain our methodology in Section 7.2.1.5.3 of
the Draft RIA to this rule. The reader interested in the details of this calculation is referred to
that discussion.

                                         7-142

-------
                                      Estimated Costs of Low-Sulfur Fuels
                              Figure 7.2-7
500 ppm
NRIM
in2007
   New
500ppmHT-
    19
None-H$i Sulfur
  NRLMFuel-
      7
IJse Existing
   New
SOOppmHT-
    1
 15 ppm
 NRLM
 in2010
                                               HTRevanp-
                                                   1
500 ppm
 NRLM
 in2010
 15 ppm
 IMA
 in2012
 15 ppm
 NRLM
 In2014
                    HT=Ffydroteater
                    L^M=Locomlive andlVferine diesel fuel
                    Nurrber in box e^|^iuiTber of refineries.

-------
Final Regulatory Support Document
                                      Figure 7.2-8
                   "High Sulfur" Refineries NRLMHydrotrealer Modifications

500 ppm
NRLM
in 2007
15 ppm
NRLM
in 2010
500 ppm
NRLM
in 20 10
15 ppm
L8M
in 20 12
15 ppm
NRLM
In 20 14


1

v
\
New 500 ppm
HT-
8*




New 15
ppmHT-
2
A r

^
2-

1


\

\

i
None High Sulfur
NRLMFuel-
12

/
Revamp to
15ppm-
6

1

\l/
New 500
ppmHT-
9

/
Revamp to
15ppm-
2



. _ _ _ _ ^
New 15
ppmHT-
1

— I —
9



\

/
Use Existing
HwyHT-
2


2



Revamp to
15ppm-
9






\




/
Exits
Market-
3




\




i
Exits
Market -
1
                    HT=Hydrotreater
                    Hvyy=Highway
                    L&M=Locomotive and Marine diesel fuel
                    Number in box equals number of refiners.
                    * One refinery installs a new HT and also uses it's existing Highway HT
                      to make 500 ppm fuet?-144

-------
                                                Estimated Costs of Low-Sulfur Fuels
High Sulfur Refineries:  These refineries are projected to produce little or no 15 ppm highway
fuel in 2010 in response to the 2007 highway diesel rule.  Therefore, we assume that any 500
ppm NRLM fuel produced would require a grass-roots hydrotreater. The production of 15 ppm
NRLM fuel was assumed to be a revamp of this 500 ppm hydrotreater, given that the 500 ppm
unit was designed knowing that the nonroad and L&M caps would soon be 15 ppm. Thus, there
are no presumed synergies between the highway and NRLM fuel programs for these refineries.

   One exception to this approach is a set of three refineries which currently produce highway
diesel fuel, but project in their pre-compliance reports to cease highway fuel production in 2006.
Because they produce no highway fuel after 2006, by definition these refineries fall into the high
sulfur refinery category.  However, they clearly have the hydrotreating capacity to produce 500
ppm fuel up to their current highway fuel production.  We assumed that this hydrotreating
capacity was available at no capital cost to produce 500 ppm NRLM fuel in 2007. We also
assumed that a grass roots hydrotreater would be needed to produce 15 ppm fuel in either 2010
for nonroad or for 2012 for L&M, as these refiners' decisions to leave the highway market likely
indicated an inability to produce 15 ppm fuel via a revamp. As it turns out, only two of these
three refineries had sufficient hydrotreating capacity from the highway hydrotreater to treat all
their distillate production. Thus, we assumed that the third refiner would have to construct a new
grass roots hydrotreater to produce 500 ppm NRLM fuel.

   Figure 7.2-8 presents a flowchart of this process for high sulfur refineries.

   We presume that these refineries must build a new hydrotreater in 2007 to desulfurize their
current high-sulfur distillate to 500 ppm.  However, due to the significant amount of lead time
available, we project that these refiners can design a revamp to desulfurize all their distillate fuel
to 15 ppm in 2010 or 2012 if they choose to do so.

   Summary of Results: Overall, for the final NRLM fuel program, we project that 63
refineries will invest to make 15 NRLM diesel fuel by 2014. Table 7.2.1-40 summarizes the
steps which we expect refineries affected by the NRLM rule to take in meeting the highway and
NRLM sulfur caps in the relevant time periods. We have separated refineries into three
categories, depending on the relative proportion of highway and high sulfur distillate fuel that
they produce after the 2007 highway fuel program, but prior to this NRLM fuel rule.
                                         7-145

-------
                                                           Table 7.2.1-40
                      Interaction Between Compliance with the 2007 Highway and Final NRLM Fuel Programs:
              Refiners Projected to Produce Some High Sulfur Distillate Fuel in 2007 Prior to the NRLM Fuel Program
Refineries that
Modifications to
comply with the 1 5
ppm Highway
Standard (Baseline)*
New Modifications
to comply with
NRLM Standards.
Year and
Fuel Control
2006
2010
Total
2007
500 ppm NRLM
2010
500 ppm NRLM
2010
ISppmNR
2012
15ppmL&M
2014
15 ppm NRLM
Highway
Refiners
Units
3
0
3
2
0
3
0
0
Mix 2006 Refiners"
New
Units
13(6)a

Revamp
Units
26

None


39
19(2)
4(2)
9(1)
6(0)
1(0)
0
0
11(3)
7(0)
5(2)
4
0
0
0
0
Mix 20 10 Refiners"
New
Units


Revamp
Units

1
None


1
1(1)
0
0
0
0
0
0
0
0
0
0
0
1
1
0
High Sulfur Refiners3
New
Units


Revamp
Units


None


22
8
9
2
0
1
0
0
6
2
9
2
0
0
0
0
Total


65
36b
13
32
15
16
a Numbers in parentheses are a subset for each category and represent mix refineries that currently have no highway diesel fuel hydrotreater.
b Two high sulfur refiners use their "idled" hwy hydrotreater to make 500 ppm NRLM fuel and exit the NRLM market when the NRLM sulfur standard is lowered to 15
ppm.

-------
                                                Estimated Costs of Low-Sulfur Fuels
   As shown in Table 7.2.1-40, we project that 36 refiners would produce 500 ppm NRLM fuel
in 2007.  Of these 36 refineries:
       28 will install new hydrotreaters
       2 "highway" refiners would perform a relatively minor revamp to their highway distillate
       hydrotreaters, and
       7 refineries could produce 500 ppm NRLM diesel fuel with an "idled" highway
       hydrotreater..
Twenty-six of the refineries that produce 500 ppm NRLM fuel have indicated that they will
produce 15 ppm highway fuel in 2006 and are categorized as follows; twenty-three 2006 mix
refineries, 2 highway refineries and one 2010 mix refinery. The seven refiners who use their
"idled" treaters to produce NRLM are categorized as follows; four were projected to build a new
hydrotreater to produce 15 ppm highway diesel fuel and will use their old highway treater to
produce 500 ppm NRLM fuel.  The other three refineries currently produce 500 ppm highway
fuel, but indicated in their pre-compliance report that they would no longer produce highway
diesel fuel starting in 2006. (Thus, these refineries were categorized as high sulfur refineries for
the purpose of this analysis).  One of these three refineries was also projected to install a new
hydrotreater to process additional high sulfur distillate, as the  capacity of their existing
hydrotreater was not sufficient to  process all their high sulfur distillate volume.

   For all of the refineries using their "idled" highway unit, we used their operating cost to
desulfurize each refineries high sulfur distillate to 500 ppm as the cost for complying with
NRLM standard. Additionally, four refineries in PADD's 1&3 were assumed to invest to fulfill
supply shortfalls in PADD 2.  We also assumed that excess hydrotreater capacity from the
highway fuel program in PADD's 1&3 is used to  supply 500 ppm NRLM volume demand.  This
amounted to about 20 percent of the national NRLM demand.

   In 2010, we project that 32 refineries will produce 15 ppm nonroad fuel while 26 refineries
will produce 500 ppm NRLM (one refinery produces 15 ppm nonroad and 500 ppm L&M fuel).
Thus, a total of 57 refineries produce NRLM fuel which is 21  more than produced 500 ppm
NRLM fuel in 2007, despite the volume of fuels being similar. There are two reason for the
additional refinery participation in 2010. One, the increase in the number of refineries affected
is the availability of idled "highway" hydrotreaters for 500 ppm fuel  production in 2007.  The
capacity of these hydrotreaters is relatively large,  so a few of these refineries can produce a large
volume of 500 ppm NRLM fuel in 2007. However, these refineries'  costs to produce 15 ppm is
not always competitive with other refineries in their PADD. Thus, many of these refineries are
not projected to produce 15 ppm nonroad fuel in 2010.  Their volume of nonroad fuel is replaced
by other refineries producing less volume per refinery.  Two, small refineries invest to produce
500 ppm NRLM fuel due to the expiration of the small refiners provisions which allow high
sulfur distillate to be sold to the 500 ppm NRLM  market. Thus, the total number of refineries
producing 15 nonroad fuel and 500 ppm L&M in  2010 increases.

   In 2012, we project that an additional 15 refineries will invest to produce 15 ppm fuel when
the L&M sulfur cap is lowered to 15 ppm.  This is 15 additional refineries producing 15 ppm
fuel than in 2010. Fifteen refineries continue to produce 500 ppm NRLM fuel.
                                         7-147

-------
Final Regulatory Support Document
    In 2014, with the expiration of the small refiner provisions, and additional 16 refineries
invest to produce 15 ppm NRLM fuel.

    7.2.1.4 Summary of Cost Estimation Factors

    This section presents a variety of costs, such as those for electricity and natural gas, as well
as cost adjustment factors.

    7.2.1.4.1 Capital Cost Adjustment Factors

    Unit Capacity:  The capital costs supplied by the vendors of desulfurization technologies
apply to a particular volumetric capacity. We adjust these costs to represent units with lower or
higher volumetric capacity using the "sixth tenths rule."00  According to this rule, commonly
used in the refining industry, the capital cost of a piece of equipment varies in proportion to the
ratio of the new capacity to the base capacity taken to some power, typically 0.6.  This allows us
to estimate how the capital cost might vary between refineries due to often large differences in
the amount of distillate fuel they are desulfurizing.

    Stream Day Basis:  The EIA data for the production of distillate by various refineries are on
a calendar basis.  In other words, it is simply the annual distillate production volume of the
period of interest divided by the number of days in the period. However, refining units are
designed on a stream day basis. A stream day is a calendar day in which the unit is operational,
or is expected to be operational. Refining units must be able to process more than the average
daily throughput due to changes in day-to-day operations, to be able to handle seasonal
difference in diesel fuel production and to be able to re-treat off-specification batches. The
capital costs for the desulfurization technologies were provided on a stream day basis.

    Actual refining units often operate 90 percent of the time, or in other words, can process 90
percent of their design capacity over the period of a year. However, when designing a new unit,
it is typical to assume a lower operational percentage. We assumed that a desulfurization unit
will be designed to meet its annual production target while operating only 80 percent of the time.
This means that the unit capacity in terms of stream days must be 20 percent greater than the
required calendar day production.

    Off-site and Construction Location Costs: The capital costs provided by vendors do not
include off-site costs, such as piping, tankage, wastewater treatment, etc. They also generally
assume construction on the Gulf Coast,  which are the lowest in the nation. Off-site costs are
typically assumed to be a set percentage of the on-site costs.
    DD The capital cost is estimated at this other throughput using an exponential equation termed the "six-tenths
rule." The equation is as follows: (Sb/Sa)exCa=Cb, where Sa is the size of unit quoted by the vendor, Sb is the size
of the unit for which the cost is desired, e is the exponent, Ca is the cost of the unit quoted by the vendor, and Cb is
the desired cost for the different sized unit. The exponential value "e" used in this equation is 0.9 for splitters and
0.65 for desulfurization units (Peters and Timmerhaus, 1991).

                                           7-148

-------
                                                 Estimated Costs of Low-Sulfur Fuels
   The off-site cost factors and construction location cost factors used in this analysis were
taken from Gary and Handewerk.37 The off site factors provided by Gary and Handewerk apply
to a new desulfurization unit.  Off-site costs are much lower for a revamped unit, as the existing
unit is already connected to the other units of the refinery, utilities, etc.  Thus, we reduced the
off-site factors for revamped units by 50 percent.38

   The off-site factors vary by refinery capacity, while the construction location factors vary
between regions of the country.39 In our analysis of the costs for the Tier 2 gasoline sulfur rule,
we estimated the average of each factor for each PADD.  There, all the naphtha desulfurization
units were new units.  Thus, the PADD-average off-site factors developed for that rule were
simply divided by two to estimate PADD-average factors for revamped units here. The resulting
factors are summarized in Table 7.2.1-41.
                                      Table 7.2.1-41
                         Offsite and Construction Location Factors

Offsite Factor
- New Unit
- Revamped Unit
Construction Location Factor
PADD 1
1.26
1.13
1.5
PADD 2
1.26
1.13
1.3
PADD 3
1.20
1.10
1
PADD 4
1.30
1.15
1.4
PADD 5
1.30
1.15
1.2
   Additional Capital Costs:  There are also likely some capital costs associated with equipment
not included in either the vendor's estimates, nor the general off-sites.  Examples include
expansions of the amine and sulfur plants to address the additional sulfur removed, a new sulfur
analyzer. Additionally, there are other capital  costs that occur due to unpredictable events, such
as material and product price changes, cost data inaccuracies, errors in estimation and other
unforseen expenses. In the NPRM, we accounted for these costs, by increasing the capital costs
(after off-sites adjustment) by 18 percent. A factor of 15 percent is often used for this type of
analysis.40  However, we increased this factor to 18 percent to include the costs of starting up a
new unit.41

    We received comment that this factor was not sufficient to include the more sizeable
increases in sulfur plant capacity associated with this NRLM sulfur control. In several recently
developed fuel programs, such as the Tier 2 gasoline and 2007 highway diesel fuel programs, the
sulfur reduction per gallon was only roughly 300 ppm. Here, the reduction is more than 3000
ppm.  Therefore, the cost of expanded sulfur processing capacity was sufficient small in these
previous programs to be appropriately accounted for within the 18 percent factor. In this rule,
much more sulfur is being removed from the fuel in the form of hydrogen sulfide, which needs to
be converted to elemental sulfur in the refinery. In Section 6.2 of the Summary and Analysis of
Comments, we evaluated the cost of sulfur plant expansions and developed a new set of capital
cost contingency factors which more appropriately account for these costs. These revised
contingency factors are shown in Table 7.2.1-42 below.
                                          7-149

-------
Final Regulatory Support Document
                                       Table 7.2.1-42
     Final Capital Cost Contingency Factors (% of Hydrotreater Costs Including Off-Sites)

Capital Contingency Factor for
Debottleneck Sulfur Plant
Capital Contingency Factor tor New
Sulfur Plant
NRLM fuel Desulfurized from Uncontrolled Sulfur to 500 ppm Standard
Conventional - New Unit
Process Dynamics - New Unit
29
34
53
69
NRLM fuel Desulfurized from Uncontrolled Sulfur to 15 ppm Standard
Conventional - New Unit
Process Dynamics - New Unit
22
26
38
49
NRLM fuel Desulfurized from SOOppm to 15 ppm Standard
Conventional - Revamped Unit
Conventional - New Unit *
Process Dynamics - Revamp Unit
18
17
1R
25
21
31
* Current highway hydrotreater was used to produce 500 ppm NRLM Fuel
   We applied the above contingency factors to each refinery depending on whether or not it
had an existing sulfur plant. We obtained this information from the 2002 EIA Petroleum Supply
Annual.

   Capital Amortization: The economic assumptions used to amortize capital costs over
production volume and the resultant capital amortization factors are summarized below in
Table 7.2.1-43.42 These inputs to the capital amortization equation are used in the following
section on the cost of desulfurizing diesel fuel to convert the capital cost to an equivalent per-
gallon cost.EE

                                       Table 7.2.1-43
          Economic Cost Factors Used in Calculating the Capital Amortization Factor
Amortization
Scheme
Societal Cost
Capital Payback
Depreciation
Life
10 Years
10 Years
Economic and
Project Life
15 Years
15 Years
Federal and
State Tax Rate
0%
39%
Return on
Investment
(ROI)
7%
6%
10%
Resulting Capital
Amortization
Factor
0.11
0.12
0.16
    The capital amortization scheme labeled Societal Cost is used most often in our estimates of
cost made below. It excludes the consideration of taxes. The other two cost amortization
schemes include corporate taxes, to represent the cost as the regulated industry might view it.
The lower rate of return, 6 percent, represents the rate of return for the refining industry over the
    EE The capital amortization factor is applied to a one-time capital cost to create an amortized annual capital cost
that occurs each year for the 15 years of the economic and project life of the unit. This implicitly assumes that
refiners will reinvest in desulfurization capacity after 15 years at the same capital cost, amortized annual cost, and
amortized cost per gallon.
                                           7-150

-------
                                                 Estimated Costs of Low-Sulfur Fuels
past 10 to 15 years.  The higher rate of return, 10 percent, represents the rate of return expected
for an industry having the general aspects of the refining industry.

    7.2.1.4.2 Fixed Operating Costs

   Operating costs based on the cost of capital are called fixed operating costs.  These costs are
termed fixed, because they are normally incurred whether or not the unit is operating or
shutdown.  Fixed operating costs normally include maintenance needed to keep the unit
operating, building costs for the control room and any support staff, supplies stored such as
catalyst, property taxes and insurance.

   We included fixed operating costs equal to 6.7 percent of the otherwise fully adjusted capital
cost (i.e., including offsite costs and adjusting for location factor and including the capital cost
contingency) and this factor was adjusted upwards using the operating cost contingency factor.43
The breakdown of the base fixed operating cost percentage is as follows:
   Maintenance costs: 3  percent
   Buildings:  1.5 percent
   Land: 0.2 percent
   Supplies: 1 percent
   Insurance:  1 percent.

   Annual labor costs were taken from the refinery model developed by the Oak Ridge National
Laboratory (ORNL).44  This model has often been used by the Department of Energy to estimate
transportation fuel quality and the impact of changes in fuel quality on refining costs.  Labor
costs are very small,  on the  order of one thousandth of a cent per gallon.

    7.2.1.4.3 Utility and Fuel Costs

   Utility and fuel costs, which comprise the bulk of what is usually called variable operating
costs, only accrue as  the unit is operating and are zero when the unit is not operating.  These
costs are usually based on calendar day capacity and include utility and fuel costs associated
with operating a hydrotreater. Additionally, we assign diesel product losses (diesel that is
cracked to gas and gasoline) that occur during hydrotreating to the variable operating costs.
These losses where described in Section 7.2.1.2  above along with the other aspects of
conventional and IsoTherming hydrotreating technologies.

   We received comments that the utility and fuels (primarily natural gas) prices did not reflect
future prices that will likely exist due to the changing supply and demand balance for ths fuel. In
the NPRM, we based future natural gas prices on the five year average price between 1995 and
2001. It now appears that the high natural gas prices existing over the past few years are likely
to remain, at least to  some degree. Prices have shifted from the $1.5-2.25 per mmBTU range
existing during the 1990's to much higher levels.

   Thus, for the final rule, we decided to base natural gas prices,  as well as those for other fuels
and utilities on EIA's price projections contained in their 2003 AEO. These price projections are

                                          7-151

-------
Final Regulatory Support Document
based on long term economic modeling and consider various market impacts of supply and
demand dynamics on fuels and utility prices, i.e. growth in GDP, known fuels regulations, costs
of refining products, increased industrial uses, etc. AEO 2003 presents these prices for every
year from 2000 to 2025.  For simplicity, we chose to use 2014 as a reasonable approximation of
the range of prices likely to occur throughout the period of this analysis.  This is also the same
year for which we project refinery fuel production volumes.  Table 7.2.1-44 presents these AEO
prices.
                                     Table 7.2.1-44
                         Fuel and Utility Prices in 2014: 2003 AEO

Fuel and Utility
LPG
Gasoline
Highway Diesel
High Sulfur Diesel
Electricity
Natural Gas
2003 AEO - Future Prices
Price
$35.49 per bbl
$1.406 per gallon*
$1.3 90 per gallon*
$0.865 per gallon
$0.0440 per kilowatt-hour
$4.15permmBTU
AEO Table No.
12
12
12
12
8
3
 Includes excise taxes.
   These fuel and utility prices represent national averages. The highway fuels include excise
taxes. We removed these taxes in our analysis.FF Also, we desired to reflect differences in fuel
and utility costs across the various PADDs. Therefore, we developed a methodology to adjust
these national average prices to reflect this variability, while still producing the same national
average price when re-averaged across the U.S.

   To do this, we evaluated how prices (excluding taxes) varied by PADD in 2001. For LPG,
gasoline and diesel fuels, this information was available by PADD. However, for natural gas
and electricity, it was available by state.  Thus, for these two fuels, we averaged the prices for all
the states within each PADD.  In all cases, we then assumed that these PADD-specific variations
would be maintained in the future on a relative basis.

   For LPG, motor gasoline and diesel fuels, we obtained prices (excluding taxes) from EIA's
2001 Petroleum Marketing Annual.  Table 7.2.1-45 provides a summary of the specific places
within the EIA 2001 report where we obtained the 2001 pricing information.  Future prices were
determined assuming that each PADD's price in 2001 would change in direct proportion to the
change in the AEO national average price (including taxes) from 2001 to 2014. The results are
presented in Table 7.2.1-45.
   FF Table EN-1 EIA Petroleum Marketing Annual 2002.

                                         7-152

-------
                                                Estimated Costs of Low-Sulfur Fuels
                                     Table 7.2.1-45
               2001 Fuel Prices: Petroleum Marketing Annual: 2001 ($/gallon)

PMA Table No.
PADD1
PADD2
PADD3
PADD4
PADD5
National Avg.
LPG
38 (Industrial Users)
0.626
0.589
0.502
0.588
0.658
0.556
Gasoline
3 1 (Sales for Resale)
0.862
0.898
0.814
0.943
1.003
0.888
Highway Diesel
Fuel
41 (Sales for Resale)
0.768
0.829
0.742
0.875
0.826
0.794
High Sulfur Diesel
Fuel
41 (Sales for Resale)
0.761
0.820
0.730
0.851
0.794
0.771
   We also obtained state-specific electricity prices and natural gas prices data from the EIA.
Electricity prices were obtained from EIA's Electricity Power Annual, 2000 and 2001.00
Natural gas prices were obtained EIA's Natural Gas Navigator.™1  In order to smooth out
significant price volatility between various regions, we averaged electricity prices across two
years (2000-2001) and averaged natural gas prices across 5 years (1997-2001).  We estimated
the average price for refineries in each PADD by weighting the state-specific prices by the
volume of crude oil that refiners process in each state.  This approach reflects geographic
breakdown of the relative electricity and natural gas usage that would occur from additional
hydrotreating. We obtained refinery raw crude throughput from EIA's 2001 Petroleum Supply
Annual. We assumed that these historical PADD-specific price differentials would be
maintained in the future. The PADD-specific historical prices for electricity and natural gas are
summarized in Table 7.2.1-46.
   GG Table 7.4 and Figure 7.7.

   HH Industrial prices.
                                         7-153

-------
Final Regulatory Support Document
                                     Table 7.2.1-46
                               Historical Fuel Prices: EIA

PADDl
PADD2
PADD3
PADD4
PADD5
National Avg.
Electricity (c/kW-hr)
6.4
4.4
4.6
3.7
6.6
5.1
Natural Gas ($ per mmBTU)
4.65
4.64
3.33
4.16
4.39
3.96
   The national average fuel and utility prices shown in Table 7.2.1-47 below were then
multiplied by the ratios of the historical PADD-specific differences to the historical national
average price shown in Tables 7.2.1-45 and 7.2.1-46.

   Finally, we assumed that steam was generated from natural gas at an efficiency of 50
percent.45  We assumed that natural gas feedstocks costs dominated the overall cost, so that on a
BTU basis steam cost twice that of natural gas. The steam cost per pound was estimated by
dividing this cost per mmBTU by the heat content of steam at 300 psi (809 BTU per pound).
The resultant PADD-specific future fuel and utility prices are shown in Table 7.2.1-47.

                                     Table 7.2.1-47
       Summary of 2014 Fuel and Utility Prices for Variable Operating Cost Estimations

Electricity (cents per kilowatt-hour)
LPG (dollars per barrel)
Highway Diesel (cents per gallon)
Non-highway Diesel (cents per gallon)
Gasoline (dollars per barrel)
Steam (cents per pound @ 300 psi)
Natural Gas ($/Mmbtu)
PADDl
5.51
20.98
79.1
72.4
31.9
0.35
4.9
PADD2
3.78
19.74
85.4
78.1
33.7
0.35
4.8
PADD3
3.99
16.82
76.4
69.5
31.2
0.25
3.5
PADD4
3.24
19.71
90.1
81.1
35.6
0.31
4.4
PADD5
5.77
22.05
85.1
75.6
41.5
0.33
4.6
 ; Prices using EIA's AEO 2003.
    7.2.1.4.4 Hydrogen Costs

    Hydrogen costs were estimated for each PADD based on the capital and operating costs of
installing or revamping a hydrogen plant fueled with natural gas. The primary basis for these
                                         7-154

-------
                                                 Estimated Costs of Low-Sulfur Fuels
costs is a technical paper published by Air Products, which is a large provider of hydrogen to
refineries and petrochemical plants.46 The particular design evaluated was a 50 million scf/day
steam methane reforming hydrogen plant installed on the Gulf Coast. The capital cost includes a
20 percent factor for offsites.  The process design parameters from this paper are summarized in
the Table 7.2.1-48.

                                     Table 7.2.1-48
                   Process Design Parameters for Hydrogen Production *
Cost Component
Natural Gas
Utilities
Electricity
Water
Steam
Capital/Fixed Operating Charges
Total Product Cost
Dollars per thousand standard cubic feet (S/MSCF)
1.18
0.03
0.03
-0.07
0.83
2.00
    ' Natural Gas @ $2.75/MMBTU; Steam @ S4.00/M Ibs; Electricity @ $0.045 KWH
   The estimates shown in Table 7.2.1-48 were adjusted to reflect natural gas and utility costs in
each PADD (shown in Table 7.2.1-46).  Changes in the value of steam production and the cost of
water were ignored, as these costs are very small.  The capital cost and fixed operating costs
were increased by 8 percent to reflect inflation from 1998 to 2001.

   We also adjusted the capacity of the hydrogen plant to reflect the capacity which would be
typical for each PADD.  The hydrogen plant capacity for PADD 3 represents the average of the
existing hydrogen plants in the PADD and several third party units producing 100 million
scf/day of hydrogen. For other PADDs, the average plant size was based on the average of
refinery-based hydrogen plants within that PADD, obtained from the Oil and Gas Journal.47 We
incorporated PADD-specific offsite and construction location factors from Table 7.2.1-41, again
assuming a 50-50 mix of new and revamped units.  Table 7.2.1-49 summarizes the average plant
size and the offsite and location factors for the installation of hydrogen plant capital for each
PADD.
                                         7-155

-------
Final Regulatory Support Document
                                     Table 7.2.1-49
        Summary of Capital Cost Factors used for Estimating Hydrogen Costs by PADD
PADD
1
2
3
4
5 Excluding CA
andAK
Alaska
Capacity (million
scf/day)
15
34
65
19
15
15
Off site Factor
1.19
1.19
1.15
1.38
1.23
1.23
Construction Location
Factor
1.5
1.3
1.0
1.4
1.2
2.0
   The adjusted hydrogen costs in each PADD are summarized in Table 7.2.1-50.

                                     Table 7.2.1-50
                           Estimated Hydrogen Costs by PADD
PADD
1
2
3
4
5 Excluding CA and AK
AK
Cost ($71000 scf)
3.56
3.01
2.09
3.33
3.19
3.97
    7.2.1.4.5 Other Operating Cost Factors

    Similar to the 15 percent contingency factor for capital costs, we included a 10 percent
contingency factor to account for operating costs beyond those directly related to operating the
desulfurization unit.48  This factor accounts for the operating cost of processing additional
hydrogen sulfide in the amine plant, additional sulfur in the sulfur plant, and other costs that may
be incurred but not explicitly accounted for in our cost analysis.  We then increased this factor by
2 percent to account for reprocessing of off-specification material (actual "off-spec" allowance is
1/2-1 percent). We adjusted the operating costs to account for as much as 5 percent of all
batches to be re-processed.  However, this is a conservative assumption for this cost analysis.
Furthermore, since this material will have been desulfurized to a level close to the 15 ppm cap,
the operating costs for reprocessing it should be much lower the second time around.
                                         7-156

-------
                                                 Estimated Costs of Low-Sulfur Fuels
   We also believe refinery managers will have to place a greater emphasis on the proper
operation of other units within their refineries, not just the new diesel fuel desulfurization unit, to
consistently deliver diesel fuel under the new standards. For example, meeting a stringent sulfur
requirement will require that the existing diesel hydrotreater and hydrocracker units operate as
expected. Also, the purity and volume of hydrogen coming off the reformer and the hydrogen
plant are important for effective desulfurization. Finally, the main fractionator of the FCC unit
must be carefully controlled to avoid significant increases in the distillation endpoint, as this can
increase the amount of sterically hindered compounds sent to the diesel hydrotreater.

   Improved control of each of these units may involve enhancements to computer-control
systems,  as well as improved maintenance practices.49 Refiners may be able to recoup some or
all of these costs through improved throughput. However, even if they cannot do so, these costs
are expected to be less than 1 percent of those estimated below for diesel fuel desulfurization.50 51
No costs were included in the cost analysis for these potential issues.

   7.2.1.5 Projected Use of Advanced Desulfurization Technologies

   In Chapter 5, we projected the mix of technologies used to comply with a program being
implemented in any year. This projection took into account the factors that affect the decisions
by refiners in choosing a new technology. The projected mix of technologies for certain
important years is summarized in Table 7.2.1-51 for the reader's benefit.

                                      Table 7.2.1-51
          Projected Use of Advanced Desulfurization Technologies for  Future Years

Conventional Technology
Process Dynamics Isotherming
2007
100
0
2010
40
60
2012+
40
60
7.2.2 Refining Costs

   In this section, we present the refining costs for the final NRLM diesel fuel program.  As
described in Section 7.2.1, the costs to produce 500 ppm fuel were estimated using conventional
technology, while those for 15 ppm fuel were projected using both conventional and advanced
desulfurization technologies.  All costs assume the economies of scale for the production of
refineries projected to exist in 2014.  Each refinery's projected costs consider their projected
production of highway diesel fuel under the 2007 highway fuel program, as well as estimates of
its distillate blendstock composition and location (i.e., PADD). Per gallon refining costs assume
a 7 percent before tax rate of return on capital. The sensitivity of these costs to 6 percent and  10
percent after tax rates of return are also evaluated.

   The refining costs for the 15 ppm  sulfur cap on highway diesel fuel are presented first.
While the determination of most of the refineries projected to produce highway fuel was made
                                          7-157

-------
Final Regulatory Support Document
using the refiners' highway fuel pre-compliance reports, additional highway fuel was needed in
PADDs 4 and 5.  This was determined using the projected refinery-specific costs of producing
15 ppm fuel. As these costs incorporate several updates since the publication of the Final RIA
for the 2007 highway diesel rule, we thought it appropriate to summarize these updated costs
here.

   The next section presents refining costs for the final NRLM fuel program. First, the overall
costs of the program are summarized.  Then, refining costs for the four main time periods of the
program are presented: 1) 2007-2010, 2) 2010-2012, 3) 2012-2014, and 4) 2014 and beyond.  All
of these costs are based on NRLM fuel production volumes expected to exist in 2014, the mid-
point of the life of desulfurization equipment built in 2007.  All per gallon costs presented in this
section are then applied to the volume of NRLM diesel fuel actually being desulfurized under the
final fuel program.  These costs would not apply to NRLM diesel fuel already meeting highway
diesel fuel sulfur standards (i.e., spillover  fuel).

   In addition, we also present refining costs for a number of sensitivity  cases:

   1)  Increasing the rate of return on capital to 6-10 percent after taxes,
   2)  No assumed use of advanced desulfurization technology,
   3)  A long term 500 ppm cap for NRLM fuel (i.e., no subsequent 15 ppm cap),
   4)  Nonroad fuel at 15 ppm and locomotive and marine fuel at 500 ppm indefinitely, and
   5)  The final NRLM fuel program with lower NRLM fuel demand.

   Finally, we present the stream of capital costs which would be required by the NRLM fuel
program, in the context of other environmental requirements facing refiners in the same
timeframe, due to the Tier 2 gasoline sulfur program and the 2007 highway diesel fuel  program.

   7.2.2.1 15 ppm Highway Diesel Fuel Program

   The refining costs associated with compliance with the 15 ppm highway diesel cap were
estimated for 2006 and 2010. As the methodology used to project these costs differs somewhat
from that used in the Final RIA for the 2007 highway diesel rule, the costs presented here also
differ and represent an update to those costs. The projected costs for producing 15 ppm highway
diesel fuel are summarized in Table 7.2.2-1.
                                         7-158

-------
                                                Estimated Costs of Low-Sulfur Fuels
                                      Table 7.2.2-1
            Highway Diesel Desulfurization Costs to Meet a 15 ppm Cap Standard
                              ($2002, 7% ROI before taxes)*

Number of Refineries
15 ppm Fuel Production (million gal/yr in 2014)
Total Capital Cost (SMillion)
Average Capital Cost per Refinery (SMillion)
Average Operating Cost per Refinery ($Million/yr)
Total Cost (c/gal)
Refineries Initially Producing 15 ppm Fuel in:
2006
96
53,495
6,060
63.1
15.3
4.0
2010
4
2,022
120
30.9
10.6
3.2
All
Refineries
100
55,517
6,180
61.8
15.1
4.0
 : Includes impact of highway fuel that is down graded in the distribution system.
   As can be seen, we project that 96 refiners will invest to produce 15 ppm highway fuel in
2006, with a total capital cost of $6.06 billion ($63.1 million per refinery).  The average cost to
produce 15 ppm highway diesel fuel is 4.0 cents per gallon.  These costs assume that all the 15
ppm fuel is being produced using conventional hydrotreating.

   We project that 4 additional refineries will invest to produce 15 ppm highway diesel fuel in
2010, as the temporary compliance option expires. The required capital cost will be $120
million ($30.9 million per refinery). The average cost for 15 ppm fuel newly produced in 2010
is 3.2 cents per gallon, which is 0.8 cents lower than 15 ppm fuel first produced in 2006. The
use of advanced technology acts to lower the cost of refiners initially entering the market in
2010. Additionally, 3 of the 4 refineries entering in 2010 desulfurize their high sulfur distillate
and existing highway diesel volume in a single hydrotreater, resulting in lower costs due to
economies of scale.

   Overall, 100 refineries produce the 15 ppm diesel fuel under the 2007 highway diesel fuel
program, with a total capital cost of $6.18 billion ($61.8 million per refinery).  The average
refining cost in 2010 will be 4.0 cents per gallon of fuel.

   7.2.2.2 Costs for Final Two Step Nonroad Program

   The final NRLM fuel program requires that NRLM fuel meet a 500 ppm sulfur cap in 2007,
with a further reduction to 15 ppm in 2010 for nonroad and 2012 for L&M.  Small refiners have
until 2010 to meet the 500 ppm cap, and until 2014 to meet the  15 ppm cap for NRLM fuels.
However, "small refiner" fuel cannot be sold in a designated region basically comprising the
Northeast and Mid-Atlantic regions.  Small refiners can also choose to produce NRLM fuel
which meets the above standards on time and sell "credits" to other refiners, who can then sell
NRLM fuel under the delayed standards. Also, 15 ppm fuel which is contaminated during
                                         7-159

-------
Final Regulatory Support Document
distribution and still meets a 500 ppm cap can be sold to the NRLM market through 2014, and to
the locomotive and marine fuel markets indefinitely.

   In this section, we first present an overall summary of the costs of the entire final NRLM fuel
program. Then we present in greater detail the refining costs for the four distinct time periods of
the final NRLM fuel program: 1) the 500 ppm NRLM cap in 2007, 2) the 15 ppm nonroad cap
and 500 ppm L&M cap in 2010 (and 500 ppm cap for small refiner nonroad fuel), 3) 15 ppm
NRLM cap in 2012 (and 500 ppm ppm cap for small refiners), and 4) the 15 ppm NRLM diesel
fuel program in 2014.  Following these presentations, we present projected costs for the various
sensitivity cases.

   Overall, for the final NRLM fuel program, we project that 63 refineries will invest to make
15 NRLM diesel fuel by 2014. A summary of the projected refining costs for the various steps
in the final NRLM fuel program  is presented in Table 7.2.2-2.

                                     Table 7.2.2-2
            Number of Refineries and Refining Costs for the Final NRLM Program

Number of Refineries Producing
500 or 15 ppm NRLM Diesel
Fuel
Production Volume
(Million gallons per year in 2014)
Refining Costs (c/gal)
Year of
Program
2007-2010
2010-2012
2012-2014
2014-2020
2007-2010
2010-2012
2012-2014
2014-2020
2007-2010
2010-2012
2012-2014
2014-2020
500 ppm Fuel
All Refineries
36a
26
15
0
13,327
3,792
728
0
1.9a
2.7
2.9
0
Small
Refineries
0
13
13
0
0
393
393
0
0
3.7
3.7
0
1 5 ppm Fuel
All
Refineries
0
32
47
63
0
8,598
12,247
13,030
0
5.0
5.6
5.8
Small
Refineries
0
2
2
15
0
335
335
728
0
5.2
5.2
6.9
a In 2007-10, refinery counts do not include 500 ppm NRLM fuel from excess capacity in 15 ppm highway
hydrotreaters, and a few idled highway hydrotreaters.  However, refining costs do include this fuel.
   As can be seen, the per gallon cost of producing 500 ppm and 15 ppm diesel fuels throughout
the various phases of the NRLM fuel program will be 1.9-2.9 and 5.0-5.8 cents, respectively.
We project that the cost of the 500 ppm cap for small refiners will be 3.7 cents per gallon, or 28
                                         7-160

-------
                                                Estimated Costs of Low-Sulfur Fuels
percent greater than that for the average refiner.  We project that the cost of the 15 ppm cap for
small refiners will be 6.9 cents per gallon, or 19 percent greater than that for the average refiner.
Table 7.2.2-3 presents a summary of the capital and annual costs for average and small refiners.

                                     Table 7.2.2-3
           Refining Costs for the Final NRLM Program Fully Implemented in 2014
                              ($2002, 7% ROI before taxes)

Number of Refineries
Total Refinery Capital Cost (SMillion)
2007
2010
2012
2014
Average Refinery Capital Cost (SMillion)
Average Refinery Operating Cost ($Million/yr)
All Refineries
63
2,280
310
1,170
590
210
36.2
8.1
Small Refineries
15
250
0
150
0
100
16.7
2.2
   As can be seen, total capital costs would be $2,280 million for the entire final 15 ppm
NRLM fuel program (average of $36.2 million per refinery). Total capital costs for the 15 small
refineries would be $250 million (average of $16.7 million per refinery).

   7.2.2.2.1 Refining Costs in Year 2007

   We project that 36 refiners would produce 500 ppm NRLM fuel in 2007. The cost of the 500
ppm NRLM cap in 2007 is summarized in Table 7.2.2-4 below.
                                         7-161

-------
Final Regulatory Support Document
                                     Table 7.2.2-4
                  Refining Costs in 2007 for 500 ppm NRLM Diesel Fuel
                              ($2002, 7% ROI before taxes)3

Number of Refineries
Total Refinery Capital Cost (SMillion)
Average Refinery Capital Cost (SMillion)
Average Refinery Operating Cost ($Million/yr)
Amortized Capital Cost (c/gal)
Operating Cost (c/gal)
Cost Per Affected Gallon (c/gal)
All Refineries
36
310
8.6
4.9
0.3
1.6
1.9
   We project that the total capital cost will be $310 million (an average of $10.3 million for
each of the 30 refineries actually building new equipment).  The total refining cost for the 500
ppm NRLM diesel fuel sulfur cap is 1.9 cents per gallon of affected fuel volume, including both
operating and amortized capital costs.

   7.2.2.2.2 Refining Costs in Year 2010

   We project that 32 refineries will produce 15 ppm nonroad fuel in 2010. This is four fewer
refineries than produced 500 ppm NRLM fuel in 2007, as some refineries continue to produce
500 ppm L&M fuel.  The total refining costs to produce 15 ppm nonroad fuel in 2010 are
presented in Table 7.2.2-5.  Separate costs are shown for all refineries, refineries not owned by
small refiners, and for those owned by small refiners.
                                         7-162

-------
                                                Estimated Costs of Low-Sulfur Fuels
                                     Table 7.2.2-5
               Total Refining Costs in 2010 for 15 ppm Nonroad Diesel Fuel
                              ($2002, 7% ROI before taxes)

Number of Refineries
Incremental Capital Cost (SMillion)
Average Refinery Capital Cost (SMillion)
Average Refinery Operating Cost ($Million/yr)
Capital Cost (c/gal)
Operating Cost (c/gal)
Cost Per Affected Gallon (c/gal)
All Refineries
32
1,090
34
9.0
1.6
3.4
5.0
Non-small
Refineries
30
1,030
32.2
8.7
1.6
3.4
5.0
Small Refinery
2
59
30
10.8
1.9
3.3
5.2
   The incremental capital cost in 2010 to produce 15 ppm nonroad fuel is $1,090 million.  The
average cost of producing 15 ppm nonroad diesel fuel is 5.0 cents per gallon. This is 3.1 cents
per gallon more than the average cost to produce 500 ppm NRLM fuel in 2007. This
incremental cost of 3.1 cents per gallon is lower than the 4.0 cent per gallon cost estimated
above for the 15 ppm highway diesel fuel cap. This difference is due to several factors which
have opposing impacts. There are three factors that tend to increase the cost of 15 ppm nonroad
fuel compared to that of 15 ppm highway fuel. One, the vast majority of relatively inexpensive
hydrocrackate was assumed to used in the highway diesel pool. Two, refiners projecting to
produce 15 ppm highway fuel based on pre-compliance report data and cost projections tend to
be those that face lower costs (greater economies of scale, low LCO fractions, etc.). Three, 80
percent of current 500 ppm highway fuel hydrotreaters assumed to be revamped to produce 15
ppm diesel fuel, while the figure is lower for nonroad fuel.  While we project that all the new
hydrotreaters built in 2007 to produce 500 ppm NRLM fuel can be revamped to 15 ppm fuel
production, we assume that none of the existing highway hydrotreaters producing 500 ppm
NRLM fuel in 2007 can be revamped to produce 15 ppm fuel.  This lowers the overall revamp
percentage to less than 80 percent. However, balancing these factors is our projection that a
significant percentage of refiners will use the Process Dynamics and other advanced
desulfurization technologies in 2010, versus 2006 when the vast majority of 15 ppm highway
fuel will first be produced. This one factor essentially compensates for the other three factors in
the other direction.

   As implied in Table 7.2.2-5, most small refiners participating in the NRLM fuel market
produced 500 ppm NRLM fuel in 2010.  However, two small refiner's costs for producing 15
ppm fuel were competitive with the other refineries in producing sufficient volumes of fuel to
satisfy market demand.  These small refiners were assumed to  sell their credits to non-small
refineries, allowing them to produce 500 ppm nonroad fuel in 2010.
                                         7-163

-------
Final Regulatory Support Document
   A significant volume of 500 ppm nonroad fuel will also be produced in 2010 under the small
refiner provisions. The remaining 500 ppm fuel production is for the L&M fuel market.  The
costs of producing 500 ppm diesel fuel in 2010 are presented in Table 7.2.2-6.

                                     Table 7.2.2-6
                      Refining Costs in 2010 for 500 ppm NRLM Fuel
                              ($2002, 7% ROI before taxes)

Number of Refineries
Total Refinery Capital Cost (SMillion)
Average Refinery Capital Cost (SMillion)
Average Refinery Operating Cost ($Million/yr)
Capital Cost (c/gal)
Operating Cost (c/gal)
Cost Per Affected Gallon (c/gal)
All
Refineries
in 2010
26
197
7.6
3.7
0.5
2.2
2.7
Non-
Small
Refineries
in 2010
13
107
8.3
6.7
0.3
2.3
2.6
Small
Refineries
in 2010
13
90
6.9
0.8
1.9
2.1
3.7
   We project that 26 refineries will produce 500 ppm NRLM fuel in 2010 at an average cost of
2.7 cents per gallon.  Thirteen of these refineries are owned by small refiners and are the only
refineries that newly invest in 2010 for new hydrotreaters to produce 500 ppm fuel.  Thirteen
non-small refineries who produce 500 ppm NRLM fuel in 2007 would continue to produce 500
ppm NRLM fuel in 2010.  Two of these non-small refiners produce 500 ppm fuel using credits
generated by small refiners producing 15 ppm nonroad fuel in 2010. The small refiners per
gallon costs are 37 percent more than the average of refiners producing fuel in 2010. The costs
for refiners that enter the market in 2010 are lowered by the non-small refineries.

   7.2.2.2.3 Refining Costs in Year 2012

   In 2012, L&M fuel produced or imported must meet a 15 ppm cap. However, 500 ppm fuel
produced during the distribution of cleaner fuels can be sold to the NRLM markets which
reduces the volume of fuel that must  be desulfurized to a  15 ppm standard.  Additionally, the
provisions that allow small refiners to sell 500 ppm fuel into the NRLM markets also continue.
The cost of producing 15 ppm NRLM fuel in 2012 is shown in Table 7'.2.2-1'.
                                        7-164

-------
                                                Estimated Costs of Low-Sulfur Fuels
                                     Table 7.2.2-7
                Total Refinery Costs in 2012 to Produce 15 ppm NRLM Fuel
                              ($2002, 7% ROI before taxes)

Number of Refineries
Total Refinery Capital Cost (SMillion)
Average Refinery Capital Cost (SMillion)
Average Refinery Operating Cost ($Million/yr)
Capital Cost (c/gal)
Operating Cost (c/gal)
Cost Per Affected Gallon (c/gal)
All Refineries
47
1,980
42.1
9.6
1.8
3.8
5.6
Non-small
Refineries
45
1,920
42.7
9.8
1.8
3.8
5.6
Small
Refineries
2
59
30
5.5
1.9
3.3
5.2
   We project that 47 refineries would produce 15 ppm NRLM fuel, or 15 more than in 2010.
The total refining cost measured from today's high sulfur level would be 5.6 cents per gallon, or
0.6 cent per gallon more than in 2010. Small refineries would have average cost of 5.2 cents per
gallon, or 7 percent lower than the average non-small refineries.

   The 15 ppm costs for the 15 refineries first producing 15 ppm L&M in 2012 are presented in
Table 7.2.2-8.  All of these 15 refineries are non-small refineries and have an incremental capital
investment of $590 million. The average cost of producing 15 ppm L&M diesel fuel is 7.3 cents
per gallon.  This is 5.4 cents per gallon more than the average cost to produce 500 ppm NRLM
fuel in 2007. This incremental cost of 5.4 cents per gallon is higher than the 4.0 cent per gallon
cost estimated above for the 15 ppm highway diesel fuel cap.  As mentioned for the 2010 15 ppm
nonroad costs,  several factors tend to increase the cost to desulfurize NRLM fuels to a 15 ppm
standard compared to that of 15 ppm highway fuel.  The incremental desulfurization costs are
higher for L&M fuel because a large portion of the lowest cost refiners were selected to invest in
2010 for 15 ppm nonroad fuel production leaving higher costs refiners producing L&M and high
sulfur distillate fuels.  Thus in 2012, L&M 15 ppm fuel is produced from these remaining
refineries with higher desulfurization costs.
                                         7-165

-------
Final Regulatory Support Document
                                     Table 7.2.2-8
        Refining Costs for 15 ppm L&M Fuel for Refiners Initially Complying in 2012
                             ($2002, 7% ROI before taxes)

Number of Refineries
Incremental Refinery Capital Cost (SMillion)
Average Refinery Capital Cost (SMillion)
Average Refinery Operating Cost ($Million/yr)
Capital Cost (c/gal)
Operating Cost (c/gal)
Cost Per Affected Gallon (c/gal)
All Refineries (Non-small)
Total
15
590
39.1
11.5
1.9
5.1
7.0
   Of the 15 additional refineries producing 15 ppm L&M fuel in 2012, six will install a new
grass roots hydrotreater as they did not invest to make 500 ppm L&M fuel prior to this time.
The remaining 9 refineries will revamp their new nonroad hydrotreater built in 2007 or 2010.
The average refinery that produces 15 ppm L&M diesel fuel for the first time in 2012 will make
a capital investment of $39.1 million.

   7.2.2.2.4 Refining Costs in Year 2014

   In 2014, all NRLM diesel fuel produced must meet a 15 ppm cap.  Additionally in 2014, the
provisions allowing 15 ppm fuel that is downgraded to 500 ppm sulfur level in the distribution
system to be sold to the nonroad fuel market expire, though this fuel can continue to be sold into
the locomotive and marine market.  Thus, the volume of 15 ppm NRLM diesel fuel produced
increases over the total volume of 15 and 500 ppm NRLM fuel produced in 2010. The cost of
producing 15 ppm NRLM fuel in 2014 is shown in Table 7.2.2-9.
                                        7-166

-------
                                                Estimated Costs of Low-Sulfur Fuels
                                 Table 7.2.2-9
             Total Refinery Costs in 2014 to Produce 15 ppm NRLM Fuel
                              ($2002, 7% ROI before taxes)

Number of Refineries
Total Refinery Capital Cost (SMillion)
Average Refinery Capital Cost (SMillion)
Average Refinery Operating Cost ($Million/yr)
Capital Cost (c/gal)
Operating Cost (c/gal)
Cost Per Affected Gallon (c/gal)
All Refineries
63
2,280
36.2
8.1
1.9
3.9
5.8
Non-small
Refineries
48
2,030
42.5
10.6
1.7
4.0
5.7
Small
Refineries
15
250
16.5
2.2
3.1
3.8
6.9
   We project that 63 refineries would produce 15 ppm NRLM fuel, or 16 more than in 2010.
The total refining cost measured from today's high sulfur level would be 5.8 cents per gallon, or
0.2 cent per gallon more than in 2010.  Small refineries would have an average cost of 6.9 cents
per gallon, or 19 percent higher than the average non-small refineries.

   The 15 ppm costs for the 16 refineries first producing 15 ppm nonroad fuel in 2014 are
presented in Table 7.2.2-10. The incremental capital investment for these 16 refineries in 2014
was $210 million.  Of this $210 million, $100 million will be spent by small refiners.

                                     Table 7.2.2-10
       Refining Costs for 15 ppm NRLM Fuel for Refiners Initially Complying in 2014
                              ($2002, 7% ROI before taxes)

Number of Refineries
Total Refinery Capital Cost (SMillion)
Average Refinery Capital Cost (SMillion)
Average Refinery Operating Cost ($Million/yr)
Capital Cost (c/gal)
Operating Cost (c/gal)
Cost Per Affected Gallon (c/gal)
All
Refineries
Total
16
300
18.9
4.5
2.4
5.2
7.6
Non- small
Refineries
Total
3
110
36.9
16.5
1.4
5.8
7.2
Small
Refineries
Total
13
190
14.6
1.7
3.9
4.0
7.9
   Of the 16 additional refineries producing 15 ppm NRLM fuel in 2014, 13 are owned by small
refiners.  Two of the 16 refineries will install a new grass roots hydrotreater as they did not
                                         7-167

-------
Final Regulatory Support Document
invest to make 500 ppm NRLM fuel prior to this time.  The remaining 14 of 16 refineries will
revamp their new nonroad hydrotreater built in 2007 or 2010. The average refinery that
produces 15 ppm nonroad diesel fuel for the first time in 2014 faces a capital investment of
$18.9 million, while the investment for the average small refiner is smaller at $14.6 million.

   7.2.2.3 Refining Costs for Sensitivity Cases

   7.2.2.3.1 Total Refining Costs at Different Rates of Return on Investment

   The costs presented in the previous section all assumed a 7 percent before tax rate of return
on investment. We also estimated total refining costs for the final NRLM fuel program using
two alternative rates of return on investment: 1)6 percent per year after taxes,  and 2) 10 percent
per year after taxes. The 6 percent rate is indicative of the economic performance of the refining
industry over the past 10-15 years. The 10 percent rate is indicative of economic performance of
an industry like refining which would attract additional capital investment. The total per gallon
cost of producing 15 ppm NRLM fuel in 2014 using all three rates of return are shown in  Table
7.2.2-11.

                                     Table 7.2.2-11
       Refining Costs in 2014 for 15 ppm NRLM Fuel in 2014 (cents per gallon $2002)
Societal Cost: 7% ROI before Taxes
Capital Payback: (6% ROI, after Taxes)
Capital Payback: (10% ROI, after Taxes)
5.8
6.1
6.9
   As can be seen, the difference in the assumed rate of return on investment increases the
societal  cost by 0.3-1.1 cents per gallon.

   7.2.2.3.2 15 ppm Nonroad Diesel Fuel with Conventional Technology

   The  use of advanced technology is expected to reduce the cost of producing 15 ppm diesel
fuel compared to conventional hydrotreating.  To determine the sensitivity of our cost estimates
to the level of advanced technology projected, we developed costs for producing 15 ppm NRLM
diesel fuel with only the use of conventional hydrotreating. We did not vary the specific
refineries projected to  produce 15 ppm NRLM fuel in 2014 from those described in the previous
section.  Total  refining costs to produce 15 ppm NRLM diesel fuel in 2014 using conventional
technology are shown  in Table 7.2.2-12.
                                         7-168

-------
                                               Estimated Costs of Low-Sulfur Fuels
                                    Table 7.2.2-12
             Total Refining Costs in 2014 to Produce 15 ppm NRLM Diesel Fuel
                with Conventional Technology ($2002, 7% ROI before taxes)

Number of Refineries
Total Refinery Capital Cost (SMillion)
Average Refinery Capital Cost (SMillion)
Average Refinery Operating Cost ($Million/yr)
Capital Cost (c/gal)
Operating Cost (c/gal)
Cost Per Affected Gallon Cost (c/gal)
All Refineries
63
2,730
42.7
10.6
2.2
4.9
7.1
Small Refineries
15
290
19.2
2.6
3.7
4.5
8.2
   The total cost to produce 15 ppm nonroad diesel fuel in 2014 with conventional technology
would be 7.1 cents per gallon, or 22 percent higher than the 5.8 cent per gallon cost with a mix
of conventional and advanced technology.  Total capital costs would be $2,730 million with
conventional technology, about 20 percent higher than the $2,286 million investment including
use of advanced technology (see Table 7.2-40).  Operating costs would be 16 percent higher with
conventional technology, $10.0 million as compared to $8.6 million with use of advanced
technology. The same relative comparisons apply to the impact of advanced technology on the
capital costs faced by small refiners. All of these figures represent the total cost of producing 15
ppm diesel fuel from high sulfur diesel fuel.

   7.2.2.3.3 Proposed Two Step NRLM Program: Nonroad Fuel to 15 ppm in 2010 and
   Locomotive and Marine at 500 ppm Indefinitely

   This section presents the refining costs of the NRLM program which EPA proposed: nonroad
fuel at  15 ppm and locomotive and marine fuel at 500 ppm. The refining impacts of this
program are shown in Tables 7.2.2-13.
                                        7-169

-------
Final Regulatory Support Document
                                     Table 7.2.2-13
             Refining Impacts for the Proposed Two Step NRLM Fuel Program
     15 ppm Nonroad Fuel in 2010 and 500 ppm Locomotive and Marine Fuel Indefinitely

Number of Refineries Producing
500 or 15 ppm NRLM Diesel
Fuel
Refining Costs (c/gal)
Year of
Program
2007-2010
2010-2014
2014+
2007-2010
2010-2014
2014+
500 ppm Fuel b
All Refineries
36
26
20
1.9
2.7
2.7
Small
Refineries
0
13
8
0
3.7
3.0
1 5 ppm Fuel
All
Refineries2
0
32
40
0
5.0
5.2
Small
Refineries
0
2
7
0
5.2
7.0
a Includes small refiners.
b In 2007-10, refinery counts do not include 500 ppm NRLM fuel from excess 15 ppm highway hydrotreaters,
and a few idled highway hydrotreaters. However, refining costs do include this fuel. One refiner produces 15 & 500
ppm fuel.
   Under this sensitivity case, we project that 59 refineries would eventually invest to make
either 15 ppm nonroad or 500 ppm locomotive and marine fuel by 2014.  The total cost of
producing 500 ppm NRLM fuel in 2007 is the same as that under the final NRLM program, as
the two programs are identical.  In 2014, the cost of 500 ppm locomotive and marine fuel would
be 2.7 cents per gallon, or sightly higher than the range for 500 ppm NRLM fuel under the final
NRLM program (1.9-2.4 cents per gallon).

   The total cost for producing 15  ppm fuel in this program are lower than the final NRLM
program costs (5.8 cents per gallon in 2014). Less volume of 15 ppm fuel is produced and the
incremental per gallon costs are less than the final programs per gallon cost. This lowers the
average cost.

   Table 7.2.2-14 presents a side-by-side comparison of some of the key refining impacts of the
proposed and final NRLM fuel programs.
                                         7-170

-------
                                               Estimated Costs of Low-Sulfur Fuels
                                    Table 7.2.2-14
 Refining Costs for Two Step Program with 500 ppm Locomotive and Marine fuel versus Final
                      NRLM Program ($2002, 7% ROI before taxes)



Number of Refineries
Total Refinery Capital Cost (SMillion)
2007
2010
2012
2014
Average Refinery Capital Cost (SMillion)
Average Refinery Operating Cost ($Million/yr)
Two Step Program with 1 5 ppm
Nonroad Fuel and 500 ppm
Locomotive and Marine Fuel
All Refineries

60
1,680
310
1,240
0
130
28.5
6.8
Small
Refineries
15
180
0
140
0
40
12.1
1.6
Final NRLM program
All
Refineries
63
2,280
310
1,170
590
210
36.2
8.1
Small
Refineries
15
250
0
150
0
100
16.7
2.2
   Overall, the 15 ppm cap on locomotive and marine fuel in our final NRLM fuel program
increases total capital investment by $600 million and increases the cost of the incremental
volume of L&M fuel by 5.2 cents per gallon (from 2.7 to 7.9 cents per gallon). Table 7.2.2-15
presents the incremental refining impacts of the 15 ppm cap on locomotive and marine fuel over
those of the 500 ppm cap.

                                    Table 7.2.2-15
  Refinery Impacts in 2014 for a 15 ppm Versus 500 ppm Cap on Locomotive and Marine Fuel
                             ($2002, 7% ROI before taxes)

Number of Affected Refiners
Total Incremental Capital, $MM
Incremental Fuel Cost SOOppm to 15 ppm, (c/gal)
Total Fuel Cost , (c/gal)
All
Refineries
23
600
5.2
7.9
   The 5.2 cent per gallon cost to reduce L&M fuel sulfur from 500 to 15 ppm is higher than the
3.5 cent per gallon cost for nonroad fuel, because we assumed that the refiners facing the lowest
desulfurization costs would produce 15 ppm nonroad  fuel, if L&M  fuel sulfur remained at 500
ppm.  Thus, 15 ppm L&M fuel is produced from the remaining refineries  that are projected to
face higher desulfurization costs.
                                        7-171

-------
Final Regulatory Support Document
    7.2.2.3.4 Refining Costs for a 500 ppm NRLM Only Program

    This section presents refining costs for a long-term 500 ppm cap on NRLM fuel (i.e., no
subsequent 15 ppm cap). We evaluated costs in 2010, after any small refiner provisions would
have expired.  These costs are summarized in Table 7.2.2-16.

                                    Table 7.2.2-16
           Refining Costs for a Stand-alone 500 ppm NRLM Diesel Fuel Standard
                             ($2002, 7% ROI before taxes)3

Number of Refineries
Total Refinery Capital Cost (SMillion)
Average Refinery Capital Cost (SMillion)
Average Refinery Operating Cost ($Million/yr)
Capital Cost (c/gal)
Operating Cost (c/gal)
Cost Per Affected Gallon (c/gal)
All
Refineries
57
480
8.4
3.6
0.4
1.6
2.0
Nonsmall
Refineries
41
360
8.8
4.7
0.3
1.6
1.9
Small
Refineries
16
120
7.7
1.0
1.5
1.7
3.2
    1 Equivalent to the costs of the 500 ppm NRLM cap in 2010 without the 15 ppm nonroad cap.
   The overall refining cost of a 500 ppm NRLM fuel cap would be 2.0 cents per gallon. We
project that 57 refineries would produce this fuel with a total capital investment of $480 million.
On average, the refining cost for small refiners would be about 60 percent higher than that of
non-small refiners at 3.2 cents per gallon.

   7.2.2.3.5 EIA-Based Demand for NRLM Fuel

    In Chapter 2 of the Summary and Analysis of Comments, we discuss the uncertainty in
current and future  demand for NRLM fuel, particularly that used in land-based nonroad
equipment.  While we base our primary cost estimates on fuel demands as predicted by EPA's
NONROAD emission model, we decided to evaluate the sensitivity of both costs and benefits to
an alternative level of fuel demand. Here, we present the refining costs assuming that the EIA-
based fuel demands are more accurate than those from NONROAD.

   The total refining costs to produce 500 and 15  ppm NRLM diesel fuel from 2007-2014 for
the two sets of fuel demands are summarized in Table 7.2.2-17.
                                        7-172

-------
                                               Estimated Costs of Low-Sulfur Fuels
                                       Table 7.2.2-17
      Total Refining Costs of NRLM Fuel from 2007-2014 With Varying Fuel Demands
                      (Cents per gallon, $2002, 7% ROI before taxes)

500 ppm NRLM fuel: 2007-2010
500 ppm NRLM fuel: 2010-2012
500 ppm NRLM fuel: 2012-2014
15 ppmNonroad fuel: 2010-2012
15 ppm NRLM fuel: 2012-2014
15 ppm NRLM fuel: 2014+
EIA-Based Fuel Demand
1.9
2.8
3.0
5.0
5.6
5.7
EPA NONROAD Fuel Demand
1.9
2.7
2.9
5.0
5.6
5.8
   As can be seen, reducing NRLM fuel demand has little impact on per gallon refining costs.
The only differences shown are a slight increase in 500 ppm costs from 2010-2014 and a slight
decrease in 15 ppm fuel costs after 2014.  The former effect occurs because the incremental 500
ppm NRLM fuel volume is coming from relatively low cost Gulf Coast refineries. While the
same effect exists in 2014 with respect to 15 ppm fuel costs, the effect of the reduced demand in
reducing costs in other refining areas is larger. Table 7.2.2-18 provides a more detailed
breakdown of the final refining impacts of the 15 ppm NRLM cap in 2014 for the two sets of
fuel demands.

                                Table 7.2.2-18
        Refining Impacts of 15 ppm NRLM Fuel in 2014 With Varying Fuel Demands
                             ($2002, 7% ROI before taxes)

# of Refiners
Total Refinery Capital Cost (SMillion)
Average Capital Cost (SMillion)
Operating Cost ($Million/yr)
Capital Cost (c/gal)
Operating Cost (c/gal)
Cost Per Gallon (c/gal)
EIA-Based Fuel Demand
55
1,870
33.9
7.5
1.9
3.8
5.7
EPA NONROAD Fuel Demand
63
2,280
36.2
8.1
1.9
3.9
5.8
                                        7-173

-------
Final Regulatory Support Document
   As the EIA-based methodology reduces NRLM fuel demand, only 55 refineries would invest
to produce NRLM fuel in 2014 versus 63 using the EPA NONROAD Model estimates. The total
15 ppm NRLM fuel cost would be 5.7 cents per gallon, or 0.1 cents per gallon less than that to
satisfy NONROAD fuel demand. Total capital costs would be $1,870 million, or about 18
percent less than the $2,280 million investment needed to produce the additional fuel volume.

   7.2.2.4 Capital Investments by the Refining Industry

   Refiners must raise capital to invest in new desulfurization equipment to produce the 500
ppm and 15 ppm diesel fuel which would be required under the final NRLM fuel program.  The
previous sections estimated the total capital cost associated with the final and various sensitivity
cases. Refiners expend this capital over a several year period prior to the time which the new
equipment must be used.  This section estimates how much capital would have to be expended in
specific years under the final and alternative programs.  These yearly expenditures are then
added to those required by other fuel quality programs being implemented in the same timeframe
and compared to historic capital expenditures made by the refining industry.

   Two fuel quality regulations are being implemented in the same timeframe as this NRLM
fuel program: The Tier 2 gasoline sulfur program and the 2007 highway diesel fuel sulfur
program.  In the Tier 2 gasoline sulfur control rule, we estimated the expenditure of capital for
gasoline desulfurization by year according to the phase in schedule promulgated in the rule.11
The 2007 highway diesel rule modified that phase in schedule by provided certain  refiners more
time to meet the Tier 2 gasoline sulfur standards. In the 2007 highway diesel rule, we projected
the stream of capital investments required by the U.S. refining industry for both the modified
Tier 2 standards and the 15 ppm highway diesel fuel sulfur program.  We updated the allocation
and amount of capital expenditures for the highway diesel rule to reflect when each refiner
would invest.  The new total capital costs for the 2007 highway diesel fuel program are
discussed in section 7.2.2.1 above. In projecting the stream of capital expended for a particular
project, we assume that the capital investment would be spread evenly over  a 24 month period
prior to the date on which the unit must be on-stream. The stream of projected capital
investment related to the Tier 2 gasoline sulfur program and  the 2007 highway diesel fuel
program rule are shown in Table 7.2.2-19.
   11 Regulatory Impact Analysis - Control of Air Pollution from New Motor Vehicles: The Tier 2 Motor Vehicle
Emissions Standards and Gasoline Sulfur Control Requirements, U.S. EPA, December 1999, EPA 420-R-99-023.
Adjusted to 2002 dollars using Chemical Engineering Plant Cost Index.

                                         7-174

-------
                                                  Estimated Costs of Low-Sulfur Fuels
                                      Table 7.2.2-19
          Capital Expenditures for Gasoline and Highway Diesel Fuel Desulfurization
                                     (SBillion, $2002)a
Calendar
Year
2002
2003
2004
2005
2006
2007
2008
2009
2010
Tier 2 Gasoline
Sulfur Program
1.76
1.15
0.88
0.61
0.16
0.06
0.06
0.02

2007 Highway
Diesel Program


1.82
3.03
1.21

0.43
0.71
0.28
Total
1.76
1.15
2.70
3.64
1.37
0.06
0.49
0.73
0.28
   "2002 dollars obtained by use of Chemical Engineering
   for Tier 2 gasoline program (1997 dollars) and highway
Plant Annual Cost Index to adjust capital costs
diesel capital program (1999 dollars).
    The two diesel fuel programs have implementation dates of June 1 of various years for fuel
leaving the refinery.  For this start up date, we assumed that 30 percent of the capital cost was
expended in the calendar year two years prior to start up, 50 percent was expended in the year
prior to start up and the remaining 20 percent was expended in the year of start up. We repeated
this analysis for the final NRLM program. The results are summarized in Table 7.2.2-20 below.
                                          7-175

-------
Final Regulatory Support Document
                                     Table 7.2.2-20
                 Capital Expenditures for the Final NRLM Fuel Program with
               Tier 2 Gasoline Sulfur and 2007 Highway Diesel Fuel Programs
                                    (SBillion, $2002)
Calendar Year
2002
2003
2004
2005
2006
2007
2008
2009
2010
2011
2012
2013
2014
Final NRLM Fuel Program
Tier 2 and Highway
Diesel
1.76
1.15
2.70
3.64
1.37
0.06
0.49
0.73
0.28




NRLM Program



0.09
0.16
0.06
0.35
0.59
0.41
0.29
0.18
0.11
0.04
Total3
1.76
1.15
2.70
3.75
1.53
0.12
0.84
1.32
0.69
0.29
0.18
0.11
0.04
a2002 dollars obtained by use of Chemical Engineering Plant Annual Cost Index to adjust capital costs for Tier 2
gasoline program (1997 dollars) and highway diesel capital program (1999 dollars).
   As can be seen, capital investments peak in 2005 for the Tier 2 and Highway diesel
programs. The final NRLM program increases this peak by just $90 million, or about 2 percent.
Thereafter, capital requirements drop dramatically but peak a second time in year 2009 due to
the 15 ppm highway and nonroad standard. The second peak is less than 36 percent of the
capital outlays that occur in year 2005. Considering all programs, when capital investment
requirements are the highest, they are caused by the Tier 2 gasoline sulfur and 2007 highway
diesel fuel programs. Compared to Tier 2 and the hwy diesel program, the capital investment
requirements for the final NRLM fuel program are much smaller and are more  spread out over
time.

   Estimates of previous capital investments by the oil refining industry for the purpose of
environmental control are available from two sources: the Energy Information Administration
(EIA) and the American Petroleum Institute (API).  According to EIA, capital investment by the
24 largest oil refiners for environmental purposes peaked at $2 billion per year during the early
                                         7-176

-------
                                                 Estimated Costs of Low-Sulfur Fuels
1990's.JJ  Total capital investment by refiners for other purposes was in the $2-3 billion per year
range during this time frame. API estimates somewhat higher capital investments for
environmental purposes, with peaks of about $3 billion in 1992-1993 .^ Based on these two
sources, during the early 90's, the US refining industry invested over 20 billion dollars in capital
for environmental controls for their refining and marketing operations,  representing about one
half of the total capital expenditures made by refiners for operations.

   The capital required for the Tier 2 gasoline, 2007 highway diesel fuel and the final NRLM
fuel program is about 73 percent of the historic peak level of investment for meeting
environmental programs experienced during 1992-1994.52 Additionally, most of the capital
outlays for all of the about mentioned fuels programs are spread out over an eight  year time
period. Given that the capital required by the final NRLM fuel program contributes less than 2
percent to the required investment in the peak year of 2005, we do not expect that the industry
would have difficulty raising this amount of capital, although we recognize that it does require
the need to continue to raise and devote capital over a longer period of time.

   7.2.2.5 Other Cost Estimates for Desulfurizing Highway Diesel Fuel

   Two other studies  have estimated a cost of producing 15 ppm NRLM fuel, one by Mathpro
and another by Baker and O'Brien (BOB). These two studies are discussed below.

   Mathpro: For the Engine Manufacturers Association and with input by the American
Petroleum Institute, Mathpro used a notional refinery model to estimate the national average
costs of desulfurizing nonroad diesel fuel after implementation of the 15 ppm standard for
highway  diesel fuel. The cost estimate from this study is presented here and compared with our
costs.

   In a study conducted for the EMA, MathPro, Inc. first estimated the cost of desulfurizing
diesel fuel to meet a 15 ppm highway diesel fuel sulfur standard followed by two-step nonroad
standards of 500 ppm and 15 ppm.53'54 MathPro assumed that desulfurization will occur entirely
with conventional hydrotreating, and refining operations and costs were modeled using their
ARMS modeling system with technical and cost data provided by Criterion Catalyst Company
LP, Akzo-Nobel Chemicals Inc., and Haldor Topsoe, Inc. The Mathpro refinery model
estimated costs based on what Mathpro terms a "notional" refinery. The notional  refinery is
configured to be typical of the refineries producing highway diesel fuel for PADDs 1, 2, and 3,
and also represent the  desulfurization cost for those three PADDs based on the inputs used in the
refinery model. The Mathpro notional refinery model maintained production of highway diesel
fuel at their base levels.
   JJ "The Impact of Environmental Compliance Costs on U.S. Refining profitability," EIA, May 16, 2003.

   KK U. S. Petroleum Refining, Assuring the Adequacy and Affordability of Cleaner Fuels, A Report by the
National Petroleum Council, June 2000.

                                          7-177

-------
Final Regulatory Support Document
   Mathpro made several estimates in their study to size their diesel desulfurization units for
estimating the capital cost, and these estimates were similar to those included in our
methodology.  The calendar day volume was adjusted to stream day volume using a 10 percent
factor to account for variances in day-to-day operations, and another 10 percent to account for
variance in seasonal demand.  In addition, Mathpro applied a factor that falls somewhere in the
range of 1 to 8 percent for sizing the desulfurization unit larger for reprocessing off-spec
material to meet different sulfur targets. Since meeting a 500 ppm standard is not very stringent,
Mathpro likely assumed that a desulfurization unit will be sized larger by 1 to 4 percent.  For
meeting the 15 ppm standard,  which is relatively stringent compared with the 500 ppm sulfur
level studied, Mathpro likely assumed the desulfurization unit would be sized larger by 5 to 8
percent. On-site investment was adjusted to include offsite investment using a factor of 1.4.  In
the final report, capital costs were amortized at a 15 percent after-tax rate of return.

   The Mathpro cost study analyzed the costs to comply with the highway program based on 5
different investment scenarios. Before deriving the best nonroad desulfurization cost estimate
using the Mathpro cost study, we must describe the various investment scenarios.  The titles of
the scenarios are listed here:

   1. No Retrofitting - Inflexible
   2. No Retrofitting - Flexible
   3. Retrofitting - De-rate/Parallel
   4. Retrofitting - Series
   5. Economies of Scale

   Scenarios 1 and 2 do not allow retrofitting, which means the existing highway diesel
hydrotreater must be removed from service and a new grassroots unit desulfurizing untreated
distillate down to under  15 ppm takes its place. The difference between scenarios 1 and 2 is that
scenario 1 does not allow some flexibilities that may be available to the refining industry. One
flexibility is that the volume of hydrocracker units is not limited to the used capacity as listed in
the 1997 API/NPRA survey, but instead the throughput can be as much as 8 percent higher,
which is half the available capacity available in the API/NPRA survery.  Another flexibility is
that jet fuel exceeds specifications and instead of limiting the qualities to current levels, they are
instead allowed to become heavier by 0.5 API or by 3 points on the E375 distillation curve and
stay within the jet fuel specifications. Allowing jet fuel to get heavier allows the refinery model
to bring some of these lighter jet fuel blendstocks into the highway diesel fuel pool, which
lowers the desulfurization cost.  The flexibilities are allowed in the rest of the scenarios as well.

   Scenarios 3 and 4 allow taking advantage of the existing highway desulfurization unit by
keeping it in place and installing additional capital including additional reactor volume, which
allows the combined used and new capital to achieve the 15 ppm standard.
The difference between  scenarios 3 and 4 is that Scenario 3 derates the existing hydrotreater,
which reduces the volume treated by that unit so it can achieve 15 by itself; another unit being
fed by a low throughput is then added in parallel, which allows it to meet the 15 ppm standard.
Scenario 4 installs the new capital in series with the existing hydrotreater with both units
handling the entire feed  rate.

                                          7-178

-------
                                                 Estimated Costs of Low-Sulfur Fuels
    Scenario 5 allows the debottlenecking of existing capacity to treat a larger volume while
producing the same specifications. Scenario 5 also allows a single unit to be installed to handle
the desulfurization of multiple refineries in refining centers, which provides an important
economy of scale for the desulfurization investment costs to that group of refineries.

    While these various investment scenarios were devised to show how different investment
scenarios affect the cost for the HD2007 rule, they have implications for the nonroad rule as
well.  For meeting the standard for nonroad diesel fuel of 500 ppm, the used highway units freed
up in  Scenarios 1 and 2 can thus be converted over to nonroad service, which dramatically
reduces the capital cost of compliance; this supplements the existing nonroad capacity.
However, for Scenario 2, the installed grassroots capacity installed for the HD2007 rule
decreased after the capital was already installed and a larger volume of existing hydrotreating
capacity removed from highway desulfurization service was put into place to supplement the
nonroad hydrotreating capacity already in place.  For Scenario 3, the needed nonroad capacity is
formed by adding grassroots  capacity. For Scenario 4, the necessary  nonroad hydrotreating
capacity is formed by increasing the existing unit capacity used, relying on some expansion of
existing units and adding some processing unit capacity in series with existing capacity. The
nonroad hydrotreating capacity for meeting the 500 ppm standard is realized for Scenario 5
similar to Scenario 4, except  no expansion of existing units occurs, but instead more capacity
from existing highway units is relied upon.

    For meeting the 15 ppm cap sulfur standard for nonroad diesel fuel, the refinery model
invested in nonroad capital either along the  same lines as the 500 ppm case, or else invested
much differently.  For Scenario 1 and  2, the refinery model installed grassroots units only, even
replacing some existing hydrotreating capacity that was likely being used for some mild
desulfurization of nonroad diesel fuel. For Scenario 2, the volume of grassroots desulfurization
capacity was  slightly lower than Scenario 1, probably due to the increased flexibility granted by
the refinery model. For Scenario 3, the refinery model added some new grassroots unit capacity
compared with the 500 ppm case, probably derating the capacity of the remaining 500 ppm and
new 500 ppm capacity.  For Scenario 4, the refinery model added more series unit capacity and
more  expansion capacity. Finally for  Scenario 5, the refinery model increased the series
processing unit capacity and  added some expansion capacity.

    The new or existing hydrotreating capacity used for meeting the 500 ppm and 15 ppm
nonroad standards incremental to meeting the highway 15 ppm  sulfur standard is shown in
Table 7.2.2-21.
                                         7-179

-------
Final Regulatory Support Document
                                     Table 7.2.2-21
  Mathpro Capital Investments (bbl/day) for Desulfurizing Highway and Nonroad Diesel Fuel

Reference Case
Highway 15 ppm
Cap Std
Nonroad Meeting
a 500 ppm
Standard
Nonroad Meeting
a 1 5 ppm Standard

Existing Cap
Existing Unit
Expansion
De-rated
Series Unit
Grassroot Unit
Existing Unit
Expansion
De-rated
Series Unit
Grassroot Unit
Existing Unit
Expansion
De-rated
Series Unit
Grassroot Unit
No Retr
Inflex
34.9
8.2



30.2
16.5



30.1




50.4
No Retr
Flex
34.9
8.2



29.3
19.4



27.6




49.3
Retr
De-rate
34.9


17.8
15.4



17.8

23.7


17.8

26.5
Retr
Series
34.9
31.1


29.4

35.0
2.9

34.1

35.0
4.9

39.1

Econ of
Scale
34.9
31.1


29.4

38.0


34.0

38.0
1.9

39.1

   We next determined which Mathpro case best approximated the investment scenarios we are
using in our 500 ppm cost analysis, but we will summarize first summarize how our cost model
estimates investments will occur.  As described earlier in this section, some refineries will
comply with the highway HD2007 rule in 2006 by putting in a new hydrotreater and thus idling
an existing hydrotreater (i.e., 20 percent of the mixed highway and nonroad refineries that have a
distillate hydrotreater and comply with the highway requirements in 2006). Other refiners have
said that they will exit the highway market altogether, thus freeing up their existing  500 ppm
treater.  We believe that the refineries exiting the highway market would use these treaters to
desulfurize NRLM diesel fuel. Adding up the volumes from these two sources of existing
hydrotreating capacity, we  estimate that 30 percent of NRLM will be  desulfurized with existing
hydrotreaters. Furthermore, we estimated that 39 percent of NRLM fuel is already hydrotreated
and blended into high sulfur distillate. We project that this hydrotreating will continue with the
use of existing hydrotreaters. Thus, the fraction of NRLM diesel fuel meeting the 500 ppm
sulfur standard in 2007 with the use of existing capital is expected to be 69 percent.  The balance
of the NRLM volume, which comprises 31 percent, is expected to be desulfurized with a new
hydrotreater installed for startup in 2007.
                                         7-180

-------
                                                Estimated Costs of Low-Sulfur Fuels
   We examined the Mathpro investment cases to match the investment scenarios in our cost
analysis.  There were no cases that matched our scenario exactly, but we found two Mathpro
cases that, together, roughly matched our investment scenario.  The first is the No Retrofit
Inflexible case, which met the nonroad requirements exclusively through using existing capacity
(with half of it already in place before the standard applied, which matches our investment
scenario). The second case is the Retrofitting Derating case, which met the nonroad
requirements through new capital investment. Our analysis for complying with the 500 ppm
sulfur standard was based on 69 percent of the nonroad volume being produced by refineries
using existing hydrotreaters and 31 percent with new units, so the Mathpro costs were weighted
69 percent No Retrofit Inflexible costs and 31 percent Retrofit DeRate costs.

   We then examined the Mathpro 15 ppm cases to determine which would best match our 15
ppm scenario.  Since we already described the Mathpro cases for estimating the incremental cost
for going from meeting the 500 ppm standard to meeting the 15 ppm sulfur standard, we needed
identify the case which best matches our 500 ppm to 15 scenario. As discussed earlier in this
section, our 15 ppm scenario has new nonroad diesel fuel hydrotreating  units being installed in
2010. Since we estimated that 31 percent of the volume of NRLM in 2007 is complied with
using new units, we project that 31 percent of the NRLM diesel fuel would meet the 15 ppm
sulfur by revamping their new 2007 treaters. The balance of the NRLM volume are projected to
comply with the 15 ppm standard with grassroots units which are installed to desulfurize
uncontrolled distillate fuel down to 15 ppm, with an operating cost credit for the uncontrolled to
500 ppm step.  Of the Mathpro cases summarized above, the first two cases, which don't allow
revamps and either allow or don't allow operational flexibility, install grassroots units for
obtaining the 15 ppm standard. We decided to use Mathpro's case one,  since the second
Mathpro case apparently allowed backsliding in the highway grassroots units needed for
complying with the HD2007 rule when the 500 ppm standard was being met, which we don't
think is possible because the highway  investments will be too far along before the nonroad
program is finalized.

   Case one, however, needed to be adjusted to better model our projections on how refiners
would invest. Mathpro's case one was associated with the replacement  of the existing
hydrotreating capacity, all of which was likely used by the refinery model  for desulfurizing
nonroad down to 500 ppm. However, we believe 31 percent of the existing nonroad
desulfurization capacity can be revamped instead of having to be replaced.  Thus, we adjusted
the Mathpro capital costs to remove 31 percent of the grassroots hydrotreating capacity which
we believe would be revamped instead. We accomplished this by estimating what percent of the
capital costs is necessary for complying with 15 ppm standard and which portion was necessary
for replacing the expected portion of existing nonroad desulfurization capital.  The nonroad
diesel fuel volume needed to be treated in Mathpro's notional refinery model is 9 thousand
barrels per day.  According to Mathpro, the capital needed to be installed to treat the nonroad
pool down to 15 ppm is increased by  10 percent to handle peak throughput rates, and then by
another 10 percent to handle peak seasonal rates and then by another 8 percent to handle
reprocessing of off-spec batches. Thus, the 9,000 barrels per day nonroad volume is increased to
about 11,800 barrels per day, which represents Mathpro's estimated capital capacity. We
subtracted 11,800 bpd from the total volume of grassroots capacity added, which was 20,300

                                         7-181

-------
Final Regulatory Support Document
bpd, to yield a total of 8,500 barrels per day of replaced capital capacity; we assumed this will be
untreated to 500 ppm nonroad hydrotreated capacity. Since we projected that 69 percent of this
existing capacity to be replaced, with the 31 percent being new units in 2007 and not replaced,
we maintained 69 percent of 8,500 bpd, or an additional 5,865 barrels of the new nonroad
hydrotreating capacity. We therefore maintained 17,665 bpd of the original 20,300 bpd of
additional capacity added in Mathpro case one.  To estimate a revised cost for Mathpro's case
one we multiplied the capital charge by a ratio of 17,665/20,300.  No adjustment was necessary
for the variable operating cost.

   In addition to the differences and adjustments as described above, there are several other
differences between our cost analysis and the cost analysis made by Mathpro that were adjusted
or deserve mentioning. First, the MathPro costs as reported in their final report are based on a  15
percent return on investment (ROI) after taxes.  As stated above, our costs are calculated based
on a 7 percent ROI before taxes, so to compare  our cost analysis with the cost analysis made by
Mathpro, we adjusted the Mathpro costs to reflect the rate of return on capital investment that we
use.  Second, the MathPro estimate includes a cost add-on (called an ancillary cost) for
reblending and reprocessing offspec diesel fuel  or for storing nontreated diesel fuel. While this
is conceptually an appropriate adjustment to estimate the cost to the refining industry, it appears
that some of the reblending costs in the MathPro study appear to be transfer payments,LL not
costs. We did not include these costs in our cost comparison.  Third, MathPro assumed that all
new hydrogen demand is met with new hydrogen plants installed in the refinery, which does  not
consider the advantage of hydrogen purchased from a third party that can be produced cheaper in
many cases. As a result, their hydrogen cost may be exaggerated, which would tend to increase
costs. In fact, Mathpro's hydrogen is priced at $3.60 per million standard cubic feet ($/MSCF).
However the hydrogen costs in our analysis is about $2.70 per MSCF.  Finally, we note that the
MathPro study took into consideration the need for lubricity additives, but did not address costs
that might be incurred in the distribution system. When we compared out costs with Mathpro's,
we did not include any costs that would be incurred in the distribution system not even lubricity
additive costs. For comparing the aggregate capital costs, the Mathpro aggregate capital costs
for the chosen cases were adjusted using the undesulfurized nonroad, locomotive, and marine
diesel fuel volumes for 2007 and for undesulfurized nonroad diesel fuel for 2010.  The
undesulfurized volumes we used for making the adjustments are presented in Section 7.1. A
comparison of Mathpro's costs and our costs to desulfurize highway diesel fuel to meet a 500
ppm sulfur standard and then a 15 ppm sulfur standard is shown below in Table 7.2.2-22.
   LL A transfer payment is when money changes hands, but no real resources (labor, natural resources,
manufacturing etc.) are consumed.

                                          7-182

-------
                                                Estimated Costs of Low-Sulfur Fuels
                                     Table 7.2.2-22
             Comparison of Mathpro's and EPA's Refining Costs for Meeting a
                500 ppm and a 15 ppm Nonroad Diesel Fuel Sulfur Standard
       (7% ROI before taxes, no lubricity additive costs nor distribution costs included)
Fuel Standard
500 ppm Cap Std.
1 5 ppm Cap Std.
Incremental to 500 ppm
Std. *
Uncontrolled to 1 5 ppm
Type of Cost
Per-gallon Cost (c/gal)
Total Capital Cost (billion$)
Per-gallon Cost (c/gal)
Total Capital Cost (billion$)
Per-gallon Cost (c/gal)
Total Capital Cost (billion$)
Mathpro's Costs
No Advanced
Tech
2.1
580
3.9
2300
6.0
2870
EPA's Costs
Advanced Tech
in 2010
2.2
310
3.6
1970
5.8
2280
No Advanced
Tech
2.2
310
4.9
2420
7.1
2730
 : Fully phased-in costs in 2014
   Baker and O'Brien Study: The Baker and O'Brien (BOB) study was conducted for API to
estimate the costs and supply impacts of two possible NRLM fuel control programs.  BOB first
estimated how refiners would respond to future diesel  fuel requirements absent any NRLM fuel
controls.  These requirements included EPA's 2007 highway fuel program and the California and
Texas fuel programs.^ This was referred to as the Base Case in the report.  The two NRLM
fuel programs evaluated were:

1) Study  Case- One step NRLM fuel program:
       15 ppm cap for all NRLM fuel in 2008

2) Sensitivity Case- Two step NRLM fuel program:
       500 ppm cap for all NRLM fuel by 2008
       15 ppm cap for nonroad fuel in 2010

BOB initiated their study prior to the NPRM, so they did not know exactly what NRLM fuel
program would be proposed.  Their two cases were designed to bracket what they believed were
likely possible proposals.  As it turns out, the final NRLM fuel program reflects portions of both
cases.  The final NRLM fuel program is a two step program, like the sensitivity case.  The final
15 pm cap applies to all NRLM fuel like the study case, though in the final NRLM fuel program,
significant volumes of NRLM fuel can be 500 ppm fuel resulting from contamination in the
distribution system.
   MM BOB assumed that refiners producing diesel fuel for Texas would have to produce the same fuel as currently
being produced in California. In addition, they assumed that 100 percent of highway fuel sold in both states would
have to meet a 15 ppm cap starting in mid-2006.
                                         7-183

-------
Final Regulatory Support Document
   The fuel supply impacts of the BOB study are addressed in Section 4.6.3.1 of the Summary
and Analysis of Comments document.  The focus here is on their projected cost to produce low
sulfur NRLM fuel.  BOB did not estimate the cost of producing 500 ppm NLRM fuel under the
Sensitivity Case.  They only stated that roughly 300,000 bbl per day of 500 ppm diesel fuel
could be produced essentially for free from idled highway hydrotreaters. This is very similar to
our findings in Section 7.2.1 above.  The primary difference is that we only consider the capital
cost to be free, since these hydrotreaters would not be operated (i.e., zero operating cost) absent
this NRLM fuel program.

   BOB developed cost estimates for 15 ppm NRLM fuel, but not for 15 ppm fuel produced
under the highway program. BOB did not use projected costs per gallon of producing 15 ppm
fuel to predict which refineries would likely produce 15 ppm fuel under either the highway or
NRLM programs. Instead, as outlined in their report, BOB made first assumed that refiners
would defer USLD capital investment whenever they had a reasonable alternative, such as
selling heating oil or exporting high  sulfur diesel fuel. BOB also assumed that some refiners
would not be able to raise or justify the capital expenditures for ULSD and would discontinue
operations.  In addition, BOB predicted that a sizeable number of domestic refineries would
close as a result of the highway and NRLM fuel programs.  As a result of these assumptions,
BOB projected that domestic refiners would only produce 200,000-300,000 bbl per day of 15
ppm NRLM fuel  out their estimated  demand of 700,000 bbl per day.
   BOB presented their cost estimates for 15 ppm NRLM for both the study and sensitivity
cases. As the study case most closely approximates the fully implemented final NRLM program,
we chose to compare our fully implemented NRLM costs to those of BOB's study case. As
BOB only presented per gallon costs graphically, we present both sets of cost estimates in
graphical form in Figure 7.2.2.5-1.
                                        7-184

-------
                                                Estimated Costs of Low-Sulfur Fuels
                                    Figure 7.2.2-8-1

            Comparison of EPA and O'Brien NRLM Desulfurization Costs to a 15 ppm
                                       Standard
    20
    18
    16
  o 12
  oi 10
  Q_
  (fl

  J  8
              100
                      200
                               300
                                       400       500
                                          MBPSD
                                                        600
                                                                 700
                                                                         800
                                                                                  900
                                        -EPA 	OBRIEN
   As mentioned above, BOB projects relatively little 15 ppm NRLM fuel production compared
to demand, and compared to that projected by EPA. From the BOB report, the difference in
volume is caused by sizeable exports of high sulfur distillate from coastal refineries and a
number of refinery shutdowns in the Midwest and Mountain regions of the U.S.  From the
information provided in the report, we cannot determine which refineries were projected to
export or close. Therefore, we cannot perform any more precise comparison of per gallon costs
than that provided in Figure 7.2.2.5-1. From this comparison, it is quite possible that BOB and
EPA are projecting roughly similar costs for many individual refineries.  In this case, the
difference between the two cost curves would be the removal of a number of larger refineries
with EPA-projected costs in the 4-8 cent per gallon range. This would compress the EPA cost
curve into something more like the BOB cost curve. Even with this assumption, it appears that
BOB is projecting that some refineries with NRLM production volumes of 10-15,000 bbl per day
have costs in the 10-17 cent per gallon range.  While above 10 cents per gallon, all the refineries
in the EPA analysis have very small NRLM production volumes.

   While BOB does not present any further detail regarding their per gallon costs, they do
provide additional detail regarding their capital and operating costs. Regarding capital costs,
BOB's projected capital investments by domestic refiners are summarized in Table 7.2.2-23.
                                         7-185

-------
Final Regulatory Support Document
                                     Table 7.2.2-23
                        BOB and EPA Capital Cost of Desulfization

Capital Investment
($ billion)
Production Volume
(lOOObblperday)*
Investment per
bbl/day production
BOB
Highway
ISppmNRLM
(Study Case)
7.15
0.55
2934
208
$2437
$2644
EPA
Highway
ISppmNRLM
6.18
2.28
3605
841
$1714
$2711
* BOB volumes are in 2010, EPA volumes are in 2014
   The primary figures is this table that we want to focus on are those in the last column, which
show the capital cost to add one barrel per day of 15 ppm fuel production capacity.  As can be
seen, BOB projects significantly higher costs for 15 ppm highway fuel.  This is likely due to
different  assumptions regarding the probability that refiners will be able to revamp their existing
500 ppm  hydrotreater to produce 15 ppm fuel. However, this difference will not be discussed
further, as the cost of 15 ppm highway fuel is not the focus of this comparison.

   Moving to NRLM fuel, BOB's estimated capital cost for 15 ppm NRLM fuel production are
within a few percent of EPA's projection on a per barrel of production basis. BOB assumes that
all refiners will use conventional hydrotreating technology to produce 15 ppm highway and
NRLM fuel. EPA projects that roughly 60 percent of the volume of 15 ppm NRLM fuel
produced will utilize advanced technology for the step from 500 ppm to 15 ppm. This would
tend to reduce EPA's projected capital costs relative to those of BOB. However, our capital
costs include the cost of new hydrogen plants and expanded sulfur plant capacity. BOB treated
hydrogen as a utility and simply included the full cost of producing hydrogen (operating plus
capital costs) in the price that refiners would have to pay. This difference would tend to increase
our capital costs relative to those of BOB. Finally, BOB's source of capital costs was a study by
the National Energy Technology Laboratory for EIA. NETL used many of the same sources
which we cite in Section 7.2.1 for the capital cost of conventional hydrotreating.  However,
NETL increased their capital cost projections from these sources by 33 percent, based on
discussions with refiners.  (The details of these discussions were not  provided, so no comment
can be made about the appropriateness of this adjustment.) Therefore, it is likely that BOB's
primary capital cost inputs for conventional hydrotreating are roughly 33 percent higher than
those described in Section 7.2.1 above. As the NETL study dates from mid-2001, it was unable
to incorporate later information, such as the successful operation of the Process Dynamics
IsoTherming demonstration unit.  Overall, we believe that our capital cost estimates are
                                         7-186

-------
                                                Estimated Costs of Low-Sulfur Fuels
reasonable in light of the BOB analysis.  First, for conventional hydrotreating, we used the same
primary cost inputs.  Second, the 33 percent adjustment by NETL was based on discussions with
refiners which we cannot evaluate. Third, it is appropriate to include advanced technologies
which have been demonstrated at the commercial level. Fourth, the inclusion of capital costs for
hydrogen plants and expanded sulfur plants provides a more complete estimate of the total
capital investment required by the refining industry and their suppliers.

   Regarding operating costs, hydrogen costs tend to dominate these costs.  Thus, we will focus
our comparison there.  Hydrogen costs are a function of the volume of hydrogen needed to
desulfurize a gallon of diesel fuel and the price of hydrogen. Regarding the former, BOB based
their hydrogen consumption estimates on a number of studies, including one which we cite in
Section 7.2.1 (Figures 31 in the BOB report).  One of these estimates, that made by IFF, projects
hydrogen consumptions over twice those of the other studies.  We evaluated this estimate in our
Draft and Final RIAs for the 2007 highway diesel rule, along with a number of other estimates.
There,  based on changes in other fuel properties, we determined that this estimate was based on
very conservative assumptions concerning the level of aromatic saturation and modest cracking
that would occur when desulfurizing diesel fuel to  7 ppm sulfur and decided not to use it any
further. As four out of five vendors projected that this level of saturation would not be
necessary, we decided not to incorporate this estimate into our cost methodology.

   The IFF estimates appear to have a significant impact on the BOB hydrogen consumption
estimates, as BOB's hydrogen consumption model over-predicts all of the other data used to
develop the model. Also, subsequent discussions with IFF staff indicate that their more recent
estimates (the original estimate was made prior to 2000) are more in line with those of the other
vendors.

   In Figure 9 of the BOB study, they present their estimated hydrogen consumption for three
different diesel fuel compositions for a grass roots  conventional hydrotreater designed to produce
15 ppm diesel fuel. We used our  methodology developed in Section 7.2.1 to estimate hydrogen
consumption for these same feeds for a grass roots hydrotreater.  Table 7.2.2-24 shows both the
EPA and BOB estimates of hydrogen consumption.

                                     Table 7.2.2-24
        EPA and BOB 15 ppm Hydrogen Consumption: Grassroots Diesel Hydrotreater
BOB Feed
Case
1
2
3
Feed Composition
100% Straight Run
50% Straight Run, 35% LCO
15%LCGO
70% LCO, 30% LCGO
Hydrogen Consumption, scf/bbl
EPA
240
582
1025
BOB
510
778
1091
                                         7-187

-------
Final Regulatory Support Document
   As can be seen, the BOB estimates are significantly higher than our estimates, particularly
for the 100 percent straight run distillate. We compared BOB's 510 scf/bbl estimate for this case
with the hydrogen consumptions which BOB presents in an appendix where it compares the
predictions of its hydrogen model to the vendor estimates (Figure 31 in the BOB report). There,
BOB shows five cases where the diesel fuel being hydrotreated is 100 percent straight run.  BOB
shows that its hydrogen model predicts hydrogen consumptions of 244-268 scf/bbl for these
feedstocks. This is roughly half that which they show in Figure 9. No explanation for this
discrepancy is presented in the report. However, if the hydrogen consumptions shown in BOB's
Figure 9 were actually used in their cost estimations, then they appeared to have over-estimated
hydrogen costs even compared to their own model validations.

   With respect to hydrogen costs, BOB assumed that hydrogen would cost twice the cost of
natural gas. They did not state whether this was on a Btu basis, or a scf basis. Other information
presented in the study implies that it was on a scf basis. As BOB  projected future natural gas
prices of roughly $3 per mmBTU (equivalent to $3 per 1000 scf), this implies that BOB
projected hydrogen costs of $6 per 1000 scf. In Section 7.2.1, we describe how we estimate
hydrogen costs.  There, we use a future natural gas price of $4.15  per mmBtu, well above that
used by BOB. However, using this natural gas price, we estimate hydrogen costs of $2.20-3.90
per 1000 scf.  As described in Section 7.2.1, we base these  costs on a new hydrogen plant typical
of the size of hydrogen plants in the region today, or by an  even mix of new plants or third party
plats for the hydrogen  supplied in the Gulf Coast. We also adjusted for variations in natural gas
costs, typical plant capacities, location factors and off-site factors all differing according to the
region of the country in which the refinery is located. It is unclear where BOB obtained its rule
of thumb on hydrogen prices.  It may have been accurate when natural gas prices were much
lower than today and capital costs comprised a much larger percentage of total costs. However,
this rule of thumb does not appear to be appropriate at today's natural gas prices. Thus, it
appears, though one cannot be sure given the lack of detail  in the report, that BOB significantly
over-estimated hydrogen costs.

7.3 Cost of Lubricity Additives

   Our evaluation of the potential impact of the non-highway diesel sulfur standards on fuel
lubricity is described in Section 5.9. We conclude that the  increased need for lubricity additives
resulting from the these sulfur standards will be similar to that for highway diesel fuel meeting
the same sulfur standard.  In the HD2007 rule, we conservatively  estimated that all diesel fuel
meeting a  15 ppm sulfur standard will use lubricity additives at a cost of 0.2 cents per gallon.55
Consistent with the estimated cost from the increased use of lubricity additives in 15 ppm
highway diesel fuel, we have included a charge of 0.2 cents per gallon in our cost calculation to
account for the increased use of lubricity additives in 15 ppm NRLM diesel fuel.  This lubricity
additive cost applies to the affected NRLM diesel fuel pool beginning in 2010.

   In estimating lubricity additive costs for 500 ppm diesel fuel, we conservatively assumed that
if diesel fuel is required to have its lubricity improved through the use of additives, that the same
additive concentration will be needed both for 15 ppm and  for 500 ppm diesel fuel.  However,
the vast majority of 500 ppm diesel fuel does not require the use of lubricity additives.  We

                                         7-188

-------
                                                 Estimated Costs of Low-Sulfur Fuels
assumed that 5 percent of all 500 ppm diesel fuel would need a lubricity additive. Based on
these assumptions, we estimate that the cost of additional lubricity additives for the affected 500
ppm NRLM diesel fuel is 0.01 cents per gallon. The amount of lubricity additive needed
increases substantially as diesel fuel is desulfurized to lower levels. Also, based on the industry
input (see Section 5.9) it is likely that substantially less than 5 percent of 500 ppm diesel fuel
outside of California requires a lubricity additive.  We therefore believe 0.01 cents per gallon
represents a conservatively high estimate of the cost of lubricity additives for affected volume of
500 ppm nonroad, locomotive, and marine diesel fuel. Although the actual cost will likely be
considerably less, we have no information to better quantify the percentage of 500 ppm diesel
fuel currently treated with a lubricity additive or the appropriate additive treatment rate. The
0.01 cents per gallon cost for a lubricity additive applies to the affected non-highway diesel pool
(NRLM) until the 15 ppm sulfur standard takes effect in 2010.

   EM FOKS/AEO NRLM Fuel Demand Scenario:

   As discussed in Section 5.9, lubricity costs vary primarily with sulfur level, as the sulfur
level affects the degree of hydrotreating applied, which in turn results in changes to other fuel
properties which affect lubricity.  Thus, lubricity costs do not vary with implementation date or
type of diesel fuel market (i.e., highway, nonroad, locomotive or marine).  Thus, as the sulfur
level of various diesel fuels change under the alternative control options, the lubricity costs vary
accordingly. However, the cost per gallon for 500 ppm fuel will remain 0.01 cent per gallon and
the cost for 15 ppm fuel will remain 0.2 cent per gallon.

7.4 Cost of Distributing Non-Highway Diesel  Fuel

   A summary of the distribution costs that we project will result from the implementation of
the NRLM sulfur standards is contained in Table 7.4.-1. How we arrived at these cost estimates
is described in the following sections.
                                          7-189

-------
Final Regulatory Support Document
                                     TABLE 7.4.-1
                SUMMARY OF DISTRIBUTION COSTS (CENTS PER GALLON) *
Cause of Increase in Distribution
Costs
Distribution of Additional NRLM
Volume to Compensate for Reduction
in Volumetric Energy Content
Distillate Interface Handling
New Product Segregation as Bulk
Plants
Heating Oil and L&M Fuel Marker
Total
Time Period Over Which Costs Apply
2007-2010
0.08
0
0.1
0.01
0.2
2010-2012
0.1
0.4
0.1
0.02
0.6
2012-2014
0.1
0.4
0.1
0.01
0.6
After 20 14
0.1
0.8
0.1
0.01
1.0
 : Costs have been rounded to one significant figure.
7.4.1 New Production Segregation at Bulk Plants

    Section 5.4.1. evaluates the potential for additional product segregation in each segment of
the distribution system. As discussed in Section 5.5.1.2., approximately 1,000 bulk plants could
add an additional storage tank and demanifold their delivery truck(s) to handle an additional
diesel product.

    In its comments to the government/industry panel convened in accordance with the Small
Business Regulatory Enforcement Act (SBREFA), the Petroleum Marketers Association of
America (PMAA) stated that, depending on the location, the cost of installing a new diesel
storage tank at a bulk plant ranges from $70,000 to $100,000. To provide a conservatively high
estimate of the cost to bulk plant operators, we used an average cost of $90,000. This is
consistent with the information we obtained from a contractor working for EPA (ICF Kaiser) on
the installed cost of a 20,000-gallon storage tank, which is the typical tank size at bulk plant
facilities. Demanifolding of the bulk plant operators delivery truck involves installing an
internal bulkhead to make two tank compartments from a single compartment.  To help control
contamination concerns, we also estimated that an additional fuel delivery system will be
installed on the tank truck (i.e., that there will be a separate delivery system for each fuel carried
by the delivery truck). The cost of demanifolding a tank truck and installing an additional fuel
delivery system is estimated at $10,000, of which $6,000 is the cost of installing a new fuel
delivery system.56

    In the NPRM,  we estimated that each bulk plant that needed to install a new storage tank
would need to demanifold a single tank truck. Thus, the NPRM estimated the cost per bulk plant
would be $100,000. Fuel distributors stated that the assumptions and calculations made by EPA
                                         7-190

-------
                                                 Estimated Costs of Low-Sulfur Fuels
in characterizing costs for bulk plant operators seem reasonable. However, they also stated that
our estimate that a single tank truck would service a bulk plant is probably not accurate.  No
suggestion was offered regarding what might be a more appropriate estimate other than the
number is likely to be much greater.  Part of the reason why we estimated that only a single tank
truck would need to be demanifolded, is that we expected that due to the seasonal nature of the
demand for heating oil versus nonroad fuel, it would primarily only be at the juncture of these
two seasons that both fuels would need to be distributed in substantial quantities. We also
expected that the small demand for heating oil in the summer and the small demand for nonroad
fuel in the winter could be serviced using a single demanifolded truck.  The primary fuel
distributed during a given season would be distributed by single compartment tank  trucks.
During the crossover between seasons, bulk plant operators would switch the fuel to which such
single compartment tank trucks are used from nonroad to heating oil and back again.1"™
Nevertheless, we agree that some of the subject bulk plant operators would likely be compelled
to demanifold more that  a single tank truck.  Lacking additional specific information, we believe
that assuming that each bulk plant operator demanifolds three tank trucks will provide a
conservatively high estimate of the cost to bulk plant operators due to this rule.

   If all 1,000 bulk plants were to install a new tank and demanifold three tank trucks, the cost
for each bulk plant would be $120,000, and the total one-time capital cost would be
$120,000,000. To provide a conservatively high estimate of the costs to bulk plant operators, we
are assuming that all 1,000 bulk plants will do so. Amortizing the capital costs over 20 years,
results in a estimated cost for tankage at such bulk plants of 0.1 cents per gallon of  affected
NRLM diesel fuel supplied. Although the impact on the overall cost of the program is small, the
cost to those bulk plant operators who need to put in a separate storage tank may represent a
substantial investment.  Thus, we believe many of these bulk plants will search out other
arrangements to continue servicing both heating oil and NRLM markets such as an exchange
agreement between two bulk plants that serve a common area.

   The need for additional storage tanks at terminals to handle products produced from pipeline
interface is discussed in  Section 7.4.1.2. of this RIA.  Aside from the costs described above for
bulk plant operators, and those discussed in Section 7.4.1.2, we project that there will be no
substantial need for additional storage tanks or other facility changes to segregate additional
products.

   EIA FOKS/AEO Nonroad Fuel Volume Scenario:

   Using EIA nonroad fuel volumes rather than our primary fuel volume scenario which utilized
the EPA NONROAD model for nonroad fuel consumption does not affect our assessment of
product distribution patterns on which the above estimate of the costs to bulk plant  operators are
based.  Therefore, our estimate of the costs to bulk plant operators under the EIA nonroad fuel
volume scenario is the same as that under our primary fuel volume scenario.  However, the
   NN To avoid sulfur contamination of NRLM fuel, the tank compartment would need to be flushed with some
NRLM fuel prior to switching from carrying heating oil to NRLM fuel.

                                          7-191

-------
Final Regulatory Support Document
volume of affected NRLM to which these costs are attributed is reduced somewhat under the
EIA nonroad volume scenario, and consequently the cost per gallon is directionally higher than
under our primary fuel volume scenario. Nevertheless, because the costs are small, this does not
result in a material change to our estimate of 0.1 cents per gallon of affected NRLM diesel fuel
supplied.

   Because our assessment of product distribution patterns is not different under the EIA
nonroad volume scenario from that under our primary scenario, we also project that aside from
the costs described above for bulk plant operators, and those discussed in Section 7.4.1.2, there
will be no substantial need for additional storage tanks or other facility changes to segregate
additional products.

7.4.2 Reduction in Fuel Volumetric Energy Content

   We project that desulfurizing diesel fuel to 500 ppm will reduce volumetric energy content
(VEC) by 0.7 percent. The cost  of which is equivalent to 0.08 cent per gallon of affected NRLM
fuel. We project that desulfurizing diesel fuel to 15 ppm will reduce volumetric energy content
by an additional 0.5 percent. This will increase the cost of distributing fuel by an additional 0.05
cents per gallon, for a total cost of 0.13 cents per gallon of affected 15 ppm NRLM fuel.
Following is a discussion of how we arrived at these estimated costs.

   The reduction in VEC due to desulfurization of NRLM fuel to meet the standards in this rule
depends on the desulfurization process used. We project that conventional hydrotreating will be
the desulfurization process used  to desulfurize NRLM to meet the 500 ppm sulfur standard.
However, as discussed in Chapter 5, we project that new technology (Process Dynamics
Isotherming) will be used as well to desulfurize NRLM to meet the 15 ppm standard.  These
processes have different projected impacts on VEC, as discussed in Chapter 5.2. and shown in
Table 7.4-2.

                                      Table 7.4-2
       Impact of Desulfurization on the Volumetric Energy Content of Diesel Fuel
Process
Hydrodesulfurization
Process Dynamics
Isotherming
Overall for NRLM Pool
NRLM Fuel Volume Processed
500 ppm
Standard
100 %
0%
-
15 ppm
Standard
40%
60%
-
Reduction in VEC
High Sulfur to 500
ppm
0.7%
NA
0.7%
Reduction in VEC
500 ppm to 15 ppm
0.7%
0.4%
0.5%
   The difference between the price of non-highway diesel fuel to end-users and the price to
resellers provides an appropriate estimate of the cost of distributing non-highway diesel fuel.
The Energy Information Administration (EIA) publishes data regarding the price excluding taxes
of high-sulfur No. 2 diesel fuel to end-users versus the price to resellers. We used the five-year
                                         7-192

-------
                                                 Estimated Costs of Low-Sulfur Fuels
average of the difference between these two prices to arrive at an estimated typical cost of
distributing NRLM fuel to the end-user. In the NPRM, we used data from 1995 through 1999 to
arrive at an estimated distribution cost of 10 cents per gallon. For this final rule, we used 1997
through 2001 data to update this analysis.  The EIA data that we used to estimate the cost of
distributing NRLM fuel is presented in Table 7.4-3.

                                       Table 7.4-3
   Cost of Distributing High-Sulfur No. 2 Diesel Fuela (cents per gallon, excluding taxes)
Year
1995
1996
1997
1998
1999
2000
2001
Average of
5 Most Recent Years
Sales to Resellers
52.4
63.9
60.2
43.7
51.9
87.5
77.1
54.4
Sales to End Users
61.4
73.2
69.8
55.5
62.0
98.1
89.2
64.4
Difference Between Sales to End Users
and Sales to Resellers
9.0
9.3
9.6
11.8
10.1
10.6
12.1
10.8
a Energy Information Administration, Annual Energy Review 2003

   Based on the information in Table 7.4-3, we assumed a 10.8 cent per gallon cost of
distributing diesel for the purposes of estimating the increased distribution costs due to reduced
VEC. We derived our estimates of the increase in distribution costs under each step of the
NRLM sulfur program by multiplying the applicable percent reduction in VEC by 10.8 cents per
gallon.

   Since the difference in price at the refiner rack versus that at retail also includes some profit
for the distributor and retailer, its use provides a conservatively high estimate of distribution
costs. The fact that a slightly less dense (lighter, less viscous) fuel requires slightly less energy
to be distributed also indicates that this estimate is conservative.

   EIA FOKS/AEO NonroadFuel Volume Scenario:

   Using EIA nonroad fuel volumes rather than our primary fuel volume scenario which utilized
the EPA NONROAD model for nonroad fuel consumption does not affect our estimate of the
increased distribution costs related to the reduction in VEC.  Thus, the 0.08 and 0.13 cent per
gallon costs for 500 ppm  and 15 ppm fuel do change.
                                          7-193

-------
Final Regulatory Support Document
7.4.3 Handling of Distillate Fuel Produced from Pipeline Interface

    As discussed in Section 5.1, the shipment of 30 ppm gasoline, 15 ppm diesel fuel, jet fuel
and, in some cases, 500 ppm locomotive and marine fuel and high sulfur heating oil, will
produce commingled distillate fuel at the interfaces of each batch. In Section 5.1, we estimate
the volumes of each interface and how the fuel distribution system could dispose of each
interface in order to maximize profits (i.e., minimize costs). Basically, interfaces containing
some gasoline are presumed to go to existing transmix facilities.  The distillate fuel produced by
these transmix processors will contain a mixture of heavy naphtha, jet fuel and 15 ppm diesel
fuel. We project that this mixture will contain 500 ppm sulfur or less and can thus be sold as 500
ppm diesel fuel of high sulfur heating oil.

    The other interface which will not be able to be blended into either of the adjacent batches is
that between jet fuel and 15 ppm diesel fuel. In the Northeast and along the Colonial and
Plantation pipelines, we assume that this distillate interface will be added to the heating oil tank,
which will continue to be distributed throughout the distribution system. Elsewhere, we do not
believe that heating oil will be distributed in pipelines.  We assume the interface containing jet
fuel and  15 ppm diesel fuel will not be shipped to transmix processors. Interface processors
basically distill transmix into a lighter than average naphtha component and a lighter than
average distillate component.00 This distillate contains all of the original jet fuel and No. 2
distillate (both highway and high sulfur) fuel. Adding an interface consisting of jet fuel and No.
2 distillate to the current transmix tank and running this through a distillation column would only
result in all of this jet-distillate interface flowing to the bottoms of the column. The additional
distillate would also affect the operation of the distillation column, as they are typically designed
for a certain fraction of the feedstock going overhead.  Thus, we believe that it would be more
economical for terminals to segregate this No. I/No. 2 distillate interface from transmix in a
separate  storage tank.  As described in Section 7.1, we estimate that this interfacial material will
likewise meet a 500 ppm sulfur cap.  Thus, the terminal can ship this interface to consumers in
either the 500 ppm diesel fuel or heating oil markets.

    The disposition of this 500 ppm interface fuel  is described in  Section 5.1.  Generally, we
assumed that this material would be sold to the heating oil first, then into the 500 ppm highway
fuel market (through 2010), to the 500 ppm NRLM market (the nonroad fuel market through
2014), and finally into the L&M diesel market (after 2014). An exception to this applies in the
Northeast/Mid-Atlantic Area, where this interface cannot be sold into the nonroad fuel market
after 2010, nor into the L&M fuel market after 2012. If the volume  of this 500 ppm interface
exceeds the demand for 500 ppm diesel  fuel  and heating oil, then we assumed that it would  have
to be shipped back to a refiner and reprocessed to meet the 15 ppm cap.
    00 Normally, one thinks of transmix processing as separating transmix back into its original gasoline and
distillate components. However, the lighter compounds in original distillate fuel inevitably mix with the heavier
compounds in the original gasoline and lower the octane of this heavy gasoline dramatically. Due to the cost of
making up for this octane loss, transmix processors typically send the heavier gasoline compounds to the distillate
half of their product..

                                           7-194

-------
                                                 Estimated Costs of Low-Sulfur Fuels
    The cost of disposing of this 500 ppm distillate material will likely vary geographically,
depending on the size of the heating oil market. In the Northeast, the only cost of disposing of
this interface will be the value lost by selling former jet fuel and 15 ppm diesel fuel as heating
oil. This cost is already included in our refining costs, as there, we increased the volume of 15
ppm diesel fuel which had to be processed due to losses during distribution. We estimate that
about 80% of the diesel fuel shipped to PADD 1 is sold in areas with large heating oil markets.
In the remainder of the country, the heating oil market is more limited. Matching any high sulfur
heating oil and users of this fuel will be more difficult and costly in terms of transportation.

    Prior to mid-2010, 500 ppm interface can simply be added to the 500 ppm NRLM fuel
storage tank, which should exist at most terminals, or the 500 ppm highway fuel storage tank, if
this fuel is being stored at that terminal. Thus, there should be essentially no cost related to
disposing of this interface material.

    From mid-2010 through 2012, 500 ppm fuel can no longer be sold to the highway fuel
market. Also, we do not expect that small refiner 500 ppm nonroad fuel and 500 ppm L&M fuel
will be widely distributed. Thus, this interface material will require its own storage tank. The
500 ppm interface can be sold to users of NRLM fuel, as well as heating oil.  The only restriction
is that it cannot be used in nonroad equipment equipped with emission controls requiring 15 ppm
fuel, nor in nonroad engines in general within the Northeast/Mid-Atlantic Area. Most nonroad
fuel users only have one fuel storage tank on-site. Or, if they have more than one tank, it is
because their operations cover long distances (e.g., farms, quarries, etc.) and multiple tanks
reduce the time it takes to move the equipment to the refueling station. Thus, nonroad
equipment users which have purchased even one new  piece of equipment requiring 15 ppm fuel
will often desire to purchase 15 ppm fuel for all their equipment.  Thus, the number of NRLM
fuel users willing to accept 500 ppm fuel will gradually diminish from 2010 to 2014. This will
increase the distance that the fuel will have to be shipped to find a purchaser.

    We estimate  that the cost to store this 500 ppm fuel at a terminal will vary by terminal. At
those terminals able to receive jet fuel and 15 ppm diesel fuel from the heart of the pipeline
batches passing by it, the only distillate-distillate interface will be from washing lines to protect
jet fuel and diesel fuel quality. This material might be stored in a small tank, but will most likely
simply be added to the existing transmix tank. Thus, incremental storage costs will likely be
negligible, but transmix volume will increase.  Terminals near the end of pipeline or pipeline
branch will receive a relatively large volume of distillate-distillate interface.  Some of these
terminals will  likely be able to use the tank that was previously used to hold heating oil or 500
ppm NRLM fuel or the tank used to hold 500 ppm L&M diesel fuel from 2010-2012.  However,
in other cases it may require some new tankage. Economics will likely encourage the off-
loading at terminals with existing tankage. However,  proximity to a large 500 ppm market
(L&M fuel, heating oil) will also likely be a factor.

    Depending on the size of the tank, storage costs vary substantially. Smaller tanks can cost $5
per gallon of capacity, while very large tanks might only cost $20 per barrel ($0.5 per gallon).
Amortizing these costs over 15 years of weekly shipments of 60% of capacity at a 7% rate of
return, storage costs range from 0.2-1.6 cents per gallon in those cases requiring a new tank.  It is

                                          7-195

-------
Final Regulatory Support Document
not possible to estimate a precise distribution of tank sizes and thus, costs. We assume that the
availability of existing tankage will balance the need for smaller tanks on average and that the
average storage cost will be near the lower end of this range, 0.4 cents per gallon. In addition,
there is an inventory cost to have this stored fuel on hand. At a 7% rate of return, assuming that
the tank is half full on average, for fuel at $1 per gallon, the carrying cost is 0.1 cent per gallon.
Thus, the total storage  cost is roughly 0.5 cent per gallon.

    There is also the potential for increased storage costs  at transmix processing facilities.  The
increased volume of distillate-distillate interface added to transmix will likely be very small
relative to the total volume of gasoline-distillate interface. Thus, existing tankage should be
sufficient. However, currently, transmix processors often ship their distillate production into
tankage at terminals which are usually located adjacent to the processing facility.  After 2010,
the only 500 ppm fuel that would be stored at most of these terminals would be interface, and all
terminals after 2012, as discussed above. These terminals may have to increase their storage
capacity beyond that necessary to handle interface received directly from the pipeline and line
washing.  We project that the incremental cost to store this transmix interface will be the same
0.5 cent per gallon as that projected above for non-transmix interface.   Since all the distillate-
distillate interface will either be  stored as a distinct fuel at the terminal  or combined with
transmix and processed, the overall storage cost for all distillate-distillate interface is 0.5 cent per
gallon.

    We expect that there will be  an additional cost of shipping this 500  ppm fuel to those who
can use it. Nonroad fuel markets will likely be served by truck, as is the case today.  Locomotive
and most marine markets will likely be served by rail. Shipping this 500 ppm fuel will not have
the economies of scale of the current nonroad market or the future 15 ppm nonroad market.
Trucks will have to spend more time driving between stops or a smaller compartment will have
to be added to the tank. In either case, costs will increase. Rail shipments will also be smaller
than today, increasing handling costs. We estimate that the additional cost of delivering 500
ppm interface to these NRLM users without 2011 and later nonroad equipment will cost 1.5
cents per gallon.  This  cost is equivalent to increasing the shipping distance by 45 miles by truck
and 100 miles by rail.pp Combined with storage costs, distributing this  fuel to NRLM users will
cost 2.0 cents per gallon.

    In those cases where the 500 ppm interface is sold to the heating oil markets outside of the
Northeast, we expect that the costs will be larger. Heating oil users outside of the Northeast are
not evenly distributed geographically. The interface will also not be evenly distributed
geographically. Thus,  the interface may not be removed from the pipeline near the users of
heating oil.  Also, we expect that this fuel will have to be transported by truck. We project that
the additional mileage  will be roughly 85 miles and cost 3.0 cents per gallon. Combined with
storage costs, distributing this fuel to heating  oil users outside of the Northeast will cost 3.5 cents
per gallon.
    pp Trucking and rail costs of 0.035 and 0.012-0.2 cent per gallon, respectively from: "Costs/Impacts of
Distributing Potential Ultra Low Sulfur Diesel", Robert E. Cunningham, Thomas R. Hogan, Joseph A. Loftus, and
Charles L. Miller, Turner and Mason and Co. Consulting Engineers, February 2000.

                                           7-196

-------
                                                 Estimated Costs of Low-Sulfur Fuels
   Finally, there are some PADDs where the NRLM and heating oil markers are not large
enough to handle all of the 500 ppm interface generated.  In these cases, the interface will have
to be shipped back to a refinery by truck, reprocessed through the refiner's hydrotreater and
shipped back to the fuel market with the rest of the refiner's production. The storage cost of 0.5
cent per gallon at terminals and transmix operators will still apply, since it will still likely to be
less costly to keep this interface segregated from gasoline-distillate transmix. (Transmix will be
sent to transmix processors, while the jet-distillate interface will have to be sent to refineries
with excess hydrotreating capacity.)  We estimate that most of this distillate will be shipped
roughly 200 miles by rail and cost 3.0 cents per gallon. Desulfurizing this material to 15 ppm
will be technically  simple, since it will consist of heavy naphtha, jet fuel and 15 ppm diesel fuel.
The two lighter fuels do not contain any stearically hindered molecules. However, refiners
generally do not add material into the middle of their distillate production train. There will
likely be a tank storing diesel fuel prior to desulfurization, where straight run, LCO and other
cracked stocks are mixed.  However, there might not be easy access to this tank from outside of
the refinery.  Thus, we expect that the handling costs will far exceed the desulfurization costs.
We project a total cost for reprocessing of 4.5 cents per gallon. Finally, this re-processed fuel
must be shipped  out again, usually via pipeline. We project this last distribution cost to be 2
cents per gallon.  Thus, the total cost for interface which  must be reprocessed is 10 cents per
gallon.

   From mid-2012 through 2014, very little changes from 2010-2012.  The only change is that
downgraded distillate can no longer be sold to the L&M fuel market in the Northeast/Mid-
Atlantic Area. Instead this fuel shifts to the heating oil market. As this is a minor change, we
assume that all of the costs of distributing the downgraded distillate to the various markets from
2012-2014 remain the same as in 2010-2014.

   In 2014, when 500 ppm fuel can no longer be sold to  nonroad equipment users, we project
that the transportation distance to L&M fuel users will nearly double, as will the transportation
cost, to 2.5 cents per gallon.  Outside of PADDs 1 and 3,  we estimate that the downgraded
material will comprise 70-100% of the L&M market, so,  given the above methodology, the
downgraded material will have to move to nearly every L&M refueling site.  With storage costs
of 0.5 cents per gallon, the total cost of distributing  downgraded material to the L&M fuel
market will be 3.0 cents per gallon.

   Likewise, we project that the transportation distance to heating oil users will also increase.
However, we do not believe that these distances will double, because the increase in downgraded
material going to the heating oil market is smaller on a relative basis than for the L&M fuel
market. Thus, we project that the transportation distance to heating oil users will increase to
roughly 130 miles and cost 4.5 cents per gallon. With storage costs of 0.5 cents per gallon, the
total cost of distributing downgraded material to the heating oil market will be  5.0 cents per
gallon. The cost to reprocess distillate to meet a 15 ppm  cap will remain at 10  cents per gallon.

   In Section 7.1, we estimated the volume of downgraded jet fuel and diesel fuel which would
be sold to the nonroad, L&M and heating oil markets prior to the NRLM rule (Table 7.1.3-9),

                                          7-197

-------
Final Regulatory Support Document
from 2007-2010 (Table 7.1.3-14), from 2010-2012 (Table 7.1.3-17), from 2013-2014 (Table
7.1.3-18) and in 2014 and beyond (Table 7.1.3-19). We likewise estimate the volumes of fuel
which must be reprocessed to meet a 15 ppm cap. These volumes are summarized in Table
7.4.4, along with the cost per gallon of storing and shipping this interface to the various fuel
markets.
                                      Table 7.4.4
              Annual Costs Associated With Distribution of Distillate Interface
Jet-Distillate
Interface Sent to:
Volume Affected
(million gallons/yr)
Cost per Gallon
Annual Cost
(million)
Baseline
NRLM Market
Heating Oil Market
Reprocessed
Total
247
219
0
—
2.0 cents
3.5 cents
10.0 cents
—
$5
$8
0
$13
2010-2012
NRLM Market
Heating Oil Market
Reprocessed
Total
1,395
1,045
0
—
2.0 cents
3.5 cents
10.0 cents
—
$30
$32
0
$63
2012-2014
NRLM Market
Heating Oil Market
Reprocessed
Total
1,395
1,045
0
—
2.0 cents
3.5 cents
10.0 cents
—
$28
$37
0
$65
20 14 and beyond
NRLM Market
Heating Oil Market
Reprocessed
Total
1,336
885
335
—
3.0 cents
5.0 cents
10.0 cents
—
$40
$44
$34
$118
   Table 7.4.4 also shows the annual cost associated with each fuel market, which is simply the
product of the fuel volume and the cost per gallon (converted from cents to dollars). The annual
cost due to the NRLM rule from 2007-2010 is $47 million, which is the total cost of $61 million
less the $14 million cost occurring prior to the rule.  Likewise, the cost due to the NRLM rule in
                                         7-198

-------
                                                 Estimated Costs of Low-Sulfur Fuels
2010-2012, 2012-2014 and 2014 and beyond is $63, $65, and $102 million, respectively. The
total affected NRLM fuel volume is 12.4 billion gallons in 2010, 12.8 billion gallons in 2012 and
13.4 billion gallons in 2014 (all three figures represent fuel production and demand grown to
2014). Thus, these annual costs represent incremental costs of 0.40, 0.41 and 0.79 cent per
gallon from 2010-2012, 2012-2014, and 2014 and beyond, respectively.QQ

   We anticipate that there will be no other significant distribution costs associated with the
NRLM sulfur standards in this rule beyond those described in Sections 7.4.1, 7.4.2, and 7.4.3.
We do not expect the need for additional storage tanks beyond that discussed in Sections 7.4.1.,
and 7.4.3., or a significant increase in pipeline downgrade or transmix volumes beyond the
modest potential increase in tranmix volume discussed in Section 7.4.3. As discussed in Section
7.4.5., we are projecting costs associated with the need to install fuel marker injection equipment
at a limited number of refineries, transmix processors, and terminals

   Operators of bulk plants and tank trucks who previously handled only high-sulfur diesel fuel
will need to begin observing practices to limit sulfur contamination during the distribution of 500
ppm and 15 ppm diesel fuel. However, these practices are either well established or will be for
compliance with the 15 ppm highway standard in 2006.  Furthermore, they are primarily
associated with purging storage tanks and fuel delivery systems of high-sulfur diesel fuel before
handling 500 ppm and 15 ppm diesel fuel.  Training employees will be necessary to stress  the
importance of consistently and carefully observing practices to limit sulfur contamination.
However, we estimate the associated costs will be minimal. In addition, we are estimating that
most of the affected bulk plant operators will install dedicated storage tanks and truck delivery
systems.  This obviates the need for much of the cautionary actions necessary to limit sulfur
contamination when both low and high-sulfur diesel fuel is carried by the same marketer.

   As discussed in Section 5.6, the vast majority of the fuel distribution system (primarily
pipeline and terminal facilities) will already have optimized their facilities and procedures  to
limit sulfur contamination for distributing 15 ppm sulfur fuel due to the need  to comply with the
highway diesel fuel program in 2006. The costs associated with this optimization process  were
accounted for in  the HD2007 Regulatory Impact Analysis.57  Highway diesel  fuel and nonroad
diesel fuel meeting a 15 ppm sulfur specification will share the same distribution  system until
nonroad diesel fuel is dyed to  meet IRS requirements as it leaves the terminal. We therefore do
not expect any additional actions or costs to optimize the distribution system  to limit sulfur
contamination during the distribution of 15 ppm nonroad diesel fuel.

   EIA FOKS/AEO Nonroad Fuel Volume Scenario: We followed the same methodology  for
estimating downgrade-related distribution costs for this scenario as our primary fuel volume
scenario which utilized the EPA NONROAD model for nonroad fuel consumption.  Using EIA
nonroad fuel volumes, as described in Section 7.1 above, reduces the volume of NRLM fuel
demanded in each PADD, except PADD 3.  Consequently, the volumes of heating oil consumed
   QQ The increase in cost in 2014 is due to the inability to use downgraded material in the nonroad market.  If the
$105 million cost in 2014 is spread only over the nonroad fuel market, the cost per gallon is 1.0 cents.

                                          7-199

-------
Final Regulatory Support Document
increase everywhere except PADD 3. This reduces the contribution of the volume of
downgraded material to the NRLM and heating oil markets substantially. Particularly in PADD
2, instead of downgraded material comprising a major portion of the NRLM and heating oil
markets, it comprises roughly 33%. We believe that this will make it easier for terminals to find
heating oil consumers and reduce the transport distance to these users. Thus, for PADD 2, we
reduced the cost of distributing interface to the heating oil market to that of the NRLM or L&M
markets (depending on the time period), or 2 cents per gallon. However, the volume of NRLM
fuel over which the increased transportion costs are spread also decreases. The net result is that
the cost of distributing interface material from 2010-2014 remains unchanged at 0.4 cent per
gallon. However, the cost after 2014 decreases from 0.79 to 0.56 cents per gallon.

7.4.4 Fuel Marker Costs

    In the NPRM we estimated that the cost to blenders of the heating oil marker in bulk
quantities would translate to 0.2 cents per gallon of fuel treated with the  marker. This estimate
was based on the fee charged by a major pipeline to inject red dye at the IRS concentration into
its customers diesel fuel.  Conversations with marker manufactures prior to the publication of the
NRLM indicated that the cost to treat fuel with either of the markers considered in the NPRM
would be lower than the costs to treat non-highway diesel fuel with red dye to meet IRS
requirements.  We used this estimate because we lacked specific cost information on the
proposed marker,  there was uncertainty regarding  the specific marker that we would require, and
we believed that it provided a conservatively high estimate of cost for any of the markers under
consideration. Since the proposal, we received input from a major distributor of fuel markers
and dyes, regarding the cost of bulk deliveries of the specified fuel marker (solvent yellow 124)
to terminals which translates to a  cost of 0.03 cents per gallon of fuel treated with the marker.
The volume of heating oil that we expect will need to be marked has also decreased substantially
from that estimated in the NPRM due to the provisions applicable in the Northeast/Mid-Atlantic
Area and Alaska.  We estimate that 1.4 billion gallons of heating oil will be marked annually, for
an annual marker  cost of $425,000.^ In the NPRM, this marker cost applied to heating oil for
just three years, but then continued on for another four years for locomotive and marine diesel
fuel. Under this final rule, the marker requirement for locomotive and marine diesel  fuel is
applicable only from 2010 though 2012, and only  outside of the Northeast/Mid-Atlantic Area
and Alaska.  However, the marker requirements for heating oil continues indefinitely.

    The NPRM projected that there would be no capital costs associated with the proposed
marker requirement.  We proposed that the marker would be added at the refinery gate, and that
the current requirement that non-highway fuel be dyed red at the refinery gate be made
voluntary. Thus, we believed that the refiner's additive injection equipment that is currently
used to inject red dye into off-highway diesel fuel  could instead be used  to inject the fuel marker.
As  a result of the allowance provided in this final rule  that the marker may be added  at the
terminal rather than the refinery gate, and our reevaluation of the conditions for dye injection at
    RR The costs of the marker requirement for L&M diesel fuel are discussed at the end of this section.

                                         7-200

-------
                                                 Estimated Costs of Low-Sulfur Fuels
the refinery, we are now assessing capital costs for terminals and refiners related to compliance
with the marker requirements.

   Except for fuel that is distributed directly from a refiner's rack, this final rule allows the
marker to be added at the terminal rather than at the refinery (see Section IV.D. of the preamble
for a discussion of the fuel marker requirements).ss We expect that except for fuel dispensed
directly from the refinery rack, the fuel marker will be added to at the terminal to avoid the
potential for marked fuel to contaminate jet fuel in during distribution by pipeline. Terminals
that need to  inject the fuel marker will need to purchase a new injection system, including a
marker storage tank and a segregated line and injector for each truck loading station at which
fuel that is required to contain the marker is dispensed. Terminals will still be subject to IRS red
dye requirements, and thus will not be able to rededicate such injection equipment to inject the
fuel marker. Due to concerns regarding the need to maintain a visible evidence of the presence
of the fuel marker, this final rule also contains a requirement that any fuel which contains the
fuel marker  also contains visible evidence of red dye.  Furthermore, there is little chance to adapt
parts of the red dye injection system (such as the feed lines and injectors) for the alternate
injection of red dye and the fuel marker due to concerns that fuel which must not contain the
marker might become contaminated with the  marker.

   We received information from various sources to estimate the cost of installing new injection
equipment to handle the heating oil marker.  Our first source of information was the Independent
Fuel Terminal Operators Association (IFTOA). IFTOA stated  that the cost for new additive
injection equipment would be $40,000 per loading arm used to deliver heating oil to tank trucks
with the cost for some terminals being as much as $250,000 (for 6-7 loading arms).

   We also  sought information from manufacturers of additive injection equipment.  Titan
industries and  Lubrizol, leading manufacturers of such equipment, provided information on the
uninstalled cost of the necessary hardware which is summarized in the following Table 7.4.S.58.

                                       Table 7.4.5
                     Uninstalled Cost of Additive Injection Hardware
Item
500 gallon Skid Storage Tank
Rack Mounted Pump Assembly
Chemical Injector
Total
Cost
$3,700 - $8,000
$5,000 - $9,000!
$2,500-$2,900
$11,200-$19,900
1. Depending on whether a single or a double pump assembly is used.  The second pump serves
as a back-up.
   SSA refinery rack functions similar to a terminal in that it distributes fuel by truck to wholesale purchaser
consumers and retailers.

                                          7-201

-------
Final Regulatory Support Document
    The lower end tank cost was more consistent with our previous experience regarding tank
costs. Consequently we elected to use $4,000 as a reasonable estimate of the uninstalled cost of
an additive storage tank.  We elected to use the higher cost estimate of $9,000 for the pump
assembly because we believe that many additive blenders would wish to have a double pump
assembly to prevent their fueling arm from being shut down when maintenance must be
performed on the primary pump.  This also provides something of a conservatively high cost
estimate. We also elected to use $3,000 as the estimated uninstalled cost of an injector unit for
this same reason. This results in an total uninstalled cost of $16,000 for the equipment necessary
to equip one injection loading arm: $13,000 for the tank and pump, and $3,000 for each injector.

    We  estimated the installed costs by two means. Our primary means was to apply the rule for
such projects of multiplying the equipment costs by 2 to arrive at the installed cost and then by
increasing this result by an  additional 50 percent to ensure that the estimated cost would be
sufficient to account for areas in the U.S. where labor costs are higher that the average (such as
the Northeast). Since the Northeast/Mid-Atlantic Area was defined to exclude terminals in the
Northeast from the marker requirement, this step might be expected to provide a conservatively
high estimate of installation costs for those facilities that do need to install new injection
equipment.  Following this  method results in an estimated installed cost of the equipment
necessary to provide marker injection at one loading arm of $50,000 ($40,000 for the tank and
pump assembly, and $10,000 for the injector assembly.  Thus, for each additional loading arm at
a terminal the cost would increase by $10,000. As a double check on these results we employed
an in-house expert to estimate the time required of various skilled tradesmen at their respective
hourly pay rates: e.g. instrumentation specialist, welder, welder's helper,  concrete installer,
engineer, and laborers. The estimate that we arrived at using this means supported the estimates
described above.  We believe that these estimates are more accurate than  those provided by
IFTOA, and therefore are using them to calculate the costs under this rule.

    Terminal operators expressed concern regarding the potential burden  of installing new
additive injection equipment. In response to these comments, this rule includes provisions that
exempt  terminal operators from the fuel marker requirements in a geographic "Northeast/Mid-
Atlantic Area" and Alaska.11 These provisions provide that any heating oil or 500 ppm sulfue
L&M diesel fuel produced by a refiner or imported that is delivered to a retailer or wholesale-
purchaser consumer inside the Northeast/Mid-Atlantic Area and Alaska does not need to contain
the marker.  The Northeast/Mid-Atlantic Area was defined to include the region where the
majority of heating oil in the country is projected to continue to be supplied though the bulk
distribution system (the Northeast and Mid-Atlantic). The vast majority of heating oil
consumption in the U.S. will be within the Northeast/Mid-Atlantic Area.  Outside of the
   "Small refiner and credit high sulfur NRLM will not be permitted to be sold in the area where terminals are not
required to add the fuel marker to heating oil and 500 ppm sulfur L&M diesel fuel produced by refiners or imported
(the "Northeast/Mid-Atlantic Area"). See Section IV.D. of the preamble. See Section 5.5.1.4 regarding our
determination of the boundary of the Northeast/Mid-Atlantic Area to minimize the number of facilities that would
need in to install new injection equipment for the fuel marker and to limit the volume of fuel that will need to be
marked.

                                          7-202

-------
                                                 Estimated Costs of Low-Sulfur Fuels
Northeast/Mid-Atlantic Area, we expect that only limited quantities of heating oil will be
supplied, primarily from certain refiner's racks.  Based on our analysis of the number of
refineries that we expect will continue to produce heating oil and information from transmix
processors on the number of such facilities, we estimate that 30 refineries and transmix processor
facilities outside of the Northeast/Mid-Atlantic Area will distribute heating oil from their racks
(in limited volumes) on a sufficiently frequent basis to warrant the installation of a marker
injection system at a total one time cost of $1,500,000.

   Terminals outside of the Northeast/Mid-Atlantic Area will mostly be located in areas without
continued production and/or bulk shipment of heating oil.  Consequently, any high sulfur diesel
fuel they sell will typically be NRLM. Terminals located within the Northeast/Mid-Atlantic
Area will not need to mark their heating oil, except for those few that choose to ship heating oil
outside of the Northeast/Mid-Atlantic Area.  The terminals most likely to install marker injection
equipment will therefore be those in states outside the Northeast/Mid-Atlantic Area with modest
markets for heating oil after the implementation  of this program.

   A few terminals inside the Northeast/Mid-Atlantic Area and near the border may choose to
install marker injection equipment so that they can serve customers outside of the
Northeast/Mid-Atlantic Area. However, based on our review of the proximity of terminals
inside the Northeast/Mid-Atlantic Area to potential heating oil markets outside of the
Northeast/Mid-Atlantic Area, we project that no more than 15 terminals will be induced to do so.
Given the relatively low level of the potential demand for marked heating oil, we believe that the
boundary area terminals that install marker injection equipment would provide for the loading of
marked heating oil into trucks at only one loading bay (at $50,000 per terminal).

   Some terminals outside  of the Northeast/Mid-Atlantic Area that are supplied by the pipeline
system which supplies the Northeast/Mid-Atlantic Area are likely to carry heating oil.
Considering the relatively low volume of heating oil demand in the states in which these
terminals are located, we estimate that only 15 terminals in this area will choose to install marker
injection equipment so they can handle heating oil. We believe that such terminals would likely
feel the need to have two loading bays at which marked heating oil could be delivered to a truck.
Considering the added cost  of a second injection station, the cost of new injection equipment
would be $60,000 for each of these terminals. Except for heating oil distributed from these
terminals, we project that the small quantities of fuel that are sold as heating oil outside  of the
Northeast/Mid-Atlantic Area will often meet a 500 ppm sulfur specification.1111 Therefore, we
expect that the other terminals outside of the  Northeast/Mid-Atlantic Area will typically not need
to distribute marked heating oil. For the infrequent instances in where terminals do receive >500
ppm fuel that they wish to distribute as heating oil (rather than blending it down to meet a 500
ppm standard using 15 ppm diesel fuel) we except that the terminal operator will elect to add the
marker by hand, thereby avoiding the cost of installing new additive injection equipment.
However, to provide a conservatively high estimated cost, we assumed that an additional 30
   uu Fuel sold as heating oil outside of the Northeast/Mid-Atlantic Area will primarily be generated as a by-
product of the distribution of 15 ppm diesel fuel by pipeline.

                                          7-203

-------
Final Regulatory Support Document
terminals outside of the Northeast/Mid-Atlantic Area will install new equipment to allow the
injection of fuel marker at one truck loading bay (at $50,000 per terminal).

   In analyzing the various situations as discussed above, we project that fewer than 60
terminals nationwide will choose to install injection equipment to add the marker to heating oil
at a total cost of $4,150,000.  The total capital cost to refiners and terminals to install injection
equipment to add the marker to heating oil is estimated to be $5,650,000.  Thus, the
Northeast/Mid-Atlantic Area provisions in this rule minimize the number of terminals that will
need to install additive injection equipment and its associated cost to comply with the fuel
marker requirements.

   Because heating oil is being marked to prevent its use in NRLM engines, for the purposes of
estimating the impact of the marker requirement on the cost of the NRLM program we have
spread the cost of adding the marker to heating oil over NRLM diesel fuel. Amortizing the
capital costs of marker injection equipment over 20 years,  results in an estimated cost of just
0.006 cents per gallon of affected NRLM diesel fuel  supplied.  Spreading the cost of the marker
for heating oil over the volume of affected NRLM fuel results in an estimated cost of 0.003
cents per gallon of affected NRLM fuel.  Adding the amortized cost of the injection equipment
and the cost or the marker results in a total estimated cost of the marker requirement for heating
oil in this rule of 0.01 cents per gallon of affected NRLM fuel.

   In addition to heating oil, 500 ppm L&M fuel produced at refineries must also be marked
from 2010 to 2012. As discussed in Section 7.2.2, we project that 6 refineries will produce this
fuel. These refineries will have to install equipment to mark the fuel, unless they already have
the equipment to mark heating oil. We assume that all 6 refineries will have to install new
equipment. We do not expect that 500 ppm L&M fuel will be distributed by common carrier
pipeline. Thus, it can be marked at the refinery and shipped to the final user by rail, truck or
barge already marked. Therefore, we expect that very few terminals will add marking equipment
exclusively for this fuel. To cover the few terminals that could do so, we have increased the
number of new marking installations to 15. At $60,000, the total capital  cost is $900,000. The
cost of the marker is 0.03 cent per gallon of marked fuel.  As described in Appendix 8B, we
estimate that 2.975 billion gallons of 500 ppm L&M fuel will be produced in 2011.  Thus, the
cost of marking two years of 500 ppm L&M fuel production will be $1.875 million. Amortizing
the $900,000 capital cost over 2 years  of 15 and 500  ppm NRLM fuel production at 7 percent
before taxes and adding in the marker  costs yields a cost of 0.01 cents per gallon of NRLM fuel
over this two year period for the marker requirement for L&M diesel fuel.

   EIA FOKS/AEO Nonroad Fuel Volume Scenario:

   Since using EIA nonroad fuel volumes rather than our  primary fuel volume scenario (which
utilized the EPA NONROAD model for nonroad fuel consumption) does not affect our
assessment of product distribution patterns, our projections of the number of facilities that will
need to install new injection equipment is the same under both scenarios.  However, there are
two factors that do have the potential to affect our per gallon cost estimate. The heating oil
volume under the EIA nonroad volume scenario is greater than that under our primary volume

                                         7-204

-------
                                               Estimated Costs of Low-Sulfur Fuels
scenario and the NRLM volume is smaller than under our primary volume scenario. The greater
volume of heating oil under the EIA volume scenario means that it is likely that the volume of
heating oil marked would be larger relative to our primary scenario, and the volume of NRLM to
which this cost (and the capitol cost of the injection equipment) would be attributed would be
smaller.  Both of these criteria direct!onally increase the per gallon marker costs under the EIA
volume scenario relative to our primary volume scenario. Because of these changes, the cost of
adding the marker increases to 0.02 cent per gallon of affected NRLM diesel fuel supplied. The
cost of marking L&M fuel stays at 0.01 cent per gallon from 2010-2012.

7.4.5 Distribution and Marker Costs Under Alternative  Sulfur Control Options

   EIA FOKS/AEO NonroadFuel Volume Scenario:

   The distribution and marker costs assuming a reduced volume  of nonroad fuel demand,
resulting from deriving this demand from  information in EIA's FOKS and AEO 2003  reports are
summarized in Table 7.4-6 below. The derivation of each cost component was discussed in the
previous sub-sections of Section 7.4.

                                    TABLE 7.4-6
 DISTRIBUTION COSTS FOR EIA FOKS/AEO FUEL DEMAND SCENARIO (CENTS PER GALLON)
Cause of Increase in Distribution Costs
New Product Segregation as Bulk Plants
Distribution of Additional NRLM Volume to
Compensate for Reduction in Volumetric
Energy Content
Distillate Interface Handling
Heating Oil and L&M Fuel Marker
Total
Time Period Over Which Costs Apply
2007-2010
0.1
0.08
0
0.03
0.2
2010-2014
0.1
0.1
0.4
0.03
0.6
After 20 14
0.1
0.1
0.6
0.03
0.8
 : Costs have been rounded to one significant figure.
   Other Fuel Control Options: The other fuel control options analyzed in this Final RIA are: 1)
500 ppm NRLM cap in 2007 with no subsequent control to 15 ppm, and 2) the proposed fuel
program of 500 ppm NRLM in 2007 and 15 ppm nonroad fuel in 2010.  The distribution costs
for the 500 ppm NRLM only program are the same as those for the final NRLM fuel program in
2007.
                                        7-205

-------
Final Regulatory Support Document
   Under the proposed fuel program, the distribution costs are essentially the same as those for
the final rule when the costs are spread over all NRLM fuel. However, when the costs of
distributing downgraded distillate are assigned to the only 15 ppm nonroad cap, as this is the
incremental step in fuel control which causes these costs, the cost per gallon is of higher.  In this
case, the cost from 2010-2014 and in 2014 and beyond increase to 0.54 and 1.0 cent per gallon,
respectively. In this case, the cost assigned to L&M fuel of distributing downgraded distillate is
zero.

7.5 Total Cost of Supplying NRLM Fuel Under the Two-Step Program

   The estimated refining, additive, and distribution costs from Sections 7.2 - 7.4 for the final
NRLM fuel program and the other fuel control options considered are summarized in Table 7.5-
1.  Estimated costs during the various phases of these programs are also shown. Note that these
fuel costs include the impacts of the small-refiner provisions. Also, in the case of the final
NRLM fuel program, we spread the downgrade distribution costs across all NRLM fuel from
2010-2012, even though L&M fuel is still at 500 ppm. We did so to avoid a higher apparent cost
of 15 ppm nonroad fuel from 2010-2012 than from 2012-2014.  However, in the case of the
proposed NRLM fuel program, we  assigned all of the downgrade distribution cost to nonroad
fuel, since the long term standard for L&M fuel is 500 ppm in this scenario. These cost
estimates do not include the costs associated with testing, labeling,  reporting, and recordkeeping
to  satisfy the compliance assurance provisions  of the final rule, but these costs are small enought
such that they would not change the values in Table 7.5-1  due to round-off.
                                         7-206

-------
                                       Estimated Costs of Low-Sulfur Fuels
                              Table 7.5-1
Summary of Fuel Costs for NRLM Fuel Control Options (cents per gallon, $2002)
Option
Final Rule
Proposed NRLM
Program: 500 ppm
NRLM in 2007, 15 ppm
Nonroadin2010
500 ppm NRLM in 2007
only (no 1 5 ppm fuel
control)
Final Rule with NRLM
Volume Derived from
EIA FOKS/AEO
Reports
Specification
500 ppm NRLM
5 00 ppm NRLM
5 00 ppm NRLM
1 5 ppm Nonroad
15 ppm NRLM
15 ppm NRLM
5 00 ppm NRLM
500 ppm L & M
500 ppm L & M
1 5 ppm Nonroad
1 5 ppm Nonroad
5 00 ppm NRLM
5 00 ppm NRLM
5 00 ppm NRLM
5 00 ppm NRLM
5 00 ppm NRLM
1 5 ppm Nonroad
15 ppm NRLM
15 ppm NRLM
Year
2007-10
2010-12
2012-14
2010-12
2012-14
2014+
2007-10
2010-14
2014+
2010-14
2014+
2007-10
2010+
2007-10
2010-12
2012-14
2010-12
2012-14
2014+
Refining
Costs
(c/gal)
1.9
2.7
2.9
5.0
5.6
5.8
1.9
2.7
2.7
5.0
5.2
1.9
2.0
1.9
2.8
3.0
5.0
5.6
5.7
Distribution &
Additive Costs
(c/gal)
0.2
0.6
0.6
0.8
0.8
1.2
0.2
0.2
0.2
1.0
1.4
0.2
0.2
0.2
0.6
0.6
0.8
0.8
1.2
Total
Costs
(c/gal)
2.1
3.3
3.5
5.8
6.4
7.0
2.1
2.9
2.9
6.0
6.6
2.1
2.2
2.1
3.4
3.6
5.8
6.4
6.9
                                 7-207

-------
Final Regulatory Support Document
    Our projected total cost for supplying 500 ppm fuel is slightly less than the historical price
differential between 500 ppm highway diesel fuel and uncontrolled high-sulfur diesel fuel. This
differential has averaged about 2.5 cents per gallon for the five-year period from 1995 to 1999.
Market prices may be either higher or lower than the societal costs estimated here as discussed in
the next section.  Thus, such  comparisons can only be considered approximate.  The primary
reason that our projected costs for 500 ppm NRLM fuel might be lower than those for highway
fuel is the ability to use existing hydrotreaters which are no longer being used to produce 500
ppm highway fuel in the 2007-2010 timeframe.

7.6 Potential Fuel Price  Impacts

    Transportation fuel prices are dependent on a wide range of factors, such as world crude oil
prices, economic activity at the national level, seasonal demand fluctuations, refinery capacity
utilization levels, processing  costs (including fuel-quality specifications), and the cost of
alternative energy sources (e.g., coal, natural gas).  Only a few of these factors, namely fuel
processing costs and refinery capacity utilization, may be affected by the NRLM fuel program.

    Fuel processing and distribution costs will clearly be affected due to the cost of desulfurizing
NRLM diesel fuel to either the 500 or 15 ppm sulfur cap.  Refinery utilization levels may be
affected as the capacity to produce 500 ppm or 15 ppm NRLM diesel fuel will depend on
refiners' investment in desulfurization capacity.  The potential impact of increased fuel
processing and distribution costs on the prices is assessed below.  The impact of the NRLM fuel
program on refinery utilization levels is beyond the scope of this analysis.  In the long run,
refiners will clearly invest to produce adequate volumes of NRLM diesel fuels, as well as other
distillate fuels. In the shorter term, the issue of refiners' adequate investment in desulfurization
capacity is addressed in Section 5.9.

    Two approaches to projecting future price impacts are evaluated here.  The most direct
approach to estimating the impact of the NRLM fuel program on prices is to observe the price
premiums commanded by similar products in the marketplace.  This is feasible for 500 ppm
NRLM diesel fuel, as both 500 ppm highway diesel fuel and high-sulfur diesel fuel are both
marketed today.  As discussed in Section 7.2.2 above, the historical price premium of 500 ppm
highway diesel fuel is 2.5 cents per gallon over that of high-sulfur distillate.  As this premium is
almost identical to our projected average total cost of the supplying 500 ppm NRLM diesel fuel,
it represents one reasonable estimate of the future price impact of the 500 ppm NRLM diesel fuel
standard.

    It is not possible to use this methodology to project the price impact of the 15 ppm nonroad
diesel fuel cap. Only a very limited amount of diesel fuel meeting a 15 ppm sulfur cap is
currently marketed in the United States. This fuel is designed to be used in vehicle fleets
retrofitted with particulate traps.  The fuel is produced in very limited quantities using equipment
designed to meet the current  EPA and California highway diesel fuel standards. It is also much
more costly to distribute due  to its extremely low volume. Thus, the current market prices for 15
ppm diesel fuel in the United States are not at all representative of what might be expected in
2010 and 2012 under the NRLM program.

                                         7-208

-------
                                                  Estimated Costs of Low-Sulfur Fuels
   A greater volume, though still not large quantities, of 10 ppm sulfur diesel fuel is currently
being sold in Europe.  The great majority of this fuel is Swedish Class 1 (so-called City) diesel
fuel, which is effectively a number one diesel fuel with very low aromatic content. The low
aromatic specification significantly affects the cost of producing this fuel. Also, this fuel is
generally produced using equipment not originally designed to produce 10 to 15 ppm sulfur fuel.
Thus, as in the United States, the prices paid for this fuel are not representative of what will
occur in the United States in 2010 and 2012. We therefore did not attempt to use current fuels,
which have sulfur levels similar to the standards in this final rule, to evaluate our cost estimate
for meeting the 15  ppm standard.

   The other approach to project potential price impacts utilizes the projected costs to meet the
500 ppm and 15 ppm NRLM fuel sulfur caps. Both sulfur caps will affect fuel processing and
distribution costs across the nation. (The exception will be California, where we presume that
sulfur caps at least as stringent as those in this final rule will already be in effect.) However,
these costs appear to vary significantly from  region to region. Because of the cost of fuel
distribution and limited pipeline capacities (pipelines  are the most efficient means of transporting
fuel), the NRLM fuel markets (and those for other transportation fuels) are actually regional in
nature.  Price differences can and usually do  exist between the various regions of the country.
Because of this, we have performed our assessment of potential price impacts on a regional
basis. For the regions in our analysis, we have chosen PADDs.  Practically speaking, there are
probably more than five fuel markets in the United States with distinct prices. However,
analyzing five distinct refining regions appears to provide a reasonable range of price impacts
without adding precision that significantly exceeds our ability to project costs.

   We  made one exception to the PADD structure. PADD 3 (the Gulf Coast) supplies more
high-sulfur distillate to PADD 1, particularly the Northeast, than is produced by PADD 1
refineries. Two large pipelines connect PADD 3 refineries to the Northeast, the Colonial and the
Plantation.  Because of this low-cost transportation connection, prices between the two PADDs
are closely linked.  We therefore combined our price analysis for PADDs 1 and 3.

   As mentioned above, it is very difficult to predict  fuel prices, either in the short term or long
term. Over the past three years, transportation fuel prices (before excise taxes) have varied by a
factor of two. Therefore, we have avoided any attempt to project absolute fuel prices.  Because
of the wide swings in absolute fuel prices, it is very difficult to assess the impact of individual
factors on fuel price.  The one exception is the price of crude oil, for two reasons.  One, the cost
of crude oil is the dominant factor in the overall  cost of producing transportation fuels. Two, the
pricing  of almost all crude oils is tied to the "world" market price of crude oil.  While the cost of
producing crude oil in each region of the world is independent of those of other crude oil,
contract prices are  tied to crude oils traded on the open market,  such as West Texas Intermediate
and North Sea Brent crude oils. Thus, as the price of world crude oil climbs, the price of
gasoline and  diesel fuel climb across the United  States, and vice versa.  There is also a very
rough correlation between refinery capacity utilization levels and fuel price. However, an
unusually high availability of imports can cause  prices to be relatively low despite high refinery
capacity utilization rates in the United States.
                                          7-209

-------
Final Regulatory Support Document
   For example, fuel prices, as a function of crude oil price, have varied widely over the past
decade. Refiner records supplied to EIA indicate that refiners' net refining margin has ranged
from a low of $0.45 per barrel in 1992 to a high of 2.78 per barrel in 2001.59 Thus, fuel prices
have varied between being so low that refineries are barely covering their cash expenses to high
enough to justify moderate cost increases in refining capacity (but not new refineries). The
NRLM program will very unlikely have a major impact on factors such as these. Thus,
projecting the likely price impact of the NRLM program is highly speculative. The best that can
be done is to develop a wide range of potential price impacts indicative of the types of conditions
that have existed in the past.

   In order to do this, we developed three projections for the potential impact of the  NRLM
program on fuel prices.  The lower end of the range assumes a very competitive NRLM fuel
market with excess refining capacity. In this case, fuel prices within a PADD are generally low
and reflect only incremental operating costs. Consistent with this assumption, we project that the
price of NRLM diesel fuel within a PADD will increase by the operating cost of the refinery
with the highest operating cost in that PADD.  This assumes that the refinery facing the highest
operating cost in producing NRLM diesel fuel  is setting the price of NRLM diesel fuel before
this rule. This may or may not be the case.  If not, the price increase may be even lower than that
projected below. Under this "low -cost" set of assumptions, the refiner with the highest
operating cost will not recover any of his invested capital related to desulfurizing NRLM diesel
fuel, but all other refiners will recover some of their investment^ Note that this scenario is only
viable  in the short run, since refineries need to  recover both operating and fixed costs in the long
run.

   The mid-range estimate of price impacts can be termed the "full-cost" scenario.  It assumes
that prices within a PADD increase by the average refining and distribution  cost within that
PADD, including full recovery of capital (at the societal rate of return of 7 percent per annum
before taxes). This scenario represents a case where there is full cost pass through to consumers
under a competitive market setting. It should be noted that there are instances when this full-cost
scenario produces lower costs than the maximum operating cost scenario. This occurs when the
bulk of the low sulfur fuel can be produced at a relatively low cost compared to a few refineries
facing  relatively high operating costs.

   Under this full-cost price scenario, lower cost refiners will recover their  capital investment
plus economic profit, while those with higher than average costs will recover some of their
invested capital, but not all of it (i.e., at a rate of return lower than 7 percent annually).

   The high-end estimate of price impacts assumes a NRLM fuel market that is constrained with
respect to fuel production capacity. Prices rise to the point necessary to encourage additional
desulfurization capacity.  Also, prices are assumed to remain at this level in  the long  term,
meaning that any additional desulfurization capacity barely fulfills demand and does  not create
    vv Theoretically, some refiners might recover all their invested capital if their operating costs were sufficiently
lower than those of the high cost refiner. However, practically, in the case of desulfurizing NRLM diesel fuel, this is
highly unlikely.

                                          7-210

-------
                                                 Estimated Costs of Low-Sulfur Fuels
an excess in capacity that would tend to reduce prices. However, prices should not increase
beyond this level in the long run, as this would encourage the construction of additional
desulfurization capacity, lowering prices. Consistent with this, prices within a PADD increase
by the maximum total refining and distribution cost of any refinery within that PADD, including
full recovery of capital (at 7 percent per annum before taxes).  All other refiners will recover
more than their capital investment.

    Table 7.6-1 presents the refining costs for the four phases of the NRLM fuel program under
the three potential price scenarios.

                                      Table 7.6-1
                 NRLM Fuel Refining Costs by Region (cents per gallon)

Maximum Operating Cost
Average Total Cost
Maximum Total Cost
500 ppm Sulfur Cap: Nonroad, Locomotive and Marine Diesel Fuel (2007-2010)
PADDs 1 and 3
PADD 2
PADD 4
PADDS
2.7
2.8
3.5
1.0
1.6
2.8
3.3
1.3
4.3
3.6
5.9
1.3
500 ppm Sulfur Cap: Nonroad, Locomotive and Marine Diesel Fuel (2010-2012)
PADDs 1 and 3
PADD 2
PADD 4
PADDS
2.3
2.9
3.9
1.6
3.7
2.9
8.9
2.8
5.0
3.8
8.9
2.9
500 ppm Sulfur Cap: Nonroad, Locomotive and Marine Diesel Fuel (2012-2014)
PADDs 1 and 3
PADD 2
PADD 4
PADDS
2.7
2.7
3.9
2.2
2.5
3.7
9.0
3.5
5.9
5.7
9.0
4.2
15 ppm Sulfur Cap: NRLM Fuel (2010-2012
PADDs 1 and 3
PADD 2
PADD 4
PADDS
4.7
5.0
7.1
3.6
4.6
7.1
11.6
4.3
8.5
8.5
12.7
4.3
15 ppm Sulfur Cap: NRLM Fuel (2012-2014)
PADDs 1 and 3
PADD 2
PADD 4
PADD 5
4.8
6.4
7.0
3.6
4.8
7.8
11.7
4.3
8.6
10.0
12.7
4.3
15 ppm Sulfur Cap: NRLM Fuel (fully implemented program: 2014 +)
PADDs 1 and 3
PADD 2
PADD 4
PADD 5
6.5
6.4
7.0
3.9
5.1
7.8
11.8
5.6
8.6
10.0
12.7
6.0
                                         7-211

-------
Final Regulatory Support Document
   Table 7.6-2 shows these same cost projections including distribution and lubricity additive
costs. The wholesale price of high-sulfur distillate fuel has varied widely even over the past
twelve months. The March 2003 heating oil futures price alone has ranged from 60-110 cents
per gallon since early 2002. Assuming a base cost of NRLM fuel of one dollar per gallon, the
increase in NRLM fuel prices will be equivalent to the price increase in terms of cents per gallon
shown below.

                                       Table 7.6-2
           Range of Possible Total Diesel Fuel Price Increases (cents per gallon)"

Maximum Operating Cost
Average Total Cost
Maximum Total Cost
500 ppm Sulfur Cap: Nonroad, Locomotive and Marine Diesel Fuel (2007-2010)
PADDs 1 and 3
PADD2
PADD4
PADDS
2.9
3.0
3.7
1.2
1.8
2.5
3.5
1.5
4.5
3.8
6.1
1.5
500 ppm Sulfur Cap: Nonroad, Locomotive and Marine Diesel Fuel (2010-2012)
PADDs 1 and 3
PADD2
PADD4
PADDS
2.9
3.5
4.5
2.2
4.3
3.5
9.5
3.4
5.6
4.4
9.5
3.5
500 ppm Sulfur Cap: Nonroad, Locomotive and Marine Diesel Fuel (2012-2014)
PADDs 1 and 3
PADD2
PADD4
PADDS
3.3
3.3
4.5
2.8
3.1
4.3
9.6
4.1
6.5
6.3
9.6
4.8
15 ppm Sulfur Cap: NRLM Fuel (2010-2012
PADDs 1 and 3
PADD2
PADD4
PADDS
5.5
5.8
7.9
4.4
5.4
6.8
12.4
5.1
9.3
9.3
13.5
5.1
15 ppm Sulfur Cap: NRLM Fuel (2012-2014)
PADDs 1 and 3
PADD2
PADD4
PADDS
5.6
7.2
7.8
4.4
5.6
8.5
12.5
5.1
9.4
10.8
13.5
5.1
15 ppm Sulfur Cap: NRLM Fuel (fully implemented program: 2014 +)
PADDs 1 and 3
PADD2
PADD4
PADDS
7.7
7.6
8.2
5.1
6.3
7.9
13.0
6.8
9.8
11.2
13.9
7.2
 Notes: a At a wholesale price of approximately $1.00 per gallon, these values also represent the percentage
 increase in diesel fuel price.
                                          7-212

-------
                                                 Estimated Costs of Low-Sulfur Fuels
   There are a number of assumptions inherent in these price projections.  First, both the lower
and upper limits of the projected price impacts described above assume that the refinery facing
the highest compliance costs is currently the price setter in their market. If this is not the case,
the price impacts would be lower than those shown in the previous tables.  Many factors affect a
refinery's total costs of fuel production.  Most of these factors, such as crude oil cost, labor costs,
age of equipment, etc., are not considered in projecting the incremental costs associated with
lower NRLM diesel fuel sulfur levels. Thus, current prices may very well be set in any specific
market by a refinery facing lower incremental compliance costs than other refineries. This point
was highlighted in a study by the National Economic Research Associates (NERA) for AAM of
the potential price impacts of EPA's 2007 highway diesel fuel program.ww In that study, NERA
criticized the above referenced study performed by Charles River Associates, et. al. for API,
which projected that prices will increase nationwide to reflect the total cost faced by the U.S.
refinery with the maximum total compliance cost of all the refineries in the U.S. producing
highway diesel fuel.  To reflect the potential that the refinery with the highest projected
compliance costs under the maximum price scenario is not the current price setter, we included
the mid-point price impacts above. It is possible that even the lower limit price impacts are too
high, if the conditions exist where prices are set based on operating costs alone.  However, these
price impacts are sufficiently low that considering even lower price impacts was not considered
critical to estimating the potential economic impact of this rule.

   Second, we assumed in some cases that a single refinery's costs could affect fuel prices
throughout an entire PADD.  While this is a definite improvement over analyses which assume
that a single refinery's costs could affect fuel prices throughout the entire nation, it is still
conservative, since one refinery's fuel can rarely have such a widespread influence.  For
example, Chicago and Detroit have experienced unusually high gasoline prices at times over the
past 4 years, but prices in St. Louis, Cincinnati, Minneapolis, etc. were not similarly  affected.
High cost refineries are more likely to have a more limited geographical impact on market
pricing than an entire PADD. In many cases, high cost refiners are able to operate profitably
because they are in a niche location where transportation costs limit competition.

   Third, by focusing solely on the cost of desulfurizing NRLM diesel fuel, we assume that the
production of NRLM  diesel fuel is independent of the production of other refining products,
such as gasoline, jet fuel and highway diesel fuel. However, this is clearly not the case. Refiners
have some  flexibility to increase the production of one product without significantly affecting
the others, but this flexibility is quite limited. It is possible that the relative economics of
producing other products could influence a refiner's decision to increase or decrease the
production of NRLM  diesel fuel under the fuel program in this rule. It is this price response that
causes fuel supply to match fuel demand. And, this response in turn could increase or decrease
the price impact relative to those projected above.
   ™ "Potential Impacts of Environmental Regulations on Diesel Fuel Prices," NERA, for AAM, December
2000.

                                          7-213

-------
Final Regulatory Support Document
   Fourth, all three of the above price projections are based on the projected cost for U.S.
refineries of meeting the NRLM fuel sulfur caps. Thus, these price projections assume that
imports of NRLM fuel, which are currently significant in the Northeast, are available at roughly
the same cost as those for U.S. refineries in PADDs 1 and 3.  We have not performed any
analysis of the cost of lower sulfur caps on diesel fuel produced by foreign refiners.  However,
there are reasons to believe that imports of 500 and 15 ppm NRLM diesel fuel will be available
at prices in the ranges of those projected for U.S. refiners.

   One recent study analyzed the relative cost of lower sulfur caps for Asian refiners relative to
those in the U.S., Europe and Japan.^ It concluded that costs for Asian refiners will be
comparatively higher, due to the lack of current hydrotreating capacity at Asian refineries.  This
conclusion is certainly valid when evaluating lower sulfur levels for highway diesel  fuels which
are already at low levels in the U.S., Europe and Japan and for which refineries in these areas
have already invested in hydrotreating capacity. It appears to be less valid when assessing the
relative cost of meeting lower sulfur standards for NRLM fuels and heating oils which are
currently at much higher sulfur levels in the U.S., Europe and Japan.  All refineries face
additional  investments to remove sulfur from these fuels and so face roughly comparable control
costs on a per gallon basis.

   One factor arguing for competitively priced imports is the fact that refinery utilization rates
are currently higher in the U.S. and Europe than in the rest of the world.  The primary issue is
whether overseas refiners will invest to meet tight sulfur standards for U.S., European and
Japanese markets. Many overseas refiners will not invest, instead focusing on local, higher
sulfur markets.  However, many overseas refiners focus on exports. Both Europe and the U.S.
are moving towards highway and nonroad diesel fuel sulfur caps in the 10-15 ppm range.
Europe is currently and projected to continue to need to import large volumes of highway diesel
fuel. Thus, it seems reasonable to expect that a number of overseas refiners will invest in the
capacity to produce  some or all of their diesel fuel at these levels.  Many overseas refiners also
have the flexibility to produce 10-15 ppm  diesel fuel from their cleanest blendstocks, as most of
their available markets have less stringent sulfur standards. Thus,  there are reasons to believe
that some capacity to produce 10-15 ppm diesel fuel will be available overseas at competitive
prices. If these refineries were operating well below capacity, they might be willing to supply
complying product at prices which only reflect incremental operating costs.  This could hold
prices down in areas where importing fuel is economical. However, it is unlikely that these
refiners could supply sufficient volumes to hold prices down nationwide. Despite this
expectation, to be conservative, in the refining cost analysis conducted earlier in this chapter, we
assumed no imports of 500 ppm or  15 ppm NRLM diesel fuel.  All 500 ppm and 15  ppm NRLM
fuel was produced by domestic refineries.  This raised the average and maximum costs of 500
ppm and 15 ppm NRLM diesel fuel and increased the potential price impacts projected above
beyond what would have been projected had we projected that 5-10 percent of NRLM diesel fuel
will be imported at competitive prices.
   xx "Cost of Diesel Fuel Desulfurization In Asian Refineries," Estrada International Ltd., for the Asian
Development Bank, December 17, 2002.

                                          7-214

-------
      Estimated Costs of Low-Sulfur Fuels
7-215

-------
Final Regulatory Support Document
Chapter 7 References

1. Fuel Oil and Kerosene Sales 2001, Energy Information Administration, Office of Oil and Gas,
November 2002.

2. NESCAUM, Characterization of Construction Equipment Activity, Assistance Agreement No.
X-991575-01, unpublished draft data.

3. Office of Highway Information Management, A Guide to Reporting Highway Statistics,
Federal Highway Administration, U.S. Department of Transportation, 1997.

4. Personal communication with Ralph Erickson, Office of Highway Policy Information, Federal
Highway Administration, U.S. Department of Transportation, August 2001.

5. Personal communication with Dan Walzer, Energy Information Administration, September
2001.

6. Personal communication with Dan Walzer, Energy Information Administration, September
2001.

7. Personal communication with Mark Stehly, Assistant Vice President Environmental,
Burlington Northern Santa Fe Railroad, September 2001.

8. Personal communication with Dan Walzer, Energy Information Administration, September
2001.

9. Petroleum Marketing Annual 2001, Energy Information Administration, Office of Oil and
Gas, September 2002.

10. U.S. Petroleum Refining: Assuring the Adequacy and Affordability of Cleaner Fuels,
National Petroleum Council Committee on Refining, June 2000.

11. Confidential discussion with pipeline companies, October 2003.

12. Annual Energy Outlook 2003, Energy Information Administration, Office of Integrated
Analysis and Forecasting, January 2003.

13. "Summary and Analysis of the Highway Diesel Fuel 2003 Pre-compliance Reports", EPA
420-R-03-103, October 2003.

14. Johnson,  Jeff, Boeing Company, "Sulfur in Jet Fuel," Presentation to the Sulfur Workshop,
QinetiQ and Farnborough, September 2 & 3, 2002.

15. "Regulatory Impact Analysis: Heavy-Duty Engine and Vehicle Standards and Highway
Diesel Fuel Sulfur Control Requirements," EPA420-R-00-026, December 2000, Docket A-2001-
28, Document No. II-A-01.
                                        7-216

-------
                                               Estimated Costs of Low-Sulfur Fuels
16.Confidential Information Submission from Diesel Desulfurization Vendor A, August 1999.

17.UOP Information Submission to the National Petroleum Council, August 1999.

18."The Lower it Goes, The Tougher it Gets," Bjorklund, Bradford L., UOP, Presentation at the
National Petroleum Council Annual Meeting, March 2000.

19.U.S. Petroleum Refining Draft Report, Appendix H, National Petroleum Council, March 30,
2000.

20. "Regulatory Impact Analysis: Heavy-Duty Engine and Vehicle Standards and Highway
Diesel Fuel Sulfur Control Requirements," EPA420-R-00-026, December 2000, Docket A-2001-
28, Document No. II-A-01.

21. Chemical Engineering Plant Cost Index, Chemical Engineering, February 2003.

22. Interpolation of the hydrogen consumption from the desulfurization data by Vendors A and
B.

23.Conversation with Jim Kennedy, Manager Project Sales, Distillate and Resid Technologies,
UOP, November 2000.

24. Conversation with Tim Heckel, Manager of Distillate Technologies Sales, UOP, March 2000.

25.Conversation with Tom W. Tippett et al, Refining Technology Division, Haldor Topsoe,
March 2000.

26. Very-Low-Sulfur Diesel Distribution Cost, Engine Manufacturers Association, August 1999.

27.Moncrief, T. L, Mongomery, W. D., Ross, M.T., Charles River Associates, Ory, R. E.,
Carney, J. T., Baker and O'Brien Inc., Ann Assessment of the Potential Impacts of Proposed
Environmental Regulations on U.S. Refinery Supply of Diesel Fuel, Charles River and Baker
and O'Brien for the American Petroleum Institute, August 2000.

28. Christie, David A.,  Advanced Controls Improve Reformulated Fuel Yield and Quality, Fuel
Reformulation, July/August 1993.

29. Personal conversation with Debbie Pack, ABB Process Analytics Inc., November 1998.

30. Ackerson, Michael; Skeds, Jon, Presentation to the  Clean Diesel Independent Review Panel,
Process Dynamics and  Linde Process Plants,  July 30, 2002.

31. Conversation with Jon Skeds, Director of Refining, Linde BOC Process Plants, August 26,
2003.

32. Conversation with Jon Skeds of Linde Process Plants at the 2002 National Petrochemical and
Refiners Association Question and Answer Meeting, October, 2002.

                                        7-217

-------
Final Regulatory Support Document
33.American Automobile Manufacturers Association Diesel Fuel Survey, Summer 1997.

34.Conversation with Cal Hodge, A Second Opinion, February 2000.

35. EIA Petroleum Supply Annual, 2002, excluding California refineries.

36..Energy Information Agency 2003 Annual Energy Outlook, Table PSA17.

37.Gary, James H., Handwerk, Glenn E., Petroleum Refining: Technology and Economics,
Marcel Dekker, New York (1994).

38.Conversation with Lyondel-Citgo refinery staff, April 2000.

39.Gary, James H., Handwerk, Glenn E., Petroleum Refining: Technology and Economics,
Marcel Dekker, New York (1994).

40.Peters, Max S., Timmerhaus, Klaus D., Plant Design and Economics for Chemical Engineers,
Third Edition, McGraw Hill Book Company, 1980.

41.Waguespack, Kevin, Review of DOE/Ensys Reformulated Diesel Study-Draft for Discussion,
Price-Waterhouse Coopers for the American Automobile Manufacturers, October 5, 2000.

42.U.S. Petroleum Refining, Meeting Requirements for Cleaner Fuels and Refineries, Volume V
- Refining Capability Appendix, National Petroleum Council, 1993.

43.Waguespack, Kevin, Review of DOE/Ensys Reformulated Diesel Study-Draft for Discussion,
Price-Waterhouse Coopers for the American Automobile Manufacturers, October 5, 2000.

44.Hadder, Gerry and Tallet, Martin; Documentation for the Oak Ridge National Laboratory
Refinery Yield Refinery Model (ORNL-RYM), 2001.

45.Perry, Robert H., Chilton, Cecil H., Chemical Engineer's Handbook, McGraw Hill 1973.

46. Hydrogen and Utility Supply Optimization, Shahani, Gouton et al, Technical Paper by Air
Products presented at the National Petrochemical and Refiners Assoc. 1998 Annual Meeting
(AM-98-60).

47.1999 Worldwide Refining Survey, Oil and Gas Journal, December 20, 1999.

48.Peters, Max S., Timmerhaus, Klaus D., Plant Design and Economics for Chemical Engineers,
Third Edition, McGraw Hill Book Company, 1980.

49. Jena, Rabi, Take the PC-Based Approach to Process Control, Fuel Reformulation,
November/December 1995.

SO.Sutton, IS., Integrated Management Systems Improve Plant Reliability, Hydrocarbon
Processing, January 1995.

                                        7-218

-------
                                              Estimated Costs of Low-Sulfur Fuels
5I.King, M. J., Evans, H. N., Assessing your Competitors' Application of CIM/CIP,
Hydrocarbon Processing, July 1993.

52.U. S. Petroleum Refining, Assuring the Adequacy and Affordability of Cleaner Fuels, A
Report by the National Petroleum Council, June 2000.

53. "Refining Economics of Diesel Fuel Sulfur Standards," study performed for the Engine
Manufacturers Association by MathPro, Inc. October 5, 1999.

54. "Refining Economics of Diesel Fuel Sulfur Standards, Supplemental Analysis of 15 ppm
Sulfur Cap," study performed for the Engine Manufacturers Association by Mathpro Inc.,
August 16, 2000.

55. "Regulatory Impact Analysis: Heavy-Duty Engine and Vehicle Standards and Highway
Diesel Fuel Sulfur Control Requirements," Section V.C.4, EPA420-R-00-026, December 2000,
Docket A-2001-28, Document No. II-A-01.

56. Phone conversation in mid-2002 with Massey's Truck and Tank Repair, Pheonix Arizona.

57. "Regulatory Impact Analysis: Heavy-Duty Engine and Vehicle Standards and Highway
Diesel Fuel Sulfur Control Requirements," EPA420-R-00-026, December 2000, Docket A-2001-
28, Document No. II-A-01.

58. Letter from Titan Industries Inc. to Jeff Herzog, EPA, November 21, 2003.  Phone
conversation with Ron Wilson, Lubrizol Additive Injection Equipment, November 21, 2003

59. The Impact of Environmental Compliance Costs on U.S. Refining Profitability 1995-2001,
Table A2, Energy Information Administration, May  16, 2003.
                                       7-219

-------
CHAPTER 8: Estimated Aggregate Cost and Cost per Ton of Reduced Emissions
   8.1 Projected Sales and Cost Allocations  	8-2
   8.2 Aggregate Engine Costs	8-3
      8.2.1 Aggregate Engine Fixed Costs	8-3
      8.2.2 Aggregate Engine Variable Costs	8-7
   8.3 Aggregate Equipment Costs  	8-11
      8.3.1 Aggregate Equipment Fixed Costs	8-11
      8.3.2 Aggregate Equipment Variable Costs	8-13
   8.4 Aggregate Fuel Costs and Other Operating Costs	8-15
      8.4.1 Aggregate Fuel Costs 	8-16
      8.4.2 Aggregate Oil-Change Maintenance Savings	8-18
      8.4.3 Aggregate CDPF Maintenance, CDPF Regeneration, and CCV Maintenance Costs
           	8-20
      8.4.4 Summary of Aggregate  Operating Costs  	8-22
      8.4.5 Summary of Aggregate  Operating Costs Associated with a Fuel-only Scenario  8-24
   8.5 Summary of Aggregate Costs of the Final Rule 	8-26
   8.6 Emission Reductions	8-29
   8.7 Cost per Ton	8-30
      8.7.1 Cost per Ton for the NRT4 Final Rule	8-30
      8.7.2 Cost per Ton for the NRLM Fuel-only Scenario 	8-34
      8.7.3 Costs and Costs per Ton for Other Control Scenarios  	8-37
          8.7.3.1 Costs and Costs per Ton of a 500 ppm NRLM Fuel-only Scenario  	8-37
          8.7.3.2 Costs and Costs per Ton of the 15 ppm L&M Fuel Increment	8-42
      8.7.4 Costs per Ton Summary	8-63

Appendix 8A: Estimated Aggregate Cost and Cost per Ton of Sensitivity Analyses	8-65

Appendix 8B: Fuel Volumes used throughout Chapter 8  	8-89

-------
                                                   Aggregate Cost and Cost per Ton
  CHAPTER 8: Estimated Aggregate Cost and Cost per Ton
                            of Reduced Emissions
   This chapter aggregates the estimated incremental engine costs, operating costs, equipment
costs, and fuel costs of the final rule.  This chapter also presents detailed information on the
calculation for the cost per ton of pollutant.  Chapter 6 details the estimated fixed and variable
costs for modifying new nonroad engines and equipment to meet new emission standards;
Chapter 6 also discusses the effects of the new low-sulfur diesel fuels on operating costs for
land-based nonroad diesel engines, locomotive engines, and marine diesel engines.  Chapter 7
describes our estimates of the costs associated with the fuel requirements in this final rule.

   We have calculated the cost per ton of emission reductions for this final rule based on the net
present value of all costs incurred and all emission reductions generated over a 30-year time
window after the program takes effect. This approach captures all the costs and emission
reductions from the final rule, including those costs incurred and emission reductions generated
by the existing fleet. The point of comparison for this evaluation is the existing set of fuel  and
engine standards (i.e., unregulated fuel and the Tier 2/Tier 3 program).  The 30-year time
window is meant to capture both the early period of the program when there are a small number
of compliant engines in  the fleet, and the later period when there is nearly complete turnover to
compliant engines. Note that all  costs and emission reductions presented here are 30-year
numbers (the net present values in 2004 of the stream of costs/reductions occurring from 2007
through 2036, expressed in $2002).

   While there is a broad consensus among economists that future benefits and costs of
regulatory programs should be discounted, there is no consensus in the literature regarding the
most appropriate discounting concept and rate to apply. In particular, the theoretical literature is
divided between two alternative approaches.  The first approach is referred to as the "demand-
side approach" (see Arrow et al,  1996), which defines the appropriate discount rate as the rate at
which society would collectively trade off current versus future consumption.  This  rate is
difficult to establish empirically,  but estimates in the literature commonly range from 1 to 4
percent. EPA's economic Guidelines suggest using a value of two to three percent.1 The second
approach is referred to as the "cost-side approach" (see Lind, 1982), and discount rates
associated with this concept reflect trade-offs between current and future consumption derived
by market rates driven by the marginal productivity of capital. This rate is also difficult to derive
from empirical data, but estimates typically fall in the range of 4 to 10 percent. OMB's circular
A-94 expresses a preference for the cost-side approach and specifies a seven percent rate.

   Given both the lack  of consensus  in the literature on the most appropriate concept and the
uncertainty surrounding the associated empirical estimates, EPA's Economic Guidelines and the
two key outside expert groups which  advise EPA on economic analytical issues all recommend
evaluating benefits and costs using a range of discount rates.  Consistent with this advice, we
have analyzed the benefits and costs of the nonroad Tier 4 rule using  both a three percent rate

                                          8-1

-------
Final Regulatory Impact Analysis
and a seven percent rate.  We present the results based on a three percent discount rate as our
primary estimates.

8.1 Projected Sales and Cost Allocations

   Projected nonroad engine and equipment sales estimates are used in several portions of this
analysis. We have used two sources for our projected sales numbers—the PSR database for the
2000 model year, and our Nonroad Model.2'3 The PSR database has been used as the basis for
our current fleet mix; i.e., which equipment types were  sold in 2000 and with engines from
which power category. The sales estimates and growth rates used throughout this analysis are
shown in Table 8.1-1.4

                                       Table 8.1-1
                   Estimated 2000 Engine Sales and Future Sales Growth
Power range
0750
Total
2000 Model Year
Sales
119,159
132,981
93,914
68,665
112,340
61,851
34,095
2,752
2,785
628,542
Annual Growth in
Engines Sold
4,116
3,505
2,046
1,499
2,321
1,414
436
50
51
15,438
Linear Growth Rate
3.5%
2.6%
2.2%
2.2%
2.1%
2.3%
1.3%
1.8%
1.8%
2.5%
   Because the new emission standards will reduce emissions of several different pollutants
(i.e., NOx, PM, NMHC, and SOx), we have attempted to allocate the estimated costs to emission
reductions of specific pollutants.  This apportionment of costs by pollutant allows us to calculate
the average cost per ton of emission reduction resulting from this rule.  Table 8.1-2 summarizes
the allocations we have used in the final rule. Deciding how to apportion costs can be difficult
even in the case of technologies that, on the surface, seem to have an obvious split by which their
costs should be attributed. For instance, we have apportioned 100 percent of the cost for CDPF
technology to PM even though CDPFs are expected to reduce NMHC emissions significantly.A
For fuel-related costs where no technology enablement occurs (i.e., fuel-derived emissions
     A CDPF is a catalyzed diesel participate filter; a DOC is a diesel oxidation catalyst; CCV is a closed crankcase
ventilation system; Regen is short for regeneration; EGR is exhaust gas recirculation; NRLM refers to nonroad,
locomotive, and marine.
                                           8-2

-------
                                                        Aggregate Cost and Cost per Ton
reductions where no new engine standards exist that rely on the new fuel), we have apportioned
one-third of the costs to PM and two-thirds to SOx.  This is different than how we allocated costs
in the proposal where we allocated 100 percent of such costs to SOx control. We believe the
allocation used here is more appropriate given that the lower sulfur fuel provides for substantial
PM reductions even without new engine standards.8  The estimated costs for 15 ppm fuel are
apportioned one-half to technology enablement (i.e.,  engine-derived emissions reductions) and
one-half to fuel-derived emissions reductions.  Respectively, these halves are allocated 50%/50%
to NOx+NMHC/PM and 33%/67% to PM/SOx.  This latter split is consistent with the fuel-
derived allocation described above.  This is different than the proposal where we allocated 15
ppm costs entirely to technology enablement. We believe the allocations used here in the final
rule are more appropriate given the substantial PM and SOx reductions that occur solely because
the fuel sulfur level has been reduced. We note throughout the discussion to which pollutant we
have attributed costs.

8.2 Aggregate Engine Costs

   This section presents aggregate engine fixed costs (recovered costs) and variable costs.
These costs are discussed in detail in Section 6.2.

8.2.1 Aggregate Engine Fixed Costs

   Chapter 6 presents the aggregate engine fixed costs, along with our best estimate of how
those costs might be recovered (i.e., on which engines), for engine R&D, tooling, and
certification, respectively (see Tables 6.2-4, 6.2-6, and  6.2-8).c Table 8.2-1 presents the
combined total of all engine fixed costs in the indicated years for each power category. Table
8.2-2 shows to what pollutant the total costs by year are allocated. Note that the cost allocations
shown in Table 8.2-1 are not generated assuming any simple split of costs between NOx and PM
control.  Some engine fixed costs are solely attributed to PM control (for example, costs
associated with the 2008 standards and costs associated with the 2013 standards for  50 to 75 hp
engines). Therefore, the costs presented in Table 8.2-2 for PM do not represent the total fixed
costs of the program if there were no new NOx standards; the same is true of NOx costs if there
     A 50/50 split between PM/SOx could be argued, but that seems inappropriate given that 98 percent of fuel borne
sulfur is exhausted as SOx and only two percent is exhausted as PM. Given that, a 2/98 split between PM/SOx could be
argued, but that seems inappropriate given the importance of PM reductions—which have much higher human health
benefits—relative to SOx reductions.  The 33/67 split between PM/SOx that we have chosen here seems to provide an
appropriate balance.

    /-i
     We have estimated a "recovered" cost for all engine and equipment fixed costs to present a per-production-unit
analysis of the cost of the final rule (see Section 6.4.3 or Chapter 10 for our estimate of engine costs on a per-unit basis).
In general, in environmental economics, it is more conventional to simply count the total costs of the program (i.e.,
opportunity costs) in the year they occur. However, this approach does not directly estimate a per-unit cost, since fixed
costs occur before the standards take effect, resulting in costs that do not correspond to units certified to the new emission
standards. As a result, we grow fixed costs until they can be "recovered" on complying units. Note that the approach
used here results in a higher estimate of the total costs of the program, since the recovered costs include  a seven percent
interest rate to reflect the time value of money (i.e., the lost opportunity cost of that capital).


                                              8-3

-------
Final Regulatory Impact Analysis
were no new PM standards. Refer to Section 6.2 for detail on how we have estimated engine
fixed costs and their recovery, and to Table 8.1-2 for how they are allocated among each
pollutant.
                                         8-4

-------
                                                   Table 8.1-2
             Summary of How Cost are Allocated Among Pollutants under the NRT4 Final Program
Item
Fuel Costs - incremental cent/gallon
Operating Costs - Oil-Change Savings
Operating Costs - CDPF Maintenance
Operating Costs - CDPF Regen (FE impact)
Operating Costs - CCV Maintenance
Engine Variable Costs
Engine Fixed Costs - R&D
Engine Fixed Costs - Tooling
Engine Fixed Costs - Certification
Equipment Variable Costs
Equipment Fixed Costs
NOx+NMHC PM SOx
500 ppm Affected NRLM
1 5 ppm Affected NR
1 5 ppm Affected L&M
500 ppm Affected NRLM
15 ppm Affected NR
1 5 ppm Affected L&M
1 5 ppm NR in new CDPF engines
1 5 ppm NR in new CDPF engines
All NR in new CCV engines
CDPF System
NOx Adsorber System
DOC
Fuel-Injection System
Regeneration System
Cooled EGR
Closed Crankcase Ventilation Sys
CDPF+NOx Adsorber
CDPF-only
DOC-only
CDPF+NOx Adsorber
CDPF-only
DOC-only
Cooled EGR
<75 hp 2008
25-50 hp 20 13
50-75 hp 20 13
75-750 hp at start of phase-in
75-750 hp at end of phase-in
>750 hp
<25hp; 25-75 hp 2008-2012
25-50 hp 20 13+
50-75 hp 20 13+
75-750 hp at start of phase-inb
75-750 hp at end of phase-in
>750 hp
<75 hp 2008 standards
25-75 hp 20 13 standards
75-750 hp at start of phase-in
75-750 hp at end of phase-in
>750hp2011
>750hD2015

50% of 50%


50% of 50%



50%

100%

50%

100%
50%
67%


50%


100%

50%

50%
100%
50%

50%

25%
50%


50%
50%
100%
100%

33%
50% of 50%
33% of 50%
33%
33%
50% of 50%
33% of 50%
33%
100%
100%
50%
100%

100%
50%
100%

50%
33%
100%
100%
50%
100%
100%

100%
50%
100%
50%

50%
100%
50%
100%
75%
50%
100%
100%
50%
50%


100%
67%
67% of 50%
67%
67%
67% of 50%
67%



































 All engines meet the new PM standard and half meet the new NOx standard. For NOx phase-in engines, the allocation is 50/50 to
PM/NOx. For PM-only engines, the allocation is 100% PM. The resultant allocation is 75/25 to PM/NOx.
                                                        3-5

-------
                    8.2-1
Aggregate Engine Fixed Costs by Power Category
          (SMillions of 2002 dollars)
Year
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
Total
30YrNPVat3%
30YrNPVat7%
0750hp
$
$
-
$ 1.9
$ 1.9
$ 1.9
$ 1.9
$ 5.1
$ 3.2
$ 3.2
$ 3.2
$ 3.2
$
$ 25.7
$ 18.3
$ 11.9
Total
$ 19.3
$ 19.3
$ 19.3
$ 62.9
$ 80.7
$ 83.8
$ 108.2
$ 111.4
$ 67.8
$ 50.0
$ 27.6
$ 3.2
$
$ 653.4
$ 491 .8
$ 343.6

-------
                                                    Aggregate Cost and Cost per Ton
                                       Table 8.2-2
                        Aggregate Engine Fixed Costs by Pollutant
                                (SMillions of 2002 dollars)
Year
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
Total
30YrNPVat3%
30YrNPVat7%
Recovery of PM
Costs
$ 19.3
$ 19.3
$ 19.3
$ 40.9
$ 49.8
$ 51.3
$ 51.3
$ 54.1
$ 32.5
$ 23.6
$ 2.8
$ 2.8
$
$ 366.9
$ 281.6
$ 201.8
Recovery of
NOx Costs
-
$
$
$ 22.0
$ 30.9
$ 32.5
$ 56.9
$ 57.3
$ 35.3
$ 26.4
$ 24.8
$ 0.4
-
$ 286.4
$ 210.3
$ 141.9
Recovery of
Fixed Costs
$ 19.3
$ 19.3
$ 19.3
$ 62.9
$ 80.7
$ 83.8
$ 108.2
$ 111.4
$ 67.8
$ 50.0
$ 27.6
$ 3.2
$
$ 653.4
$ 491.8
$ 343.6
   We have assumed that all engine R&D expenditures occur over a five-year span preceding
the first year any emission-control device is introduced into the market, with the exception of
R&D for the 2008 standards which occurs over a four-year span preceding the standards as
described in Chapter 6. Where a phase-in exists (for example, for NOx standards on engines
between 75 and 750 hp), expenditures are assumed to occur over the five years preceding the
first year that NOx adsorbers will be introduced, then continuing during the phase-in years; the
expenditures will be incurred consistent with the phase-in of the standard. All R&D
expenditures are then recovered by the engine manufacturer over an identical time span
following the introduction of the technology. We include a cost of seven percent when
amortizing engine R&D expenditures.

   We have assumed that all tooling and certification costs are incurred one year in advance of
the new standard and are recovered over a five-year period after the new standards take effect;
we include a cost of seven percent when amortizing engine tooling costs.

   We have calculated the net present value of the engine fixed costs over the 30-year period
following implementation of the program as $492 million. This value assumes a three percent
social discount rate.

8.2.2 Aggregate Engine Variable Costs

   Engine variable costs are discussed in detail in Section 6.2.2. As explained there, we have
generated cost estimation equations to calculate engine variable costs.  These cost estimation
equations are summarized in Table 6.4-2. Using these equations, we have calculated the engine
                                           3-7

-------
Final Regulatory Impact Analysis
variable costs during the years 2008 through 2036 as shown in Tables 8.2-3 and 8.2-4 (refer to
Table 8.1-2 for how costs have been allocated to PM and NOx).  Because of their nature,
variable costs are proportional to engine sales and are projected to increase in the future as
engine sales increase.  We have calculated the net present value of the engine variable costs over
the 30-year period following implementation of the program as $13.6 billion. This value
assumes a three percent social discount rate.

-------
Table 8.2-3
Aggregate Engine Variable Costs by Power Category ($Millions of 2002 dollars)
Year
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2031
2032
2033
2034
2035
2036
30YrNPVat3%
30YrNPVat7%
0750hp
$
$
-
$ 6.6
$ 6.7
$ 5.3
$ 5.4
$ 29.6
$ 30.0
$ 24.6
$ 24.9
$ 25.3
$ 25.6
$ 26.0
$ 26.3
$ 26.7
$ 27.0
$ 27.3
$ 27.7
$ 28.0
$ 28.4
$ 28.7
$ 29.1
$ 29.4
$ 29.7
$ 30.1
$ 30.4
$ 30.8
$ 31.1
$ 348.3
$ 168.5
Total
$ 61.8
$ 63.2
$ 61.1
$ 340.2
$ 636.8
$ 798.3
$ 864.4
$ 838.5
$ 852.0
$ 859.6
$ 873.1
$ 886.5
$ 899.9
$ 913.3
$ 926.8
$ 940.2
$ 953.6
$ 967.0
$ 980.4
$ 993.9
$ 1 ,007.3
$ 1 ,020.7
$ 1,034.1
$ 1 ,047.6
$ 1,061.0
$ 1 ,074.4
$ 1 ,087.8
$ 1,101.2
$ 1,114.7
$ 13,562.1
$ 6,871.3
8-9

-------
                                    Table 8.2-4
Aggregate Engine Variable Costs by Technology and by Pollutant (SMillions of 2002 dollars)
Year
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2031
2032
2033
2034
2035
2036
30YrNPVat3%
30YrNPVat7%
Fuel System
-
-
-
-
$
$ 53.3
$ 54.3
$ 41.6
$ 42.4
$ 43.1
$ 43.8
$ 44.6
$ 45.3
$ 46.1
$ 46.8
$ 47.6
$ 48.3
$ 49.1
$ 49.8
$ 50.5
$ 51.3
$ 52.0
$ 52.8
$ 53.5
$ 54.3
$ 55.0
$ 55.8
$ 56.5
$ 57.2
$ 657.0
$ 323.5
Cooled EGR
-
-
$
$ 6.2
$ 6.3
$ 29.2
$ 29.8
$ 24.4
$ 24.8
$ 25.2
$ 25.7
$ 26.1
$ 26.5
$ 27.0
$ 27.4
$ 27.8
$ 28.3
$ 28.7
$ 29.2
$ 29.6
$ 30.0
$ 30.5
$ 30.9
$ 31.3
$ 31.8
$ 32.2
$ 32.7
$ 33.1
$ 33.5
$ 391.7
$ 194.8
CCV
$ 0.5
$ 0.6
$ 0.4
$ 7.1
$ 13.4
$ 11.8
$ 10.3
$ 10.4
$ 10.6
$ 10.8
$ 10.9
$ 11.1
$ 11.2
$ 11.4
$ 11.6
$ 11.7
$ 11.9
$ 12.0
$ 12.2
$ 12.4
$ 12.5
$ 12.7
$ 12.8
$ 13.0
$ 13.2
$ 13.3
$ 13.5
$ 13.6
$ 13.8
$ 175.8
$ 90.7
DOC
$ 61.2
$ 62.6
$ 60.7
$ 62.0
$ 63.3
$ 21.2
$ 21.7
$ 22.2
$ 22.7
$ 23.2
$ 23.7
$ 24.2
$ 24.7
$ 25.2
$ 25.7
$ 26.2
$ 26.8
$ 27.3
$ 27.8
$ 28.3
$ 28.8
$ 29.3
$ 29.8
$ 30.3
$ 30.8
$ 31.3
$ 31.8
$ 32.3
$ 32.8
$ 611.1
$ 377.0
CDPF System
-
-
$
$ 168.8
$ 338.4
$ 414.1
$ 380.3
$ 381 .3
$ 387.3
$ 388.0
$ 393.9
$ 399.8
$ 405.7
$ 411.7
$ 417.6
$ 423.5
$ 429.5
$ 435.4
$ 441 .3
$ 447.3
$ 453.2
$ 459.1
$ 465.1
$ 471 .0
$ 476.9
$ 482.8
$ 488.8
$ 494.7
$ 500.6
$ 6,127.5
$ 3,102.8
CDPF Regen
System
-
-
$
$ 28.7
$ 73.2
$ 137.4
$ 128.8
$ 115.6
$ 117.5
$ 118.9
$ 120.8
$ 122.8
$ 124.7
$ 126.6
$ 128.5
$ 130.5
$ 132.4
$ 134.3
$ 136.2
$ 138.2
$ 140.1
$ 142.0
$ 143.9
$ 145.8
$ 147.8
$ 149.7
$ 151.6
$ 153.5
$ 155.5
$ 1,860.1
$ 933.3
NOx Adsorber
System
-
-
$
$ 67.4
$ 142.1
$ 131.2
$ 239.3
$ 243.0
$ 246.8
$ 250.5
$ 254.2
$ 257.9
$ 261 .6
$ 265.4
$ 269.1
$ 272.8
$ 276.5
$ 280.2
$ 284.0
$ 287.7
$ 291 .4
$ 295.1
$ 298.8
$ 302.6
$ 306.3
$ 310.0
$ 313.7
$ 317.5
$ 321 .2
$ 3,738.8
$ 1 ,849.0
Total PM Costs
$ 61.5
$ 62.9
$ 60.9
$ 263.1
$ 481 .7
$ 605.3
$ 563.1
$ 545.1
$ 554.0
$ 557.0
$ 565.8
$ 574.6
$ 583.4
$ 592.3
$ 601.1
$ 609.9
$ 618.7
$ 627.5
$ 636.3
$ 645.1
$ 653.9
$ 662.8
$ 671 .6
$ 680.4
$ 689.2
$ 698.0
$ 706.8
$ 715.6
$ 724.4
$ 9,015.3
$ 4,620.3
Total
NOx+NMHC
Costs
$ 0.3
$ 0.3
$ 0.2
$ 77.1
$ 155.1
$ 193.0
$ 301.4
$ 293.4
$ 298.0
$ 302.6
$ 307.2
$ 311.9
$ 316.5
$ 321.1
$ 325.7
$ 330.3
$ 334.9
$ 339.5
$ 344.1
$ 348.7
$ 353.3
$ 358.0
$ 362.6
$ 367.2
$ 371.8
$ 376.4
$ 381.0
$ 385.6
$ 390.2
$ 4,546.9
$ 2,251.0
Total Costs
$ 61.8
$ 63.2
$ 61.1
$ 340.2
$ 636.8
$ 798.3
$ 864.4
$ 838.5
$ 852.0
$ 859.6
$ 873.1
$ 886.5
$ 899.9
$ 913.3
$ 926.8
$ 940.2
$ 953.6
$ 967.0
$ 980.4
$ 993.9
$ 1 ,007.3
$ 1 ,020.7
$ 1,034.1
$ 1 ,047.6
$ 1,061.0
$ 1 ,074.4
$ 1 ,087.8
$ 1,101.2
$ 1,114.7
$ 13,562.1
$ 6,871.3
                                       8-10

-------
                                                   Aggregate Cost and Cost per Ton
8.3 Aggregate Equipment Costs

   This section aggregates the amortized fixed and variable cost for equipment estimated in
Section 6.3.

8.3.1 Aggregate Equipment Fixed Costs

   In Table 6.3-4 we presented the aggregate equipment fixed costs, along with our best
estimate of how those costs might be recovered, for equipment redesign and revisions to product
literature.  Table 8.3-1 presents aggregate equipment fixed costs and Table 8.3-2 shows to what
pollutant these costs are attributed. Note that the cost allocations shown in Table 8.3-2 are not
generated assuming any simple split of costs between NOx and PM control. Some equipment
fixed costs are solely attributed to PM control (for example, costs associated with the 2008
standards and costs associated with the 2013 standards for 50 to 75 hp engines).  The costs
presented in Table 8.3-1 for PM therefore do not represent the total fixed costs of the program if
there were no new NOx standards; the same is true of NOx costs if there were no new PM
standards.  Refer to Section 6.3 for detail on how we have estimated equipment fixed costs and
their recovery, and to Table 8.1-2 for how they are allocated among each pollutant.

   We have assumed that all equipment fixed costs (redesign and product literature) occur over
a two-year span preceding the first year any emission-control device is introduced into the
market. Where a phase-in exists (for example, for NOx standards on engines over 75 hp
engines), expenditures are assumed to  occur over the two  years preceding the first year that NOx
adsorbers will be introduced, then continuing during the phase-in years; the expenditures will be
incurred consistent with the phase-in of the standard. All  expenditures are then recovered by the
equipment manufacturer over 10 years following the introduction of the technology. We include
a cost of seven percent when amortizing equipment fixed  costs.

   We have calculated the net present value of the equipment fixed costs over the 30-year
period following implementation of the program as $847 million.  This value assumes a three
percent social discount rate.
                                         8-11

-------
                             Table 8.3-1
Aggregate Equipment Fixed Costs by Power Range (SMillions of 2002 dollars)
Year Recovered
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
Total
30YrNPVat3%
30YrNPVat7%
0750hp
$
$
$
$ 0.6
$ 0.6
$ 0.6
$ 0.6
$ 4.9
$ 4.9
$ 4.9
$ 4.9
$ 4.9
$ 4.9
$ 4.3
$ 4.3
$ 4.3
$ 4.3
$ 48.9
$ 31.5
$ 18.1
Total
$ 4.5
$ 4.5
$ 4.5
$ 53.1
$ 86.4
$ 97.0
$ 117.7
$ 122.0
$ 122.0
$ 122.0
$ 117.5
$ 117.5
$ 117.5
$ 68.9
$ 35.6
$ 25.0
$ 4.3
$ 1,219.9
$ 847.2
$ 536.6

-------
                                                   Aggregate Cost and Cost per Ton
                                      Table 8.3-2
                      Aggregate Equipment Fixed Costs by Pollutant
                               (SMillions of 2002 dollars)
Year
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
Total
30YrNPVat3%
30YrNPVat7%
Recovery of
PM Costs
$ 4.5
$ 4.5
$ 4.5
$ 28.5
$ 45.1
$ 50.4
$ 50.4
$ 54.7
$ 54.7
$ 54.7
$ 50.2
$ 50.2
$ 50.2
$ 26.2
$ 9.6
$ 4.3
$ 4.3
$ 547.3
$ 384.9
$ 247.9
Recovery of
NOx+NMHC
Costs
$
$
$
$ 24.6
$ 41.2
$ 46.5
$ 67.3
$ 67.3
$ 67.3
$ 67.3
$ 67.3
$ 67.3
$ 67.3
$ 42.7
$ 26.0
$ 20.7
$
$ 672.5
$ 462.2
$ 288.7
Recovery of
Fixed Costs
$ 4.5
$ 4.5
$ 4.5
$ 53.1
$ 86.4
$ 97.0
$ 117.7
$ 122.0
$ 122.0
$ 122.0
$ 117.5
$ 117.5
$ 117.5
$ 68.9
$ 35.6
$ 25.0
$ 4.3
$ 1,219.9
$ 847.2
$ 536.6
8.3.2 Aggregate Equipment Variable Costs

   The equipment variable costs, such as sheet metal costs, mounting hardware, and labor, were
estimated by power category in Section 6.3.  The aggregate equipment variable costs through
2036 are presented in Table 8.3-3. Table 8.3-4 shows the total aggregate equipment variable
costs allocated by  pollutant (refer to Table 8.1-2 for how costs have been allocated to PM and
NOx).  We have calculated the net present value of the equipment variable costs over the 30-year
period following implementation of the program as $434 million.  This value assumes a three
percent social discount rate.
                                         8-13

-------
                                Table 8.3-3
Aggregate Equipment Variable Costs by Power Category (SMillions of 2002 dollars)
Year
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2031
2032
2033
2034
2035
2036
30YrNPVat3%
30 YR NPVat7%
0750hp
$
$
$
$
-
$
$
$ 0.4
$ 0.4
$ 0.4
$ 0.4
$ 0.4
$ 0.4
$ 0.4
$ 0.4
$ 0.4
$ 0.4
$ 0.4
$ 0.4
$ 0.4
$ 0.4
$ 0.4
$ 0.4
$ 0.4
$ 0.4
$ 0.4
$ 0.4
$ 0.4
$ 0.5
$ 4.8
$ 2.3
Total
$
$
$
$ 9.1
$ 19.6
$ 24.2
$ 27.9
$ 27.5
$ 27.9
$ 28.2
$ 28.7
$ 29.1
$ 29.5
$ 29.9
$ 30.4
$ 30.8
$ 31.2
$ 31.6
$ 32.1
$ 32.5
$ 32.9
$ 33.4
$ 33.8
$ 34.2
$ 34.6
$ 35.1
$ 35.5
$ 35.9
$ 36.3
$ 434.2
$ 217.4

-------
                                                  Aggregate Cost and Cost per Ton
                                     Table 8.3-4
                    Aggregate Equipment Variable Costs by Pollutant
                               (SMillions of 2002 dollars)
Year
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2031
2032
2033
2034
2035
2036
30YrNPVat3%
30YrNPVat7%
PM Costs
-
$
$
$ 6.8
$ 14.7
$ 19.7
$ 17.1
$ 16.5
$ 16.8
$ 17.0
$ 17.2
$ 17.5
$ 17.7
$ 18.0
$ 18.3
$ 18.5
$ 18.8
$ 19.0
$ 19.3
$ 19.6
$ 19.8
$ 20.1
$ 20.4
$ 20.6
$ 20.9
$ 21.1
$ 21.4
$ 21.7
$ 21.9
$ 268.9
$ 136.3
NOx Costs
-
$
$
$ 2.3
$ 4.9
$ 4.5
$ 10.8
$ 11.0
$ 11.1
$ 11.3
$ 11.4
$ 11.6
$ 11.8
$ 11.9
$ 12.1
$ 12.3
$ 12.4
$ 12.6
$ 12.8
$ 12.9
$ 13.1
$ 13.3
$ 13.4
$ 13.6
$ 13.8
$ 13.9
$ 14.1
$ 14.3
$ 14.4
$ 165.3
$ 81.1
Total Variable
Costs
-
$
$
$ 9.1
$ 19.6
$ 24.2
$ 27.9
$ 27.5
$ 27.9
$ 28.2
$ 28.7
$ 29.1
$ 29.5
$ 29.9
$ 30.4
$ 30.8
$ 31.2
$ 31.6
$ 32.1
$ 32.5
$ 32.9
$ 33.4
$ 33.8
$ 34.2
$ 34.6
$ 35.1
$ 35.5
$ 35.9
$ 36.3
$ 434.2
$ 217.4
8.4 Aggregate Fuel Costs and Other Operating Costs

   Aggregate costs presented here are used in the calculation of costs per ton of emission
reductions resulting from this final rule.  This includes a two-step fuel sulfur control program
consisting of a NRLM sulfur cap of 500 ppm beginning in 2007 to be followed by a nonroad
(NR) sulfur cap of 15 ppm beginning in 2010 and a locomotive and marine (L&M) sulfur cap of
15 ppm beginning in 2012. Refer to Chapters 5 and 7 for more information about the fuel
program and how the costs for that portion of the NRT4 final rule were estimated.

   As noted, the second step in the fuel program limits NR sulfur levels to 15 ppm beginning in
2010.  This fuel program enables the introduction of advanced emission-control
technologies—CDPFs and NOx adsorbers—that will enable nonroad engines to meet the new
Tier 4 standards, and it also achieves additional emissions reductions from the fuel control itself
(i.e., independent of new engine standards). The combination of the two-step NRLM fuel
                                        8-15

-------
Final Regulatory Impact Analysis
program and the new diesel engine standards represents the full engine and fuel program (i.e.,
the NRT4 final rule). Section 8.4.1 presents our estimate of the aggregate fuel costs associated
with the NRT4 final rule. Sections 8.4-2 through 8.4-4 present estimates of other operating
costs—CDPF and CCV maintenance, fuel economy impacts, and oil change
maintenance—associated with the NRT4 final rule. Section 8.4-5 presents the cost of the fuel
program absent any new engine standards. These costs differ from the costs associated with the
fuel program costs of the NRT4 final rule in that no CDPF and CCV maintenance costs, and no
fuel economy impacts would be realized. We present these costs because they are used in
calculations of $/ton associated with  such a "fuel-only" scenario.

8.4.1 Aggregate Fuel Costs

   Fuel costs, described in detail in Chapter 7, are developed on a cent-per-gallon basis. Table
8.4-1 summarizes cent-per-gallon fuel costs (see Table 7.5-1), estimated fuel volumes for NR,
L&M, and the resultant annual fuel costs associated with the two-step NRT4 final rule fuel
program.  Table 8.4-1 shows that the 30-year net present value of the new lower sulfur
requirements is estimated at $16.3 billion.  This assumes a three percent social discount rate.
Note that the affected fuel volumes presented in Table 8.4-1 are gallons consumed in both new
and existing engines since both new and existing engines will have to pay for the higher cost
fuel. We have not included spillover gallons or other such gallons that would have entered the
NRLM fuel pool with a sulfur level below the new cap absent the new requirements since these
gallons do not represent an incremental increase in costs associated with the NRT4 final rule.
                                          8-16

-------
                                                                           Table 8.4-1
                                                Aggregate Fuel Costs of the Two-Step Fuel Program ($2002)
Year
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2031
2032
2033
2034
2035
2036
30YrNPVat3%
30YrNPVat7%
Affected NR
500 ppm
(106aallons)
4,790
8,406
8,599
4,014
614
528
468
199
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
24,054
20,174
15 ppm
(106aallons)
-
-
-
6,189
8,145
8,420
8,671
9,713
10,539
10,747
10,955
11,162
1 1 ,370
1 1 ,578
1 1 ,786
1 1 ,994
12,201
12,409
12,617
12,823
13,030
13,236
13,442
13,649
13,855
14,061
14,268
14,474
14,680
14,887
180,224
92,196
Affected L&M
500 ppm
(106aallons)
1,990
3,454
3,498
3,185
2,975
1,396
247
104
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
1 4,363
1 1 ,729
15 ppm
(106aallons)
-
-
-
0
0
1,965
3,397
3,081
2,860
2,888
2,918
2,953
2,995
3,024
3,052
3,093
3,125
3,161
3,195
3,230
3,265
3,301
3,336
3,371
3,406
3,441
3,476
3,512
3,547
3,582
44,087
22,124
Fuel Cost*
500 ppm
($/aal)
$ 0.021
$ 0.021
$ 0.021
$ 0.028
$ 0.033
$ 0.034
$ 0.035
$ 0.035
























15 ppm
($/aal)
$
$
$
$ 0.058
$ 0.058
$ 0.062
$ 0.064
$ 0.068
$ 0.070
$ 0.070
$ 0.070
$ 0.070
$ 0.070
$ 0.070
$ 0.070
$ 0.070
$ 0.070
$ 0.070
$ 0.070
$ 0.070
$ 0.070
$ 0.070
$ 0.070
$ 0.070
$ 0.070
$ 0.070
$ 0.070
$ 0.070
$ 0.070
$ 0.070


NR Fuel Costs
500 ppm
(106 dollars)
$ 101
$ 177
$ 181
$ 112
$ 20
$ 18
$ 16
$ 7
$
$
$
-
$
$
$
-
$
$
$
-
$
$
$
-
$
$
$
-
$
$
$ 547
$ 456
15 ppm
(106 dollars)
-
-
-
359
472
518
555
656
738
752
767
781
796
810
825
840
854
869
883
898
912
927
941
955
970
984
999
1,013
1,028
1,042
$ 12,360
$ 6,261
Total
(106 dollars)
$ 101
$ 177
$ 181
$ 471
$ 493
$ 536
$ 571
$ 663
$ 738
$ 752
$ 767
$ 781
$ 796
$ 810
$ 825
$ 840
$ 854
$ 869
$ 883
$ 898
$ 912
$ 927
$ 941
$ 955
$ 970
$ 984
$ 999
$ 1,013
$ 1 ,028
$ 1 ,042
$ 12,907
$ 6,717
L&M Fuel Costs
500 ppm
(106 dollars)
$ 42
$ 73
$ 73
$ 89
$ 98
$ 48
$ 9
$ 4
$
$
$
-
$
$
$
-
$
$
$
-
$
$
$
-
$
$
$
-
$
$
$ 368
$ 297
15 ppm
(106 dollars)
$
$
$
$n
u
$ 0
$ 121
$ 217
$ 208
$ 200
$ 202
$ 204
$ 207
$ 210
$ 212
$ 214
$ 217
$ 219
$ 221
$ 224
$ 226
$ 229
$ 231
$ 233
$ 236
$ 238
$ 241
$ 243
$ 246
$ 248
$ 251
$ 3,052
$ 1 ,524
Total
(106 dollars)
$ 42
$ 73
$ 73
$ 89
$ 98
$ 169
$ 226
$ 212
$ 200
$ 202
$ 204
$ 207
$ 210
$ 212
$ 214
$ 217
$ 219
$ 221
$ 224
$ 226
$ 229
$ 231
$ 233
$ 236
$ 238
$ 241
$ 243
$ 246
$ 248
$ 251
$ 3,419
$ 1,821
NRLM Annual
Fuel Costs
(106 dollars)
$ 142
$ 249
$ 254
$ 561
$ 591
$ 704
$ 797
$ 874
$ 938
$ 954
$ 971
$ 988
$ 1 ,006
$ 1 ,022
$ 1 ,039
$ 1 ,056
$ 1 ,073
$ 1 ,090
$ 1,107
$ 1,124
$ 1,141
$ 1,158
$ 1,174
$ 1,191
$ 1 ,208
$ 1 ,225
$ 1 ,242
$ 1 ,259
$ 1 ,276
$ 1 ,293
$ 16,326
$ 8,538
* Fuel costs are relative to uncontrolled fuel and assume that, during the transitional years of 2010, 2012, & 2014, the first 5 months are at the previous year's cost and the remaining 7 months are at the
next year's cost.
See Appendix 8B for how these fuel volumes were developed.

-------
Final Regulatory Impact Analysis
8.4.2 Aggregate Oil-Change Maintenance Savings

   Maintenance savings associated with extended oil-change intervals are developed on a cent-
per-gallon basis, as described in Section 6.2.3.1. The cent-per-gallon savings for nonroad
engines is the fleet weighted value for nonroad engines presented in Section 6.2.3.1.  This fleet
weighted value is derived using data presented in Table 6.2-28 as discussed in that section.  The
cent-per-gallon savings for locomotive and marine engines is taken directly from Table 6.2-28.
Table 8.4-2 summarizes the annual maintenance savings and associated fuel volumes for
nonroad, locomotive, and marine engines.  Note that the fuel volumes used for oil change
maintenance savings are the same affected volumes presented in Table 8.4-1.  We have not
included savings associated with unaffected gallons (i.e., low sulfur gallons that would have
entered the NRLM fuel pool absent the new requirements) since we assume that engines
consuming those gallons benefit from the low sulfur fuel absent the NRT4 final rule. As shown
in Table 8.4-2, the net present value of the oil change maintenance savings is estimated at $7.1
billion.  This assumes a three percent social discount rate.
                                          8-18

-------
                                 Table 8.4-2
Oil-Change Maintenance Savings Associated with the Two-Step Fuel Program ($2002)
Year
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2031
2032
2033
2034
2035
2036
30YrNPVat3%
30YrNPVat7%
Affected NR
500 ppm
(106aallons)
4,790
8,406
8,599
4,014
614
528
468
199
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
24,054
20,174
15 ppm
(106aallons)
-
-
-
6,189
8,145
8,420
8,671
9,713
10,539
1 0,747
10,955
11,162
1 1 ,370
1 1 ,578
1 1 ,786
1 1 ,994
12,201
1 2,409
12,617
12,823
13,030
13,236
1 3,442
1 3,649
13,855
14,061
1 4,268
1 4,474
1 4,680
1 4,887
180,224
92,196
Affected L&M
500 ppm
(106aallons)
1,990
3,454
3,498
3,185
2,975
1,396
247
104
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
14,363
1 1 ,729
15 ppm
(106aallons)
-
-
-
0
0
1,965
3,397
3,081
2,860
2,888
2,918
2,953
2,995
3,024
3,052
3,093
3,125
3,161
3,195
3,230
3,265
3,301
3,336
3,371
3,406
3,441
3,476
3,512
3,547
3,582
44,087
22,124
NR Savings
savings=$0.029/gal
(106 dollars)
$ 140
$ 246
$ 251
$ 117
$ 18
$ 15
$ 14
$ 6
$
$
$
-
$
$
$
-
$
$
$
-
$
$
$
-
$
$
$
-
$
$
$ 703
$ 590
savings=$0.032/gal
(106 dollars)
$
$
$
$ 198
$ 261
$ 270
$ 278
$ 311
$ 338
$ 344
$ 351
$ 358
$ 364
$ 371
$ 377
$ 384
$ 391
$ 397
$ 404
$ 411
$ 417
$ 424
$ 431
$ 437
$ 444
$ 450
$ 457
$ 464
$ 470
$ 477
$ 5,772
$ 2,953
L&M Savings
savings=$0.010/gal
(106 dollars)
$ 21
$ 36
$ 37
$ 33
$ 31
$ 15
$ 3
$ 1
$
$
$
-
$
$
$
-
$
$
$
-
$
$
$
-
$
$
$
-
$
$
$ 150
$ 123
savings=$0.011/gal
(106 dollars)
$
$
$
$ 0
$ 0
$ 23
$ 39
$ 35
$ 33
$ 33
$ 33
$ 34
$ 34
$ 35
$ 35
$ 35
$ 36
$ 36
$ 37
$ 37
$ 37
$ 38
$ 38
$ 39
$ 39
$ 39
$ 40
$ 40
$ 41
$ 41
$ 506
$ 254
NRLM
Total Savings
(106 dollars)
$ 161
$ 282
$ 288
$ 349
$ 310
$ 322
$ 333
$ 353
$ 370
$ 377
$ 384
$ 391
$ 399
$ 406
$ 412
$ 420
$ 427
$ 434
$ 441
$ 448
$ 455
$ 462
$ 469
$ 476
$ 483
$ 490
$ 497
$ 504
$ 511
$ 518
$ 7,132
$ 3,919

-------
Final Regulatory Impact Analysis
8.4.3 Aggregate CDPF Maintenance, CDPF Regeneration, and CCV Maintenance Costs

   Costs associated with CDPF maintenance and CCV maintenance are developed on a cent-
per-gallon basis as described in Section 6.2.3. Table 8.4-3 summarizes the CDPF maintenance
and CDPF regeneration costs associated with the NRT4 fuel program. The fuel volumes shown
in Table 8.4-3 differ from those shown in Tables 8.4-1 through 8.4-2 because here we want only
those gallons consumed in new CDPF equipped engines.  Therefore, fuel consumed in existing
engines and fuel consumed in new engines not yet equipped with a CDPF are not included in
Table 8.4-3.

   The cent-per-gallon costs shown for CDPF maintenance are taken from data presented in
Table 6.2-29.  As engines in  different power categories add CDPFs, the weighted $/gallon
number changes until all new engines have added a CDPF and the fleet weighted average
becomes the 0.6 cents/gallon value presented in Section 6.2.3.2.  The cent-per-gallon costs
shown for CDPF regeneration are taken from information presented in Section 6.2.3.3.2.  The
weighted value shown accounts for the 60 cent/gallon base fuel cost for diesel fuel and the NOx
phase-in on different engines—engines equipped with a CDPF and no NOx adsorber incur a 2%
fuel economy impact associated with regeneration while engines equipped with both  a CDPF and
a NOx adsorber incur a 1% fuel economy impact. This weighted number also accounts for the
different 15 ppm fuel cost during the years 2010-2014 and then for 2015 and later.

   As shown in Table  8.4-3, the 30-year net present value of these two CDPF-related operating
costs is estimated at $2.3  billion.  This assumes a three percent social discount rate.
                                         8-20

-------
                                                     Aggregate Cost and Cost per Ton
                                        Table 8.4-3
 CDPF Maintenance and CDPF Regeneration Costs Associated with the Two-Step Fuel Program
                                         ($2002)
Year
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2031
2032
2033
2034
2035
2036
30YrNPVat3%
30YrNPVat7%
Fuel Consumed in New
CDPF Equipped Engines
(106 gallons)
-
-
-
-
559
1,543
2,774
4,010
5,343
6,630
7,842
8,966
10,006
10,975
1 1 ,848
12,631
13,358
14,044
14,697
15,304
15,852
16,351
16,825
17,277
17,704
18,116
18,521
18,913
19,287
19,645
1 64,697
74,092
Weighted
Maintenance
Cost
($/aal)
$
$
$
$
$ 0.002
$ 0.003
$ 0.005
$ 0.006
$ 0.006
$ 0.006
$ 0.006
$ 0.006
$ 0.006
$ 0.006
$ 0.006
$ 0.006
$ 0.006
$ 0.006
$ 0.006
$ 0.006
$ 0.006
$ 0.006
$ 0.006
$ 0.006
$ 0.006
$ 0.006
$ 0.006
$ 0.006
$ 0.006
$ 0.006


Weighted
Regeneration
Cost
($/aal)
$
$
$
$
$ 0.010
$ 0.010
$ 0.010
$ 0.007
$ 0.008
$ 0.008
$ 0.008
$ 0.008
$ 0.008
$ 0.008
$ 0.008
$ 0.008
$ 0.008
$ 0.008
$ 0.008
$ 0.008
$ 0.008
$ 0.008
$ 0.008
$ 0.008
$ 0.008
$ 0.008
$ 0.008
$ 0.008
$ 0.008
$ 0.008


CDPF Maintenance
Cost
(106 dollars)
$
$
$
$
$ 1
$ 5
$ 14
$ 23
$ 31
$ 40
$ 47
$ 55
$ 61
$ 67
$ 72
$ 77
$ 82
$ 86
$ 90
$ 94
$ 97
$ 100
$ 103
$ 106
$ 109
$ 111
$ 113
$ 116
$ 118
$ 120
$ 997
$ 445
CDPF Regeneration
Cost
(106 dollars)
$
$
$
$
$ 6
$ 15
$ 28
$ 30
$ 41
$ 52
$ 62
$ 72
$ 81
$ 89
$ 97
$ 103
$ 109
$ 114
$ 120
$ 125
$ 129
$ 133
$ 137
$ 141
$ 144
$ 148
$ 151
$ 154
$ 157
$ 160
$ 1 ,343
$ 605
Total Costs
(106 dollars)
$
$
$
$
$ 6
$ 20
$ 42
$ 53
$ 73
$ 92
$ 110
$ 127
$ 142
$ 156
$ 169
$ 180
$ 191
$ 200
$ 210
$ 218
$ 226
$ 234
$ 240
$ 247
$ 253
$ 259
$ 264
$ 270
$ 275
$ 280
$ 2,340
$ 1 ,050
* Note that fuel used in CDPF engines includes some highway spillover fuel.
"Weighted Regeneration Cost ($/gal) changes year-to-year due to different fuel economy impacts with a NOx adsorber (1 percent)
and without a NOx adsorber (2 percent) matched with the phase-in schedules of the emission standards.
    The cent-per-gallon costs for CCV maintenance are taken from data presented in Table 6.2-
30. Table 8.4-4 presents the annual costs associated with CCV maintenance. The gallons shown
in Table 8.4-4 are gallons of fuel consumed in engines in power ranges for which the new CCV
requirements have gone into effect.  However, these are not necessarily equal to the gallons
consumed in new CCV equipped engines since only the turbocharged engines will be adding a
CCV system.  Therefore, the cent-per-gallon costs in early years is essentially zero since so few
engines in the <75hp range are turbocharged and, hence, so few are adding a CCV system and
incurring the associated maintenance costs. As shown in Table  8.4-4, the 30-year net present
value of the CCV maintenance costs are estimated at $275 million. This assumes a three percent
social discount rate.
                                           8-21

-------
Final Regulatory Impact Analysis
                                       Table 8.4-4
            CCV Maintenance Costs Associated with the Two-Step Fuel Program
                                         ($2002)
Year
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2031
2032
2033
2034
2035
2036
30YrNPVat3%
30YrNPVat7%
Fuel Consumed in Power
Categories Adding CCV
System
(106aallons)
-
242
248
254
927
2,023
3,369
4,716
6,160
7,552
8,857
10,042
11,139
12,161
13,084
13,913
14,680
15,402
16,088
16,724
17,301
17,827
18,327
18,805
19,258
19,695
20,125
20,543
20,940
21,323
182,540
82,865
Weighted
Maintenance
Cost
($/aal)
$
$ 0.000
$ 0.000
$ 0.000
$ 0.001
$ 0.001
$ 0.002
$ 0.002
$ 0.002
$ 0.002
$ 0.002
$ 0.002
$ 0.002
$ 0.002
$ 0.002
$ 0.002
$ 0.002
$ 0.002
$ 0.002
$ 0.002
$ 0.002
$ 0.002
$ 0.002
$ 0.002
$ 0.002
$ 0.002
$ 0.002
$ 0.002
$ 0.002
$ 0.002


Total Costs
(106 dollars)
$
$ 0
$ 0
$ 0
$ 1
$ 3
$ 5
$ 7
$ 9
$ 11
$ 13
$ 15
$ 17
$ 18
$ 20
$ 21
$ 22
$ 23
$ 24
$ 25
$ 26
$ 27
$ 28
$ 28
$ 29
$ 30
$ 30
$ 31
$ 32
$ 32
$ 275
$ 124
                 * Weighted Maintenance
                 implementation schedule
Cost ($/gal) changes year-to-year due to the
for engines adding the CCV system.
8.4.4 Summary of Aggregate Operating Costs

   The net operating costs include the incremental costs for fuel (Table 8.4-1), cost savings
from reduced oil changes (Table 8.4-2), costs for CDPF maintenance and regeneration (Table
8.4-3), and costs for CCV maintenance (Table 8.4-4).  The results of this summation for the two-
step NRT4 program are shown in Table 8.4-5. The oil-change maintenance savings, CDPF
maintenance and regeneration costs, and the CCV maintenance costs are added together in Table
8.4-5 and presented as "Other Operating Costs." The other operating costs are presented as
negative values because the oil change maintenance savings (negative costs) outweigh the other
                                          8-22

-------
                                                   Aggregate Cost and Cost per Ton
operating costs and, thus, their summation represents a net savings. The "Net Operating Cost" is
the sum of the incremental fuel costs shown in Table 8.4-1 and the other operating costs shown
in Tables 8.4-2 through 8.4-4.  As shown in Table 8.4-5, the 30-year net present value of the net
operating costs is estimated at $11.8 billion consisting of the $16.3 billion fuel cost and the $4.5
billion savings associated with other operating costs. These net present values assume a three
percent social discount rate.

   Also included in Table 8.4-5  are the costs by pollutant (refer to Table 8.1-2 for how these
costs have been allocated). The sum of the SOx cost, the PM cost, and the NOx+NMHC cost is
the value presented in the "Net Operating Cost" column.
                                      Table 8.4-5
                   Aggregate Net Operating Costs and
                           Associated with the NRT4
                                        ($2002)
Costs by Pollutant
Program
Year
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2031
2032
2033
2034
2035
2036
30YrNPVat3%
30YrNPVat7%
Fuel Costs
(106 dollars)
$ 142
$ 249
$ 254
$ 561
$ 591
$ 704
$ 797
$ 874
$ 938
$ 954
$ 971
$ 988
$ 1 ,006
$ 1 ,022
$ 1 ,039
$ 1 ,056
$ 1 ,073
$ 1 ,090
$ 1,107
$ 1,124
$ 1,141
$ 1,158
$ 1,174
$ 1,191
$ 1 ,208
$ 1 ,225
$ 1 ,242
$ 1 ,259
$ 1 ,276
$ 1 ,293
$ 16,326
$ 8,538
Other Operating
Costs
(106 dollars)
$ (161)
$ (282)
$ (288)
$ (349)
$ (302)
$ (299)
$ (286)
$ (294)
$ (288)
$ (274)
$ (261)
$ (250)
$ (240)
$ (231 )
$ (224)
$ (219)
$ (214)
$ (210)
$ (207)
$ (204)
$ (202)
$ (201 )
$ (201 )
$ (201 )
$ (201 )
$ (201 )
$ (202)
$ (203)
$ (204)
$ (205)
$ (4,517)
$ (2,745)
Net Operating
Costs
(106 dollars)
$ (18)
$ (33)
$ (34)
$ 212
$ 289
$ 406
$ 512
$ 581
$ 650
$ 680
$ 710
$ 738
$ 766
$ 791
$ 815
$ 838
$ 859
$ 880
$ 900
$ 920
$ 938
$ 956
$ 974
$ 991
$ 1 ,007
$ 1 ,024
$ 1 ,040
$ 1 ,056
$ 1 ,072
$ 1 ,088
$ 1 1 ,809
$ 5,793
SOx Related
Costs
(106 dollars)
$ (12)
$ (22)
$ (23)
$ 88
$ 117
$ 172
$ 217
$ 232
$ 245
$ 249
$ 253
$ 257
$ 261
$ 265
$ 268
$ 272
$ 276
$ 280
$ 284
$ 288
$ 292
$ 296
$ 300
$ 304
$ 308
$ 312
$ 316
$ 320
$ 324
$ 328
$ 3,934
$ 1 ,976
PM Related
Costs
(106 dollars)
$ (6)
$ (11)
$ (11)
$ 84
$ 118
$ 170
$ 223
$ 259
$ 300
$ 324
$ 347
$ 368
$ 389
$ 408
$ 425
$ 441
$ 456
$ 470
$ 484
$ 497
$ 509
$ 521
$ 532
$ 543
$ 553
$ 563
$ 573
$ 583
$ 593
$ 602
$ 6,091
$ 2,928
NOx+HC
Related Costs
(106 dollars)
-
$ 0
$ 0
$ 40
$ 54
$ 64
$ 72
$ 90
$ 105
$ 108
$ 111
$ 114
$ 116
$ 119
$ 122
$ 124
$ 127
$ 129
$ 132
$ 134
$ 137
$ 139
$ 141
$ 144
$ 146
$ 148
$ 151
$ 153
$ 155
$ 157
$ 1 ,784
$ 889

-------
Final Regulatory Impact Analysis
8.4.5 Summary of Aggregate Operating Costs Associated with a Fuel-only Scenario

   The aggregate operating costs of a fuel-only scenario would be essentially the same as those
presented above for the full NRT4 program with the exception of those operating costs
associated with maintenance or regeneration of new engine hardware.  These operating cost
elements would not be incurred because without new engine standards the new engine hardware
would not be added. However, the oil change maintenance savings would still be realized just as
they would under the full NRT4 program.

   As noted several times throughout this chapter, Table 8.1-2 shows how we allocated costs to
each pollutant under the full engine and fuel program. However, the allocations shown in that
table assume an engine program to which a  portion of the fuel-related costs are allocated.
Specifically, the 15 ppm NR fuel, which enables aftertreatment devices and, thus, new NR
engine standards, is split evenly between engine derived benefits and fuel derived benefits.
Subsequently, the costs allocated to fuel derived benefits were split one-third to PM and two-
thirds to SOx.

   Under the fuel-only scenario, there are no new engine standards. As  a result, all the fuel
costs are allocated to fuel-derived benefits.  Consistent with the approach taken in the full engine
and fuel program, we have allocated one-third of those costs to PM and two-thirds of those costs
to SOx. Table 8.4-6 shows the cost allocations under the fuel-only scenario.

                                      Table 8.4-6
                       Cost Allocations under the Fuel-only Scenario
Item
Fuel Costs - incremental cent/gallon
Operating Costs - Oil-Change Savings
Operating Costs - CDPF Maintenance
Operating Costs - CDPF Regen (FE impact)
Operating Costs - CCV Maintenance

500 ppm Affected NRLM
1 5 ppm Affected NR
15ppmAffectedL&M
500 ppm Affected NRLM
15 ppm Affected NR
15ppmAffectedL&M
NOx+HC

PM
33%
SOx
67%
None
   Note that there are no costs associated with CDPF and CCV maintenance or with CDPF
regeneration since there would be no new engine standards under the fuel-only scenario. Note
also that the oil change maintenance savings would still be realized absent any new engine
standards.

   Table 8.4-7 presents the net operating costs associated with a fuel-only scenario.  The costs
presented in Table 8.4-7 include the incremental costs for fuel (Table 8.4-1) and costs for oil-
change maintenance savings (Table 8.4-2).  The oil-change maintenance savings are presented in
the table as "Other Operating Costs," and, thus represent a net savings.  The "Net Operating
                                          8-24

-------
                                                    Aggregate Cost and Cost per Ton
Cost" is the sum of the incremental fuel costs and the other operating costs. Table 8.4-7 also
presents these costs by pollutant (refer to Table 8.4-6 for how these costs have been allocated).
Since there are no new engine standards under a fuel-only scenario there are no costs associated
with technology enablement and, hence, no costs allocated to NOx+NMHC. As shown in Table
8.4-7, the 30-year net present value of costs associated with a fuel-only scenario is estimated at
$9.2 billion consisting of the $16.3 billion fuel cost and a $7.1 billion savings associated with oil
change maintenance. These values assume a three percent social discount rate.

                                       Table 8.4-7
                   Aggregate Net Operating Costs and Costs by Pollutant
                           Associated with a Fuel-Only Scenario
                                         ($2002)
Year
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2031
2032
2033
2034
2035
2036
30 Yr NPVat3%
30YrNPVat7%
Fuel Costs
(106 dollars)
$ 142
$ 249
$ 254
$ 561
$ 591
$ 704
$ 797
$ 874
$ 938
$ 954
$ 971
$ 988
$ 1 ,006
$ 1 ,022
$ 1 ,039
$ 1 ,056
$ 1 ,073
$ 1 ,090
$ 1,107
$ 1,124
$ 1,141
$ 1,158
$ 1,174
$ 1,191
$ 1 ,208
$ 1 ,225
$ 1 ,242
$ 1 ,259
$ 1 ,276
$ 1 ,293
$ 16,326
$ 8,538
Other Operating
Costs
(106 dollars)
$ (161)
$ (282)
$ (288)
$ (349)
$ (31 0)
$ (322)
$ (333)
$ (353)
$ (370)
$ (377)
$ (384)
$ (391)
$ (399)
$ (406)
$ (41 2)
$ (420)
$ (427)
$ (434)
$ (441)
$ (448)
$ (455)
$ (462)
$ (469)
$ (476)
$ (483)
$ (490)
$ (497)
$ (504)
$ (511)
$ (51 8)
$ (7,132)
$ (3,919)
Net Operating
Costs
(106 dollars)
$ (18)
$ (33)
$ (34)
$ 212
$ 281
$ 382
$ 464
$ 521
$ 568
$ 577
$ 587
$ 597
$ 607
$ 617
$ 626
$ 636
$ 646
$ 656
$ 666
$ 676
$ 686
$ 696
$ 706
$ 716
$ 725
$ 735
$ 745
$ 755
$ 765
$ 775
$ 9,194
$ 4,618
SOx Related
Costs
(106 dollars)
$ (12)
$ (22)
$ (23)
$ 141
$ 187
$ 255
$ 310
$ 347
$ 378
$ 385
$ 391
$ 398
$ 405
$ 411
$ 417
$ 424
$ 431
$ 437
$ 444
$ 451
$ 457
$ 464
$ 470
$ 477
$ 484
$ 490
$ 497
$ 503
$ 510
$ 517
$ 6,130
$ 3,079
PM Related
Costs
(106 dollars)
$ (6)
$ (11)
$ (11)
$ 71
$ 94
$ 127
$ 155
$ 174
$ 189
$ 192
$ 196
$ 199
$ 202
$ 206
$ 209
$ 212
$ 215
$ 219
$ 222
$ 225
$ 229
$ 232
$ 235
$ 239
$ 242
$ 245
$ 248
$ 252
$ 255
$ 258
$ 3,065
$ 1 ,539
NOx+HC Related
Costs
(106 dollars)
$
$
$
$
$
$
$
-
$
$
$
-
$
$
$
-
$
$
$
-
$
$
$
-
$
$
$
-
$
$
$
$
                                          8-25

-------
Final Regulatory Impact Analysis
8.5 Summary of Aggregate Costs of the Final Rule

   Table 8.5-1 presents a summary of all the costs presented above for the NRT4 final rule
engine and fuel program.  Engine costs are the summation of costs presented in Tables 8.2-1 and
8.2-3, equipment costs are the summation of costs presented in Tables 8.3-1 and 8.3-3, and fuel
costs, other operating costs, and net operating costs are presented in Table 8.4-5. The "Total
Program Costs" are the summation of engine costs,  equipment costs, and net operating costs. As
shown, the 30-year net present value of the NRT4 program is estimated at $27.1 billion
consisting of $14.1 billion in  engine costs, $1.3 billion in equipment costs, $16.3 billion in fuel
costs, and a savings of $4.5 billion in other operating costs.  These values assume a three
percent social discount rate.

   Table 8.5-2 presents the summary of all the costs presented above by pollutant (refer to Table
8.1-2 for how we have allocated costs among the various pollutants).

   Note that a similar summary of aggregate costs associated with a fuel-only scenario are
presented in full in Table 8.4-6  since there are no new engine or equipment costs associated with
that scenario.
                                          8-26

-------
                                      Aggregate Cost and Cost per Ton
                          Table 8.5-1
Summary of Aggregate Costs for the NRT4 Final Engine and Fuel Program
                    (SMillions of 2002 dollars)
Year
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2031
2032
2033
2034
2035
2036
30YrNPVat3%
30 Yr NPVat7%
Engine
Costs
-
$ 81
$ 82
$ 80
$ 403
$ 718
$ 882
$ 973
$ 950
$ 920
$ 910
$ 901
$ 890
$ 900
$ 913
$ 927
$ 940
$ 954
$ 967
$ 980
$ 994
$ 1,007
$ 1,021
$ 1,034
$ 1,048
$ 1,061
$ 1,074
$ 1,088
$ 1,101
$ 1,115
$14,054
$ 7,215
Equipment
Costs
-
$ 5
$ 5
$ 5
$ 62
$ 106
$ 121
$ 146
$ 149
$ 150
$ 150
$ 146
$ 147
$ 147
$ 99
$ 66
$ 56
$ 36
$ 32
$ 32
$ 33
$ 33
$ 33
$ 34
$ 34
$ 35
$ 35
$ 35
$ 36
$ 36
$ 1 ,281
$ 754
Fuel Costs
$ 142
$ 249
$ 254
$ 561
$ 591
$ 704
$ 797
$ 874
$ 938
$ 954
$ 971
$ 988
$ 1 ,006
$ 1 ,022
$ 1 ,039
$ 1 ,056
$ 1 ,073
$ 1 ,090
$ 1,107
$ 1,124
$ 1,141
$ 1,158
$ 1,174
$ 1,191
$ 1 ,208
$ 1 ,225
$ 1 ,242
$ 1 ,259
$ 1 ,276
$ 1 ,293
$ 16,326
$ 8,538
Other
Operating
Costs
$ (161)
$ (282)
$ (288)
$ (349)
$ (302)
$ (299)
$ (286)
$ (294)
$ (288)
$ (274)
$ (261 )
$ (250)
$ (240)
$ (231 )
$ (224)
$ (219)
$ (214)
$ (210)
$ (207)
$ (204)
$ (202)
$ (201 )
$ (201 )
$ (201 )
$ (201 )
$ (201 )
$ (202)
$ (203)
$ (204)
$ (205)
$(4,517)
$ (2,745)
Net
Operating
Costs
$ (18)
$ (33)
$ (34)
$ 212
$ 289
$ 406
$ 512
$ 581
$ 650
$ 680
$ 710
$ 738
$ 766
$ 791
$ 815
$ 838
$ 859
$ 880
$ 900
$ 920
$ 938
$ 956
$ 974
$ 991
$ 1,007
$ 1,024
$ 1,040
$ 1,056
$ 1,072
$ 1,088
$ 1 1 ,809
$ 5,793
Total
Annual
Costs
$ (18)
$ 53
$ 53
$ 297
$ 754
$ 1 ,229
$ 1,515
$ 1 ,699
$ 1 ,749
$ 1 ,750
$ 1 ,770
$ 1 ,785
$ 1 ,802
$ 1 ,838
$ 1 ,827
$ 1 ,830
$ 1 ,855
$ 1 ,869
$ 1 ,899
$ 1 ,932
$ 1 ,965
$ 1 ,997
$ 2,028
$ 2,059
$ 2,089
$ 2,119
$ 2,149
$ 2,179
$ 2,209
$ 2,239
$ 27,144
$ 13,762
                             8-27

-------
Final Regulatory Impact Analysis
                                    Table 8.5-2
                 Summary of Aggregate Costs for the NRT4 Final Engine
                            and Fuel Program by Pollutant
                             (SMillions of 2002 dollars)
Year
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2031
2032
2033
2034
2035
2036
30YrNPVat3%
30YrNPVat7%
PM Costs
$ (6)
$ 74
$ 75
$ 169
$ 458
$ 761
$ 949
$ 940
$ 970
$ 982
$ 999
$ 1 ,004
$ 1 ,034
$ 1 ,059
$ 1 ,061
$ 1 ,070
$ 1 ,088
$ 1,112
$ 1,130
$ 1,153
$ 1,174
$ 1,195
$ 1,215
$ 1 ,235
$ 1 ,254
$ 1 ,273
$ 1 ,292
$ 1,311
$ 1 ,330
$ 1 ,348
$ 16,041
$ 8,134
NOx+NMHC
Costs
-
$ 0
$ 0
$ 40
$ 179
$ 296
$ 348
$ 526
$ 534
$ 519
$ 518
$ 524
$ 507
$ 515
$ 497
$ 488
$ 490
$ 477
$ 484
$ 491
$ 498
$ 506
$ 513
$ 520
$ 527
$ 534
$ 541
$ 548
$ 555
$ 562
$ 7,169
$ 3,652
SOx Costs
$ (12)
$ (22)
$ (23)
$ 88
$ 117
$ 172
$ 217
$ 232
$ 245
$ 249
$ 253
$ 257
$ 261
$ 265
$ 268
$ 272
$ 276
$ 280
$ 284
$ 288
$ 292
$ 296
$ 300
$ 304
$ 308
$ 312
$ 316
$ 320
$ 324
$ 328
$ 3,934
$ 1 ,976
Total Costs
$ (18)
$ 53
$ 53
$ 297
$ 754
$ 1 ,229
$ 1,515
$ 1 ,699
$ 1 ,749
$ 1 ,750
$ 1 ,770
$ 1 ,785
$ 1 ,802
$ 1 ,838
$ 1 ,827
$ 1 ,830
$ 1 ,855
$ 1 ,869
$ 1 ,899
$ 1 ,932
$ 1 ,965
$ 1 ,997
$ 2,028
$ 2,059
$ 2,089
$ 2,119
$ 2,149
$ 2,179
$ 2,209
$ 2,239
$ 27,144
$ 13,762
                                       8-28

-------
                                                      Aggregate Cost and Cost per Ton
8.6 Emission Reductions
    Table 8.6-1 presents the emission reductions estimated to result from the fuel program in
conjunction with the new engine standards. Also presented are reductions associated with a fuel-
only scenario. A complete discussion of these emission reductions and how they were generated
can be found in Chapter 3.

                                         Table 8.6-1
        Emission Reductions Associated with the NRT4 Final Fuel and Engine Program
                              and the Fuel-only  Scenario (tons)
Year
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2031
2032
2033
2034
2035
2036
30YrNPVat3%
30YrNPVat7%
NRT4 Fuel and Engine Program
PM
10,700
19,500
20,400
22,300
25,900
32,100
39,200
46,900
54,900
62,400
69,600
76,400
82,800
88,800
94,400
99,700
104,600
109,400
113,900
118,200
122,300
125,900
129,500
132,900
136,000
139,100
142,100
145,000
147,800
150,500
1 ,430,500
690,800
NOx+NMHC
0
200
400
700
19,100
49,600
84,400
143,600
203,000
261,100
316,900
368,500
417,300
463,000
504,400
542,400
578,100
611,100
642,300
671 ,400
698,200
723,200
746,900
768,500
788,800
808,400
827,300
845,600
863,100
880,100
7,077,900
3,142,700
SOx
133,000
235,400
240,100
255,500
268,600
277,800
285,700
291,600
297,400
302,600
307,700
312,900
318,300
323,300
328,300
333,600
338,800
344,000
349,200
354,400
359,600
364,800
370,000
375,300
380,500
385,800
391,000
396,300
401 ,600
406,900
5,725,900
3,164,100
NRLM Fuel-only Program
PM
10,700
19,000
19,400
20,600
21,600
22,400
23,000
23,500
24,000
24,400
24,800
25,200
25,600
26,000
26,400
26,900
27,300
27,700
28,100
28,500
28,900
29,400
29,800
30,200
30,600
31,000
31,500
31,900
32,300
32,700
461 ,000
254,800
SOx
133,000
235,400
240,100
255,500
268,600
277,700
285,500
291 ,500
297,300
302,400
307,500
312,700
318,000
323,100
328,000
333,400
338,500
343,700
348,900
354,100
359,300
364,500
369,700
374,900
380,100
385,400
390,600
395,900
401 ,200
406,400
5,722,100
3,162,300
               b Note that the SOx reductions for the Final program and the fuel-only scenario are
               nearly identical while the PM reductions are very different. This is a result of there being
               no new engine standards under the fuel-only scenario and, therefore, no CDPFs added
               to new engines.
                                            8-29

-------
Final Regulatory Impact Analysis
8.7 Cost per Ton

   We have calculated the cost per ton of the final rule based on the net present value of all
costs incurred and all emission reductions generated over a 30-year time window following
implementation of the program.  This approach captures all the costs and emission reductions
from the final rule, including costs incurred and emission reductions generated by both the new
and the existing fleet.

   The baseline (i.e., the point of comparison) for this evaluation is the existing set of engine
standards (i.e., the Tier 2/Tier 3 program) and fuel standards (i.e., unregulated sulfur level). The
30-year time window is meant to capture both the early period of the program when there are a
small number of compliant engines in the fleet, and the later period when there is nearly
complete turnover to compliant engines. The final rule also requires reduced sulfur content in
NRLM diesel fuel with a 500 ppm cap beginning in 2007, a 15 ppm NR cap beginning in 2010,
and a 15 ppm L&M  cap beginning in 2012.

   In Section 8.7.1 we present the cost per ton for the NRT4 final engine and fuel
program—this represents the cost per ton of this final rule including all costs and emissions
reductions associated with the new fuel standards and the new engine standards.  In Section 8.7.2
we present the cost per ton for the fuel-only scenario—this scenario would include  the same fuel
standards as the full  engine and fuel program but no new engine standards.  In Section 8.7.3 we
present two different sets of cost per ton information—cost per ton of a 500 ppm fuel scenario
should it remain in place forever with no new engine standards, and the incremental cost per ton
of the 15 ppm L&M portion of the fuel program. In Section 8.7.4, we summarize all the cost per
ton calculations presented in Sections 8.7.1 through 8.7.3. In Appendix 8A, we present the cost
per ton of two sensitivity cases—the case 1 sensitivity shows the cost per ton using future
projections of fuel demand developed by the Energy Information Administration; and, the case 2
sensitivity shows the cost per ton if we increase the percentage of mobile versus stationary
generator sets (i.e., increase the number of generator sets that will meet the new standards) and
increase the usage rates for some >750hp equipment. The rationale for choosing these two
sensitivity cases is presented in  section 8A. 1.

8.7.1 Cost per Ton for the NRT4 Final Rule

   The NRT4 final rule adopts fuel requirements in two steps—reducing NRLM sulfur levels
from current uncontrolled levels to 500 ppm in 2007 and then controlling NR fuel and L&M fuel
to 15 ppm in 2010 and 2012, respectively. Beginning June 1, 2007, refiners must produce
NRLM diesel fuel that meets a maximum sulfur level of 500 ppm. Then, beginning in June 1,
2010, NR fuel must meet a maximum sulfur level of 15 ppm and, beginning in June 1, 2012,
L&M fuel must meet a maximum sulfur level of 15  ppm. This program also adopts new Tier 4
engine standards for nonroad diesel engines that begin in different years for different power
categories.  See Table 1 in the Executive Summary for details on the new engine standards and
when they are implemented. All nonroad diesel-fueled engines with a CDPF must be refueled
with the new 15 ppm diesel fuel.
                                          8-30

-------
                                                    Aggregate Cost and Cost per Ton
   The costs of the final rule include costs associated with both steps in the fuel program (500
ppm and  15 ppm) and costs for the engine standards including equipment modifications.
Maintenance costs  and savings realized by both the existing fleet (nonroad, locomotive, and
marine), future locomotive and marine engines, and the new fleet of nonroad engines complying
with the new emissions standards are included. Figure 8.7-1 presents in graphic form the cost of
the final rule. These costs are summarized in Table 8.5-1. The cost streams include the
amortized capital (fixed) costs and variable costs.
                                      Figure 8.7-1
                     Estimated Aggregate Cost of the NRT4 Final Rule
   $2,500
                                                                                    2036
   -$1,000
                                              Year
     •Engine Costs —^Equipment Costs
Costs  —^Other Operating Costs  -^Total Program Costs
   Figure 8.7-1 shows that total annual costs are estimated to be $50 million in the first year the
new engine standards apply, increasing to $2.2 billion in 2036 as increasing numbers of engines
become subject to the new standards and an ever increasing amount of fuel is consumed.  As
shown in Table 8.5-1, the 30-year net present value of the costs for this program is estimated as
$27.1 billion using a three percent discount rate.
                                          8-31

-------
Final Regulatory Impact Analysis
   The calculations of cost per ton of each emission reduced under the final program divides the
net present value of the annual costs assigned to each pollutant (see Table 8.5-2 for costs by
pollutant and Table 8.1-2 for how we have allocated costs by pollutant) by the net present value
of the total annual reductions of each pollutant - NOx+NMHC, PM and SOx (see Table 8.6-1).

   The net present value of the costs associated with each pollutant, calculated with a three
percent discount rate, are shown in Table 8.5-1  as $7.2 billion for NOx+NMHC, $16.0 billion for
PM and $3.9 billion for SOx.  The 30-year net present value, with a three percent discount rate,
of emission reductions are 7.1 million tons for NOx+NMHC, 1.4 million tons for PM and 5.7
million tons for SOx (see Table 8.6-1). Our air quality analysis, emissions reduction analysis,
and benefits analysis are found in Chapters 2, 3, and 9, respectively.

   The cost per ton of emissions reduced for the NRT4 final rule is calculated by dividing the
net present value of the annualized costs of the program through 2036 by the net present value of
the annual emission reductions through 2036.  These results are shown in Table 8.7-1.
                                          8-32

-------
                                                    Aggregate Cost and Cost per Ton
                                       Table 8.7-1
                 Aggregate Costs and Costs per Ton for the NRT4 Final Rule
              30-year Net Present Values at a 3% and 7% Discount Rate ($2002)
Item
SOOppm at $0.021/gal, 2007-2010
SOOppm at $0.033/gal, 2010-2012
SOOppm at $0.035/gal, 2012-2014
ISppm at $0.058/gal, 2010-2012
15ppm at $0.064/gal, 2012-2014
ISppm at $0.070/gal, 2014+
SOOppm Fuel Cost
ISppmFuel Cost
Other Operating Costs*
Engine Costs
Equipment Costs
Total Program Costs
NOx+NMHC Costs
PM Costs
SOx Costs
NOx+NMHC Reduction
PM Reduction
SOx Reduction
Cost per Ton NOx+NMHC
Cost per Ton PM
Cost per Ton
Units
(106 gallons)
(106 gallons)
(106 gallons)
(106 gallons)
(106 gallons)
(106 gallons)
(Smillion)
(Smillion)
(Smillion)
(Smillion)
(Smillion)
(Smillion)
(Smillion)
(Smillion)
(Smillion)
(106tons)
(106tons)
(106tons)
($/ton)
($/ton)
($/ton)
3% discount rate
29,690
7,068
1,660
15,223
17,998
191,091
$915
$15,411
-$4,517
$14,054
$1,281
$27,144
$7,169
$16,041
$3,934
7.1
1.4
5.7
$1,010
$11,200
$690
7% discount rate
25,207
5,500
1,196
11,715
12,800
89,805
$753
$7,785
-$2,745
$7,215
$754
$13,762
$3,652
$8,134
$1,976
3.1
0.7
3.2
$1,160
$11,800
$620
Source
Table 8.4-1
Table 8.4-1
Table 8.4-1
Table 8.4-1
Table 8.4-1
Table 8.4-1
Table 8.4-1
Table 8.4-1
Table 8.4-5
Table 8.5-1
Table 8.5-1
Table 8.5-1
Table 8.5-2
Table 8.5-2
Table 8.5-2
Table 8.6-1
Table 8.6-1
Table 8.6-1
Calculated
Calculated
Calculated
* Other operating costs include oil change maintenance savings, CDPF and CCV maintenance costs, and CDPF
regeneration costs.
       We have also calculated the cost per ton of emissions reduced in the year 2030 using the
annual costs and emission reductions in that year alone. This number, shown in Table 8.7-2,
approaches the long-term cost per ton of emissions reduced after all fixed costs of the program
have been recovered by industry leaving only the variable costs of control (and maintenance
costs), and after most (though not all) of the pre-control fleet has been retired.
                                          8-33

-------
Final Regulatory Impact Analysis
                                      Table 8.7-3
                     Long-Term Cost per Ton of the NRT4 Final Rule
                        Annual Values without Discounting ($2002)
Pollutant
NOx+NMHC
PM
SOx
Long-Term Cost per
Ton in 2030
$680
$9,300
$810
8.7.2 Cost per Ton for the NRLM Fuel-only Scenario

   The costs of the fuel-only scenario include costs associated with both steps in the fuel
program absent any new engine standards. Oil change maintenance savings would be realized
by both the existing fleet and the new fleet of engines as these savings are not dependent on any
new engine standards.  Figure 8.7-2 presents in graphic form the cost of the fuel-only scenario.
These costs are summarized in Table 8.4-7. The cost streams include the amortized capital
(fixed) costs and variable costs.
                                         8-34

-------
                                                     Aggregate Cost and Cost per Ton
                                       Figure 8.7-2
                 Estimated Aggregate Cost of the NRLM Fuel-only Scenario
   $2,500
   $2,000
   $1,500
   $1,000
o
O
E    $500
D)
O
     $-
    $(500)
   $(1,000)
        2004
                                                  2036
                                               Year
                   •Fuel Costs
•Other Operating Costs
•Net Costs
   Figure 8.7-2 shows that total annual costs are estimated to be -$33 million in the first full
year of the new fuel standards (i.e., a $33 million savings), increasing to $775 million in 2036 as
an ever increasing amount of fuel is consumed.  As shown in Table 8.4-7, the 30-year net present
value of the fuel-only scenario is estimated as $9.2 billion using a three percent discount rate.

   The calculations of cost per ton of each emission reduced under the fuel-only scenario
divides the net present value of the annual costs assigned to each pollutant (see Table 8.4-7 for
costs by pollutant and Table 8.4-6 for how we have allocated costs by pollutant) by the net
present value of the total annual reductions of each pollutant.  The 30-year net present value of
the costs associated with each pollutant, calculated with a three percent discount rate, are shown
in Table 8.4-7 as $3.1 billion for PM and  $6.1 billion for SOx.  If we exclude the oil change
maintenance savings, the costs of the fuel-only scenario would be $5.4 billion for PM and $10.9
billion for SOx.  The 30-year net present value, with a three percent discount rate, of emission
reductions are 461 thousand tons for PM and 5.7 million tons for SOx. Our air quality analysis,
emissions reduction analysis, and benefits analysis are found in Chapters 2, 3, and 9,
                                           8-35

-------
Final Regulatory Impact Analysis
respectively.  Table 8.7-4 presents the cost per ton results for the fuel-only scenario including the
oil change maintenance savings and excluding those savings.

                                        Table 8.7-4
                Aggregate Costs and Costs per Ton for the Fuel-only Scenario
              30-year Net Present Values at a 3% and 7% Discount Rate ($2002)
Item
SOOppm at $0.021/gal, 2007-2010
SOOppm at $0.033/gal, 2010-2012
SOOppm at $0.035/gal, 2012-2014
ISppm at $0.058/gal, 2010-2012
ISppm at $0.064/gal, 2012-2014
ISppm at $0.070/gal, 2014+
SOOppm Fuel Cost
15ppm Fuel Cost
Other Operating Costs*
Total Costs (w/ maintenance savings)
Total Costs (w/o maintenance savings)
PM Costs (w/ maintenance savings)
PM Costs (w/o maintenance savings)
SOx Costs (w/ maintenance savings)
SOx Costs (w/o maintenance savings)
PM Reduction
SOx Reduction
Cost per Ton PM (w/ maintenance savings)
Cost per Ton PM (w/o maintenance savings)
Cost per Ton SOx (w/ maintenance savings)
Cost per Ton Sox (w/o maintenance savings)
Units
(106 gallons)
(106 gallons)
(106 gallons)
(106 gallons)
(106 gallons)
(106 gallons)
(Smillion)
(Smillion)
(Smillion)
(Smillion)
(Smillion)
(Smillion)
(Smillion)
(Smillion)
(Smillion)
(106tons)
(106tons)
($/ton)
($/ton)
($/ton)
($/ton)
3% discount
rate
29,690
7,068
1,660
15,223
17,998
191,091
$915
$15,411
-$4,517
$9,194
$16,326
$3,065
$5,442
$6,130
$10,884
0.46
5.7
$6,600
$11,800
$1,070
$1,900
7% discount
rate
25,207
5,500
1,196
11,715
12,800
89,805
$753
$7,785
-$2,745
$4,618
$8,538
$1,539
$2,846
$3,079
$5,692
0.26
3.2
$6,000
$11,200
$970
$1,800
Source
Table 8.4-1
Table 8.4-1
Table 8.4-1
Table 8.4-1
Table 8.4-1
Table 8.4-1
Table 8.4-1
Table 8.4-1
Table 8.4-7
Table 8.4-7
Table 8.4-7
Table 8.4-7
Calculated**
Table 8.4-7
Calculated* *
Table 8.6-1
Table 8.6-1
Calculated
Calculated
Calculated
Calculated
* Other operating costs include oil change maintenance savings.
** Calculated as one-third (PM) or two-thirds (SOx) of the Total Scenario Costs w/o maintenance savings.
       We have also calculated the cost per ton of emissions reduced in the year 2030 using the
annual costs and emission reductions in that year alone. This number, shown in Table 8.7-5,
approaches the long-term cost per ton of emissions reduced.
                                           8-36

-------
                                                   Aggregate Cost and Cost per Ton
                                      Table 8.7-5
                 Long-Term Cost per Ton of the NRT4 Fuel-only Scenario
                       Annual Values without Discounting ($2002)
Pollutant
PM (with maintenance savings)
PM (without maintenance savings)
SOx (with maintenance savings)
SOx (without maintenance savings)
Long-Term Cost per Ton
in 2030
$7,900
$13,200
$1,270
$2,100
8.7.3 Costs and Costs per Ton for Other Control Scenarios

   Here we look at the costs and costs per ton of other control scenarios.  Specifically, we look
at the cost per ton of the 500 ppm NRLM fuel scenario should it continue forever without any
new engine standards. We also look at the incremental cost per ton of the 15 ppm L&M fuel
scenario.

   8.7.3.1 Costs and Costs per Ton of a 500 ppm NRLM Fuel-only Scenario

   A 500 ppm NRLM fuel-only scenario would mirror the fuel-only scenario discussed above
with the exception that no 15 ppm fuel step would occur. The incremental fuel cost would be
$0.021  per gallon during the years 2007 through 2010 and then $0.022 per gallon thereafter (see
Table 7.5-1).  The oil change maintenance savings would be $0.029 per gallon for NR and
$0.010 per gallon for L&M (see Table 8.4-2). Tables 8.7-6 and 8.7-7 present the fuel costs and
oil change maintenance savings, respectively, associated with a 500 ppm NRLM fuel-only
scenario.
                                         8-37

-------
Final Regulatory Impact Analysis
                                            Table 8.7-6
             Aggregate Fuel Costs of a 500 ppm NRLM Fuel-only Scenario ($2002)
Year
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2031
2032
2033
2034
2035
2036
30YrNPVat3%
30 Yr NPVat7%
Affected NR Fuel
500 ppm
(106 gallons)
4,790
8,406
8,599
8,400
8,300
8,479
8,659
8,839
9,018
9,196
9,374
9,552
9,730
9,907
10,085
10,263
10,441
10,619
10,797
10,973
11,150
1 1 ,326
1 1 ,503
1 1 ,679
1 1 ,856
12,032
12,209
12,386
12,562
12,739
179,520
99,928
Affected L&M Fuel
500 ppm
(106 gallons)
1,990
3,454
3,498
3,457
3,450
3,489
3,518
3,552
3,586
3,623
3,659
3,699
3,747
3,781
3,812
3,859
3,897
3,939
3,980
4,022
4,064
4,106
4,148
4,190
4,232
4,275
4,318
4,360
4,403
4,447
68,639
38,879
Fuel Cost*
500 ppm
($/gal)
$ 0.021
$ 0.021
$ 0.021
$ 0.022
$ 0.022
$ 0.022
$ 0.022
$ 0.022
$ 0.022
$ 0.022
$ 0.022
$ 0.022
$ 0.022
$ 0.022
$ 0.022
$ 0.022
$ 0.022
$ 0.022
$ 0.022
$ 0.022
$ 0.022
$ 0.022
$ 0.022
$ 0.022
$ 0.022
$ 0.022
$ 0.022
$ 0.022
$ 0.022
$ 0.022


NRLM Fuel Costs
(106 dollars)
$ 142
$ 249
$ 254
$ 256
$ 258
$ 263
$ 268
$ 273
$ 277
$ 282
$ 287
$ 292
$ 296
$ 301
$ 306
$ 311
$ 315
$ 320
$ 325
$ 330
$ 335
$ 340
$ 344
$ 349
$ 354
$ 359
$ 364
$ 368
$ 373
$ 378
$ 5,428
$ 3,027
       * Fuel costs are relative to uncontrolled fuel and assume that, during the transitional years of 2010 & 2014, the first 5
       months are at the previous year's cost and the remaining 7 months are at the next year's cost.
       See Appendix 8B for how these fuel volumes were developed.
                                               8-38

-------
                                                   Aggregate Cost and Cost per Ton
                                      Table 8.7-7
    Oil-Change Maintenance Savings Associated with a 500 ppm NRLM Fuel-only Scenario
                                        ($2002)
Year
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2031
2032
2033
2034
2035
2036
30YrNPVat3%
30YrNPVat7%
Affected NR Fuel
500 ppm
(106 gallons)
4,790
8,406
8,599
8,400
8,300
8,479
8,659
8,839
9,018
9,196
9,374
9,552
9,730
9,907
10,085
10,263
10,441
10,619
10,797
10,973
11,150
1 1 ,326
1 1 ,503
1 1 ,679
1 1 ,856
12,032
12,209
12,386
12,562
12,739
179,520
99,928
Affected L&M Fuel
500 ppm
(106 gallons)
1,990
3,454
3,498
3,457
3,450
3,489
3,518
3,552
3,586
3,623
3,659
3,699
3,747
3,781
3,812
3,859
3,897
3,939
3,980
4,022
4,064
4,106
4,148
4,190
4,232
4,275
4,318
4,360
4,403
4,447
68,639
38,879
NR Savings
savings=$0.029/gal
(106 dollars)
$ 140
$ 246
$ 251
$ 246
$ 243
$ 248
$ 253
$ 258
$ 264
$ 269
$ 274
$ 279
$ 284
$ 290
$ 295
$ 300
$ 305
$ 310
$ 316
$ 321
$ 326
$ 331
$ 336
$ 341
$ 347
$ 352
$ 357
$ 362
$ 367
$ 372
$ 5,248
$ 2,921
L&M Savings
savings=$0.010/gal
(106 dollars)
$ 21
$ 36
$ 37
$ 36
$ 36
$ 37
$ 37
$ 37
$ 38
$ 38
$ 38
$ 39
$ 39
$ 40
$ 40
$ 40
$ 41
$ 41
$ 42
$ 42
$ 43
$ 43
$ 43
$ 44
$ 44
$ 45
$ 45
$ 46
$ 46
$ 47
$ 719
$ 407
NRLM
Total Savings
(106 dollars)
$ 161
$ 282
$ 288
$ 282
$ 279
$ 284
$ 290
$ 296
$ 301
$ 307
$ 312
$ 318
$ 324
$ 329
$ 335
$ 340
$ 346
$ 352
$ 357
$ 363
$ 369
$ 374
$ 380
$ 385
$ 391
$ 397
$ 402
$ 408
$ 413
$ 419
$ 5,967
$ 3,328
   Table 8.7-8 presents the annual net operating costs (Tables 8.7-6 and 8.7-7) along with the
costs by pollutant associated with a 500 ppm NRLM fuel-only scenario.  Because a 500 ppm
NRLM fuel-only scenario is analogous to the NRT4 fuel-only scenario discussed above (i.e., no
new engine standards and, thus, only fuel-derived benefits will occur), we would allocate costs
to PM and SOx the same way as the NRT4 fuel-only scenario (see Table 8.4-6) except that costs
for 15 ppm fuel would clearly be zero. Table 8.7-8 also presents the emission reductions that
would result from a 500 ppm NRLM fuel-only scenario.
                                         8-39

-------
Final Regulatory Impact Analysis
Table 8.7-8
Aggregate Net Operating Costs, Costs by Pollutant, and Emissions Reductions
Associated with a 500 ppm NRLM Fuel-only Scenario ($2002)
Year
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2031
2032
2033
2034
2035
2036
30YrNPVat3%
30 Yr NPVat7%
Fuel Costs
(Smillion)
$ 142
$ 249
$ 254
$ 256
$ 258
$ 263
$ 268
$ 273
$ 277
$ 282
$ 287
$ 292
$ 296
$ 301
$ 306
$ 311
$ 315
$ 320
$ 325
$ 330
$ 335
$ 340
$ 344
$ 349
$ 354
$ 359
$ 364
$ 368
$ 373
$ 378
$ 5,428
$ 3,027
Other Operating
Costs ($million)
$ (161)
$ (282)
$ (288)
$ (282)
$ (279)
$ (284)
$ (290)
$ (296)
$ (301 )
$ (307)
$ (312)
$ (318)
$ (324)
$ (329)
$ (335)
$ (340)
$ (346)
$ (352)
$ (357)
$ (363)
$ (369)
$ (374)
$ (380)
$ (385)
$ (391 )
$ (397)
$ (402)
$ (408)
$ (413)
$ (419)
$ (5,967)
$ (3,328)
Net Operating
Costs ($million)
$ (18)
$ (33)
$ (34)
$ (26)
$ (20)
$ (21)
$ (22)
$ (23)
$ (24)
$ (25)
$ (26)
$ (26)
$ (27)
$ (28)
$ (29)
$ (30)
$ (31)
$ (31)
$ (32)
$ (33)
$ (34)
$ (35)
$ (35)
$ (36)
$ (37)
$ (38)
$ (39)
$ (39)
$ (40)
$ (41)
$ (539)
$ (301)
SOx Costs
(Smillion)
$ (12)
$ (22)
$ (23)
$ (17)
$ (14)
$ (14)
$ (15)
$ (15)
$ (16)
$ (17)
$ (17)
$ (18)
$ (18)
$ (19)
$ (19)
$ (20)
$ (20)
$ (21)
$ (21)
$ (22)
$ (23)
$ (23)
$ (24)
$ (24)
$ (25)
$ (25)
$ (26)
$ (26)
$ (27)
$ (27)
$ (359)
$ (201)
PM Costs
(Smillion)
$ (6)
$ (11)
$ (11)
$ (9)
$ (7)
$ (7)
$ (7)
$ (8)
$ (8)
$ (8)
$ (9)
$ (9)
$ (9)
$ (9)
$ (10)
$ (10)
$ (10)
$ (10)
$ (11)
$ (11)
$ (11)
$ (12)
$ (12)
$ (12)
$ (12)
$ (13)
$ (13)
$ (13)
$ (13)
$ (14)
$ (180)
$ (100)
PM Reduction
(tons)
10,700
19,000
19,400
19,700
20,000
20,400
20,800
21,100
21 ,500
21 ,900
22,200
22,600
23,000
23,300
23,700
24,100
24,500
24,800
25,200
25,600
25,900
26,300
26,700
27,100
27,400
27,800
28,200
28,600
29,000
29,300
419,800
233,800
SOx Reduction
(tons)
133,000
235,400
240,100
244,000
248,500
253,100
257,600
262,200
266,700
271 ,300
275,800
280,400
285,200
289,700
294,200
299,000
303,600
308,200
312,900
317,500
322,200
326,800
331 ,500
336,200
340,800
345,500
350,200
354,900
359,700
364,400
5,210,600
2,901 ,700
   The calculations of cost per ton of each emission reduced under the 500 ppm NRLM fuel-
only scenario divides the net present value of the annual costs assigned to each pollutant (see
Table 8.7-8) by the net present value of the total annual reductions of each pollutant (Table 8.7-
8). The 30-year net present value of the costs (remember that negative costs are actually
savings) associated with each pollutant, calculated with a three percent discount rate, are shown
in Table 8.7-8 as -$107 million for PM and -$213  million for SOx.  If we exclude the oil change
maintenance savings, the costs of the fuel-only scenario would be $1.9 billion for PM and $3.8
billion for SOx. The 30-year net present value, with a three percent discount rate, of emission
reductions are 420 thousand tons for PM and 5.2 million tons for SOx.  Our air quality analysis,
                                          8-40

-------
                                                     Aggregate Cost and Cost per Ton
emissions reduction analysis, and benefits analysis are found in Chapters 2, 3, and 9,
respectively.  Table 8.7-9 presents the cost per ton results for the 500 ppm NRLM fuel-only
scenario including the oil change maintenance savings and excluding those savings.

                                       Table 8.7-9
             Aggregate Cost per Ton for the 500 ppm NRLM Fuel-only Scenario
              30-year Net Present Values at a 3% and 7% Discount Rate ($2002)
Item
SOOppm at $0.021/gal, 2007-2010
SOOppm at $0.022/gal, 2010+
SOOppm Fuel Cost
Other Operating Costs*
Total Costs (w/ maintenance savings)
Total Costs (w/o maintenance savings)
PM Costs (w/ maintenance savings)
PM Costs (w/o maintenance savings)
SOx Costs (w/ maintenance savings)
SOx Costs (w/o maintenance savings)
PM Reduction
SOx Reduction
Cost per Ton PM (w/ maintenance savings)
Cost per Ton PM (w/o maintenance savings)
Cost per Ton SOx (w/ maintenance savings)
Cost per Ton Sox (w/o maintenance savings)
Units
(106 gallons)
(106 gallons)
(Smillion)
(Smillion)
(Smillion)
(Smillion)
(Smillion)
(Smillion)
(Smillion)
(Smillion)
(106tons)
(106tons)
($/ton)
($/ton)
($/ton)
($/ton)
3% discount
rate
31,316
216,843
$5,428
-$5,967
-$539
$5,428
-$180
$1,809
-$359
$3,619
0.42
5.2
-$400
$4,300
-$70
$690
7% discount
rate
26,500
112,307
$3,027
-$3,328
-$301
$3,027
-$100
$1,009
-$201
$2,018
0.23
2.9
-$400
$4,400
-$70
$700
Source
Table 8.7-6
Table 8.7-6
Table 8.7-6
Table 8.7-7
Table 8.7-7
Table 8.7-7
Table 8.7-8
Calculated* *
Table 8.7-8
Calculated* *
Table 8.7-8
Table 8.7-8
Calculated
Calculated
Calculated
Calculated
* Other operating costs include oil change maintenance savings.
** Calculated as one-third (PM) or two-thirds (SOx) of the Total Scenario Costs w/o maintenance savings.
   We have also calculated the cost per ton of emissions reduced in the year 2030 using the
annual costs and emission reductions in that year alone. This number, shown in Table 8.7-10,
approaches the long-term cost per ton of emissions reduced.
                                          8-41

-------
Final Regulatory Impact Analysis
                                      Table 8.7-10
             Long-Term Cost per Ton of the 500 ppm NRLM Fuel-only Scenario
                        Annual Values without Discounting ($2002)
Pollutant
PM (with maintenance savings)
PM (without maintenance savings)
SOx (with maintenance savings)
SOx (without maintenance savings)
Long-Term Cost per
Ton in 2030
-$400
$4,300
-$70
$690
   8.7.3.2 Costs and Costs per Ton of the 15 ppm L&M Fuel Increment

   In this section, we evaluate the incremental cost per ton of the  15 ppm L&M fuel cap in 2012
(final NRLM program) relative to retaining the 500 ppm cap on L&M fuel (the proposed NRLM
program) indefinitely. Nonroad  diesel fuel is assumed to be subject to a 15 ppm cap starting in
2010 in both cases. We assume that the emission standards applicable to nonroad engines are
the same regardless of the sulfur cap applicable to L&M fuel. Therefore, the only differences
between the 500 and 15 ppm cap on L&M fuel are in emissions of SO2 and sulfate PM, fuel costs
and engine maintenance savings. The cost of complying with emission  standards for land-based
nonroad equipment, as well as HC, NOx,  and non-sulfate PM emissions from this equipment are
unaffected.

   The difference in costs between the two L&M fuel caps are primarily related to the
production 15 ppm L&M fuel. The differences in sulfurous emissions arise from differences in
the sulfur content of both L&M fuel and,  in the Northeast/Middle Atlantic area, heating oil.
While the difference in heating oil sulfur  content is a direct result of the final NRLM fuel
provisions for the Northeast/Middle Atlantic area, heating oil sulfur content is not directly
regulated by this final rule. Therefore, we develop estimates of the incremental cost
effectiveness of the 15 ppm L&M fuel cap both with and without the changes in heating oil
sulfur.  However, we believe that the most appropriate estimate of the incremental cost
effectiveness of the 15 ppm L&M fuel cap is that including the change in heating oil sulfur
content.

   The key inputs to this sensitivity analysis are: 1) the volumes and sulfur contents of each type
of distillate fuel being produced  and consumed in the 2012-2036 timeframe, and 2) the cost of
supplying these fuels over the  same timeframe. The fuels produced prior to June 1,  2012 are
identical under the two scenarios being evaluated here. Thus, we ignore all emissions and costs
prior to June 1, 2012. This incremental analysis models the U.S. minus  California.,  although it
would also apply for the total U.S. as well since California's fuel quality is not expected to
change with the requirement that L&M fuel meet a 15 ppm cap.
                                          8-42

-------
                                                   Aggregate Cost and Cost per Ton
   The process for estimating the annual production volumes of each fuel was described in
Chapter 7. The first step in the process was to develop a comprehensive description of fuel
production and demand in 2001 for non-highway and highway diesel fuel which accounted for
the spillover of low sulfur, highway fuel into the non-highway markets. The analysis also
considered the downgrade of jet fuel and highway diesel fuel, along with some gasoline, to lower
quality fuels during pipeline distribution.

   We then  developed a set of analogous estimates for 2014, starting with demand.  Fuel
demand in 2014 was projected using the EPA draft NONROAD2004 model and EIA's AEO
2003. We also estimated the volume of highway diesel fuel demand considering the highway
diesel fuel requirements being implemented in 2006 and 2010. Spillover of highway fuel into
the non-highway markets was assumed to remain constant (in terms of the percentage of each
non-highway market represented by spillover).  The volume of gasoline, jet fuel and highway
diesel fuel in 2014 downgraded to 500 ppm and high sulfur distillate was projected to increase in
proportion to the growth in jet fuel demand and the supply of highway diesel fuel. This
downgraded fuel was first distributed to the non-highway fuel markets assuming sulfur controls
on highway fuel only, followed by the 500 ppm standard on NRLM fuel in 2007 and subsequent
15 ppm standards on nonroad fuel and L&M fuel in 2010 and 2012, respectively.  NRLM fuel
not already complying with the required sulfur limit prior to the NRLM rule from  spillover of
highway fuel or downgrade, had to be desulfurized at refineries.

   We then  used these 2014 estimates of fuel production, downgrade and spillover to develop
similar estimates for individual calendar years starting with 2007 and going through 2040
consistent with the phase of NRLM program in place at the time. These individual, annual
estimates were based on a  slightly more approximate methodology  which assumed that the
fraction of each non-highway distillate fuel's market demand represented by spillover and
downgrade remained constant at its 2014 level.  Regarding spillover, this is the  same assumption
made in developing our estimate of spillover in 2014. However, with respect to downgrade, this
assumption differs from that used in the more comprehensive 2014  analysis.  Because the
demand for jet fuel and highway diesel fuel is projected to grow faster than that for NRLM fuel
and heating oil, the percentage of downgrade in the NRLM and heating oil markets is higher in
2014 than in 2001.  Thus, the net effect of assuming that the percentage of downgrade remains
constant at 2014 levels underestimates the percentage of downgrade in the non-highway fuel
markets after 2014, and overestimates it prior to 2014.

   The effect of assuming constant downgrade percentages in the non-highway markets on the
estimated costs and benefits of the overall rule is very small, given  that it affects only a small
portion of the overall fuel demanded, that none of the benefits of the engine emission standards
are involved and that the changes in costs and benefits are offsetting. However, it has a larger
impact on this incremental analysis, as about half of the 30-year sulfur dioxide emission benefits
of the 15 ppm L&M cap are due to a shift in downgrade from the L&M fuel market to the
heating oil market in the Northeast/Middle Atlantic area. Thus, for this incremental  analysis, we
revised the assumption that the downgrade fraction of the demand for the various non-highway
fuels for years other than 2014 will remain constant at their 2014 levels. Instead, we estimated
the volume of downgrade generated each year, based on future highway diesel fuel supply and

                                         8-43

-------
Final Regulatory Impact Analysis
jet fuel demand. We made one simplifying assumption: that highway diesel fuel supply grew at
the same rate as highway fuel demand.  Highway fuel supply includes spillover to the non-
highway fuel markets. While nonroad fuel demand is projected to grow at roughly the  same rate
as highway fuel, L&M fuel and heating oil demand are expected to grow much more slowly.
Thus, this simplifying assumption overestimates highway fuel supply.  However, the degree of
overestimation is slight,  since only about 10% of highway diesel fuel supply is spillover to the
non-highway pool, and about 70% of that goes to the nonroad fuel market.

   Estimates of the demand for highway and jet fuel through 2025 are taken from EIA's AEO
2003. After 2025 the yearly projected demand for both highway diesel fuel and jet fuel are
estimated from the average projected growth from AEO 2003 between 2020 and 2025.  The
year-over-year growth rates for highway and jet fuel from 2020 to 2025 were 1.019 and 1.021,
respectively. The annual demand for highway and jet fuel from 2012 to 2036 and the volume
ratios to the projected 2014 volumes are summarized in Table 8.7-11. In last column of Table
8.7-11 an average set of volume ratios are shown which represents the combined growth for
highway and jet-based downgrade in heating oil. The relative volume of highway and jet-based
downgrade was similar in NRLM diesel fuel, so these volume ratios  were used for estimating
non-2014 downgrade volumes for NRLM diesel fuel as well.
                                        8-44

-------
                                                 Aggregate Cost and Cost per Ton
                                    Table 8.7-11
       Projected Highway Diesel Fuel and Jet Fuel Demand - AEO 2003 (Trillion BTU)
Year

2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2031
2032
2033
2034
2035
2036
Highway
Fuel Demand
7,500
7,670
7,840
7,980
8,110
8,250
8,390
8,560
8,700
8,850
9,020
9,200
9,400
9,580
9,762
9,947
10,140
10,330
10,530
10,730
10,930
11,140
11,350
11,560
11.780
Ratio to 20 14
0.957
0.978
1.000
1.018
1.034
1.052
1.070
1.092
1.110
1.129
1.151
1.173
1.199
1.222
1.245
1.269
1.293
1.317
1.343
1.368
1.394
1.421
1.447
1.475
1.503
Jet Fuel
Fuel Demand
4,140
4,260
4,380
4,500
4,620
4,730
4,860
4,970
5,090
5,200
5,310
5,430
5,540
5,660
5,780
5,900
6,020
6,150
6,280
6,410
6,550
6,680
6,820
6,970
7.110
Ratio to 20 14
0.945
0.973
.000
.027
.055
.080
.110
.135
.162
.187
.212
.240
.265
.292
.319
.347
.375
.404
.434
.464
.495
.526
.558
.591
.624
Avg Ratio to
2014
0.954
0.977
1.000
1.020
1.039
1.059
1.079
1.102
1.122
1.142
1.165
1.189
1.214
1.238
1.262
1.287
1.312
1.338
1.364
1.390
1.417
1.445
1.473
1.502
1.531
   The next step is to estimate the annual demand, spillover, downgrade and production
volumes for NRLM fuel from 2012 to 2036 for both the proposed and final rule NRLM
programs.  Starting with the proposed NRLM fuel program, we estimated the jet and highway-
based downgrade in the nonroad, locomotive and marine fuel markets from mid-2012 to mid-
2014 by multiplying the 2014 highway and jet-based downgrade volumes shown in Table 7.1.4-1
by the ratio of highway and jet fuel demand in each year to 2014 from Table 8.7-11,
respectively. For the years following 2014, we multiplied the 2014 highway and jet-based
downgrade volumes shown in Table 7.1.4-2 by the ratio of highway and jet fuel demand in each
year to 2014 from Table 8.7-11, respectively. Annual demand for NRLM fuel, and the
contribution of spillover and small refiner fuel to these markets, were estimated by multiplying
the 2014 estimates of these volumes in Tables 7.1.4-1 (for 2012-2014) and 7.1.4-2 (for 2015 and
beyond) by the growth in NRLM fuel demand contained in Tables 7.1.5-1 (for nonroad and
locomotive fuel) and  7.1.5-2 (for marine fuel). Annual production volumes of NRLM fuel were
                                        8-45

-------
Final Regulatory Impact Analysis
estimated by subtracting the downgrade, spillover and small refiner fuel volumes from total
demand.  The resulting estimates of downgrade, spillover, small refiner fuel, and 15 and 500
ppm production volumes for nonroad, locomotive and marine diesel fuel for the proposed rule
program are summarized in Tables 8.7-12, 8.7-13 and 8.7-14, respectively. The highway-based
and jet fuel-based downgrade volumes are combined together into one column.

                                     Table 8.7-12
            Nonroad Fuel Supply Under the Proposed NRLM Fuel Program With
            the Shift of Downgrade to the Heating Oil Market (million gallons) *
Year
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2031
2032
2033
2034
2035
2036
Downgrade
1,061
1,085
463
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
Small Refiner Fuel
627
640
272
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
Spillover
2,760
2,818
2,940
3,047
3,107
3,167
3,227
3,288
3,352
3,408
3,468
3,528
3,588
3,648
3,708
3,767
3,827
3,887
3,946
4,006
4,066
4,125
4,185
4,245
4,304
New 15 ppm Fuel
8,327
8,501
9,641
10,539
10,747
10,955
11,162
11,370
11,578
11,786
11,993
12,201
12,409
12,616
12,823
13,029
13,236
13,443
13,649
13,855
14,062
14,268
14,474
14,681
14,887
Total Volume
12,774
13,045
13,316
13,586
13,854
14,122
14,390
14,658
14,926
15,193
15,461
15,729
15,997
16,265
16,531
16,797
17,063
17,329
17,595
17,861
18,127
18,393
18,659
18,925
19,191
    : Excludes NRLM fuel demand in California
                                        8-46

-------
                                                Aggregate Cost and Cost per Ton
                                   Table 8.7-13
Locomotive Volumes Under the Proposed NRLM Fuel Program With the Shift of Downgrade to
                      the Heating Oil Market (million gallons) *
Year
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2031
2032
2033
2034
2035
2036
Downgrade
579
593
,176
,614
,644
,675
,707
,743
,775
,807
,843
,881
,921
,959
,997
2,036
2,076
2,116
2,157
2,199
2,242
2,286
2,330
2,376
2,422
New 500 ppm Fuel
1,705
1,710
1,190
804
800
791
777
766
751
731
719
703
686
673
656
638
619
600
580
559
537
515
491
467
442
Spillover
602
607
566
539
544
549
554
559
563
566
571
576
581
586
591
596
601
605
610
615
619
624
629
634
638
Total Volume
2,886
2,909
2,932
2,956
2,988
3,015
3,038
3,067
3,089
3,104
3,132
3,160
3,187
3,218
3,244
3,270
3,295
3,321
3,347
3,373
3,399
3,425
3,450
3,476
3,502
   : Excludes NRLM fuel demand in California
                                       8-47

-------
Final Regulatory Impact Analysis
                                     Table 8.7-14
 Marine Volumes Under the Proposed NRLM Fuel Program With the Shift of Downgrade to the
                         Heating Oil Market (million gallons) *
Year
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2031
2032
2033
2034
2035
2036
Downgrade
446
456
451
436
445
453
462
471
480
489
498
509
519
530
540
551
561
572
583
595
606
618
630
642
655
New 500 ppm Fuel
,333
,338
,369
,409
,419
,431
,451
,473
,488
,503
,526
,538
,555
,568
,585
,602
,618
,634
,650
,666
,682
,697
,712
,727
,742
Spillover
280
283
281
280
283
286
290
295
299
302
307
311
315
319
323
327
331
335
339
343
347
352
356
360
364
Total Volume
2,059
2,078
2,103
2,126
2,146
2,170
2,203
2,240
2,266
2,294
2,331
2,357
2,389
2,417
2,448
2,479
2,510
2,542
2,573
2,604
2,635
2,667
2,698
2,729
2,760
   * Excludes NRLM fuel demand in California
   Annual estimates of downgrade, spillover, small refiner fuel, and 15 and 500 ppm production
volumes under the final NRLM fuel program in years other than 2014 were estimated from the
estimates for 2014 in the same manner. The only difference is a new set of 2014 estimates.  The
2014 estimates of downgrade, spillover, small refiner fuel, and total demand for NRLM fuel for
mid-2012 to mid-2014 were taken from Table 7.1.3-19. The 2014 estimates of downgrade,
spillover, small refiner fuel, and total demand for NRLM fuel for 2015 and beyond were taken
from Table 7.1.3-20.  The resulting estimates of downgrade, spillover, small refiner fuel, and 15
and 500 ppm production volumes for nonroad, locomotive and marine diesel fuel for the
proposed rule program are summarized in Tables 8.7-15, 8.7-16 and 8.7-17, respectively.
                                         8-48

-------
                                                Aggregate Cost and Cost per Ton
                                   Table 8.7-15
Nonroad Fuel Supply Under the Final Rule Fuel Program With the Shift of Downgrade to the
                       Heating Oil Market (million gallons) *
Year
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2031
2032
2033
2034
2035
2036
Downgrade
1,061
1,085
463
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
Small Refiner Fuel
528
468
199
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
Spillover
2,760
2,818
2,941
3,047
3,107
3,167
3,227
3,288
3,352
3,408
3,468
3,528
3,588
3,648
3,708
3,767
3,827
3,887
3,946
4,006
4,066
4,125
4,185
4,245
4,304
New 15 ppmFuel
8,426
8,674
9,713
10,539
10,747
10,955
11,162
11,370
11,578
11,786
11,993
12,201
12,409
12,616
12,823
13,029
13,236
13,443
13,649
13,855
14,062
14,268
14,474
14,681
14,887
Total Volume
12,774
13,045
13,316
13,586
13,854
14,122
14,390
14,658
14,926
15,193
15,461
15,729
15,997
16,265
16,531
16,797
17,063
17,329
17,595
17,861
18,127
18,393
18,659
18,925
19,191
  : Excludes NRLM fuel demand in California
                                      8-49

-------
Final Regulatory Impact Analysis
                                    Table 8.7-16
Locomotive Fuel Supply Under the Final Rule Fuel Program With the Shift of Downgrade to the
                         Heating Oil Market (million gallons) *
Year
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2031
2032
2033
2034
2035
2036
Downgrade
397
274
849
,281
,304
,329
,355
,383
,408
,434
,462
,492
,524
,554
,585
,616
,647
,679
,712
,745
,779
,814
,849
,885
,922
Small Refiner Fuel
761
99
42
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
Spillover
602
607
589
577
583
589
593
599
603
606
611
617
622
628
633
638
643
648
653
658
663
668
674
679
684
New 15 ppmFuel
,127
,930
,476
,099
,100
,098
,090
,086
,069
,053
,058
,051
,041
,035
,026
,016
,005
994
982
969
956
942
928
912
897
Total Volume
2,841
2,909
2,932
2,956
2,988
3,015
3,038
3,067
3,089
3,104
3,132
3,160
3,187
3,218
3,244
3,270
3,295
3,321
3,347
3,373
3,399
3,425
3,450
3,476
3,502
    : Excludes NRLM fuel demand in California
                                        8-50

-------
                                               Aggregate Cost and Cost per Ton
                                  Table 8.7-17
Marine Fuel Supply Under the Final Rule Fuel Program With the Shift of Downgrade to the
                       Heating Oil Market (million gallons) *
Year
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2031
2032
2033
2034
2035
2036
Downgrade
285
173
155
141
143
146
149
152
155
158
161
164
168
171
174
178
181
185
188
192
196
199
203
207
211
Small Refiner Fuel
636
148
62
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
Spillover
280
283
281
280
283
286
290
295
299
302
307
311
315
319
323
327
331
335
339
343
347
352
356
360
364
New 15 ppmFuel
874
1,474
1,605
,705
,720
,738
,763
,793
,813
,834
,863
,883
,906
,927
,951
,975
,998
2,022
2,046
2,069
2,092
2,116
2,139
2,162
2,185
Total Volume
2,059
2,078
2,103
2,126
2,146
2,170
2,203
2,240
2,266
2,294
2,331
2,357
2,389
2,417
2,448
2,479
2,510
2,542
2,573
2,604
2,635
2,667
2,698
2,729
2,760
 : Excludes NRLM fuel demand in California
                                      8-51

-------
Final Regulatory Impact Analysis
   The cost of supplying NRLM fuel under the final NRLM program and for the proposed
NRLM program are developed in Chapter 7 and summarized in Table 7.5-1.  The engine
maintenance savings associated with reduced sulfur contents are developed in Chapter 6 and
summarized in Table 6.2-29. We assume that the per gallon costs developed for 2014 apply
through 2036.  With the increase in downgrade volume, the cost of reprocessing downgrade
which occurs in some regions would increase. However, this increase occurs both with and
without the 15 ppm L&M fuel cap.  Thus, we did not update the estimated cost of reprocessing
downgrade.  The per gallon costs and savings under both L&M fuel caps are summarized here in
Table 8.7-18.

                                      Table 8.7-18
              Total Diesel Fuel Costs Under 500 and 15 ppm L&M Fuel Caps*

Refining
Cost
Additive and
Distribution Cost
Maintenance
Savings
Total w/o
Maintenance Savings
Total with
Maintenance Savings
Final NRLM Fuel Program
2012-2014
15 ppmNonroad
Small Refiner 500
jpm Nonroad
Small Refiner 500
jpm L&M
2014 +
15 ppmNonroad
15 ppm L&M
5.6
2.9
2.9

5.8
5.8
0.8
0.2
0.2

1.2
1.2
-3.2
-2.9
-1

-3.2
-1.1
6.4
3.1
3.1

7
7
3.2
0.2
2.1

3.8
5.9
500 ppm NRLM Fuel Cap in 2007 and 15 ppm Nonroad Fuel Cap in 2010 (proposed rule program)
2012-2014
15 ppmNonroad
Small Refiner 500
jpm Nonroad
500 ppm L&M
5
2.7
2.7
0.8
0.2
0.3
-3.2
-2.9
-1.0
5.8
2.9
3
2.6
0
2.0
2014 +
15ppmNR
500 ppm L&M
5.2
2.7
1.2
0.2
-3.2
-1.0
6.4
2.9
3.2
1.9
* Fuel costs are relative to uncontrolled fuel and assume that, during the transitional years of 2012 & 2014, the first 5
months are at the previous year's cost and the remaining 7 months are at the next year's cost.
   We then multiplied the production volume of each fuel in a given calendar year by the net
cost of using that fuel from Table 8.7-18. For this incremental analysis, we only present
estimated annual costs including the maintenance savings because, on the increment, these
maintenance savings are minor (0.1 c/gal) compared to the incremental cost of producing 15 ppm
L&M fuel.  Little information would be gained from presenting costs without the maintenance
savings, as is done for the final rule analysis and the other sensitivity cases. We do present the
                                          8-52

-------
                                                  Aggregate Cost and Cost per Ton
final discounted costs without the maintenance savings, as well as the cost-effectiveness based
on the costs without maintenance savings, in Table 8.7-24. The resulting annual costs are shown
in Table 8.7-19.
                                     Table 8.7-19
   Annual Fuel Costs & Oil Change Maintenance Savings With the Shift of Downgrade to the
                           Heating Oil Market ($2002 million)
Year
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2031
2032
2033
2034
2035
2036
Final NRLM Fuel
Program
$ 268
$ 472
$ 524
$ 566
$ 575
$ 583
$ 592
$ 602
$ 610
$ 618
$ 628
$ 637
$ 645
$ 654
$ 663
$ 671
$ 680
$ 688
$ 697
$ 705
$ 714
$ 722
$ 731
$ 739
$ 747
1 5 ppm NR Cap
and 500 ppm L&M Cap
$ 162
$ 282
$ 337
$ 379
$ 386
$ 393
$ 400
$ 406
$ 413
$ 420
$ 426
$ 433
$ 440
$ 446
$ 453
$ 459
$ 466
$ 473
$ 479
$ 486
$ 492
$ 499
$ 505
$ 511
$ 518
Total 30-Year Costs (2007-2036)
Undiscounted
30 Yr NPV at 3%
30YrNPVat7%
$ 15,731
$ 8,640
$ 4,249
$ 10,664
$ 5,829
$ 2.847
15 ppm L&M
Incremental Costs
$ 107
$ 190
$ 187
$ 187
$ 189
$ 191
$ 193
$ 195
$ 197
$ 198
$ 202
$ 203
$ 206
$ 208
$ 210
$ 212
$ 214
$ 216
$ 218
$ 220
$ 222
$ 224
$ 226
$ 227
$ 229

$ 5,068
$ 2,811
$ 1.402
   The absence of the shift of downgrade to the heating oil market in the Northeast/Middle
Atlantic area has no impact on the supply of NRLM fuel under the proposed NRLM fuel
program. Thus, the various volumes of NRLM fuel shown in Tables 8.7-12 through 8.7-14 still
apply.  Without the shift of downgrade to heating oil, the production volumes of NRLM fuel
under the final NRLM fuel program become very similar to those for the proposed NRLM fuel
program, except that L&M fuel produced after mid-2012 would have to meet a 15 ppm cap
instead of a 500 ppm cap. The volumes of spillover, downgrade and demand are identical. The
only difference is that the volume of 500 ppm, small refiner fuel is slightly greater with a long-
term 500 ppm L&M fuel cap than with a 15 ppm cap.  Thus, the incremental volume of 15 ppm
fuel from mid-2012 to mid-2014 under the 15 ppm L&M fuel cap is slightly higher than simply
                                        8-53

-------
Final Regulatory Impact Analysis
the volume of 500 ppm L&M fuel which must be produced under the 500 ppm L&M cap. Table
8.7-20 shows the breakdown of nonroad, locomotive and marine fuel supply for the final NRLM
fuel program without a shift in downgrade to heating oil.
                                     Table 8.7-20
    NRLM Fuel Supply Under the Final NRLM Fuel Program Without a Downgrade Shift
                            to Heating Oil (million gallons) *
Year

2012
2013
2014
2015 +

2012
2013
2014
2015 +

2012
2013
2014
2015 +
Downgrade
Small Refiner Fuel
Spillover
New 15 ppmFuel| Total Volume
Nonroad Diesel Fuel
1,061
1,085
463
528
468
199
2,760
2,818
2,940
8,426
8,674
9,713
12,774
13,045
13,316
Same as for Proposed NRLM Fuel Program
Locomotive Diesel Fuel
579
593
1,176
761
99
42
602
607
566
944
1610
1148
2,886
2,909
2,932
Same as for Proposed NRLM Fuel Program
Marine Diesel Fuel
446
456
451
636
148
62
280
283
280
697
1191
1310
2,059
2,078
2,103
Same as for Proposed NRLM Fuel Program
   * Excludes NRLM fuel demand in California
   The per gallon costs shown in Table 8.7-19 are unaffected by the absence of a shift in
downgrade to the heating oil market.0 Thus, the annual costs with a 500 ppm L&M cap are the
same as before.  The annual costs under the final NRLM program decrease slightly, as 15 ppm
L&M fuel does not need to replace downgrade shifted from the L&M market to the heating oil
market in the Northeast/Middle Atlantic exclusion area.  The annual costs under both programs
are shown in Table 8.7-21.
   D The reduced volume of 15 ppm L&M fuel under the final NRLM fuel program could
reduce the per gallon cost of 15 ppm fuel, as those refiners facing the highest costs might be the
first to avoid producing this fuel. However, as indicated by the sensitivity analysis of potentially
lower nonroad fuel demand (Case 1 Sensitivity) discussed in Section 3 of Appendix 8A,
significantly lowering the demand for 15 ppm NRLM fuel has little effect on the cost per gallon.
                                         8-54

-------
                                                   Aggregate Cost and Cost per Ton
                                     Table 8.7-21
   Annual Fuel Costs & Oil Change Maintenance Savings Without Shift of Downgrade to the
                           Heating Oil Market ($2002 million)
Year
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2031
2032
2033
2034
2035
2036
Final NRLM Fuel
Program
$249
$432
$488
$531
$539
$547
$556
$564
$572
$580
$588
$596
$604
$612
$619
$627
$635
$643
$650
$658
$665
$673
$680
$687
$695
Proposed NRLM Fuel
Program
$162
$282
$337
$379
$386
$393
$400
$406
$413
$420
$427
$433
$440
$446
$453
$460
$466
$473
$479
$486
$492
$499
$505
$512
$518
Incremental Cost of 15 ppm L&M
Cap
$88
$150
$151
$152
$153
$155
$156
$158
$159
$160
$162
$163
$164
$165
$166
$168
$169
$170
$171
$172
$173
$174
$175
$176
$177
Total 30-Year Costs (2007 - 2036)
Undiscounted
30-Year NPV at 3%
30-Year NPV at 7%
$14,690
$8,070
$3,969
$10,665
$5,830
$2,847
$4,025
$2,240
$1,121
   Moving to emission reductions, we used the methodology used in the draft 2004 NONROAD
model to estimate SO2 and sulfate PM emissions from NRLM engines (Section 3.1 of the Final
RIA). To calculate the emission reductions, we needed estimates for the sulfur levels for
nonroad, locomotive and marine diesel fuel.

   In Section 7.1.6 of the Final RIA, we present our estimate of the sulfur levels of on-purpose
produced diesel fuel, spillover, and downgrade. These sulfur levels, spillover (11 ppm), small
refiner fuel (340 ppm), and non-small refiner fuel (either 340 or 11 ppm), are unaffected by
changing the volume of downgrade projected to be generated during fuel distribution. For
downgrade, in Section 7.1, we estimated that jet-based downgrade contained 400-470 ppm sulfur
and highway-based downgrade contained 25-35 ppm sulfur.  The relative volumes of these
downgrades varies by region. We calculated a national average sulfur content for combined
                                         8-55

-------
Final Regulatory Impact Analysis
highway-based and jet-based downgrade used in the L&M markets by weighting the sulfur
contents of each downgrade type in each region.  The result was an average downgrade sulfur
content of 101 ppm for the proposed NRLM program and 172 ppm for the final NRLM program.
These sulfur levels were used for downgrade volumes for all the years of the incremental
analysis.E We also applied these downgrade sulfur contents to the small volume of downgrade
used in the nonroad fuel market from mid-2012 to mid-2014.  The resulting overall sulfur levels
for NRLM fuel are summarized in Table 8.7-22.
   E The downgrade comprised of highway diesel fuel and jet fuel likely changes in sulfur level throughout the
period as the relative volume of highway and jet fuel varies relative to each other. However, the growth of highway
diesel fuel and jet fuel is very similar so very little change is expected throughout the analysis period. Thus, this
assumption seems reasonable.

                                           8-56

-------
                                                   Aggregate Cost and Cost per Ton
                                      Table 8.7-22
 Sulfur Levels of NRLM Diesel Fuel Based on Revised Downgrade Estimates (million gallons)

Year

2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2031
2032
2033
2034
2035
2036
2037
2038
2039
2040
48 State Analysis
Proposed Rule
NR
36
36
21
11
11
11
11
11
11
11
11
11
11
11
11
11
11
11
11
11
11
11
11
11
11
11
11
11
11
L&M
236
235
215
201
200
199
198
198
197
195
193
192
191
190
189
188
187
187
186
185
184
183
181
180
179
178
177
176
175
Final Rule
NR
32
29
19
11
11
11
11
11
11
11
11
11
11
11
11
11
11
11
11
11
11
11
11
11
11
11
11
11
11
L&M
122
43
49
55
55
55
56
56
56
57
57
57
58
58
59
59
59
60
60
61
61
62
62
62
63
63
64
64
65
50 State Analysis
Proposed Rule
NR
36
36
22
11
11
11
11
11
11
11
11
11
11
11
11
11
11
11
11
11
11
11
11
11
11
11
11
11
11
L&M
237
237
214
198
197
196
195
194
194
192
190
189
188
187
186
185
184
183
182
181
180
179
178
177
176
175
174
173
171
Final Rule
NR
33
30
19
11
11
11
11
11
11
11
11
11
11
11
11
11
11
11
11
11
11
11
11
11
11
11
11
11
11
L&M
125
47
51
54
54
54
55
55
55
56
56
57
57
57
58
58
59
59
59
60
60
61
61
62
62
62
63
63
64
   We developed these for 50-state and 48-state regions, as this was done for the other
alternatives evaluated in Chapter 3.  We use the 50-state sulfur levels here, even though the
volumes developed above are for the U.S. excluding California.  Thus, the total sulfur dioxide
and sulfate PM emissions resulting from combining the fuel volumes with the sulfur contents are
not correct. However, as the 15 ppm L&M cap has no impact on sulfur levels in California, the
difference in sulfurous emissions between the two L&M fuel caps is correct. To avoid any
possible mis-use of the absolute emissions under either L&M cap, we only present the
differential emission estimates below.
                                         8-57

-------
Final Regulatory Impact Analysis
   In Section 7.1.6, we also estimate the sulfur content of heating oil by assuming that heating
oil has the same sulfur content as NRLM fuel prior to the final NRLM rule.  That is acceptable
for the analysis of the overall NRLM rule, since the emission reductions related to changes in the
sulfur content of heating oil are minor relative to emission reductions related to changes in sulfur
content of NRLM. However,  in analyzing the incremental step of reducing L&M fuel sulfur
from 500 to 15 ppm, heating oil related emission represent a significant portion of the emission
reductions and therefore warrant closer scrutiny. The impacts on the sulfur content of heating oil
occur in the overall program primarily as a result of changes in where spillover and downgrade
are projected to be used. With the imposition of the 15 ppm limit, downgrade product in
particular is forced from other markets into the heating oil market. When this downgraded
distillate cannot be used in NR or in L&M fuel, it will shift to the heating oil market.  The main
impact of this is felt as the last increment of diesel fuel, the L&M portion, is required to meet a
15 ppm limit, and primarily in the Northeast and Mid-Atlantic area where the majority of heating
oil is marketed, and when under the provisions of the final rule downgraded material  cannot
continue to be sold into the NRLM markets.  The downgrade contains between 31 (highway-
based) and 435 ppm (jet-based) sulfur, well below that of heating oil. Thus, the sulfur content of
heating oil decreases significantly in the Northeast/Mid-Atlantic area with a 15 ppm cap on
L&M fuel.

   In the Northeast and Middle Atlantic area of the U.S., certain states regulate the sulfur
content of heating oil, so some of the heating oil in this area contains much less sulfur than
NRLM fuel.  As  a result, the sulfur level estimates based on high sulfur diesel fuel may not be
entirely accurate for representing the sulfur level of heating oil, particularly in this area of the
country.  Given that the majority  of the impact on emissions from heating oil for analyzing the
L&M increment to 15 ppm are in this part of the country, we looked to see what other data might
be available to better assess the sulfur levels. We obtained heating oil surveys from TRW 5'6.
TRW surveys covers heating oil produced in the U.S.  TRW's districts A and B match the
Northeast/Mid-Atlantic area area quite closely.  In 2001 and 2002, heating oil produced by
refineries for this market averaged 1385 ppm sulfur. (As was described in Section 7.1.6, we
exclude sulfur measurements less than 500 ppm, as these likely represent spillover from the
highway fuel supply.) This is less than half that of average NRLM fuel  in PADD 1 (2925 ppm,
see Table 7.1.6-3).

   One difficulty in using the heating oil survey results directly is that the heating oil may be
marketed as a single high sulfur distillate fuel to both the diesel fuel and heating oil markets.
Thus, much of the intended sales for heating oil purposes could have been used as diesel fuel.
The TRW surveys for both diesel fuel and heating oil cover only a small fraction of the total
volume of fuel sold in the U.S. It is not clear whether the heating oil not covered by the data
submitted by refiners to TRW resembles the high sulfur diesel fuel containing roughly 2900 ppm
sulfur or the heating oil containing roughly 1400 ppm  sulfur. Because of this uncertainty, we
assume that the average heating oil in the Northeast/Mid-Atlantic Area contains 2155 ppm
sulfur, the average of the TRW survey estimates for high sulfur diesel fuel and heating oil.

   With the imposition of the 15 ppm L&M standard in 2012, and because of the
Northeast/Mid-Atlantic area provisions of the final NRLM fuel program, 616 million gallons of

                                          8-58

-------
                                                  Aggregate Cost and Cost per Ton
downgrade is shifted from the NRLM market to the heating oil market in 2014.  Of this, 143
million gallons is jet-based downgrade and 473 million gallons is highway-based downgrade. In
PADD 1, jet-based downgrade is estimated to contain 470 ppm sulfur, while highway-based
downgrade contains 35 ppm sulfur.  Thus, the average sulfur content of both downgrades is 129
ppm.  Shifting this downgrade from the NRLM fuel market to the heating oil market reduces the
sulfur content of the 616 million gallons of heating oil by 2026 ppm (2155 ppm minus 129 ppm).
The volume of downgrade used in heating oil is estimated for the years before and after 2014
using the overall downgrade growth rates shown in Table 8.7-11. The resulting incremental
volume of downgrade estimated to be shifted over to heating oil from 2012 to 2036 due to the
final NRLM program is summarized in Table 8.7-23.  The same 2026 ppm sulfur reduction due
to the shift of downgrade to the heating oil pool is used for all the years.
                                     Table 8.7-23
   Incremental Volume of Downgrade Forced into Heating Oil by the Final NRLM Program
                                   (Million gallons)
Year
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2031
2032
2033
2034
2035
2036
Volume
343
602
616
628
640
652
665
679
691
704
718
732
748
763
778
793
808
824
840
856
873
890
907
925
943
                                        8-59

-------
Final Regulatory Impact Analysis
    We estimate that 99% of the sulfur in heating oil is emitted in the form of sulfur dioxide and
1% in the form of sulfate PM.7  Otherwise, the reductions in sulfur dioxide and sulfate PM
emissions due to this shift of downgrade to the PADD 1 heating oil market were estimated using
the formula described in Chapter 3.F Table 8.7-16 presents the annual sulfur dioxide and sulfate
PM emission reductions from NRLM fuel and heating oil.  The reductions in NRLM emissions
represent the difference in sulfur dioxide and sulfate PM emissions under the proposed and final
NRLM fuel programs. These emissions under each fuel program are derived from combining
the sulfur contents shown in Table 8.7-24 for the 50-state region with the NRLM fuel demands
shown in Tables 8.7-12 through 8.7-17.
      As described in Chapter 3, sulfur dioxide has twice the mass of sulfur contained within it. Diesel fuel and heating
oil are both assumed to have a density of 7.1 pounds per gallon. Thus, the formula for calculating the sulfur dioxide
emission reduction from heating oil consumption in 2014 is: 616 million gallons * 7.1 Ib/gal * 2026 parts sulfur per
million parts heating oil by mass  * 99% conversion of sulfur to SO2 * 2 Ibs SO2 per Ib sulfur / 2000 Ib/ton. Sulfate PM in
the atmosphere is estimated to have 7 times the mass of the sulfur contained within it.  Thus, the formula for calculating
the sulfate PM emission reduction from heating oil consumption in 2014 is: 616 million gallons * 7.1 Ib/gal * 2026 parts
sulfur per million parts heating oil by mass * 1% conversion of sulfur to sulfate PM * 7 Ibs sulfate PM per Ib sulfur /
2000 Ib/ton.

                                              8-60

-------
                                                   Aggregate Cost and Cost per Ton
                                     Table 8.7-24
  Annual Sulfur Dioxide and Sulfate PM Emission Reductions:
                                   Cap (tons per year)
15 ppm Versus 500 ppm L&M
Year

2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2031
2032
2033
2034
2035
2036
Sulfur Dioxide
NRLM Fuel
4,305
7,450
6,264
5,319
5,332
5,342
5,353
5,381
5,385
5,346
5,327
5,309
5,310
5,310
5,309
5,311
5,305
5,300
5,294
5,283
5,274
5,258
5,245
5,226
5,209
Heating Oil
4,884
8,572
8,772
8,944
9,108
9,276
9,453
9,649
9,822
9,999
10,195
10,404
10,627
10,836
11,046
11,261
11,479
11,702
11,929
12,160
12,396
12,637
12,882
13,132
13,387
Total
9,189
16,022
15,036
14,263
14,440
14,618
14,806
15,030
15,207
15,345
15,522
15,713
15,937
16,146
16,355
16,572
16,784
17,002
17,223
17,443
17,670
17,895
18,127
18,358
18,596
Sulfate PM
NRLM Fuel
372
709
580
415
416
417
418
420
420
417
416
414
414
414
414
414
414
414
413
412
412
410
409
408
407
Heating Oil
173
303
310
316
322
328
334
341
347
354
360
368
376
383
391
398
406
414
422
430
438
447
455
464
473
Total
545
1012
890
731
738
745
752
761
767
771
776
782
790
797
805
812
820
828
835
842
850
857
864
872
880
30-Year (2007-2036) Emission Reduction
Undiscounted
30YrNPVat3%
30 Yr NPV at 7%
134,700
76,800
39.700
264,600
144,600
70.800
399,300
221,400
110.500
10,760
6,180
3.230
9,350
5,110
2.500
20,100
11,300
5.730
   If no shift in downgrade to heating oil is assumed, the sulfur dioxide and sulfate PM emission
reductions due to the 15 ppm L&M fuel cap are simply the differences in the emissions in the
two columns of Table 8.7-24 labeled NRLM fuel.0

   The 30-year cost effectiveness of the 15 ppm L&M cap is the ratio of the 30-year costs
shown in Tables 8.7-19 and 8.7-21 divided by the  30-year emission reductions of sulfur dioxide
and sulfate PM shown in Table 8.7-24. We have allocated 67 percent of the costs to sulfur
dioxide emission control and 33 percent to  sulfate PM control consistent with our allocation of
   G We ignored the small change in L&M fuel sulfur content which would occur if the
downgrade remained in the L&M market.
                                         8-61

-------
Final Regulatory Impact Analysis
costs associated with fuel-derived benefits throughout our analysis.  The results are presented in
Table 8.7-25.

                                      Table 8.7-25
             Incremental Cost Effectiveness of the 15 ppm L&M Fuel Sulfur Cap
                 30-year Net Present Values at a 3% Discount Rate ($2002)


3% Discount Rate
SOx
PM
7% Discount Rate
SOx
PM
With Shift of Downgrade to Heating Oil
Cost ($ million)
Emissions Reduction (tons)
Cost per ton ($/ton)
$ 1,870
221,400
$ 8,450
$ 940
11,300
$ 83,200
$ 935
110,500
$ 8,460
$ 467
5,730
$ 81,500
Without Shift of Downgrade to Heating Oil
Cost ($ million)
Emissions Reduction (tons)
Cost per ton ($/ton)
$1,493
76,800
$ 19,400
$747
6,180
$ 120,700
$747
39,700
$ 18,800
$374
3,230
$ 115,800
   As can be seen, the incremental cost effectiveness of the 15 ppm L&M fuel cap worsens
without the shift in downgrade to the heating oil market. This indicates that the cost
effectiveness of shifting downgrade from the L&M market to the heating oil market and
replacing it with 15 ppm L&M fuel is more cost effective than simply reducing L&M fuel sulfur
from 500 ppm to 15 ppm. The shift in downgrade itself is environmentally neutral from sulfur
perspective, since all of the sulfur is emitted regardless of whether it is burned in a locomotive or
marine diesel engine or a furnace or stationary diesel engine. The conversion of sulfur to PM is
less for heating oil, but as the majority of the sulfur is emitted as sulfur dioxide in either case,
sulfur dioxide emissions are the same. The difference is that, with the downgrade shift to
heating oil, the new 15 ppm L&M  fuel replaces high sulfur heating oil. Without the shift, the
new 15 ppm L&M fuel replaces 500 ppm L&M fuel.  The cost of producing 15 ppm L&M fuel
from high sulfur fuel are higher than from 500 ppm fuel, 8.3 cents per gallon versus 3.1 cents per
gallon. However, the sulfur reduction is also higher and to a much greater degree.  With heating
oil at 2155 ppm, the in-use reduction is 2144 ppm, while that from 500 ppm L&M fuel is only
329 ppm.  Thus, the sulfur benefits are a factor of seven time higher, while costs are less than a
factor of three higher.

   While we evaluate the incremental cost effectiveness of the 15 ppm L&M cap with and
without the shift of downgrade to the heating oil market, we believe that the former is the most
appropriate way to evaluate this fuel control step as it is consistent with the design of the
program which reflects the characteristics of the distribution system.  The prohibition on using
downgrade in the NRLM markets in the Northeast/Middle Atlantic area eliminates the marking
of the significant volume of heating oil in this area beginning in 2007. This is an important and
                                          8-62

-------
                                                       Aggregate Cost and Cost per Ton
valuable aspect of the final NRLM fuel program which was made regardless of any decision to
control L&M fuel to  15 ppm. Thus, it is appropriate to include the effect of this provision on the
cost effectiveness of  15 ppm L&M fuel control.

8.7.4 Costs per Ton  Summary

    Table 8.7-26 presents a summary of the cost per ton calculations presented in Sections 8.7.1
through 8.7.4.

    As noted in section 8.1, we have allocated costs slightly differently in the final analysis than
we did in the proposed analysis.11 Table 8.7-27 presents the costs per ton using the allocations
used in the proposal.  To clarify, Table 8.7-27 does not present the costs per ton from the
proposed analysis.  Instead, the values presented in Table 8.7-27 are the  costs per ton using the
final rule's costs and  emissions reductions but allocating the costs using the method used in the
proposal. As such, Table 8.7-27 provides a comparison of how the new  cost allocations affect
the costs per ton and  does not provide a comparison of the final costs per ton to the proposed
costs per ton.
    TT
     The cost allocations used in the proposal differed slightly in that costs associated with fuel-derived benefits were
allocated entirely to SOx (FRM allocations split them one-third to PM and two-thirds to SOx) and costs of 15 ppm fuel
were allocated entirely to engine-derived benefits (FRM allocations split them one-half to fuel-derived benefits and one-
half to engine-derived benefits).


                                            8-63

-------
Final Regulatory Impact Analysis
                                     Table 8.7-26
           Summary of Costs and Cost per Ton Estimates based on 30 Year NPVs
                                       ($2002)
NRT4 Full Program
NPV of Total Cost ($millions)
$/ton PM
$/ton NOx+NMHC
$/ton SOx
15ppm NRLM Fuel-only Scenario
NPV of Total Cost w/ Savings ($millions)
NPV of Total Cost w/o Savings ($millions)
$/ton PM w/ Savings
$/ton PM w/o Savings
$/ton SOx w/ Savings
$/ton SOx w/o Savings
SOOppm NRLM Fuel-only Scenario
NPV of Total Cost w/ Savings ($millions)
NPV of Total Cost w/o Savings ($millions)
$/ton PM w/ Savings
$/ton PM w/o Savings
$/ton SOx w/ Savings
$/ton SOx w/o Savings
15 ppm L&M Fuel-only Scenario (Increment) *
NPV of Incremental Cost w/ Savings ($millions)
$/ton PM w/ Savings (incremental)
$/ton SOx w/ Savings (incremental)
3% discount rate
$ 27,100
$ 1 1 ,200
$ 1,010
$ 690

$ 9,200
$ 16,300
$ 6,600
$ 1 1 ,800
$ 1 ,070
$ 1 ,900

$ (500)
$ 5,400
$ (400)
$ 4,300
$ (70)
$ 690

$ 2,810
$ 83,200
$ 8,450
7% discount rate
$ 13,800
$ 1 1 ,800
$ 1,160
$ 620

$ 4,600
$ 8,500
$ 6,000
$ 1 1 ,200
$ 970
$ 1 ,800

$ (300)
$ 3,000
$ (400)
$ 4,300
$ (70)
$ 700

$ 1 ,400
$ 81 ,500
$ 8,460
       * Includes shift of downgrade to heating oil in the Northeast/Middle Atlantic area

                                     Table 8.7-27
                    Costs and Costs per Ton of the NRT4 Full Program
                          using the Proposal's Cost Allocations
                     30 Year NPVs using a 3% Discount Rate ($2002)
NRT4 Full Program
NPV of Total Cost ($millions)
$/ton PM
$/ton NOx+NMHC
$/ton SOx
3% discount rate
$ 27,100
$ 1 1 ,000
$ 1 ,250
$ 460
                                         8-64

-------
                                                 Aggregate Cost and Cost per Ton
 Appendix 8A: Estimated Aggregate Cost and Cost per Ton
                          of Sensitivity Analyses

8A.1 What Sensitivity Analyses Have Been Performed?

   This Appendix contains two sensitivity analyses EPA performed regarding the emissions
inventory predictions from the NONROAD model, as well as cost and cost per ton analysis
which correspond to these two NONROAD model sensitivities. In the NONROAD model
sensitivity Case 1, we have adjusted the emissions predictions so thatNONROAD's fuel
consumption estimates match the predictions of fuel volume from the Energy Information
Agency. In the NONROAD model sensitivity Case 2, we have increased the fraction of diesel
generators sold in the U.S. which are considered "mobile" (and therefore decreased the
percentage which are "stationary") and we have increased the annual hours of use for several
categories of nonroad equipment in the >750 hp category.

   In the remainder of section 8 A. 1, we describe why we have included these sensitivity
analyses in the final rule. In section 8A.2, we describe what changes were made to the
NONROAD model, how each of the sensitivities were performed, and the emission inventory
impacts of Case 1 and Case 2. In section 8A.3, we describe how we have altered our engine and
fuel program cost methodology to match Case 1 and Case 2, what the resulting program cost
estimates are using Case 1 and Case 2, and finally what the cost-per-ton estimates are for Case 1
and Case 2.  In section 8A.4, we summarize the results presented in sections 8A.I through 8A.3.

   8A.1.1 What is the Case 1 Sensitivity Analysis?

   The Case 1 sensitivity analysis results from comments we received on the proposal which
suggested that the NONROAD model over-predicts the growth rate of the nonroad fleet. The
commenters suggested that the NONROAD model's growth rates should be adjusted downward
so that overall fuel consumption matches the predictions made by the Department of Energy's
Energy Information Agency (EIA). As described in detail in the Summary and Analysis of
Comments for this rule, we disagree with these comments and we have not made a change to the
NONROAD model as a result of these comments (see section 2.3.2.2.3 of the Summary and
Analysis of Comments document for this final rule).  However, we are performing a sensitivity
analysis (Case 1) which estimates what the impact of such a change would have on our estimates
of the emissions reduction of this rule, the costs of this rule, and our cost-per-ton estimates.

   8A.1.2 What is the Case 2 Sensitivity Analysis?

   The Case 2 sensitivity analysis results from information we received during the development
of the rule on two issues which indicates NONROAD is under-predicting emissions from some
nonroad engines.  One of these issues is the partitioning of generator sets into mobile and
stationary. The second issue is the annual hours of use estimates for large engines (those >750
hp).
                                        8-65

-------
Final Regulatory Impact Analysis
   8A. 1.2.1 Information Regarding Mobile & Stationary Generator Sets

   During our discussions with several engine manufacturers who produce the >750 hp diesel
engines, three manufacturers (who together represent a majority of the market), provided EPA
with recent year  sales estimates of engines used in mobile machines in the >750 hp category
(e.g., mining trucks, dozers, wheel loaders, etc.) and generator sets. These manufacturers
produce engines for generator sets which are certified to the existing Tier  1 nonroad standards, as
well as engines which are not certified to the nonroad standards because the engines are designed
for stationary power generation and therefore are not subject to EPA's nonroad standards. Many
of the >750 hp nonroad certified engines which are used in generator sets  are used in
applications such as the large portable power generators that are contained in a Class 8 truck
trailer, where power generation ratings of 1,  1.5 and 2 megawatts are common. These products
are designed to be portable and are used by rental companies and in other  industries where large
amounts of power are needed for a relatively short duration of time.  The data from the engine
manufacturers  indicates that approximately 30 percent of the >750 hp diesel generator sets sold
in the U.S. are  portable and subject to EPA nonroad diesel standards.  In addition, manufacturers
build some stationary engines to nonroad certified configurations to simplify their product base
and thus the nonroad engine standards yield an added indirect, yet real, emission benefit.

   The data which is used to estimate the nonroad equipment population in NONROAD comes
from the PSR database. This database does not distinguish between mobile and stationary diesel
generator sets.  As documented in EPA report EPA420-P-02-004, we estimate for all of the
diesel generators what percent of the PSR database diesel generator sets are mobile (and
therefore subject to the EPA's nonroad standards) and what percent is stationary. These
estimates vary  by power range, with the percent that are considered stationary  increasing with
increasing rated power. For example, for <25 hp engines we estimate 10 percent are stationary,
and for >600 hp, we estimate that 100 percent are stationary. Once these  percentages are
applied to the PSR database data to remove the estimated stationary generator  sets,  the remaining
generator set data is used to estimate the population of generator sets in NONROAD.

   The recent  information we received from the engine manufacturers (-30 percent of
generator sets >750 hp are mobile/portable) is substantially different from the current
assumptions which go into NONROAD (no generators >600 hp are mobile/portable). Because at
this time we do not have reference-able industry-wide information on this  issue, we have not
performed a new analysis to update NONROAD. However, it is clear that the recent confidential
information from the engine companies indicates NONROAD is underestimating the number of
nonroad diesel generators.  As discussed in Chapter 8A.2, we have performed a sensitivity
analysis which includes a higher percentage of mobile diesel generator sets based on the
information we received from the engine manufacturers.

   8A.1.2.2 Information Regarding Usage Factors for >750hp Mobile Machines and
   Generators

   As discussed in the preamble for this final rule, we have recognized  some of the unique
features of the  >750 hp mobile machines. Most of the >750 hp engines used in the mobile

                                         8-66

-------
                                                   Aggregate Cost and Cost per Ton
machine category are used in mining applications, such as mining trucks, dozers, excavators and
loaders. As part of our feasibility analysis, we spent a considerable amount of time with a
number of engine manufacturers and equipment manufacturers to understand the applications
these large engines are used in.  In addition, several manufacturers provided EPA with data
regarding the >750 hp mobile machine applications. One of the pieces of data which we noticed
was the high annual hours of use for this equipment, which in some cases was greater than 4,000
hours per year.  During our discussions with both engine and equipment manufacturers,
companies made the point these large pieces of equipment are very expensive (in excess of $1
million), and that mining operations are often run 7 days a week, "around the clock".  Because of
these two factors, the large mobile machines are operated at higher annual usage rates than most
nonroad applications.

   While we received this type of information from multiple companies, the most convincing
data we received came from one of the industry's larger equipment companies.  This equipment
company provided EPA with confidential data for mobile machines >750 hp which included
sales and annual hours of use estimates.  The equipment types covered by the data included
applications such as off-highway trucks, dozers, wheel loaders, and off-highway tractors. The
data was representative of 10 years worth of sales, and several thousand pieces of equipment.
On average, the manufacturer estimated the annual hours of use for this equipment was > 3,500
hours per year.  We also received information from several engine  and equipment companies
which indicates the annual hours of use for >750 portable generator sets are on the order of 1,000
hours per year.

   The NONROAD model contains estimates of annual hours of use which are used in the
process of estimating annual emissions.  The annual hours of use values are documented in EPA
report EPA420-P-02-014. The annual hours of use do not vary by  power category, therefore the
estimate for a 250 hp dozer is the same as the estimate for a 1,000 hp dozer. For the >750 hp
applications on which we received new data, the highest annual hours of use value in
NONROAD is 1,641  hours/year for off-highway trucks, and for generator sets the value is 338
hours/year. These values are substantially lower than the usage information we received from
engine and equipment manufacturers. While we now believe NONROAD underestimates the
emissions impact of the >750hp equipment based on the new information we have received, we
have not changed NONROAD at this time. The information we received, though useful for this
sensitivity analysis, is not adequately reference-able and may not be sufficiently representative.
In addition, while we believe it is directionally correct, we have not had an opportunity to
independently verify the information or collect additional data from other sources. As a result,
though not reflected in the NONROAD model results for this final rule, the Case 2 sensitivity
analysis does include higher annual hours of use values for several categories  of mobile
machines >750hp and for generator sets >750hp, which is based on the information we received
from engine and equipment companies.
                                         8-67

-------
Final Regulatory Impact Analysis
8A.2 What Emissions Modeling was Done?

8A.2.1  Case 1: Inventories Adjusted to Match Fuel Consumption Derived from EIA
Sources

   To represent the emissions inventory for Case 1, we did not perform additional NONROAD
runs. Rather, we adjusted the NONROAD fuel consumption and emissions estimates so that
estimated fuel consumption matched fuel consumption estimates derived from EIA sources. We
performed the adjustment by applying ratios to the NONROAD fuel consumption and emissions
outputs. Specifically,, we calculated an adjustment ratio r as
                              ry =
                    NONROAD.j


                     -* T7TA ,,
where
       NONROADj;
is a national fuel consumption estimate as generated by Draft NONROAD2004
for year y, and FElAj, is a corresponding estimate derived from EIA's Annual Energy Outlook
2003 (AEO 2003). These reports provide distillate fuel consumption projections by economic
sector.
   The derivation of FEIA is based on a linear projection of nonroad diesel fuel consumption
from 2002 to 2040, as described below. To establish a basis for estimaton of a growth rate, we
derived estimates for the years 2002 and 2014 from AEO 2003, the derivation of which is
described in Chapter 7.1 of the RIA. These two estimates, along with corresponding estimates
from Draft NONROAD2004, are  shown in Table 8A.2-1.

                                     Table 8A.2-1
 Nonroad Fuel Consumption: Draft NONROAD2004 and Estimates derived from EIA Sources
                               (Million gallons per year)
Year
2002
2014
Draft NONROAD2004
10,625
14,433
Derived from EIA Sources
8,428
9,814
   Using the following equation, we estimated a 1.4%/year average linear growth rate (without
compounding) in fuel consumption gEIA over this 12-year period:
                               _
                           &E1A ~
                                    2014-2002    FRTA.,m,
                                                   EIA,2002
Using the resulting growth rate (0.014/year), we projected fuel consumption from 2002 to 2040,
based on the expression

                     ^EIA^^EIA^^+O-
                                        8-68

-------
                                                  Aggregate Cost and Cost per Ton
   The resulting EIA-derived fuel consumption estimates are shown in Table 8A.2-2, along with
fuel consumption estimates from Draft NONROAD2004.  The ratio of the two fuel consumption
estimates in each year are also shown.

   Table 8A.2-3 shows projected land-based nonroad diesel fuel consumption and associated
emissions inventories (NOX, SO2, PM10) at the national level for selected years between 2001 and
2040, as estimated by NONROAD and from EIA sources. Results are shown for both the base
and control cases. These results are also presented graphically in Figures 8A.2-1 - 8A.2-4.
                                        8-69

-------
Final Regulatory Impact Analysis
                                  Table 8A.2-2
             2001 -2040 Nonroad Fuel Consumption (Million gallons per year)
Calendar Year
2001
2002
2003
2004
2005
2006
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2031
2032
2033
2034
2035
2036
2037
2038
2039
2040
Draft NUNRUAD2004
(-^NONROAD)
10,625
10,919
11,213
11,507
11,801
12,092
12,384
12,676
12,968
13,259
13,553
13,846
14,139
14,433
14,726
15,016
15,307
15,597
15,887
16,178
16,468
16,759
17,049
17,339
17,630
17,918
18,206
18,495
18,783
19,071
19,360
19,648
19,936
20,225
20,513
20,801
21,090
21,378
21,666
21.955
lilA FUKS/AEU Derived
(^K,A)
9,080
8,428
8,544
8,659
8,775
8,890
9,006
9,121
9,237
9,352
9,468
9,583
9,699
9,814
9,930
10,045
10,160
10,276
10,391
10,507
10,622
10,738
10,853
10,969
11,084
11,200
11,315
11,431
11,546
11,662
11,777
11,892
12,007
12,123
12,239
12,354
12,470
12,585
12,701
12.816
Ratio
(r)
1.170
1.296
1.312
1.329
1.345
1.360
1.375
1.390
1.404
1.418
1.431
1.445
1.458
1.471
1.483
1.495
1.507
1.518
1.529
1.540
1.550
1.561
1.571
1.581
1.591
1.600
1.609
1.618
1.627
1.635
1.644
1.652
1.660
1.668
1.676
1.684
1.691
1.699
1.706
1.713
                                      8-70

-------
                            Aggregate Cost and Cost per Ton
               Table 8A.2-3
 Case 1: Adjustment to Match EIA Projections
Projected Nonroad Diesel Emissions Inventories
Year(y)
2002
2005
2010
2015
2020
2025
2030
2035
2040
National Emissions Inventory (thousand tons)
NOx
Base
1,184
1,096
906
781
731
722
733
754
780
Control
1,184
1,096
906
650
442
334
282
259
251
S02
Base
133
139
140
145
154
162
171
179
188
Control
133
139
10.7
0.673
0.644
0.644
0.660
0.684
0.712
PM10
Base
128
111
94.0
87.9
86.8
88.1
90.2
92.8
96.3
Control
128
111
82.4
55.4
33.7
21.2
13.7
9.5
7.5
                   8-71

-------
Final Regulatory Impact Analysis
8A.2.2  Case 2: Large Equipment Population and Activity

   To represent Case 2, we performed NONROAD runs with modified inputs for selected
equipment types. Specifically, we used modified activity for large equipment (>750 hp) in five
equipment types, as shown in Table 8A.2-4. This change represents the use of large equipment
on a continuous shift basis. Additionally, we modified the fractions of generators assumed to be
mobile, as opposed to stationary equipment, as shown in Table 8A.2-5.  The modified fractions
increased populations for generators of size 100 hp and greater, resulting in an increase in the
total generator population of approximately 135,000 pieces. As in Case 1, we repeated the
analysis for both the base and control  cases.

   Table 8A.3-6 shows projected land-based nonroad diesel fuel consumption and associated
emissions inventories (NOX, SO2, PM10) at the national level for selected years between 2001 and
2040, for both the base and control  cases. These results are also presented graphically in Figures
8A.2-1 - 8A.2-4.

                                     Table 8A.2-4
                     Case 2: Large Equipment Population and Activity
                 Annual Activity Estimates for Large Equipment (>750 hp)
Equipment Type
Excavators
Off-Highway Trucks
Rubber Tire Loaders
Crawler Tractors/Dozers
Off-Highway Tractors
Generators
Activity (hours/year)
FRM Base
1,092
1,641
761
936
855
338
Sensitivity Case
3,800
3,800
3,800
3,800
3,800
1,000
                                         8-72

-------
                                     Aggregate Cost and Cost per Ton
                         Table 8A.2-5
         Case 2: Large Equipment Population and Activity
Modified Mobile-Equipment Population Fractions for Diesel Generators
Hp Class
<25
25-40
40-50
50-75
75-100
100-175
175-300
300-600
600-750
>750
FRM Base
Mobile Fraction
0.90
0.90
0.70
0.70
0.70
0.20
0.15
0.10
0.0
0.0
Mobile Population
240,180
121,050
16,530
61,000
74,240
25,340
14,090
7,320
0
0
Total 559,750
Sensitivity Case
Mobile Fraction
0.90
0.90
0.70
0.70
0.70
0.62
0.54
0.46
0.38
0.30
Mobile Population
240,180
121,050
16,530
61,000
74,240
78,560
50,720
33,660
6,260
12,290
694,490
                            8-73

-------
                         Table 8A.2-6
         Case 2: Large Equipment Population and Activity:
Projected Nonroad Diesel Fuel Consumption and Emissions Inventories
Year (y)
2001
2005
2010
2015
2020
2025
2030
2035
2040
Fuel Consumption (million gal)
NONROAD
FRM 50-state
Base
10,630
11,800
13,260
14,730
16,180
17,630
19,070
20,510
21,950
NONROAD
Sensitivity-Case
12,550
13,960
15,710
17,470
19,220
20,970
22,710
24,440
26,180
National Emissions Inventory (thousand tons)
NOx
Base
1,817
1,759
1,519
1,409
1,393
1,434
1,502
1,585
1,678
Control
1,817
1,759
1,518
1,132
848
692
916
595
594
SO2
Base
198
220
234
256
282
307
333
358
384
Control
198
220
17.9
1.15
1.16
1.21
1.28
1.37
1.45
PM10
Base
189
165
148
145
149
156
164
173
183
Control
189
165
128
89.3
57.0
38.2
26.4
19.3
16.0

-------
                                                    Aggregate Cost and Cost per Ton
       30,000
                                                                      • FRM: Base
                                                                       Case 2: Large Equipment
                                                                      • Case 1: Adjustment to EIA
             2005     2010
       2015    2020
       Time (years)
                2025    2030
Figure 8A.2-1. Projected land-based nonroad diesel fuel consumption at the national level for the FRM
base and two sensitivity cases. Case 1 represents Draft NONROAD2004 estimates adjusted to match
EIA-based projections; Case 2 represents modified population and activity estimates for large
equipment (>750 hp).
       2,000,000
       1,800,000 -
    _ 1,600,000 -
    (/>           *
    § 1,400,000-
    +2.
    w 1,200,000-
    •5; 1,000,000-
    1   800,000 -
    UJ
    x   600,000 -
    o
    z   400,000 -
         200,000 -
               2005
                                              -FRM: Base
                                               Case 2: Base
                                              -Case 1: Base
                                              -FRM: Control
                                               Case 2: Control
                                              -Case 1: Control
2010
2015     2020
 Time (years)
2025
2030
 Figure 8A.2-2. Projected land-based nonroad NOx inventories at the national level for the
 FRM base and two sensitivity cases. Case 1 represents Draft NONROAD2004 estimates
 adjusted to match EIA-based projections; Case 2 represents modified population and activity
 estimates for large equipment (>750 hp).
                                         8-75

-------
Final Regulatory Impact Analysis
           350,000

           300,000
        =  250,000
                                 *****************
                                                          • FRM: Base
                                                          Case 2: Base
                                                          • Case 1: Base
                                                          • FRM: Control
                                                          Case 2: Control
                                                          • Case 1: Control
                 2005      2010       2015      2020
                                       Time (years)
                                      2025
                               2030
    Figure 8A.2-3.  Projected land-based nonroad SO2 inventories at the national level for the FRM base and
    two sensitivity cases. Case 1 represents Draft NONROAD2004 estimates adjusted to match EIA-based
    projections; Case 2 represents modified population and activity estimates for large equipment (>750 hp).
              200,000
          •3T  150,000
           c
           o
           W
           O  100,000
IhSfrHf
          .8
          LLI
               50,000--
                                                       • FRM Base
                                                        Case 2: Base
                                                       • Case 1: Base
                                                       • FRM Control
                                                        Case 2: Control
                                                       • Case 1: Control
                    2005
        2010
2015     2020
 Time (years)
2025
2030
      Figure 8A.2-4. Projected land-based nonroad PM10 inventories at the national level for the FRM base
      and two sensitivity cases. Case 1 represents Draft NONROAD2004 estimates adjusted to match EIA-
      based projections; Case 2 represents modified population and activity estimates for large equipment
      (>750 hp).
                                             8-76

-------
                                                   Aggregate Cost and Cost per Ton
8A.3 What Are the Costs and Costs per Ton?

   Here we look at the cost per ton of two sensitivity cases—a Case 1 sensitivity using future
fuel consumption projections developed by the Energy Information Administration (EIA); and, a
Case 2 sensitivity that incorporates more generator sets in both the costs and emissions
reductions estimates than are incorporated under NRT4 full engine and fuel program (i.e., the
NRT4 final rule estimates).

8A.3.1 Costs and Costs per Ton for the Case 1 Sensitivity

   Under the Case 1 sensitivity we use future fuel projections developed by EIA rather than
using the projections generated in our NONROAD model as discussed in Section 8A.1. Doing
this results in lower fuel-related costs (including all operating costs expressed throughout this
Regulatory Impact Analysis on a cent-per-gallon basis) since the EIA projections are lower than
our model's projections. Doing this also results in lower emissions reductions as discussed in
Section 8A.2. The engine and equipment costs under the Case 1 sensitivity would be identical to
those under the full engine and fuel program since all engine standards would still be
implemented. Tables 8A.3-1 though 8A.3-4 show all fuel-related costs associated with the Case
1 sensitivity.  All these tables are analogous to Tables 8.4-1 through 8.4-4 presented above for
the NRT4 final program. The cent per gallon fuel costs are presented in Table 7.5-1.
                                          8-77

-------
                                                                    Table 8A.3-1
                                            Aggregate Fuel Costs of the Case 1 Sensitivity ($2002)
Year
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2031
2032
2033
2034
2035
2036
30YrNPVat3%
30YrNPVat7%
Affected NR Fuel
500 ppm
(106aallons)
3,671
6,373
6,454
3,086
631
531
460
194
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
18,602
15,567
15 ppm
(106aallons)
-
-
-
3,873
6,721
6,574
6,488
7,153
7,662
7,751
7,840
7,929
8,018
8,107
8,196
8,285
8,374
8,464
8,553
8,642
8,731
8,820
8,909
8,998
9,087
9,176
9,265
9,354
9,444
9,533
124,895
64,783
Affected L&M Fuel
500 ppm
(106aallons)
1,981
3,438
3,483
3,069
2,785
1,243
120
50
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
13,818
11,317
15 ppm
(106aallons)
-
-
-
-
-
2,116
3,657
3,527
3,441
3,476
3,511
3,551
3,598
3,632
3,663
3,709
3,747
3,788
3,828
3,868
3,909
3,949
3,989
4,029
4,070
4,110
4,150
4,190
4,231
4,271
52,202
26,078
Fuel Cost *
500 ppm
($/aal)
$ 0.021
$ 0.021
$ 0.021
$ 0.029
$ 0.034
$ 0.035
$ 0.036
$ 0.036
























15 ppm
($/aal)
-
$
$
$ 0.058
$ 0.058
$ 0.062
$ 0.064
$ 0.067
$ 0.069
$ 0.069
$ 0.069
$ 0.069
$ 0.069
$ 0.069
$ 0.069
$ 0.069
$ 0.069
$ 0.069
$ 0.069
$ 0.069
$ 0.069
$ 0.069
$ 0.069
$ 0.069
$ 0.069
$ 0.069
$ 0.069
$ 0.069
$ 0.069
$ 0.069


NR Fuel Costs
500 ppm
(106 dollars)
$ 77
$ 134
$ 136
$ 88
$ 21
$ 18
$ 16
$ 7
$
$
$
$
-
$
$
$
-
$
$
$
-
$
$
$
-
$
$
$
-
-
$ 430
$ 357
15 ppm
(106 dollars)
-
-
-
225
390
404
415
479
529
535
541
547
553
559
566
572
578
584
590
596
602
609
615
621
627
633
639
645
652
658
$ 8,447
$ 4,342
Total
(106 dollars)
$ 77
$ 134
$ 136
$ 313
$ 411
$ 423
$ 431
$ 485
$ 529
$ 535
$ 541
$ 547
$ 553
$ 559
$ 566
$ 572
$ 578
$ 584
$ 590
$ 596
$ 602
$ 609
$ 615
$ 621
$ 627
$ 633
$ 639
$ 645
$ 652
$ 658
$ 8,877
$ 4,698
L&M Fuel Costs
500 ppm
(106 dollars)
$ 42
$ 72
$ 73
$ 88
$ 95
$ 43
$ 4
$ 2
-
$
$
$
-
$
$
$
-
$
$
$
-
$
$
$
-
$
$
$
-
-
$ 354
$ 287
15 ppm
HO6 dollars)
-
$
$
$
$
$ 130
$ 234
$ 236
$ 237
$ 240
$ 242
$ 245
$ 248
$ 251
$ 253
$ 256
$ 259
$ 261
$ 264
$ 267
$ 270
$ 272
$ 275
$ 278
$ 281
$ 284
$ 286
$ 289
$ 292
$ 295
$ 3,570
$ 1 ,776
Total
(106 dollars)
$ 42
$ 72
$ 73
$ 88
$ 95
$ 173
$ 238
$ 238
$ 237
$ 240
$ 242
$ 245
$ 248
$ 251
$ 253
$ 256
$ 259
$ 261
$ 264
$ 267
$ 270
$ 272
$ 275
$ 278
$ 281
$ 284
$ 286
$ 289
$ 292
$ 295
$ 3,924
$ 2,063
NRLM Annual
Fuel Costs
(106 dollars)
$ 119
$ 206
$ 209
$ 401
$ 506
$ 596
$ 670
$ 723
$ 766
$ 775
$ 783
$ 792
$ 802
$ 810
$ 818
$ 828
$ 836
$ 845
$ 854
$ 863
$ 872
$ 881
$ 890
$ 899
$ 908
$ 917
$ 926
$ 935
$ 944
$ 952
$ 12,801
$ 6,762
*Fuel costs are relative to uncontrolled fuel and assume that, during the transitional years of 2010, 2012, & 2014, the first 5 months are at the previous year's cost and the remaining 7
months are at the next year's cost.
See Appendix 8B for information on how these fuel volumes were developed.

-------
                              Table 8A. 3-2
Oil Change Maintenance Savings Associated with the Case 1 Sensitivity ($2002)
Year
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2031
2032
2033
2034
2035
2036
30YrNPVat3%
30YrNPVat7%
Affected NR Fuel
500 ppm
(106 gallons)
3,671
6,373
6,454
3,086
631
531
460
194
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
18,602
15,567
15 ppm
(106 gallons)
-
-
-
3,873
6,721
6,574
6,488
7,153
7,662
7,751
7,840
7,929
8,018
8,107
8,196
8,285
8,374
8,464
8,553
8,642
8,731
8,820
8,909
8,998
9,087
9,176
9,265
9,354
9,444
9,533
124,895
64,783
Affected L&M Fuel
500 ppm
(106 gallons)
1,981
3,438
3,483
3,069
2,785
1,243
120
50
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
13,818
11,317
15 ppm
(106 gallons)
-
-
-
-
-
2,116
3,657
3,527
3,441
3,476
3,511
3,551
3,598
3,632
3,663
3,709
3,747
3,788
3,828
3,868
3,909
3,949
3,989
4,029
4,070
4,110
4,150
4,190
4,231
4,271
52,202
26,078
NR Savings
savings=$0.029/gal
(106 dollars)
$ 107
$ 186
$ 189
$ 90
$ 18
$ 16
$ 13
$ 6
$
$
$
$
$
$
-
$
$
$
-
$
$
$
-
$
$
$
-
$
$
$
$ 544
$ 455
savings=$0.032/gal
(106 dollars)
$
$
$
$ 124
$ 215
$ 211
$ 208
$ 229
$ 245
$ 248
$ 251
$ 254
$ 257
$ 260
$ 263
$ 265
$ 268
$ 271
$ 274
$ 277
$ 280
$ 282
$ 285
$ 288
$ 291
$ 294
$ 297
$ 300
$ 302
$ 305
$ 4,000
$ 2,075
L&M Savings
savings=$0.010/gal
(106 dollars)
$ 21
$ 36
$ 36
$ 32
$ 29
$ 13
$ 1
$ 1
$
$
$
$
$
$
-
$
$
$
-
$
$
$
-
$
$
$
-
$
$
$
$ 145
$ 118
savings=$0.011/gal
(106 dollars)
$
$
$
$
$
$ 24
$ 42
$ 40
$ 39
$ 40
$ 40
$ 41
$ 41
$ 42
$ 42
$ 43
$ 43
$ 43
$ 44
$ 44
$ 45
$ 45
$ 46
$ 46
$ 47
$ 47
$ 48
$ 48
$ 49
$ 49
$ 599
$ 299
NRLM
Total Savings
(106 dollars)
$ 128
$ 222
$ 225
$ 246
$ 263
$ 263
$ 264
$ 276
$ 285
$ 288
$ 291
$ 295
$ 298
$ 301
$ 305
$ 308
$ 311
$ 315
$ 318
$ 321
$ 324
$ 328
$ 331
$ 334
$ 338
$ 341
$ 344
$ 348
$ 351
$ 354
$ 5,287
$ 2,948

-------
Final Regulatory Impact Analysis
                                           Table 8A. 3-3
    CDPF Maintenance and CDPF Regeneration Costs Associated with the Case 1 Sensitivity
                                             ($2002)
Year
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2031
2032
2033
2034
2035
2036
30YrNPVat3%
30YrNPVat7%
Fuel Consumed in New
CDPF Equipped
Engines
(106aallons)
-
-
-
-
461
1,204
2,076
2,953
3,885
4,782
5,612
6,369
7,056
7,685
8,239
8,726
9,168
9,579
9,962
10,314
10,622
10,896
11,151
1 1 ,390
11,612
1 1 ,823
12,027
12,224
12,407
12,579
1 1 1 ,737
50,796
Weighted
Maintenance
Cost
($/aal)
$
$
$
$
$ 0.002
$ 0.003
$ 0.005
$ 0.006
$ 0.006
$ 0.006
$ 0.006
$ 0.006
$ 0.006
$ 0.006
$ 0.006
$ 0.006
$ 0.006
$ 0.006
$ 0.006
$ 0.006
$ 0.006
$ 0.006
$ 0.006
$ 0.006
$ 0.006
$ 0.006
$ 0.006
$ 0.006
$ 0.006
$ 0.006


Weighted
Regeneration
Cost
($/aal)
$
$
$
$
$ 0.010
$ 0.010
$ 0.010
$ 0.007
$ 0.008
$ 0.008
$ 0.008
$ 0.008
$ 0.008
$ 0.008
$ 0.008
$ 0.008
$ 0.008
$ 0.008
$ 0.008
$ 0.008
$ 0.008
$ 0.008
$ 0.008
$ 0.008
$ 0.008
$ 0.008
$ 0.008
$ 0.008
$ 0.008
$ 0.008


CDPF Maintenance
Cost
(106 dollars)
$
$
$
$
$ 1
$ 4
$ 10
$ 17
$ 23
$ 29
$ 34
$ 39
$ 43
$ 47
$ 50
$ 53
$ 56
$ 59
$ 61
$ 63
$ 65
$ 67
$ 68
$ 70
$ 71
$ 72
$ 74
$ 75
$ 76
$ 77
$ 675
$ 305
CDPF Regeneration
Cost
(106 dollars)
$
$
$
$
$ 5
$ 12
$ 21
$ 22
$ 30
$ 38
$ 45
$ 51
$ 57
$ 62
$ 67
$ 71
$ 75
$ 78
$ 81
$ 84
$ 87
$ 89
$ 91
$ 93
$ 95
$ 96
$ 98
$ 100
$ 101
$ 102
$ 911
$ 415
Total Costs
(106 dollars)
$
$
$
$
$ 5
$ 16
$ 32
$ 39
$ 53
$ 66
$ 79
$ 90
$ 100
$ 110
$ 118
$ 124
$ 131
$ 137
$ 142
$ 147
$ 152
$ 156
$ 159
$ 163
$ 166
$ 169
$ 172
$ 174
$ 177
$ 180
$ 1 ,587
$ 720
* Note that fuel used in CDPF engines includes some highway spillover fuel.
"Weighted Regeneration Cost ($/gal) changes year-to-year due to different fuel economy impacts with a NOx adsorber (1 percent)
and without a NOx adsorber (2 percent) matched with the phase-in schedules of the emission standards.
                                               8-80

-------
                                                     Aggregate Cost and Cost per Ton
                                       Table 8A. 3-4
               CCV Maintenance Costs Associated with the Case 1 Sensitivity
                                         ($2002)
Year
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2031
2032
2033
2034
2035
2036
30YrNPVat3%
30YrNPVat7%
Fuel Consumed in
Engines Adding CCV
System
(106aallons)
-
183
186
173
778
1,606
2,561
3,496
4,478
5,447
6,339
7,133
7,855
8,516
9,099
9,612
10,076
10,505
10,905
11,271
1 1 ,593
1 1 ,879
12,146
12,398
12,631
12,853
13,069
13,277
13,471
13,654
124,105
56,982
Weighted
Maintenance
Cost
($/aal)
$
$ 0.000
$ 0.000
$ 0.000
$ 0.001
$ 0.001
$ 0.002
$ 0.002
$ 0.002
$ 0.002
$ 0.002
$ 0.002
$ 0.002
$ 0.002
$ 0.002
$ 0.002
$ 0.002
$ 0.002
$ 0.002
$ 0.002
$ 0.002
$ 0.002
$ 0.002
$ 0.002
$ 0.002
$ 0.002
$ 0.002
$ 0.002
$ 0.002
$ 0.002


Total Costs
(106 dollars)
$
$ 0
$ 0
$ 0
$ 1
$ 2
$ 4
$ 5
$ 7
$ 8
$ 10
$ 11
$ 12
$ 13
$ 14
$ 14
$ 15
$ 16
$ 16
$ 17
$ 18
$ 18
$ 18
$ 19
$ 19
$ 19
$ 20
$ 20
$ 20
$ 21
$ 187
$ 86
                   * Weighted Maintenance
                   implementation schedule
Cost ($/gal) changes year-to-year due to the
for engines adding the CCV system.
   Using Tables 8A.3-1 through 8A.3-4 and Table 8.2-2 (engine fixed costs by pollutant), Table
8.2-4 (engine variable costs by pollutant), Table 8.3-2 (equipment fixed costs by pollutant), and
Table 8.3-4 (equipment variable costs) we can generate the annual costs and costs by pollutant
for the Case 1 sensitivity.  Table 8A.3-5 shows these results (this table is analogous to Tables
8.5-1 and 8.5-2 for the NRT4 final program). Note that the pollutant allocations for the Case 1
sensitivity are identical to those used for the NRT4 final program (see Table 8.1-2). Also shown
in Table 8A.3-5 are the emissions reductions associated with the Case 1 sensitivity (these values
are analogous to Table 8.6-1 for the NRT4 final program).
                                           8-81

-------
Table 8A.3-5
Summary of Aggregate Costs, Costs by Pollutant, and Emissions Reductions Associated with the Case 1 Sensitivity ($2002)
Year
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2031
2032
2033
2034
2035
2036
30 Yr NPVat3%
30YrNPVat7%
Engine Costs
(Smillion)
-
$ 81
$ 82
$ 80
$ 403
$ 718
$ 882
$ 973
$ 950
$ 920
$ 910
$ 901
$ 890
$ 900
$ 913
$ 927
$ 940
$ 954
$ 967
$ 980
$ 994
$ 1 ,007
$ 1 ,021
$ 1 ,034
$ 1 ,048
$ 1 ,061
$ 1 ,074
$ 1 ,088
$ 1,101
$ 1,115
$ 14,054
$ 7,215
Equipment
Costs
(Smillion)
-
$ 5
$ 5
$ 5
$ 62
$ 106
$ 121
$ 146
$ 149
$ 150
$ 150
$ 146
$ 147
$ 147
$ 99
$ 66
$ 56
$ 36
$ 32
$ 32
$ 33
$ 33
$ 33
$ 34
$ 34
$ 35
$ 35
$ 35
$ 36
$ 36
$ 1 ,281
$ 754
Fuel Costs
(Smillion)
$ 119
$ 206
$ 209
$ 401
$ 506
$ 596
$ 670
$ 723
$ 766
$ 775
$ 783
$ 792
$ 802
$ 810
$ 818
$ 828
$ 836
$ 845
$ 854
$ 863
$ 872
$ 881
$ 890
$ 899
$ 908
$ 917
$ 926
$ 935
$ 944
$ 952
$ 12,801
$ 6,762
Other Operating
Costs
(Smillion)
$ (128)
$ (222)
$ (225)
$ (246)
$ (256)
$ (245)
$ (229)
$ (232)
$ (225)
$ (214)
$ (203)
$ (194)
$ (1 86)
$ (1 79)
$ (173)
$ (1 69)
$ (1 65)
$ (162)
$ (1 59)
$ (157)
$ (1 55)
$ (1 54)
$ (1 53)
$ (1 53)
$ (1 53)
$ (1 53)
$ (1 53)
$ (1 53)
$ (154)
$ (1 54)
$ (3,514)
$ (2,143)
Net Operating
Costs
(Smillion)
$ (9)
$ (16)
$ (16)
$ 154
$ 250
$ 351
$ 440
$ 491
$ 541
$ 561
$ 580
$ 598
$ 615
$ 631
$ 645
$ 659
$ 671
$ 683
$ 695
$ 706
$ 717
$ 727
$ 737
$ 746
$ 755
$ 764
$ 773
$ 781
$ 790
$ 798
$ 9,286
$ 4,619
Total Annual
Costs
(Smillion)
$ (9)
$ 69
$ 71
$ 239
$ 715
$ 1,174
$ 1 ,444
$ 1,610
$ 1 ,640
$ 1 ,631
$ 1 ,640
$ 1 ,645
$ 1 ,652
$ 1 ,678
$ 1 ,657
$ 1 ,651
$ 1 ,667
$ 1 ,672
$ 1 ,694
$ 1,719
$ 1 ,743
$ 1 ,767
$ 1 ,791
$ 1,814
$ 1 ,837
$ 1 ,860
$ 1 ,882
$ 1 ,905
$ 1 ,927
$ 1 ,949
$ 24,622
$ 12,588
PM Costs
(Smillion)
$ (3)
$ 80
$ 81
$ 145
$ 441
$ 735
$ 912
$ 893
$ 911
$ 914
$ 924
$ 922
$ 944
$ 962
$ 957
$ 959
$ 972
$ 989
$ 1 ,002
$ 1,019
$ 1 ,035
$ 1 ,050
$ 1 ,065
$ 1 ,080
$ 1 ,094
$ 1,109
$ 1,123
$ 1,137
$ 1,151
$ 1,165
$ 14,505
$ 7,429
NOx+NMHC
Costs
(Smillion)
-
$ 0
$ 0
$ 25
$ 170
$ 282
$ 330
$ 501
$ 503
$ 487
$ 485
$ 489
$ 471
$ 477
$ 458
$ 448
$ 448
$ 433
$ 439
$ 445
$ 451
$ 457
$ 463
$ 469
$ 474
$ 480
$ 486
$ 492
$ 497
$ 503
$ 6,590
$ 3,372
SOx Costs
(Smillion)
$ (6)
$ (11)
$ (11)
$ 69
$ 104
$ 157
$ 201
$ 215
$ 226
$ 229
$ 231
$ 234
$ 237
$ 239
$ 241
$ 244
$ 247
$ 250
$ 252
$ 255
$ 258
$ 260
$ 263
$ 265
$ 268
$ 271
$ 273
$ 276
$ 279
$ 281
$ 3,527
$ 1 ,787
PM Reduction
(tons)
7,100
12,700
13,200
15,800
19,000
23,200
27,900
32,800
37,900
42,500
47,000
51,100
54,900
58,500
61 ,700
64,700
67,400
70,000
72,400
74,700
76,800
78,600
80,400
82,100
83,600
85,000
86,400
87,800
89,000
90,200
919,400
450,100
NOx+NMHC
Reduction
(tons)
-
100
200
400
13,200
34,100
57,500
97,100
136,200
173,800
209,400
241 ,700
271 ,800
299,400
324,000
346,100
366,500
385,000
402,200
417,900
432,200
445,200
457,300
468,000
477,900
487,400
496,300
504,900
512,900
520,600
4,421 ,600
1,981,700
SOx
Reduction
(tons)
87,600
153,200
1 54,400
186,500
201,600
209,500
215,500
218,500
221,300
223,400
225,500
227,700
230,100
232,200
234,200
236,600
238,800
241,100
243,300
245,600
247,900
250,100
252,400
254,700
257,100
259,400
261,700
264,100
266,400
268,800
4,032,300
2,240,800

-------
                                                        Aggregate Cost and Cost per Ton
    The calculations of cost per ton of each emission reduced under the Case 1 sensitivity divides
the net present value of the annual costs assigned to each pollutant (Table 8 A. 3-5) by the net
present value of the total annual reductions of each pollutant (Table 8A.3-5). The 30-year net
present value of the costs associated with each pollutant, calculated with a three percent discount
rate, are  shown in Table 8A.3-5 as $6.6 billion for NOx+NMHC, $14.5 billion for PM, and $3.5
billion for SOx with the total cost of the program estimated at $24.6 billion. The 30-year net
present value, with a three percent discount rate, of emission reductions are estimated at 4.4
million tons of NOx+NMHC, 919 thousand tons of PM and 4.0 million tons of SOx (see Table
8A.3-5).  How these emissions reductions were developed is described in Section 8A.2 (see
Table 8A.2-3).1 The results of the cost per ton calculations are shown in  Table 8A.3-6.

                                         Table 8A. 3-6
                       Aggregate Cost per Ton for the Case 1  Sensitivity
               30-year Net Present Values at a 3% and 7% Discount Rate ($2002)
Item
Cost per Ton NOx+NMHC
Cost per Ton PM
Cost per Ton SOx
3% discount rate
$1,490
$15,800
$870
7% discount rate
$1,700
$16,500
$800
Source
Calculated
Calculated
Calculated
       We have also calculated the cost per ton of emissions reduced in the year 2030 using the
annual costs and emission reductions in that year alone.  This number, shown in Table 8A.3-7,
approaches the long-term cost per ton of emissions reduced.
     Note that the emissions reductions shown in Table 8A.3-5 are not identical to the reductions one would get using
the inventories presented in Table 8A.2-3. The emissions inventories in Table 8A.2-3 are for land based nonroad engines
only and do not include emissions associated with locomotive and marine engines.  To make the comparison between the
Case 1 $/ton and the NRT4 full program $/ton, the Case 1 locomotive and marine emissions reductions are included with
the Case 1 nonroad land based emissions reductions. Because the emissions reductions associated with locomotive and
marine engines are directly proportional to gallons of fuel consumed (because no new emission control hardware is being
added to those engines), we have calculated the locomotive and marine emissions reductions by taking the ratio of the
Case 1 locomotive and marine gallons consumed to the NRT4 full program locomotive and marine gallons consumed and
multiplied that ratio by the NRT4 locomotive and marine emissions reductions to arrive at the Case 1 locomotive and
marine emissions reductions. These Case 1 locomotive and marine emissions reductions were then added to the Case 1
nonroad land based emissions reductions to arrive at the Case 1 emissions reductions shown in Table 8A.3-5.
                                             8-83

-------
Final Regulatory Impact Analysis
                                      Table 8A. 3-7
                     Long-Term Cost per Ton of the Case 1 Sensitivity
                        Annual Values without Discounting ($2002)
Pollutant
NOx+NMHC
PM
SOx
Long-Term Cost per Ton
in 2030
$1,000
$13,200
$1,050
8A.3.2 Costs and Costs per Ton of the Case 2 Sensitivity

   Under the Case 2 sensitivity, more generator sets are assumed to be mobile than are assumed
under NRT4 full engine and fuel program, as described in Section 8A. 1. This results in higher
engine and equipment variable costs since more generator sets (gensets) add NOx adsorbers and
CDPFs and more equipment fixed costs since more machines must undergo redesign and product
support literature changes.  Engine fixed costs would not change since we believe that the R&D
work estimated for the NRT4 full program would cover these additional gensets. Fuel-related
costs would also increase because more machines would incur CDPF and CCV maintenance
costs and CDPF regeneration costs. Increased costs for the incrementally higher cost fuel and
savings associated with that fuel should not change since our earlier calculations for the NRT4
full engine and fuel program would have included these costs (i.e., those costs and savings are
included in the NRT4 final rule).

   We have calculated the increased engine variable costs using the equations shown in Table
6.4-2 and have applied those costs to the same nonroad engine fleet with the exception that more
gensets are included. Likewise, we followed the same process for developing equipment costs as
described in Chapter 6 to generate the higher equipment fixed and variable costs.

   Because more machines are adding the new engine  hardware (CDPFs and NOx adsorbers),
the emissions reductions associated with the Case 2 sensitivity would be higher than under the
NRT4 final program. These higher emissions reductions were generated using our NONROAD
model as discussed in section 8A.2.2.  These emissions reductions are directly proportional to
the increased amount of fuel that would be consumed in these additional engines and, likewise,
to the increased fuel-related costs under this sensitivity. Using that direct relationship, we can
estimate the incremental fuel-related costs by calculating the ratio of fuel-related costs  under the
full engine and fuel program to the emissions reductions under the full engine and fuel  program
and then applying that ratio to the emissions reductions under the Case 2 sensitivity. Table
8A.3-8 presents the annual costs, the  costs by pollutant, and the emissions reductions of the Case
2 sensitivity. Note that costs have been allocated as done under the NRT4 full  engine and fuel
program (see Table 8.1-2). Note also that the emissions reductions  shown in Table 8A.3-8
include the higher reductions from gensets and the nonroad, locomotive, and marine reductions
that would occur under the full engine and fuel program.
                                         8-84

-------
                                                 Table 8A.3-8
Summary of Aggregate Costs, Costs by Pollutant, and Emissions Reductions Associated with the Case 2 Sensitivity ($2002)
Year
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2031
2032
2033
2034
2035
2036
30 Yr NPVat3%
30YrNPVat7%
Engine Costs
(Smillion)
-
$ 81
$ 82
$ 80
$ 423
$ 745
$ 906
$ 1 ,000
$ 989
$ 959
$ 947
$ 939
$ 928
$ 939
$ 953
$ 967
$ 981
$ 995
$ 1 ,009
$ 1 ,023
$ 1 ,037
$ 1 ,051
$ 1 ,065
$ 1 ,079
$ 1 ,093
$ 1,107
$ 1,121
$ 1,135
$ 1,149
$ 1,163
$ 14,628
$ 7,502
Equipment
Costs
(Smillion)
-
$ 5
$ 5
$ 5
$ 67
$ 111
$ 126
$ 151
$ 157
$ 158
$ 158
$ 154
$ 155
$ 155
$ 103
$ 70
$ 60
$ 39
$ 33
$ 33
$ 34
$ 34
$ 34
$ 35
$ 35
$ 36
$ 36
$ 37
$ 37
$ 38
$ 1 ,344
$ 791
Fuel Costs
(Smillion)
$ 161
$ 281
$ 286
$ 602
$ 811
$ 902
$ 984
$ 1,041
$ 1,103
$ 1,113
$ 1,124
$ 1,136
$ 1,149
$ 1,162
$ 1,177
$ 1,193
$ 1,210
$ 1 ,227
$ 1 ,245
$ 1 ,263
$ 1,281
$ 1 ,300
$ 1,318
$ 1 ,337
$ 1 ,356
$ 1 ,374
$ 1 ,393
$ 1,412
$ 1,431
$ 1 ,450
$ 18,772
$ 9,891
Other Operating
Costs ($million)
$ (181)
$ (31 8)
$ (324)
$ (375)
$ (41 5)
$ (383)
$ (352)
$ (349)
$ (339)
$ (31 9)
$ (302)
$ (287)
$ (274)
$ (262)
$ (254)
$ (247)
$ (241)
$ (237)
$ (233)
$ (229)
$ (227)
$ (226)
$ (225)
$ (225)
$ (225)
$ (226)
$ (227)
$ (228)
$ (229)
$ (230)
$ (5,236)
$ (3,198)
Net Operating
Costs ($million)
$ (21)
$ (37)
$ (38)
$ 228
$ 396
$ 520
$ 631
$ 691
$ 764
$ 793
$ 822
$ 849
$ 875
$ 900
$ 923
$ 946
$ 969
$ 991
$ 1,012
$ 1 ,034
$ 1 ,054
$ 1 ,074
$ 1 ,093
$ 1,112
$ 1,130
$ 1,149
$ 1,167
$ 1,185
$ 1 ,202
$ 1 ,220
$ 13,535
$ 6,693
Total Annual
Costs
(Smillion)
$ (21)
$ 49
$ 49
$ 312
$ 885
$ 1 ,376
$ 1 ,663
$ 1 ,842
$ 1,910
$ 1,911
$ 1 ,927
$ 1 ,941
$ 1 ,958
$ 1 ,994
$ 1 ,979
$ 1 ,984
$ 2,010
$ 2,025
$ 2,054
$ 2,090
$ 2,125
$ 2,159
$ 2,192
$ 2,226
$ 2,259
$ 2,291
$ 2,324
$ 2,356
$ 2,389
$ 2,421
$ 29,507
$ 14,986
PM Costs
(Smillion)
$ (7)
$ 73
$ 74
$ 179
$ 486
$ 803
$ 996
$ 987
$ 1 ,037
$ 1 ,052
$ 1 ,070
$ 1 ,075
$ 1,105
$ 1,130
$ 1,132
$ 1,140
$ 1,160
$ 1,184
$ 1,201
$ 1 ,224
$ 1 ,247
$ 1 ,268
$ 1 ,290
$ 1,311
$ 1,331
$ 1,351
$ 1 ,372
$ 1 ,392
$ 1,412
$ 1,431
$ 1 7,040
$ 8,635
NOx+NMHC
Costs
(Smillion)
-
$ 0
$ 0
$ 34
$ 267
$ 377
$ 421
$ 591
$ 596
$ 576
$ 571
$ 574
$ 556
$ 563
$ 542
$ 533
$ 535
$ 522
$ 529
$ 537
$ 545
$ 552
$ 560
$ 568
$ 576
$ 583
$ 591
$ 599
$ 606
$ 614
$ 7,988
$ 4,102
SOx Costs
(Smillion)
$ (14)
$ (25)
$ (26)
$ 99
$ 133
$ 195
$ 246
$ 264
$ 278
$ 282
$ 287
$ 292
$ 296
$ 301
$ 305
$ 310
$ 315
$ 319
$ 324
$ 329
$ 333
$ 338
$ 343
$ 347
$ 352
$ 357
$ 361
$ 366
$ 371
$ 375
$ 4,479
$ 2,248
PM
Reduction
(tons)
12,100
21,900
22,900
25,000
29,700
37,000
45,000
53,100
62,400
70,900
79,000
86,100
92,600
98,800
104,600
110,100
115,300
120,200
125,100
129,600
133,900
137,900
141,700
145,400
148,800
152,100
155,400
158,500
161,600
164,500
1 ,584,300
768,300
NOx+NMHC
Reduction
(tons)
-
200
300
600
45,600
100,500
152,300
215,900
281 ,600
345,100
405,600
461 ,200
514,100
563,900
609,600
652,200
692,700
730,500
766,600
800,400
831 ,200
860,200
887,700
913,100
936,700
959,800
982,000
1 ,003,700
1 ,024,500
1 ,044,800
8,662,500
3,900,200
SOx
Reduction
(tons)
149,900
265,500
270,900
289,300
304,900
315,000
323,800
330,700
337,600
343,500
349,500
355,600
361,800
367,700
373,600
379,800
385,800
391,900
398,000
404,000
410,100
416,100
422,200
428,300
434,400
440,500
446,700
452,800
458,900
465,100
6,511,100
3,593,900

-------
Final Regulatory Impact Analysis
    The calculations of cost per ton of each emission reduced under the Case 2 sensitivity divides
the net present value of the annual costs assigned to each pollutant (Table 8 A. 3-8) by the net
present value of the total annual reductions of each pollutant (Table 8A.3-8).  The 30-year net
present value of the costs associated with each pollutant, calculated with a three percent discount
rate, are shown in Table 8A.3-8 as $8.0 billion for NOx+NMHC, $17.0 billion for PM, and $4.5
billion for SOx, with the total cost of the program estimated at $29.5 billion. The 30-year net
present value, with a three percent discount rate, of emission reductions are estimated at 8.7
million tons of NOx+NMHC, 1.6 million tons of PM, and 6.5 million tons of SOx (see Table
8A.3-8). How these emissions reductions were developed is described in Section 8A.2 (see
Table 8A.2-6)/  The results of the cost per ton calculations are shown in Table 8A.3-9.

                                        Table 8A. 3-9
                       Aggregate Cost per Ton for the Case 2 Sensitivity
               30-year Net Present Values at a 3% and 7% Discount Rate ($2002)
Item
Cost per Ton NOx+NMHC
Cost per Ton PM
Cost per Ton SOx
3% discount rate
$920
$10,800
$690
7% discount rate
$1,050
$11,200
$630
Source
Calculated
Calculated
Calculated
       We have also calculated the cost per ton of emissions reduced in the year 2030 using the
annual costs and emission reductions in that year alone.  This number, shown in Table 8A.3-10,
approaches the long-term cost per ton of emissions reduced.

                                        Table 8A.3-10
                      Long-Term Cost per Ton of the Case 2 Sensitivity
                          Annual Values without Discounting ($2002)
Pollutant
NOx+NMHC
PM
SOx
Long-Term Cost per Ton
in 2030
$620
$9,000
$810
     Note that the emissions reductions shown in Table 8A.3-8 are not identical to the reductions one would get using
the inventories presented in Table 8A.2-6. The emissions inventories in Table 8A.2-6 are for land based nonroad engines
only and do not include emissions associated with locomotive and marine engines. To make the comparison between the
Case 2 $/ton and the NRT4 full program $/ton, the Case 2 locomotive and marine emissions reductions are included with
the Case 2 nonroad land based emissions reductions. The Case 2 locomotive and marine emissions reductions would be
identical to those under the NRT4 full program since nothing about the Case 2 sensitivity would impact emissions
reductions from locomotive and marine engines. Therefore, the NRT4 full program locomotive and marine emissions
reductions have been added to the Case 2 nonroad land based emissions reductions to arrive at the Case 2 emissions
reductions shown in Table 8A. 3-8.
                                             8-86

-------
                                                   Aggregate Cost and Cost per Ton
8A.4 Summary of the Sensitivity Analyses Results

   We present here a summary of the results of the Case 1 and Case 2 sensitivity analyses, and
we compare these results to the NRT4 full engine and fuel program (i.e., the NRT4 Final Rule).

   Table 8A.4-1 shows the emission reduction comparison between the NRT4 full program and
the sensitivity cases for PM and NOx. As can be seen, the Case 1 sensitivity results in a
decrease in both PM and NOx emissions reductions on the order of 35 to 40 percent. The Case 2
sensitivity results in an increase in PM reductions on the order of 10 percent and an increase in
NOx reductions on the order of 20 percent.

                                     Table 8A.4-1
        Emissions Reduction* Comparison for Case 1 and Case 2 Sensitivity Analyses
                     30 Year Net Present Values at a 3% Discount Rate
Baseline/Control Scenario
Nonroad Tier 4 Final Rule
Case 1 Sensitivity Analysis
Case 2 Sensitivity Analysis
NOx+NMHC
(tons)
7,077,900
4,421,600
8,662,500
Percent
Relative to
NRT4FRM
-
-38%
22%
PM
(tons)
1,430,500
919,400
1,584,300
Percent
Relative to
NRT4FRM
-
-36%
11%
       See Tables 8.6-1, 8A.3-5, and 8A.3-8, respectively.
   Table 8A.4-2 summarizes the results of the two sensitivity cases with respect to cost-
effectiveness for NMHC+NOx, PM, and SOx, and compares these values to the final NRT4
program. As can be seen, the Case 1 sensitivity analysis results in an increase in the $/ton
estimates for all pollutants. However, in all cases, these estimates are still within the range of
previous mobile source control programs for NMHC+NOx and PM, and for SOx on the same
order as stationary control programs for acid rain (see Tables VI.D-3, -4, and -5 of the preamble
for this final rule). For the Case 2 sensitivity analysis, Table 8A.4-2 shows that the cost-
effectiveness for NOx+NMHC and for PM are lower than for the final Tier 4 program, and for
SOx the cost-effectiveness is the same as for the final Tier 4 program.
                                         8-87

-------
Final Regulatory Impact Analysis
                                    Table 8A.4-2
    Comparison of Aggregate Cost per Ton Estimates: NRT4 Final Rule, Case 1 Sensitivity
                        Analysis, and Case 2 Sensitivity Analysis
              30-year Net Present Values at a 3 percent Discount Rate ($2002)
Pollutant
NOx+NMHC ($/ton)
PM ($/ton)
SOx ($/ton)
Nonroad Tier 4
Final Rule
$1,010
$11,200
$690
Casel
Sensitivity
Analysis
$1,490
$15,800
$870
Case 2
Sensitivity
Analysis
$920
$10,800
$690

-------
                                                 Aggregate Cost and Cost per Ton
   Appendix 8B: Fuel Volumes used throughout Chapter 8

   The volumes in this Appendix were developed from the information contained in Section 7.1
of Chapter 7 of the RIA. Demand volumes are estimated for each EPA use category, including
nonroad, locomotive, marine, and highway diesel fuel, and heating oil, for 2014. The 2014
estimated volumes of pipeline downgrade and highway diesel fuel spillover are apportioned to
various EPA use categories depending on the regulatory scenario. By default, this analysis
estimates the volume of fuel which must be desulfurized for supplying the overall demand of
each EPA use category.  The regulatory scenarios modeled for their volumes for this Chapter
include the Final Rule Program and several sensitivity cases which are summarized here. For
each case, the table which summarizes the 2014 volumes is listed along with the case
description.

Final Rule Program:
   Period from 2007 to 2010 - NRLM must meet a 500 ppm sulfur cap standard. Small refiners
   are exempted and are assumed to produce high sulfur distillate and sell that fuel into the
   NRLM diesel fuel pool (Table 7.1.3-14).

-  Period from 2010 to 2012 - nonroad must meet a 15 ppm sulfur cap standard and locomotive
   and marine must meet a 500 ppm sulfur cap. Small refiners are exempted and can sell
   exempted fuel into the nonroad diesel fuel pool, except for most of PADD 1, providing that
   they produce 500 ppm fuel (Table 7.1.3-17).

   Period from 2012 to 2014 - NRLM must meet a 15 ppm sulfur cap standard.  Small refiners
   are exempted and can sell exempted fuel into the NRLM diesel fuel pool,  except for most of
   PADD 1, providing that they produce 500 ppm fuel (Table 7.1.3-19).

   Period from 2014 and thereafter - The small refiner provisions have expired (Table 7.1.3-20).

NRLM to 500 ppm only:
-  Period from 2007 to 2010 - Same as Final Rule Program above for the period 2007 to 2010
   (Table 7.1.3-14).

-  Period after 2010 - NRLM fuel remains at 500 ppm (Table 7.1.4-1).

Final Rule Program, EIA nonroad volumes:
   Same as Final Rule  Program  except that the nonroad volumes were developed using EIA
   information instead  of using NONROAD (Tables 7.1.4-10, 7.1.4-11, 7.1.4-12, and 7.1.4-13 ).

   All the volume streams in each case were apportioned into specific families of similar fuels
depending on the quality of the specific volume stream and whether it was regulated under the
NRLM Program.  These fuel families and the streams which comprise them are summarized in
the following table.
                                        8-89

-------
Final Regulatory Impact Analysis
                        Fuel Families and the Fuels They Represent*
Fuel Family
High Sulfur

High Sulfur
(3000 ppm)
distillate fuel
including
NRLMand
heating oil

Small refiner
fuel from
2007 to 20 10

Old 500 ppm

500 ppm
diesel fuel
meeting the
Highway
Diesel Fuel
Program
requirements




New 500
ppm
500 ppm
diesel fuel
meeting the
Nonroad
Diesel Fuel
Program
requirements

Small refiner
fuel from
2010 to 2014
Old 15 ppm

15 ppm
diesel fuel
meeting the
Highway
Diesel Fuel
Program
requirements




Reprocessed
Downgrade
Oversupply of
downgrade
into a market
which must be
reprocessed to
15 ppm





New 15 ppm

15 ppm diesel
fuel meeting
the Nonroad
Diesel Fuel
Program
requirements





Total
Volume
Total of
these
various
volumes.







* California gallons are not included. "Affected" 500 ppm gallons are labeled here as "New 500 ppm" and "Affected" 15
ppm gallons are the summation of the columns labeled "Reprocessed Downgrade" and "New 15 ppm."

The 2014 volumes are adjusted to estimate the volumes in each year from 2007 to 2040 using
growth ratios compared to 2014 based on the growth rate factors in Tables 7.1.5-1 and 7.1.5-2.

    Analyzing and categorizing the volumes in this fashion resulted in the development of the
input volumes used in this chapter. For a more complete summary of how the volumes were
calculated consult Section 7.1 of Chapter 7 of the RIA. The following tables summarize this
information.
                                          8-90

-------
                                    Aggregate Cost and Cost per Ton
                        Table 8B-1
Nationwide Nonroad Volumes Under the NRT4 Final Rule Fuel Program
Year
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2031
2032
2033
2034
2035
2036
High Sulfur
(million gallons)
4,027
584
597
255
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
Old 500 ppm
(million gallons)
239
179
183
673
1,043
1,066
1,088
463
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
New 500 ppm
(million gallons)
4,790
8,406
8,599
4,014
614
528
468
199
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
Old 15 ppm
(million gallons)
2,369
2,526
2,585
2,643
2,701
2,760
2,818
2,941
3,047
3,107
3,167
3,227
3,288
3,352
3,408
3,468
3,528
3,588
3,648
3,708
3,767
3,827
3,887
3,946
4,006
4,066
4,125
4,185
4,245
4,304
Reprocessed
Downgrade
(million gallons)
-
-
-
-
-
-
-
68
118
121
123
125
127
130
132
134
137
139
141
144
146
148
151
153
155
158
160
162
165
167
New 15 ppm
(million gallons)
-
-
-
6,189
8,145
8,420
8,671
9,645
10,421
10,626
10,832
1 1 ,037
1 1 ,243
1 1 ,448
1 1 ,654
1 1 ,859
12,064
12,270
1 2,475
12,679
12,883
13,088
13,292
1 3,496
13,700
13,904
14,108
14,312
14,516
1 4,720
Total Volume
(million gallons)
1 1 ,426
1 1 ,695
1 1 ,964
12,233
12,504
12,774
13,045
13,316
13,586
13,854
14,122
14,390
14,658
14,926
15,193
15,461
15,729
15,997
16,265
16,531
16,797
17,063
17,329
17,595
17,861
18,127
18,393
18,659
18,925
19,191
                           8-91

-------
Final Regulatory Impact Analysis
                                  Table 8B-2
        Nationwide Locomotive Volumes Under the NRT4 Final Rule Fuel Program
Year
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2031
2032
2033
2034
2035
2036
High Sulfur
(million gallons)
968
138
140
59
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
Old 500 ppm
(million gallons)
45
40
40
356
591
410
278
849
1,266
1,279
1,291
1,301
1,313
1,322
1,329
1,341
1,353
1,365
1,378
1,389
1,400
1,411
1,422
1,433
1,444
1,455
1,466
1,477
1,488
1,500
New 500 ppm
(million gallons)
1,141
1,978
2,005
1,805
1,671
761
99
42
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
Old 15 ppm
(million gallons)
539
568
576
585
596
602
607
589
577
583
589
593
599
603
606
611
617
622
628
633
638
643
648
653
658
663
668
674
679
684
Reprocessed
Downgrade
(million gallons)
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
New 15 ppm
(million gallons)
-
-
-
-
-
1,114
1,925
1,476
1,154
1,166
1,177
1,186
1,197
1,205
1,212
1,222
1,233
1,244
1,256
1,266
1,276
1,286
1,296
1,306
1,316
1,327
1,337
1,347
1,357
1,367
Total Volume
(million gallons)
2,694
2,724
2,761
2,805
2,858
2,841
2,909
2,932
2,956
2,988
3,015
3,038
3,067
3,089
3,104
3,132
3,160
3,187
3,218
3,244
3,270
3,295
3,321
3,347
3,373
3,399
3,425
3,450
3,476
3,502
                                     8-92

-------
                                   Aggregate Cost and Cost per Ton
                        Table 8B-3
Nationwide Marine Volumes Under the NRT4 Final Rule Fuel Program
Year
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2031
2032
2033
2034
2035
2036
High Sulfur
(million gallons)
806
190
192
81
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
Old 500 ppm
(million gallons)
21
23
23
269
452
292
175
222
257
259
262
266
271
202
152
154
156
158
160
162
164
166
168
170
172
175
177
179
181
183
New 500 ppm
(million gallons)
849
1,476
1,494
1,380
1,304
636
148
62
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
Old 15 ppm
(million gallons)
252
266
269
273
277
280
283
281
280
283
286
290
295
299
302
307
311
315
319
323
327
331
335
339
343
347
352
356
360
364
Reprocessed
Downgrade
(million gallons)
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
New 15 ppm
(million gallons)
-
-
-
0
0
851
1,472
1,605
1,706
1,722
1,741
1,768
1,798
1,818
1,841
1,871
1,892
1,917
1,939
1,964
1,989
2,014
2,039
2,065
2,090
2,115
2,140
2,165
2,190
2,215
Total Volume
(million gallons)
1,929
1,955
1,979
2,003
2,033
2,059
2,078
2,103
2,126
2,146
2,170
2,203
2,240
2,266
2,294
2,331
2,357
2,389
2,417
2,448
2,479
2,510
2,542
2,573
2,604
2,635
2,667
2,698
2,729
2,760
                           8-93

-------
Final Regulatory Impact Analysis
                                  Table 8B-4
            Nationwide Nonroad Volumes Under the SOOpptn NRLM Scenario
Year
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2031
2032
2033
2034
2035
2036
High Sulfur
(million gallons)
4,027
584
597
255
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
Old 500 ppm (million
gallons)
239
179
183
936
1,503
1,535
1,568
1,600
1,633
1,665
1,697
1,729
1,762
1,794
1,826
1,858
1,890
1,923
1,955
1,987
2,019
2,051
2,083
2,115
2,147
2,179
2,210
2,242
2,274
2,306
New 500 ppm
(million gallons)
4,790
8,406
8,599
8,400
8,300
8,479
8,659
8,839
9,018
9,196
9,374
9,552
9,730
9,907
10,085
10,263
10,441
10,619
10,797
10,973
11,150
1 1 ,326
1 1 ,503
1 1 ,679
1 1 ,856
12,032
12,209
12,386
12,562
12,739
Old 15 ppm
(million gallons)
2,369
2,526
2,585
2,643
2,701
2,760
2,818
2,877
2,935
2,993
3,051
3,109
3,166
3,224
3,282
3,340
3,398
3,456
3,514
3,571
3,629
3,686
3,744
3,801
3,859
3,916
3,973
4,031
4,088
4,146
Reprocessed
Downgrade
(million gallons)
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
New 15 ppm
(million gallons)
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
Total Volume
(million gallons)
1 1 ,426
1 1 ,695
1 1 ,964
12,233
12,504
12,774
13,045
13,316
13,586
13,854
14,122
14,390
14,658
14,926
15,193
15,461
15,729
15,997
16,265
16,531
16,797
17,063
17,329
17,595
17,861
18,127
18,393
18,659
18,925
19,191
                                     8-94

-------
                                  Aggregate Cost and Cost per Ton
                       Table 8B-5
Nationwide Locomotive Volumes Under the SOOpptn NRLM Scenario
Year
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2031
2032
2033
2034
2035
2036
High Sulfur
(million gallons)
968
138
140
59
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
Old 500 ppm
(million gallons)
45
40
40
211
339
342
345
347
350
354
357
360
364
366
368
371
374
378
381
385
388
391
394
398
401
404
408
411
415
418
New 500 ppm
(million gallons)
1,141
1,978
2,005
1,950
1,923
1,942
1,958
1,973
1,990
2,011
2,029
2,044
2,064
2,079
2,089
2,108
2,127
2,145
2,166
2,184
2,202
2,220
2,239
2,258
2,277
2,296
2,315
2,334
2,354
2,374
Old 15 ppm
(million gallons)
539
568
576
585
596
602
607
611
616
623
629
633
640
644
647
653
659
664
671
677
682
688
694
699
705
711
717
723
729
735
Reprocessed
Downgrade
(million gallons)
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
New 15 ppm
(million gallons)
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
Total Volume
(million gallons)
2,694
2,724
2,761
2,805
2,858
2,886
2,909
2,932
2,956
2,988
3,015
3,038
3,067
3,089
3,104
3,132
3,160
3,187
3,218
3,244
3,270
3,295
3,321
3,347
3,373
3,399
3,425
3,450
3,476
3,502
                          8-95

-------
Final Regulatory Impact Analysis
Table 8B-6
Nationwide Marine Volumes Under the 500
Year
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2031
2032
2033
2034
2035
2036
High Sulfur
(million gallons)
806
190
192
81
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
Old 500 ppm
(million gallons)
21
23
23
142
229
232
234
237
240
242
245
249
253
256
259
263
266
270
273
276
280
283
287
290
294
297
301
305
308
312
New 500 ppm
(million gallons)
849
1,476
1,494
1,508
1,527
1,546
1,560
1,579
1,597
1,612
1,630
1,654
1,682
1,702
1,723
1,751
1,770
1,794
1,815
1,838
1,862
1,885
1,909
1,932
1,956
1,979
2,003
2,026
2,050
2,073
Old 15 ppm
(million gallons)
252
266
269
273
277
280
283
286
289
292
295
300
305
309
312
317
321
325
329
333
337
342
346
350
355
359
363
367
372
376
)pm NRLM Scenario
Reprocessed
Downgrade
(million gallons)
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
New 15 ppm
(million gallons)
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
Total Volume
(million gallons)
1,929
1,955
1,979
2,003
2,033
2,059
2,078
2,103
2,126
2,146
2,170
2,203
2,240
2,266
2,294
2,331
2,357
2,389
2,417
2,448
2,479
2,510
2,542
2,573
2,604
2,635
2,667
2,698
2,729
2,760
                                    8-96

-------
                               Aggregate Cost and Cost per Ton
                   Table 8B-7
Nationwide Nonroad Volumes Under the Case 1 Sensitivity
Year
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2031
2032
2033
2034
2035
2036
High Sulfur
(million gallons)
2,996
592
600
253
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
Old 500 ppm
(million gallons)
444
153
155
66
-
358
621
262
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
New 500 ppm
(million gallons)
3,671
6,373
6,454
3,086
631
531
460
194
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
Old 15 ppm
(million gallons)
1,959
2,067
2,093
2,141
2,183
2,188
2,198
2,275
2,338
2,366
2,393
2,420
2,447
2,474
2,502
2,529
2,556
2,583
2,610
2,638
2,665
2,692
2,719
2,746
2,774
2,801
2,828
2,855
2,882
2,910
Reprocessed
Downgrade
(million gallons)
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
New 15 ppm
(million gallons)
-
-
-
3,873
6,721
6,574
6,488
7,153
7,662
7,751
7,840
7,929
8,018
8,107
8,196
8,285
8,374
8,464
8,553
8,642
8,731
8,820
8,909
8,998
9,087
9,176
9,265
9,354
9,444
9,533
Total Volume
(million gallons)
9,070
9,186
9,302
9,419
9,535
9,651
9,767
9,884
10,000
10,116
10,233
10,349
10,465
10,581
10,698
10,814
10,930
1 1 ,047
11,163
1 1 ,279
1 1 ,395
11,512
1 1 ,628
1 1 ,744
11,861
1 1 ,977
12,093
12,210
12,326
12,442
                      8-97

-------
Final Regulatory Impact Analysis
                                   Table 8B-8
              Nationwide Locomotive Volumes Under the Case 1 Sensitivity
Year
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2031
2032
2033
2034
2035
2036
High Sulfur
(million gallons)
910
162
164
70
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
Old 500 ppm
(million gallons)
116
37
38
426
715
410
188
455
651
658
664
669
675
680
684
690
696
702
709
714
720
726
731
737
743
748
754
760
766
771
New 500 ppm
(million gallons)
1,129
1,957
1,983
1,741
1,575
732
120
50
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
Old 15 ppm
(million gallons)
539
568
576
569
568
590
607
584
570
576
581
586
591
595
599
604
609
614
620
625
630
635
640
645
650
655
660
665
670
675
Reprocessed
Downgrade
(million gallons)
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
New 15 ppm
(million gallons)
-
-
-
-
-
1,155
1,995
1,842
1,735
1,754
1,770
1,783
1,801
1,813
1,822
1,838
1,855
1,871
1,889
1,904
1,919
1,934
1,950
1,965
1,980
1,995
2,010
2,025
2,041
2,056
Total Volume
(million gallons)
2,694
2,724
2,761
2,805
2,858
2,886
2,909
2,932
2,956
2,988
3,015
3,038
3,067
3,089
3,104
3,132
3,160
3,187
3,218
3,244
3,270
3,295
3,321
3,347
3,373
3,399
3,425
3,450
3,476
3,502
                                     8-98

-------
                              Aggregate Cost and Cost per Ton
                   Table 8B-9
Nationwide Marine Volumes Under the Case 1 Sensitivity
Year
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2031
2032
2033
2034
2035
2036
High Sulfur
(million gallons)
757
186
188
79
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
Old 500 ppm
(million gallons)
67
22
22
326
551
309
133
136
140
141
143
145
147
149
151
153
155
157
159
161
163
165
167
169
171
173
175
177
179
181
New 500 ppm
(million gallons)
853
1,482
1,499
1,328
1,210
511
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
Old 15 ppm
(million gallons)
252
266
269
269
271
278
283
281
280
283
286
290
295
299
302
307
311
315
319
323
327
331
335
339
343
347
352
356
360
364
Reprocessed
Downgrade
(million gallons)
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
New 15 ppm
(million gallons)
-
-
-
-
-
961
1,662
1,685
1,706
1,722
1,742
1,768
1,798
1,819
1,841
1,871
1,892
1,917
1,939
1,964
1,989
2,015
2,040
2,065
2,090
2,115
2,140
2,165
2,190
2,215
Total Volume
(million gallons)
1,929
1,955
1,979
2,003
2,033
2,059
2,078
2,103
2,126
2,146
2,170
2,203
2,240
2,266
2,294
2,331
2,357
2,389
2,417
2,448
2,479
2,510
2,542
2,573
2,604
2,635
2,667
2,698
2,729
2,760
                      8-99

-------
Final Regulatory Impact Analysis
Chapter 8 References

1.  "Guidelines for Preparing Economic Analyses," U.S. Environmental Protection Agency, EPA
240-R-00-003, September 2000.

2. Power Systems Research, OELink Sales Version, 2002.

3. Nonroad Engine Growth Estimate, Report No. NR-008b, Docket Item II-A-32.

4. "Engine Sales Used in Proposed Nonroad Tier 4 Cost Analysis," memorandum from Todd
Sherwood to Public Docket No. A-2001-28, Docket Item II-B-37.

5. Dickson, Cheryl, "Heating Oils, 2001," TRW Petroleum Technologies, TRW-221 PPS, 01/4,
July 2001.

6. Dickson, Cheryl, "Heating Oils, 2002," TRW Petroleum Technologies, TRW-226 PPS,
2002/4, September, 2002.

7.  Batey, John E. and Roger McDonald, "Advantages of Low Sulfur Home Heating Oil - Interim
Report of Complied Research,  Studies, and Data Resources, " National Oilheat Research
Alliance and the Department of Energy, December 2002.
                                       8-100

-------
CHAPTER 9: Cost-Benefit Analysis
   9.1 Time Path of Emission Changes for the Final Standards  	9-8
   9.2 Development of Benefits Scaling Factors Based on Differences in Emission Impacts
       Between the Final Standards and Modeled Preliminary Control Options	9-11
   9.3 Summary of Modeled Benefits and Apportionment Method  	9-12
       9.3.1  Overview of Analytical Approach  	9-16
       9.3.2  Air Quality Modeling	9-17
          9.3.2.1 PM Air Quality Modeling with REMSAD  	9-17
          9.3.2.2 Ozone Air Quality Modeling with CAMx	9-18
       9.3.3  Health Impact Functions	9-19
       9.3.4  Economic Values for Health Outcomes	9-23
       9.3.5  Welfare Effects	9-24
          9.3.5.1 Visibility Benefits  	9-24
          9.3.5.2 Agricultural Benefits	9-25
          9.3.5.3 Other Welfare Benefits  	9-26
       9.3.6  Treatment of Uncertainty  	9-28
       9.3.7  Model Results	9-29
       9.3.8  Apportionment of Benefits to NOx, SO2, and Direct PM Emissions Reductions 9-39
   9.4 Estimated Benefits of Final Nonroad Diesel Engine Standards  in 2020 and 2030 . . . 9-42
   9.5 Development of Intertemporal Scaling Factors and Calculation of Benefits Over Time
        	9-48
   9.6 Comparison of Costs  and Benefits	9-53

   APPENDIX 9A: Benefits Analysis of Modeled Preliminary Control Option	9-77
   APPENDIX 9B: Supplemental Analyses Addressing Uncertainties in the Concentration-
       Response and Valuation Functions  for Particulate Matter Health Effects  	9-205
   APPENDIX 9C: Sensitivity Analyses of Key Parameters in the Benefits Analysis  .... 9-250
   APPENDIX 9D:  Visibility Benefits Estimates for Individual Class I Areas	9-269

-------
                                                                  Cost-Benefit Analysis
                  CHAPTER 9: Cost-Benefit Analysis
    This chapter reports EPA's analysis of the public health and welfare impacts and associated
monetized benefits to society of the final Nonroad Diesel Engines Tier 4 Standards. EPA is
required by Executive Order 12866 to estimate the costs and benefits of major new pollution
control regulations. Accordingly, the analysis presented here attempts to answer three questions:
(1) what are the physical health and welfare effects of changes in ambient air quality resulting
from reductions in nitrogen oxides (NOx), sulfur dioxide (SO2), non-methane hydrocarbons
(NMHC), carbon monoxide (CO) and direct diesel particulate matter (PM2 5)A emissions?; (2)
how much are the changes in these effects attributable to the final rule worth to U.S. citizens as a
whole in monetary terms?; and (3) how do the monetized benefits compare to the costs over
time? It constitutes one part of EPA's thorough examination of the relative merits of this
regulation. In Chapter 12, of the Draft RIA, we provided an analysis of the benefits of several
alternatives to the selected standards  to examine their relative benefits and costs for public
comment.

    For the final  rulemaking, we rely on the air quality modeling conducted for the proposed
rule, documented in the Regulatory Impact Analysis (U.S. EPA, 2003a), available at
http ://www. epa. gov/nonroad.B  To estimate the benefits of the final rule, we use a set of scaling
factors which separately estimate a set of emission reduction profiles for NOx,  SO2, and directly
emitted diesel PM2 5.  For this analysis of the final rule, we conduct a benefits transfer analysis
using those same scaling factors, applied to the updated results of the modeled preliminary
control option which accounts for changes in the health benefits methodology adopted during the
recent proposed Interstate Air Quality Rule (IAQR) analysis.0  These methodological changes
are reflected both in the detailed estimates for 2020 and 2030 and in the time stream of total
monetized benefits. The methodological  changes are summarized in this chapter and described
in detail in Appendix  9A.

    EPA has used the  best available information and tools of analysis to quantify the expected
changes in public health, environmental and economic benefits for the modeled option.  We
   AEmissions from nonroad diesel engines include directly emitted fine particles (carbon and sulfates) as well as
gaseous pollutants that react in the atmosphere to form fine particles. This final rule will results in reductions in
ambient PM particle levels due to reductions in both directly emitted particles as well as reductions in PM precursor
emissions, including NOx and SO2.

   BAs discussed in Chapter 2, because of the long lead times to conduct complex photochemical air quality
modeling at the national scale, decisions must be made early in the process about the scenarios to be modeled.
Based on updated information and public comment, EPA has made changes to the final control program, which
results in changes in emissions as detailed in Chapter 3, section 3.6.

   GNote that the methodology for estimating visibility benefits is unchanged from proposal. The documents
related to the IAQR can be found at OAR Docket number 2003-0053.

                                            9-1

-------
Final Regulatory Impact Analysis
summarize the results of that analysis in section 9.3, and present details in Appendix 9A, directly
following this chapter.  The standards we are finalizing in this rulemaking are slightly different
in the amount of emission reductions expected to be achieved in 2020 and 2030 relative to both
the proposed standards and the preliminary modeled option. As such, we determined that
benefits would need to be scaled to reflect the differences in emission reductions between the
modeled and final standards. The results of that scaling analysis are the focus of this chapter.

   In order to characterize the benefits attributable to the Nonroad Diesel Engines standards,
given the constraints on time and resources available for the analysis, we  use a benefits transfer
method to scale the benefits of the modeled preliminary control options to reflect the differences
in emission reductions.  We also apply intertemporal scaling factors to examine the stream of
benefits over the rule implementation period.  The benefits transfer method used to estimate
benefits for the final rule is similar to that used to estimate benefits in the recent analysis of the
Large Si/Recreational Vehicles standards (see U.S. EPA 2002, Docket A-2000-01, Document V-
B-4). A similar method has also been used in recent benefits analyses for the proposed Clean
Air Act Section 112 Utility Mercury Emission Reduction  rule, the proposed Industrial Boilers
and Process Heaters National Emissions Standards for Hazardous Air Pollutants (NESHAP)
standards (Docket numbers OAR-2003-A-96-47) and the Reciprocating Internal Combustion
Engines NESHAP standards (Docket numbers OAR-2002-0059 and A-95-35). One significant
limitation to this method is the inability to scale ozone-related benefits. Because ozone is a
homogeneous gaseous pollutant formed through complex  atmospheric photochemical processes,
it is not possible to apportion ozone benefits to the precursor emissions of NOx and VOC.
Coupled with the potential for NOx reductions to either increase or decrease ambient ozone
levels, this prevents us from scaling the benefits associated with a particular combination of
VOC and NOx emissions reductions to another (a more detailed discussion is provided below).
Because of our inability to scale ozone benefits, we provide the ozone benefits results for the
modeled preliminary control options as a referent, but do not include ozone benefits as part of
the monetized benefits of the standards.  For the most part, quantifiable ozone benefits do not
contribute significantly  to the monetized benefits: thus, their omission will not materially affect
the conclusions of the benefits analysis.

   Table 9-1 lists the known quantifiable and unquantifiable effects considered for this analysis.
We quantify benefits for the contiguous 48 states. Note that this table categorizes ozone-related
benefits as unquantified effects.  Furthermore, we quantify benefits for the contiguous 48 states.
We have quantified ozone-related benefits in our modeling of the preliminary control option,
summarized in Section 9.3 and detailed in Appendix 9A.  However, as noted above, we are
unable to quantify ozone-related benefits for the final standards. It is important to note that there
are significant categories of benefits which can not be monetized (or in many cases even
quantified), resulting in a significant limitation to this analysis.  Also, EPA currently does not
have appropriate tools for modeling changes in ambient concentrations of CO or air toxics input
into a national benefits analysis. Although these pollutants have been linked to numerous
adverse health effects, we are unable to quantify the CO- or air toxics-related health or welfare
benefits of the final rule at this time. We also omitted the significant SO2 reductions from lower
sulfur in home heating oil in the Northeast.
                                          9-2

-------
                                                                 Cost-Benefit Analysis
   The benefit analysis that we performed for our rule can be thought of as having seven parts,
each of which will be discussed separately in the Sections that follow. These seven steps include
the following:

   1.  Identification of final standards and calculation of the impact that the standards will have
       on the nationwide inventories for NOx, non-methane hydrocarbons (NMHC), SO2, and
       PM emissions throughout the rule implementation period;
   2.  Calculation of scaling factors relating emissions changes resulting from the final
       standards to emissions changes from a set of preliminary control options that were used
       to model air quality and benefits (see Appendix 9A for full details).
   3.  Apportionment of modeled benefits of preliminary control options to NOx, SO2, and
       diesel PM emissions (see Appendix 9A for a complete discussion of the modeling of the
       benefits for the preliminary set of standards, including updates in the benefits
       methodology since the time of proposal).
   4.  Application of scaling factors to apportioned modeled  benefits associated with NOx, S02,
       and PM in 2020 and 2030.
   5.  Development of intertemporal scaling factors based on 2020 and 2030 modeled air
       quality and benefits results.
   6.  Application of intertemporal  scaling factors to the yearly emission changes expected to
       result from the standards from 2010 through 2030 to obtain yearly monetized benefits.
   7.  Calculation of present value of stream of benefits.

   This analysis presents estimates of the potential benefits from the final Nonroad Diesel
Engine rule occurring in future years. The predicted emissions reductions that will result from
the rule have yet to occur, and therefore the actual changes in  human health and welfare
outcomes to which economic values  are ascribed are predictions. These predictions are based on
the best available  scientific evidence and judgment, but there is unavoidable uncertainty
associated with each step in the complex process between regulation and specific health and
welfare outcomes. Uncertainties associated with projecting input and parameter values into the
future may contribute significantly to the overall uncertainty in the benefits estimates. However,
we make these projections to more completely examine the impact of the program as the
equipment fleet turns over.

   In general, the chapter is organized around the seven steps laid out above.  In Section 1, we
identify the potential standard to analyze, establish the timeframe over which benefits are
estimated, and summarize emissions impacts. In Section 2, we summarize the changes in
emissions that were used in the preliminary modeled benefits analysis and develop the ratios  of
the emissions reductions under the final standards to preliminary emissions reductions that are
used to scale modeled benefits. In Section 3, we summarize the modeled benefits associated with
the emissions changes for the preliminary control options and  apportion those benefits to the
individual emission species (NOx, SO2, and PM2 5).  In Section 4, we estimate the benefits in
2020 and 2030 for the final standards, based on scaling of the  modeled benefits of the
preliminary control options.  In Section 5, we develop  intertermporal scaling factors based on the
ratios of yearly emission changes to the emission changes in 2020 and 2030 and estimate yearly
benefits of the final standards, based on scaling of the benefits in 2020 and 2030. Finally, in

                                           9-3

-------
Final Regulatory Impact Analysis
Section 6, we compare the estimated streams of benefits and costs over the full implementation
period, 2007 to 2030, to calculate the present value of net benefits for the final standards.
                                         9-4

-------
                                     Table 9-1
Health and Welfare Effects of Pollutants Affected by the Final Nonroad Diesel Engine Rule
Pollutant/Effect
PM/Health
PM/Welfare
Quantified and Monetized Effects in
Primary Analysis
Premature mortality in adults
Infant mortality
Bronchitis - chronic and acute
Hospital admissions - respiratory
and cardiovascular
Emergency room visits for asthma
Non-fatal heart attacks (myocardial
infarction)
Asthma exacerbations (asthmatic
population)
Lower and upper respiratory illness
Respiratory symptoms (asthmatic
population)
Minor restricted activity days
Work loss days
Visibility in California,
Southwestern, and Southeastern
Class I areas
Quantified and/or
Monetized Effects in
Sensitivity Analyses


Unquantified Effects
Low birth weight
Changes in pulmonary function
Chronic respiratory diseases other than chronic bronchitis
Morphological changes
Altered host defense mechanisms
Non-asthma respiratory emergency room visits
PM reductions associated with reductions in sulfur in home heating oil
Visibility in Northeastern, Northwestern, and Midwestern Class I areas
Visibility in residential and non-Class I areas
Household soiling
Sulfate PM reductions associated with reductions in sulfur in home
heating oil
                                        9-5

-------
Pollutant/Effect
Ozone/Health
Ozone/Welfare
Nitrogen and
Sulfate
Deposition/
Welfare
SO2/Health
Quantified and Monetized Effects in
Primary Analysis




Quantified and/or
Monetized Effects in
Sensitivity Analyses




Unquantified Effects
Increased airway responsiveness to stimuli
Inflammation in the lung
Chronic respiratory damage
Premature aging of the lungs
Acute inflammation and respiratory cell damage
Increased susceptibility to respiratory infection
Non-asthma respiratory emergency room visits
Hospital admissions - respiratory
Emergency room visits for asthma
Minor restricted activity days
School loss days
Chronic Asthma2
Asthma attacks
Cardiovascular emergency room visits
Premature mortality - acute exposures'1
Acute respiratory symptoms
Decreased commercial forest productivity
Decreased yields for fruits and vegetables
Decreased yields for commercial and non-commercial crops
Damage to urban ornamental plants
Impacts on recreational demand from damaged forest aesthetics
Damage to ecosystem functions
Decreased outdoor worker productivity
Costs of nitrogen controls to reduce eutrophication in selected eastern
estuaries
Impacts of acidic sulfate and nitrate deposition on commercial forests
Impacts of acidic deposition on commercial freshwater fishing
Impacts of acidic deposition on recreation in terrestrial ecosystems
Impacts of nitrogen deposition on commercial fishing, agriculture, and
forests
Impacts of nitrogen deposition on recreation in estuarine ecosystems
Reduced existence values for currently healthy ecosystems
Hospital admissions for respiratory and cardiac diseases
Respiratory symptoms in asthmatics
9-6

-------
 Pollutant/Effect
Quantified and Monetized Effects in
        Primary Analysis
  Quantified and/or
Monetized Effects in
Sensitivity Analyses
Unquantified Effects
 NOx/Health
                                                               Lung irritation
                                                               Lowered resistance to respiratory infection
                                                               Hospital Admissions for respiratory and cardiac diseases
 CO/Health
                                                               Premature mortality
                                                               Behavioral effects
                                                               Hospital admissions - respiratory, cardiovascular, and other
                                                               Other cardiovascular effects
                                                               Developmental effects
                                                               Decreased time to onset of angina
 NMHCsc
 Health
                                                               Cancer (diesel PM, benzene, 1,3-butadiene, formaldehyde, acetaldehyde)
                                                               Anemia (benzene)
                                                               Disruption of production of blood components (benzene)
                                                               Reduction in the number of blood platelets (benzene)
                                                               Excessive bone marrow formation (benzene)
                                                               Depression of lymphocyte counts (benzene)
                                                               Reproductive and developmental effects  (1,3-butadiene)
                                                               Irritation of eyes and mucous membranes (formaldehyde)
                                                               Respiratory and respiratory tract
                                                               Asthma attacks in asthmatics (formaldehyde)
                                                               Asthma-like symptoms in non-asthmatics (formaldehyde)
                                                               Irritation of the eyes, skin, and respiratory tract (acetaldehyde)
                                                               Upper respiratory tract irritation & congestion (acrolein)
 NMHCsc
 Welfare
                                                               Direct toxic effects to animals
                                                               Bioaccumlation in the food chain
                                                               Reduced odors
a While no causal mechanism has been identified linking new development of chronic asthma to ozone exposure, two epidemiological studies shows a statistical
association between long-term exposure to ozone and development of chronic asthma in exercising children and some non-smoking men (McConnell, 2002;
McDonnell, etal, 1999).
b Premature mortality associated with ozone is not separately included in the calculation of total monetized benefits.
c All non-methane hydrocarbons (NMHCs) listed in the table are also hazardous air pollutants listed in Section 112(b) of the Clean Air Act.
                                                                          9-7

-------
Final Regulatory Impact Analysis
9.1 Time Path of Emission Changes for the Final Standards

   The final standards have various cost and emission related components, as described earlier
in this RIA. These components would begin at various times and in some cases would phase in
over time. This means that during the early years of the program there would not be a consistent
match between cost and benefits. This is especially true for the equipment control portions and
initial fuel changes required by the program, where the full equipment cost would be incurred at
the time of equipment purchase, while the fuel and maintenance costs, along with the emission
reductions and benefits resulting from all these costs would occur throughout the lifetime of the
equipment. Because of this inconsistency and our desire to more appropriately match the costs
and emission reductions of our program, our analysis examines costs and benefits throughout the
period of program implementation. This chapter focuses on estimating the stream of benefits
over time and comparing streams of benefits and costs. Detailed information on cost estimates
can be found in  chapters 6, 7 and 8 of this RIA.

   For the nonroad diesel engine standards, implementation will occur in stages: reductions in
sulfur content of nonroad diesel fuel and then adoption of controls on most new nonroad engines.
Because full turnover of the fleet of nonroad diesel engines will not occur for many years, the
emission reduction benefits of the standards will not be fully realized until several decades after
the reduction in fuel sulfur content. The timeframe for the  analysis reflects this turnover,
beginning in 2007 and extending through 2030.

   Chapter 3  discussed the development of the 1996, 2020 and 2030 baseline emissions
inventories for the nonroad sector and for the sectors not affected by this rule.  The emission
sources and the  basis for current and future-year inventories are listed in Table 9-2. Using these
modeled inventories, emissions with and without the standards are interpolated to provide
streams of emissions from the rule implementation date through full implementation in 2030.
These streams of emissions are presented in Chapter 3. NOx and VOC contribute to ambient
ozone formation, while NOx, SO2, NMHC/VOC, and directly emitted PM2 5 emissions are
precursors to ambient PM2 5 and PM10 concentrations.  Although the rule is expected to reduced
CO and air toxics emissions as well, we do not include benefits related to these reductions in the
benefits analysis due to a lack of appropriate air quality and exposure models.
                                          9-8

-------
                                                                Cost-Benefit Analysis
                                       Table 9-2
 Emissions Sources and Basis for Current and Future-Year Inventories for Air Quality Modeling
     Emissions Source
     1996 Base year
 Future-year Base Case Projections
 Utilities
1996 NEI Version 3.12
(CEM data)
Integrated Planning Model (IPM)
 Non-Utility Point and Area
 sources
1996 NEI
Version 3.12 (point)
Version 3.11 (area)
BEA growth projections
 Highway vehicles
MOBILESb model with
MOBILE6 adjustment
factors for VOC and
NOx;
PARTS model for PM
VMT projection data
 Nonroad engines (except
 locomotives, commercial
 marine vessels, and
 aircraft)	
NONROAD2002 model
BEA and Nonroad equipment
growth projections
Note: Full description of data, models, and methods applied for emissions inventory development and modeling are
provided in the Emissions Inventory TSD (U.S. EPA, 2003a).
   Table 9-3 summarizes the expected changes in emissions of key species.  SO2 emissions are
expected to be reduced by over 84 percent within the first two years of implementation.
Emissions of PM2 5, NOx, and NMHC are expected to be reduced significantly over the period of
implementation from 2007 to 2030. Table 9-4 breaks out the expected changes in emissions of
key species for the components the fuel portion of the program.
                                          9-9

-------
Final Regulatory Impact Analysis
                                     Table 9-3
                      Summary of Reduction in 48-State Emissions"
          Attributable to Final Nonroad Diesel Engine Standards and Fuel Programs

2010
2015
2020
2025
2030
Tons Reduced
(Percent of baseline from this category)"
Direct PM2 5
21,692
13%
53,072
33%
85,808
52%
110,043
64%
128,350
72%
NOx
149
0%
193,431
17%
442,061
39%
613,629
54%
734,184
62%
SO2
256,447
91%
297,513
99%
323,378
99%
349,312
99%
375,354
99%
VOC
525
0%
8,318
8%
18,141
19%
25,002
26%
30,030
31%
a NOx, VOC, and CO inventories are for land-based diesel engines only; PM and SO2 inventories include
land-based, recreational marine, commercial marine, and locomotive diesel engines.
                                     Table 9-4
                      Summary of Reduction in 48-State Emissions
            Attributable to Final Fuel Programs of the Nonroad Diesel Standards

2010
2015
2020
2025
2030
Tons Direct PM2 5 and SO2 Reduced
Fuel Only Program
Direct PM2 5
20,051
23,241
25,248
27,265
29,293
SO2
256,447
297,389
323,137
348,994
374,982
500 ppm NRLM Fuel
Program
Direct PM2 5
19,156
20,876
22,674
24,482
26,300
SO2
245,007
267,118
290,192
313,367
336,665
1 5 ppm LM Fuel Program
(no home heating oil)
Direct PM2 5
0
428
433
427
426
SO2
0
5,318
5,382
5,308
5,294
                                        9-10

-------
                                                                 Cost-Benefit Analysis
9.2 Development of Benefits Scaling Factors Based on Differences in
Emission Impacts Between the Final Standards and Modeled Preliminary
Control Options

   Based on the projected time paths for emissions reductions, we focused our detailed
emissions and air quality modeling on two future years, 2020 and 2030, which reflect partial and
close to complete turnover of the fleet of nonroad diesel engines to rule compliant engines. The
emissions changes modeled for these two years are similar to those in the final standards,
differing in the treatment of smaller engines and fuel requirements.0 Table 9-5 summarizes the
reductions in emissions of NOx, SO2, and PM2 5 from baseline for the preliminary and final
standards, the difference between the two, and the ratio of emissions reductions from the final
standards to the preliminary control options.  The ratios presented in the last column of Table 9-5
are the basis for the benefits scaling approach discussed below.
   DAs discussed in Chapter 2, emissions and air quality modeling decisions are made early in EPA's analytical
process. Since the preliminary control scenario was developed, EPA has gathered more information regarding the
technical feasibility of the standards and considered public comment. As a result, we have revised the control
scenario as described in detail in previous chapters of this document. Section 3.6 describes the changes in the inputs
and resulting emission inventories between the preliminary baseline and control scenarios used for the air quality
modeling and the final baseline and control scenarios.

                                          9-11

-------
Final Regulatory Impact Analysis
                                      Table 9-5
                     Comparison of 48-state Emission Reductions"'b
                in 2020 and 2030 Between Preliminary and Final Standards
Emissions Species
2020
NOx
S02
PM25
2030
NOx
SO2
PM7,
Reduction from Baseline
Preliminary

663,618
414,692
98,121

1,009,744
483,401
138,208
Final

442,061
323,378
85,808

734,184
375,354
128,350
Difference in
Reductions
(Final minus
Preliminary)

221,557
91,314
12,313

275,560
108,047
9,858
Ratio of
Reductions
(Final/
Preliminary)

0.67
0.78
0.87

0.73
0.78
0.93
a Includes all affected nonroad sources: land-based, recreational marine, commercial marine, and locomotives.
b We note that the magnitude of NOx reductions determined in the final rule analysis is somewhat less than what
was reported in the proposal's draft RIA, especially in the later years when the fleet has mostly turned over to Tier
4 designs. The greater part of this is due to the fact that we have deferred setting a long-term NOx standard for
mobile machinery over 750 hp to a later action. When this future action is completed, we would expect roughly
equivalent reductions between the proposal and the overall final program, though there are some other effects
reflected in the differing NOx reductions as well, due to updated modeling assumptions and the adjusted NOx
standards levels for engines over 750 hp. Preamble Section II.A.4 contains a detailed discussion of the NOx
standards we are adopting for engines over 750 hp, and the basis for those standards.
9.3 Summary of Modeled Benefits and Apportionment Method

   As a second step in the analysis, we calculated scaling factors relating emissions changes
resulting from the final standards to emissions changes from a set of preliminary control options
that were used to model air quality and benefits (see Appendix 9A for full details).  Based on the
emissions inventories developed at the time of the proposal for the preliminary control option,
we conducted a benefits analysis to determine the air quality and associated human health and
welfare benefits resulting from the reductions in emissions of NOx, SO2, NMHC/VOC, and
PM25. Based on the availability of air quality and exposure models, this summary focuses on
reporting the health and welfare benefits of reductions in ambient PM and ozone concentrations.
However, health improvements may also come from reductions in exposure to CO and air toxics.
 The full analysis is available in Appendix 9A and the benefits Technical Support Document
(TSD) (Abt Associates, 2003).

   The reductions in emissions of NOx, SO2, and PM25 from nonroad engines in the United
States are expected to result in wide-spread overall reductions in ambient concentrations of
                                        9-12

-------
                                                                  Cost-Benefit Analysis
ozone and PM2 5E.  These improvements in air quality are expected to result in substantial health
benefits, based on the body of epidemiological evidence linking PM and ozone with health
effects such as premature mortality, chronic lung disease, hospital admissions, and acute
respiratory symptoms.  Based on modeled changes in ambient concentrations of PM25 and
ozone, we estimate changes in the incidence of each health effect using concentration-response
(C-R) functions derived from the epidemiological literature with appropriate baseline
populations and incidence rates.  We then apply estimates of the dollar value of each health
effect to obtain a monetary estimate of the total PM- and ozone-related health benefits of the
rule. Welfare effects are estimated using economic models which link changes in physical
damages (e.g., light extinction or agricultural yields) with economic values.

    Since the publication of the RIA for the proposed rule, EPA has received new technical
guidance and input regarding its methodology for conducting PM- and ozone-related benefits
analysis from the Health Effects Subgroup (HES) of the Science Advisory Board (SAB) Council
reviewing the 812 blueprint (SAB-HES, 2003) and the Office of Management and Budget
(OMB) through ongoing discussions regarding methods used in conducting regulatory impact
analyses (RIAs) (e.g., see OMB Circular A-4). The SAB HES recommendations include the
following (SAB-HES, 2003):

       •  use of the updated ACS Pope et al. (2002) study rather than the ACS Krewski et al.
          study to estimate premature mortality for the primary analysis;

       •  dropping the alternative estimate used in earlier RIAs and instead including a primary
          estimate that incorporates consideration of uncertainly in key effects categories such
          as premature mortality directly into the estimates (e.g., use of the standard errors from
          the Pope et al. (2002) study in deriving confidence bounds for the adult mortality
          estimates);

       •  addition of infant mortality (children under the age of one) into the primary estimate,
          based on supporting evidence from the World Health Organization Global Burden of
          Disease study (World Health Organization, 2002) and other published studies that
          strengthen the evidence for a relationship between PM exposure and respiratory
          inflamation and infection in children leading to death;

       •  inclusion of asthma exacerbations for children in the primary estimate;

       •  expansion of the age groups evaluated for a range of morbidity effects beyond the
          narrow band of the studies to the broader (total) age group (e.g., expanding a study
          population for 7 to 11 year olds to cover the entire child age range of 6 to 18 years).
   E Reductions in NOx are expected to result in some localized increases in ozone concentrations, especially in
NOx-limited large urban areas, such as Los Angeles, New York, and Chicago. A fuller discussion of this
phenomenon is provided in Chapter 2.3.  While localized increases in ozone will result in some increases in health
impacts from ozone exposure in these areas, on net, the reductions in NOx are expected to reduce national levels of
health impacts associated with ozone.

                                           9-13

-------
Final Regulatory Impact Analysis
       •   inclusion of new endpoints (school absences [ozone], nonfatal heart attacks in adults
           [PM], hospital admissions for children under two [ozone]), and suggestion of a new
           meta-analysis of hospital admissions (PM10) rather than using a few PM25 studies;F
           and

       •   updating of populations and baseline incidences.

    Recommendations from the Office of Management and Budget (OMB) regarding EPA's
methods have focused on the approach used to characterize uncertainty in the benefits estimates
generated for RIAs, as well as the approach used to value premature mortality estimates.  The
EPA is currently in the process of developing a comprehensive, integrated strategy for
characterizing the impact of uncertainty in key elements of the benefits modeling process (e.g.,
emissions modeling, air quality modeling, health effects incidence estimation, valuation) on the
results that are generated. A subset of this effort involved an expert elicitation designed to
characterize uncertainty in the estimation of PM-related mortality resulting from both short-term
and longer-term exposure.  In its 2002 report, the NAS provides a number of recommendations
on how EPA might improve the characterization of uncertainty in its benefits analyses. One
recommendation was that "EPA should begin to move the assessment of uncertainties from its
ancillary analyses into its primary analyses by conducting probabalistic, multiple-source
uncertainty analyses.  This shift will require specification of probability distributions for major
sources of uncertainty. These distributions should be based on available data and expert
judgement."  The NAS elaborated  on this recommendation by suggesting a program of
methodological development involving review and critique of existing protocols for selection
and elicitation of experts by decision analysts, biostatisticians, and psychologists.  They
recommended the use of formally elicited expert judgements, but noted that a number of issues
must be addressed,  and that sensitivity analyses would be needed for distributions that are based
on expert judgment. They also recommended that EPA clearly distinguish between data-derived
components of an uncertainty assessment and those based on expert opinions.  As a first step in
addressing the NAS recommendations regarding expert elicitation, EPA, in collaboration with
OMB, conducted a  pilot expert elicitation to characterize uncertainties in the relationship
between ambient PM2 5 concentrations and premature mortality.  While it is premature to include
the results of the pilot in the primary analysis for this rulemaking, EPA and OMB believe this
pilot moves toward the goal of incorporating additional uncertainty analyses in its future primary
benefits analyses. The pilot expert elicitation is described in Appendix 9B and the full report is
placed in the public docket.

    We have  also modified the analysis to reflect new information in the academic literature on
the appropriate characterization of the value of reducing the risk of premature mortality (value of
    FNote that the SAB-HES comments were made in the context of a review of the methods for the Section 812
analysis of the costs and benefits of the Clean Air Act. This context is pertinent to our interpretation of the SAB-
HES comments on the selection of effect estimates for hospital admissions associated with PM (SAB-HES, 2003).
The Section 812 analysis is focused on a broad set of air quality changes, including both the coarse and fine fractions
of PM10.  As such, impact functions that focus on the full impact of PM10 are appropriate. However, for the Nonroad
Diesel Engines rule, which is expected to affect primarily the fine fraction (PM25) of PM10, impact functions that
focus primarily on PM2 5 are more appropriate.

                                           9-14

-------
                                                                 Cost-Benefit Analysis
statistical life (VSL)). In previous analyses, we used a distribution based on 26 VSL estimates
from the economics literature. For this analysis, we are characterizing the VSL distribution in a
more general fashion, based on two recent meta-analyses of the wage-risk-based VSL literature
(Mrozek and Taylor, 2000 and Viscusi and Aldy, 2003). The new distribution is assumed to be
normal, with a mean of $5.5 million and a 95 percent confidence interval between $1 and $10
million. The $1 million lower confidence limit represents the lower end of the interquartile range
from the Mrozek and Taylor (2000) meta-analysis.G The $10 million upper confidence limit
represents the upper end of the interquartile range from the Viscusi and Aldy (2003)
meta-analysis.

   The EPA has addressed many of the comments received from the SAB-HES and OMB in
developing the analytical approach for the final rule. We use an approach consistent with the
methods used in the benefits analysis of the recently proposed Interstate Air Quality rule
(IAQR). We have also reflected advances in data and methods in air quality modeling,
epidemiology, and economics in developing this analysis. Updates to the assumptions and
methods used in estimating PM25-related and ozone-related benefits since completion of the
Proposed Nonroad Diesel Rule include the following:

Health Endpoints

       •  We incorporated updated impact functions to reflect updated time-series studies of
          hospital admissions to correct for errors in application of the generalized additive
          model (GAM) functions in S-plus. More information on  this issue is available at
          http://www.healtheffects.org.

       •  The primary analysis used an all-cause mortality effect estimate based on the Pope et
          al. (2002) reanalysis of the ACS study data.

       •  Infant mortality was included in the primary analysis.

       •  Asthma exacerbations were incorporated into the primary analysis. Although the
          analysis of the proposed rule included asthma exacerbations as a separate endpoint
          outside of the base case analysis, for the final rule, we will include asthma
          exacerbations in children 6 to 18 years of age as part of the primary analysis.
Valuation
       •   In generating the monetized benefits for premature mortality in the primary analysis,
          the VSL will be entered as a mean (best estimate) of $5.5 million. Unlike the
          analysis of the proposed rule, the final rule analysis will not include a value of
   GAn alternative rationale for the low end of the range could be found in some recent stated preference studies
suggesting VSL of between $1 and $5 million (Alberini et al., forthcoming).

                                          9-15

-------
Final Regulatory Impact Analysis
          statistical life year (VSLY) estimate. This reflects the advice of the SAB-Council and
          concerns raised by commentors on the proposed rule.

   The proposed Nonroad Diesel rule included an alternative estimate in addition to the primary
estimate that was intended to evaluate the impact of several key assumptions on the estimated
reductions in premature mortality and chronic bronchitis. However, reflecting comments from
the SAB-HES, rather than including an alternative estimate in the analysis of the final rule, the
EPA will investigate the impact of key assumptions on mortality and morbidity estimates
through a series of sensitivity analyses. This advice is consistent with the NAS
recommendations as well.

9.3.1 Overview of Analytical Approach

   This section summarizes the three steps involved in our analysis of the modeled preliminary
control options: 1) Calculation of the impact that a set of preliminary fuel and engine standards
would have on the nationwide inventories for NOx, NMHC, SO2, and direct PM emissions in
2020 and 2030; 2) Air quality modeling for 2020 and 2030 to determine changes in ambient
concentrations of ozone and PM, reflecting baseline and post-control emissions inventories;  and
3) A benefits analysis to determine the changes in human health and welfare, both in terms of
physical effects and monetary value, that result from the projected changes in ambient
concentrations of various pollutants for the modeled standards.

   We follow a "damage-function" approach in calculating total benefits of the modeled
changes in environmental quality. This approach estimates changes in individual health and
welfare endpoints (specific effects that can be associated with changes in air quality) and assigns
values to those changes assuming independence of the individual values.  Total benefits are
calculated simply as the sum of the values for all non-overlapping health and welfare endpoints.
This imposes no  overall preference structure, and does not account for potential income or
substitution effects, i.e. adding a new endpoint will not reduce the value of changes in other
endpoints. The "damage-function" approach is the standard approach for most cost-benefit
analyses of regulations affecting environmental quality, and it has been used in several recent
published analyses (Banzhaf et al., 2002; Levy et al., 2001; Kunzli et al., 2000; Levy et al., 1999;
Ostro and Chestnut, 1998). Time and resource constraints prevented us from performing
extensive new research to measure either the health outcomes or their values for this analysis.
Thus, similar to these studies, our estimates are based on the best available methods of benefits
transfer. Benefits transfer is the science and art of adapting primary research from similar
contexts to obtain the most accurate measure of benefits available for the environmental quality
change under analysis.

   There are significant categories of benefits that cannot be monetized (or in many cases even
quantified), and thus they are not included in our accounting of health and welfare benefits.
These unquantified effects include low birth weight, changes in pulmonary function, chronic
respiratory diseases other than chronic bronchitis, morphological changes, altered host defense
mechanisms, non-fatal cancers, and non-asthma respiratory emergency room visits. A complete
discussion of PM -related health effects can be found in the PM Criteria Documents (U.S. EPA

                                          9-16

-------
                                                                 Cost-Benefit Analysis
1996a, U.S. EPA, 2004) and the EPA Diesel HAD (U.S. EPA 2002).  A discussion of the state of
the science as of the lastNAAQS review of ozone-related effects can be found in the Ozone
Criteria Document (U.S. EPA 1996b). Since many health effects overlap, such as minor
restricted activity  days and asthma symptoms, we made assumptions intended to reduce the
chances of "double-counting" health benefits, which may result in an underestimate of the total
health benefits of the pollution controls.

9.3.2 Air Quality Modeling

   As described in Chapter 2 and the technical support documents (TSDs), we used a national-
scale version of the REgional Modeling System for Aerosols and Deposition (REMSAD version
7) to estimate PM air quality in the contiguous United States.  We used the Comprehensive Air
Quality Model with Extensions (CAMx) to estimate ambient ozone concentrations,11 using two
domains representing the Eastern and Western U.S.  These models are discussed in the air
quality TSD for this rule.

   9.3.2.1 PM Air Quality Modeling with REMSAD

   REMSAD is appropriate for evaluating the impacts of emissions reductions from nonroad
sources, because it accounts for spatial and temporal variations as well as differences in the
reactivity of emissions. The annual county level emission inventory data described  in Chapter 3
was speciated, temporally allocated and gridded to the REMSAD modeling domain to simulate
PM concentrations for the 1996 base year and the 2020 and 2030 base and control scenarios.
Peer-reviewed for the EPA, REMSAD is a three-dimensional grid-based Eulerian air quality
model designed to estimate annual particulate concentrations and deposition over large spatial
scales (Seigneur et al., 1999).  Each of the future scenarios was simulated using 1996
meteorological data to provide daily averages and annual mean PM concentrations required for
input  to the concentration-response functions of the benefits analysis. Details regarding the
application of REMSAD Version 7 for this analysis are provided in the Air Quality  Modeling
TSD (U.S. EPA, 2003b).  This version reflects updates in the following areas to improve
performance and address comments from the 1999 peer-review:

   1.  Gas phase chemistry updates to "micro-CB4" mechanism including new treatment for the
       NO3 and N2O5 species and the addition of several reactions to better account for the wide
       ranges in temperature, pressure, and concentrations that are encountered for  regional and
       national applications.

   2.  PM chemistry updates to calculate particulate nitrate concentrations through use of the
       MARS-A  equilibrium algorithm and internal calculation of secondary organic aerosols
       from both  biogenic (terpene) and anthropogenic (estimated aromatic) VOC emissions.
   HIn the benefits analysis of the recent Heavy Duty Engine/Diesel Fuel rule, we used the Urban Airshed Model
Variable-Grid (UAM-V) to estimate ozone concentrations in the Eastern U.S.  CAMx has a number of improvements
relative to UAM and has improved model performance in the Western U.S. Details on the performance of CAMx
can be found in Chapter 2 as well as the Air Quality Modeling TSD (U.S. EPA, 2003b).

                                          9-17

-------
Final  Regulatory Impact Analysis
    3.  Aqueous phase chemistry updates to incorporate the oxidation of S02 by O3 and O2 and to
       include the cloud and rain liquid water content from MM5 meteorological data directly in
       sulfate production and deposition calculations.

As discussed earlier in Chapter 2, the model tends to underestimate observed PM2 5
concentrations nationwide, especially over the western U.S.1

    9.3.2.2 Ozone Air Quality Modeling with CAMx

    We use the emissions inputs described in Chapter 3 with a regional-scale version of CAMx
to estimate ozone air quality in the Eastern and Western U.S. CAMx is an Eulerian three-
dimensional photochemical grid air quality model designed to calculate the concentrations of
both inert and chemically reactive pollutants by simulating the physical and chemical processes
in the atmosphere that affect ozone formation.  Because it accounts for spatial  and temporal
variations as well as differences in the reactivity of emissions, the CAMx is useful for evaluating
the impacts of the nonroad diesel engine rule on U.S. ozone concentrations.  As discussed earlier
in Chapter 2, although the model tends to underestimate observed ozone, especially over the
western U.S.,  it exhibits less bias and error than any past regional ozone modeling application
conducted by  EPA (i.e., Ozone Transport Assessment Group (OTAG), On-highway Tier-2
Passenger Vehicles, and Heavy Duty Engine/Diesel Fuel 2007 program).

    Our analysis applies the modeling system separately to the Eastern and Western U.S. for five
emissions scenarios:  a 1996 baseline projection, a 2020 baseline projection and a 2020 projection
with nonroad  controls, a 2030 baseline projection and a 2030 projection with nonroad controls.
As discussed in detail in the technical support document, a 1996 base year assessment is
necessary because the relative  model predictions  are used with ambient air quality observations
from 1996 to determine the expected changes in 2020 and 2030 ozone concentrations due to the
modeled emission changes (Abt Associates, 2003). These results are used solely in the benefits
analysis.
    1 Comments from industry have stated that EPA's methodology form computing benefits over time is based on
unsupportable assumptions related to air quality modeling. Specifically, they state that EPA assumes that there will
be no interactions between precursors and directly emitted PM in the formation of secondary PM and that EPA
excludes consideration of non-linearities in its air quality modeling. The commentor is partially incorrect in the
statement that "EPA assumes no interactions between NOx, SO2, and direct PM in the formation of PM2 5." In order
to estimate benefits in years other than 2020 and 2030, it was necessary to interpolate values from 2020 and 2030.
We used sophisticated air quality modeling (using the REMSAD model) to predict changes in ambient PM2 5 in 2020
and 2030. This air quality modeling for 2020 and 2030 does incorporate the nonlinear interactions between NOx,
SO2, and direct PM.  However, in order to develop the intertemporal scaling factors, we had to make some
simplifying assumptions. We assumed that the interactions between SO2 and NOx were linear over time, rather than
assuming that there was no interaction. In other words, we assumed that the rate of change in the sulfate to SO2,
nitrate to NOx, and primary PM to direct PM ratios was a linear function of time. The rate of change is driven by
differences in the baseline emissions between 2020 and 2030 and by differences in the ratio of NOx to SO2
reductions from the nonroad sector. We verified the interpolation approach by predicting 2020 benefits using
scaling factors for sulfate, nitrate, and direct PM based on the modeled 2030 benefits. Scaled benefits were within 4
percent of the actual modeled benefits for 2020.

                                             9-18

-------
                                                                 Cost-Benefit Analysis
   As discussed in more detail in Chapter 2.3, our ozone air quality modeling showed that the
NOx emissions reductions from the preliminary modeled standards are projected to result in
increases in ozone concentrations for certain hours during the year, especially in urban, NOx-
limited areas. Most of these increases are expected to occur during hours where ozone levels are
low (and often below the one-hour ozone standard). However, most of the country experiences
decreases in ozone concentrations for most hours in the year.

9.3.3 Health Impact Functions

   Health impact functions are derived from the epidemiology literature. A standard health
impact function has four components: an effect estimate from a particular epidemiological study,
a baseline incidence rate for the health effect (obtained from either the epidemiology study or a
source of public health statistics like the Centers for Disease Control), the affected population,
and the estimated change in the relevant PM or ozone summary measure.

A typical health impact function might look like:
where y0 is the baseline incidence, equal to the baseline incidence rate times the potentially
affected population, p is the effect estimate, and Ax is the estimated change in the summary
PM2 5 or ozone measure. There are other functional forms, but the basic elements remain the
same.

    Integral to the estimation of the impact functions  are reasonable estimates of future
population projections. The underlying data used to create county-level 2010 population
projections is based on county level allocations of national population projections from the U.S.
Census Bureau (Hollman, Mulder and Kalian, 2000). County-level allocations of populations by
age, race, and sex are based on economic forecasting models developed by Woods and Poole,
Inc (WP), which account for patterns  of economic growth and migration.

    The WP projections of county level population are based on historical population data  from
1969-1999, and do not include the 2000 Census results.  Given the availability of detailed 2000
Census data, we constructed adjusted county level population projections for each future year
using a two stage process.  First, we constructed ratios of the projected WP populations in  a
future year to the projected WP population in 2000 for each future year by age, sex, and race.
Second, we multiplied the block level 2000 Census population data by the appropriate age, sex,
and race  specific WP ratio for the county containing the census block, for each future year. This
results in a set of future population projections that is consistent with the most recent detailed
census data.

    Specific populations matching the study populations in each epidemiological study are
constructed by accessing the appropriate age-specific projections from the overall population
database. For some endpoints, such as asthma attacks, we further limit the population by


                                          9-19

-------
Final Regulatory Impact Analysis
applying prevalence rates to the overall population. We do not have sufficient information to
quantitatively characterize uncertainty in the population estimates.

   Fundamental to the estimation of health benefits was our utilization of the PM epidemiology
literature. We rely upon effect estimates derived from published, peer reviewed epidemiological
studies that relate health effects to ambient concentrations of PM. The specific studies from
which effect estimates are drawn are listed in Table 9-5. While a broad range of serious health
effects have been associated with exposure to elevated PM levels, we include only a subset of
health effects in this benefit analysis due to limitations in available effect estimates and concerns
about double-counting of overlapping effects (U.S. EPA, 1996).  For the most part, we use the
same set of effect estimates as we used in the analysis of the proposed Nonroad Diesel Engines
rule. However, based on recent advice from the SAB, we use an updated effect estimate for
premature mortality and include two additional health effects, infant mortality and asthma
exacerbations. Because of their significance in the analysis,  we provide a more detailed
discussion of premature mortality and chronic illness endpoints below.

   To generate health outcomes, projected changes in ambient PM concentrations were entered
into BenMAP, a  customized geographic information system based program.  BenMAP
aggregates populations to air quality model grids and calculates changes in air pollution metrics
(e.g., daily averages) for input into health impact  functions.  BenMAP uses grid cell level
population data and changes in pollutant concentrations to estimate changes in health outcomes
for each grid cell. Details on the BenMAP program can be found in the BenMAP User's Manual
(Abt Associates,  2003).

   The baseline  incidences for health outcomes used in our analyses are selected and adapted to
match the specific populations  studied. For example, we use age- and county-specific baseline
total mortality rates in the estimation of PM-related premature mortality. County-level incidence
rates are  not available for other endpoints.  We used national incidence rates whenever possible,
because these data are most applicable to a national assessment of benefits. However, for some
studies, the only  available incidence information comes from the studies themselves; in these
cases, incidence in the study population is assumed to represent typical incidence at the national
level. Sources of baseline incidence rates are reported in Table 9-6.

   In this assessment we made analytical judgements affecting both the selection of effect
estimates and the application of those estimates in formulating health impact  functions. In
general, we selected effect estimates that 1) most closely match the pollutants of interest, i.e.
PM2 5) cover the broadest potentially exposed population (i.e. all  ages functions would be
preferred to adults 27 to 35), 3) have appropriate model  specification (e.g. control for
confounding pollutants), 4) have been peer-reviewed,  and  5) are biologically  plausible. Other
factors may also  affect our selection of effect estimates for specific endpoints, such as premature
mortality. Some of the more important of these relating to premature mortality  and chronic
illness  are discussed below and are discussed in detail in Appendix 9A. Alternative assumptions
about these judgements may lead to substantially different results and they are explored using
appropriate sensitivity analyses provided in Appendix 9B.
                                          9-20

-------
                                                                 Cost-Benefit Analysis
   While there is a consistent body of evidence supporting a relationship between a number of
adverse health effects and ambient PM levels, there is often only a single study of a specific
endpoint covering a specific age group.  There may be multiple estimates examining subgroups
(i.e. asthmatic children). However, for the purposes of assessing national population level
benefits, we chose the most broadly applicable effect estimate to more completely capture health
benefits in the general population.  Estimates for subpopulations are provided in Appendix 9A.

   There is no consensus on whether or not there is a threshold for the health effects of PM, and
if so, what the possible threshold might be.  Consistent with recent literature (Daniels et al.,
2000; Pope et al., 2002; Rossi et al., 1999; Schwartz, 2000), we chose for the purposes of this
analysis to assume that PM-related health effects occur down to natural background (i.e., there is
no health effects threshold). We assume that all of the health impact functions are continuous
and differentiable down to natural background levels. Our assumptions regarding thresholds are
considered reasonable by the National Research Council in its recent review of methods for
estimating the public health benefits of air pollution regulations. In their review, the National
Research Council concluded that there is no evidence for any departure from linearity in the
observed range of exposure to PM10 or PM25, nor any indication of a threshold. (NRC,  2002).
They cite the weight of evidence available from both short and long term exposure models and
the similar effects found in cities with low and high ambient concentrations of PM. We explore
this important assumption in a sensitivity analysis described in Appendix 9C.

   Premature Mortality

   As receommended by the NAS (2002) and the SAB-HES, and demonstrated in the  Kunzli et
al. (2000) health impact assessment, we focus on the prospective cohort long-term exposure
studies in deriving the health impact function for our base estimate of premature mortality.
Cohort analyses are better able to capture the full public health impact of exposure to air
pollution over time (Kunzli, 2001; NRC, 2002). We selected an effect estimate from the
extended analysis of the American Cancer Society (ACS) cohort (Pope et al., 2002) because it
represents the most comprehensive cohort analysis with the longest period of followup. In
addition, this study has been recommended for impact assessments by the SAB-HES (SAB-HES,
2003). This effect estimate quantifies the relationship between annual mean PM2 5 levels and all-
cause mortality in adults 30 and older. We  selected the effect estimate estimated using the
measure of PM representing average exposure over the follow-up period, calculated as the
average of 1979-1984 and  1999-2000 PM25 levels.

   In previous analyses, infant mortality has not been evaluated as part of the primary analysis
due to uncertainty in the strength of the association between exposure to PM and postneonatal
mortality. Instead, benefits estimates related to reduced infant mortality have been included as
part of the sensitivity analyses. However recently published studies have strengthened the case
for an association between PM exposure and respiratory inflamation  and infection leading to
premature mortality in infants under five years of age.  Specifically, the SAB's HES noted the
release of the World Health Organization Global Burden of Disease Study focusing on ambient
air which cites several recently-published time-series studies relating daily PM exposure to
mortality in  children. The HES also cites the study by Belanger et al., (2003) as corroborating

                                          9-21

-------
Final Regulatory Impact Analysis
findings linking PM exposure to increased respiratory inflamation and infections in children.
With regard to the cohort study conducted by Woodruff et al. (1997), the HES notes several
strengths of the study including the use of a larger cohort drawn from a large number of
metropolitan areas and efforts to control for a variety of individual risk factors in children (e.g.,
maternal educational level, maternal ethnicity, parental marital status and maternal smoking
status).  We follow the HES recommendation to include infant mortality in the primary benefits
estimate using the effect estimate from the Woodruff et al. (1997) study.

    Chronic Illness

    Although there are several  studies examining the relationship between PM of different size
fractions and incidence of chronic bronchitis, we use a study by Abbey et al. (1995) to obtain our
estimate of avoided incidences of chronic bronchitis in adults aged 25 and older, because Abbey
et al. (1995) is the only available estimate of the relationship between PM25 and chronic
bronchitis. Based on the Abbey et al. study, we estimate the number of new chronic bronchitis
cases that will "reverse" over time and subtract these reversals from the estimate of avoided
chronic bronchitis incidences.  Reversals refer to those cases of chronic bronchitis that were
reported at the start of the  Abbey et al. survey, but were subsequently not reported at the end of
the survey.  Since we assume that chronic bronchitis is a permanent condition, we subtract these
reversals. Given the  relatively high value assigned to chronic bronchitis, this ensures that we do
not overstate the economic value of this health effect.

    Non-fatal heart attacks have been linked with short term exposures to PM25 in the U.S.
(Peters et al., 2001) and other countries (Poloniecki et al., 1997). We use a recent study by
Peters et al.  (2001) as the basis for the C-R function estimating the relationship between PM25
and non-fatal heart attacks in adults. Peters et al. is the only available U.S. study to provide a
specific estimate for heart attacks. Other studies, such as Samet et al. (2000) and Moolgavkar et
al. (2000) show a consistent relationship between all cardiovascular hospital admissions,
including for non-fatal heart attacks, and PM.  Given the lasting impact of a heart attack on
longer-term health costs and earnings, we choose to provide a separate estimate for non-fatal
heart attacks based on the  single available U.S. C-R function. The finding of a specific impact
on heart attacks is consistent with hospital admission and other studies showing relationships
between fine particles and cardiovascular effects both within and outside the U.S. These studies
provide a weight of evidence for this type of effect.  Several epidemiologic studies (Liao et al.,
1999; Gold et al., 2000; Magari et al., 2001) have shown that heart rate variability (an indicator
of how much the heart is able to speed up or slow down in response to momentary stresses) is
negatively related to PM levels. Heart rate variability is a risk factor for heart attacks and other
coronary heart diseases (Carthenon et al., 2002; Dekker et al., 2000; Liao et al.,  1997, Tsuji et al.
1996).  As such, significant impacts of PM on heart rate variability are consistent with an
increased risk of heart attacks.

9.3.4 Economic Values for Health Outcomes

    Reductions in ambient concentrations of air pollution generally lower the risk of future
adverse health affects by a fairly small amount  for a large population.  The appropriate

                                          9-22

-------
                                                                 Cost-Benefit Analysis
economic measure is therefore willingness-to-pay (WTP) for changes in risk prior to the
regulation (Freeman, 1993).  For some health effects, such as hospital admissions, WTP
estimates are generally not available.  In these cases, we use the cost of treating or mitigating the
effect as a primary estimate.  These costs of illness (COI) estimates generally understate the true
value of reductions in risk of a health effect, reflecting the direct expenditures related to
treatment but not the value of avoided pain and suffering from the health effect (Harrington and
Portney, 1987; Berger, 1987). Unit values for health endpoints are provided in Table 9-7.  All
values are in constant year 2000 dollars.

    The length of the delay between reduction in chronic PM exposures and reduction in
mortality rates is unknown and yet an important parameter in the benefits analysis. The size of
such a time lag is important for the valuation of premature mortality incidences as economic
theory suggests benefits occurring in the future should be discounted relative to benefits
occurring today. Although there is no specific scientific evidence of the size  of a PM effects lag,
current scientific literature on adverse health effects associated with smoking and the difference
in the effect size between chronic exposure studies and daily premature mortality studies suggest
that all incidences of premature mortality reduction associated with a given incremental change
in PM exposure would not occur in the same year as the exposure  reduction.  This literature
implies that lags of a few years or longer are plausible.  For our current analysis, based  on
previous advice from the SAB (EPA-SAB-COUNCIL-ADV-00-001, 1999), we have assumed a
five-year distributed lag structure, with 25 percent of premature deaths occurring in the first year,
another 25 percent in the second year, and 16.7 percent in each of the remaining three years. To
account for the preferences of individuals for current risk reductions relative to future risk
reductions, we discount the value of avoided premature mortalities occurring beyond the
analytical year (2020 or 2030) using three and seven percent discount rates.

    A more recent SAB-HES report confirmed the NAS (2002) conclusion that there is  little
justification for the 5-year time course used by EPA in its past assessments, and  suggested that
future  assessments more fully and explicitly account for the uncertainty.  The SAB-HES
suggests that appropriate lag structures may be developed based on the distribution of cause
specific deaths within the overall all-cause estimate.  The SAB-HES specifically noted
understanding mechanisms of damage and developing models for  different cause of death
categories may be the key to characterizing more appropriate cessation lag functions. They note
that our current understanding of mechanisms suggests there are likely short-term (e.g., less than
six months for some cardiovascular effects), medium term (e.g., 2-5 years for COPD), and
longer term (e.g., 15 to 25 years for lung cancer). They noted that there is a current lack of direct
data to specify a lag function and recommended that information on the lag function be
considered in future expert elicitations and/or sensitivity analyses.  While we are working to
develop the underlying data to support a more appropriate segmented lag structure, for this
analysis we maintain the 5-year lag structure used in the benefits analysis for the proposed rule.
We have added an  additional sensitivity analysis to Appendix 9C examining the impact of
assuming a segmented lag of the type suggested by the  SAB-HES.  The overall impact of
moving from the 5-year distributed lag to this version of a segmented lag is relatively modest,
reducing benefits by approximately 8 percent when a three percent discount rate is used and 22
percent when a seven percent discount rate is used.

                                           9-23

-------
Final Regulatory Impact Analysis
    Our analysis accounts for expected growth in real income over time. Economic theory
argues that WTP for most goods (such as environmental protection) will increase if real incomes
increase. The economics literature suggests that the severity of a health effect is a primary
determinant of the strength of the relationship between changes in real income and WTP
(Alberini, 1997; Miller, 2000; Evans and Viscusi, 1993). As such, we use different factors to
adjust the WTP for minor health effects, severe and chronic health effects, and premature
mortality. We also adjust WTP for improvements in recreational visibility. Adjustment factors
used to account for projected growth in real income from 1990 to 2030 are 1.08 for minor health
effects, 1.27 for severe and chronic health effects, 1.23 for premature mortality, and 1.61 for
recreational visibility. Adjustment factors for 2020 are 1.07 for minor health effects, 1.23 for
severe and chronic health effects, 1.20 for premature mortality, and 1.52 for recreational
visibility. Note that due to a lack of reliable projections of income growth past 2024, we assume
constant WTP from 2024 through 2030.  This will result in an underestimate of benefits
occurring between 2024 and 2030.   Details of the calculation of the income adjustment factors
are provided in Appendix 9A.

9.3.5 Welfare Effects

    Our analysis of the preliminary control option examined two categories of welfare effects:
visibility in a subset of national parks and changes in consumer and producer surplus associated
with changes in agricultural yields.  There are a number of other environmental effects which
may affect human welfare, but due to a lack of appropriate physical effects or valuation methods,
we are unable to quantify or monetize these effects for our analysis of the nonroad  standards.

    9.3.5.1  Visibility Benefits

    Changes in the level of ambient particulate matter caused by the reduction in emissions from
the preliminary control options will change the level of visibility in much of the U.S. as
discussed in Chapter 2. Visibility directly affects people's enjoyment of a variety of daily
activities. Individuals value visibility both in the places they live, work, and recreate, in the
places they travel to for recreational purposes, and at sites of unique public value,  such as the
Grand Canyon.

    For the purposes of this analysis, visibility improvements were valued only for a limited set
of mandatory federal Class I areas. Benefits of improved visibility in  the places people live,
work, and recreate outside of these limited set of Class I areas were not included in our estimate
of total benefits, although they are examined in a sensitivity analysis presented in Appendix 9B.
All households in the U.S. are  assumed to derive some benefit from improvements in Class I
areas, given their national importance and high visitation rates from populations throughout the
U.S. However, values are assumed to be higher if the Class I area is located close to their home/
   J For details of the visibility estimates discussed in this section, please refer to the benefits technical support
document for this RIA (Abt Associates 2003).

                                           9-24

-------
                                                                 Cost-Benefit Analysis
We use the results of a 1988 contingent valuation survey on recreational visibility value
(Chestnut and Rowe, 1990a; 1990b) to derive values for visibility improvements. The Chestnut
and Rowe study measured the demand for visibility in Class I areas managed by the National
Park Service (NFS) in three broad regions of the country: California, the Southwest, and the
Southeast.  The Chestnut and Rowe study did not measure values for visibility improvement in
Class I areas outside the three regions. Their study covered 86 of the 156 Class I areas in the
U.S. We can infer the value of visibility changes in the other Class I areas by transferring values
of visibility changes at Class I areas in the study regions. However, these values are less certain
and are thus presented only as an sensitivity estimate in Appendix 9B.

   A general willingness to pay equation for improved visibility (measured in deciviews) was
developed as a function of the baseline level of visibility, the magnitude of the visibility
improvement, and household income. The behavioral parameters of this equation were taken
from analysis of the Chestnut and Rowe data. These parameters were used to calibrate WTP for
the visibility changes resulting from the Nonroad Diesel Engine rule. The method for
developing calibrated WTP functions is based on the approach developed by Smith, et al. (2002),
and is described in detail in the benefits technical support document for the proposed rule (Abt
Associates, 2003).  Major sources of uncertainty for the visibility benefit estimate include  the
quality of the underlying study and the benefits transfer process used.  Judgments used to choose
the functional form and key parameters of the estimating equation for willingness to pay for the
affected population could have significant effects on the size of the estimates. Assumptions
about how individuals respond to changes in visibility that are either very small, or outside the
range covered in the Chestnut and Rowe study, could also affect the results. EPA is considering
next steps in improving its visibility benefits estimates.

   9.3.5.2 Agricultural Benefits

   Laboratory and field experiments have shown reductions in yields for agronomic crops
exposed to  ozone, including vegetables (e.g., lettuce) and field crops (e.g., cotton and wheat).
The economic value associated with varying levels of yield loss for ozone-sensitive commodity
crops is analyzed using the AGSIM® agricultural benefits model (Taylor, et al., 1993). AGSIM®
is an econometric-simulation model that is based on a large set of statistically estimated demand
and supply equations for agricultural commodities produced in the United States.

   The model  employs biological exposure-response information derived from controlled
experiments conducted by the NCLAN (NCLAN, 1996). For the purpose of our analysis,  we
analyze changes for the six most economically significant crops for which C-R functions are
available: corn, cotton, peanuts, sorghum, soybean, and winter wheat. For some crops there are
multiple C-R functions, some more  sensitive to ozone and some less. Our base estimate assumes
that crops are evenly mixed between relatively sensitive and relatively insensitive varieties.

   The measure of benefits calculated by the AGSIM® model is the net change in consumer and
producer surplus from baseline ozone concentrations to the ozone concentrations resulting from
emission reductions. Using the baseline and post-control equilibria, the model calculates the
change in net consumer and producer surplus on a crop-by-crop basis. Dollar values are

                                          9-25

-------
Final Regulatory Impact Analysis
aggregated across crops for each standard.  The total dollar value represents a measure of the
change in social welfare associated with changes in ambient ozone.

   9.3.5.3 Other Welfare Benefits

   Ozone also has been shown conclusively to cause discernible injury to forest trees (US EPA,
1996; Fox and Mickler, 1996). In our previous analysis of the HD Engine/Diesel Fuel rule, we
were able to quantify the effects of changes in ozone concentrations on tree growth for a limited
set of species.  Due to data limitations, we were not able to quantify such impacts for this
analysis.

   An additional welfare benefit expected to accrue as a result of reductions in ambient ozone
concentrations in the U.S. is the economic value the public receives from reduced aesthetic
injury to forests. There is sufficient scientific information available to reliably establish that
ambient ozone levels cause visible injury to foliage and impair the growth of some sensitive
plant species (US EPA, 1996c, p. 5-521).  However,  present analytic tools and resources
preclude EPA from quantifying the benefits of improved forest aesthetics.

   Urban ornamentals represent an additional vegetation category likely to experience some
degree of negative effects associated with exposure to ambient ozone levels and likely to impact
large economic sectors. In the absence of adequate exposure-response functions and economic
damage functions for the potential range of effects relevant to these types of vegetation, no direct
quantitative economic benefits analysis has been conducted.

   The nonroad diesel standards, by reducing NOX emissions, will also reduce nitrogen
deposition on agricultural land and forests. There is  some evidence that nitrogen deposition may
have positive effects on agricultural output through passive fertilization. Holding all other
factors constant, farmers' use of purchased fertilizers or manure may increase as deposited
nitrogen is reduced. Estimates of the potential value of this possible increase in the use of
purchased fertilizers are not available, but it is likely that the overall value is very small relative
to other health and welfare effects.

   The nonroad diesel standards are also expected to produce economic benefits in the form of
reduced materials damage.  There are two important categories of these benefits. Household
soiling refers to the accumulation of dirt, dust, and ash on exposed surfaces. Criteria pollutants
also have corrosive effects on commercial/industrial  buildings and structures of cultural and
historical significance. The effects on historic buildings and outdoor works of art are of
particular concern because of the uniqueness and irreplaceability of many of these objects.

   Previous EPA benefit analyses have been able to provide quantitative estimates of household
soiling damage.  Consistent with SAB advice, we determined that the existing data (based on
consumer expenditures from the early 1970's) are too out of date to provide a reliable enough
estimate of current household soiling  damages (EPA-SAB-Council-ADV-003, 1998) to include
in our base estimate.  We calculate household soiling damages in a sensitivity estimate provided
in Appendix 9C.

                                          9-26

-------
                                                                 Cost-Benefit Analysis
   EPA is unable to estimate any benefits to commercial and industrial entities from reduced
materials damage.  Nor is EPA able to estimate the benefits of reductions in PM-related damage
to historic buildings and outdoor works of art.  Existing studies of damage to this latter category
in Sweden (Grosclaude and Soguel, 1994) indicate that these benefits could be an order of
magnitude larger than household soiling benefits.

   Reductions in emissions of diesel  hydrocarbons that result in unpleasant odors may also lead
to improvements in public welfare.  The magnitude of this benefit is very uncertain, however,
Lareau and Rae (1989) found a significant and positive WTP to reduce the number of exposures
to diesel odors. They found that households were on average willing to pay around $20 to $27
(2000$) per year for a reduction of one exposure to intense diesel odors per week (translating
this to a national level, for the approximately 125 million households in 2020, the total WTP
would be between $2.5 and $3.4 billion annually). Their results are not in a form that can be
transferred to the context of this analysis, but the general magnitude of their results suggests this
could be a significant welfare benefit  of the rule.

   The effects of air pollution on the  health and stability of ecosystems are potentially very
important, but are at present poorly understood and difficult to measure.  The reductions in NOX
caused by the rule could produce significant benefits. Excess nutrient loads, especially of
nitrogen, cause a variety of adverse consequences to the health of estuarine and coastal waters.
These effects include toxic and/or noxious algal blooms such as brown and red tides, low
(hypoxic) or zero (anoxic) concentrations of dissolved oxygen in bottom waters, the loss of
submerged aquatic vegetation due to the light-filtering effect of thick algal mats, and
fundamental shifts in phytoplankton community structure (Bricker et al., 1999).

   Direct C-R functions relating changes in nitrogen loadings to changes in estuarine benefits
are not available.  The preferred WTP based measure of benefits depends on the availability of
these C-R functions and on estimates  of the value of environmental responses.  Because neither
appropriate C-R functions nor sufficient information to estimate the marginal value of changes in
water quality exist at present, calculation of a WTP measure is not possible. Likewise, EPA is
unable to quantify climate-change related impacts.

   If better models of ecological effects can be defined, EPA believes that progress can be made
in estimating WTP measures for ecosystem functions. For example, if nitrogen or sulfate
loadings can be linked to measurable  and definable changes in fish  populations  or definable
indexes of biodiversity, then CV studies can be designed to elicit individuals' WTP for changes
in these effects. This is an important  area for further research and analysis, and will require
close collaboration among air quality  modelers, natural scientists, and economists.

9.3.6 Treatment of Uncertainty

   In any complex analysis, there are likely to be many sources of uncertainty.  This analysis is
no exception.  Many inputs are used to derive the final estimate of economic benefits, including
emission inventories, air quality models (with their associated parameters and inputs),
epidemiological estimates of C-R functions, estimates of values, population estimates, income

                                          9-27

-------
Final Regulatory Impact Analysis
estimates, and estimates of the future state of the world (i.e., regulations, technology, and human
behavior).  Some of the key uncertainties in the benefits analysis are presented in Table 9-8. For
some parameters or inputs it may be possible to provide a statistical representation of the
underlying uncertainty distribution.  For other parameters or inputs, the necessary information is
not available.

    In addition to uncertainty, the annual benefit estimates presented in this analysis are also
inherently variable due to the truly random processes that govern pollutant emissions and
ambient air quality in a given year.  Factors such as hours of equipment use and weather display
constant variability regardless of our ability to accurately measure them. As such, the estimates
of annual benefits should be viewed as representative of the magnitude of benefits expected,
rather than the actual benefits that would occur every year.

    We present a primary estimate of the total benefits, based on our interpretation of the best
available scientific literature and methods and supported by the SAB-HES and the NAS (NRC,
2002).     The benefits estimates generated for the final Nonroad Diesel Engine rule are subject
to a number of assumptions and uncertainties, which are discussed throughout the document.
For example, key assumptions underlying the primary estimate for the premature mortality
which accounts for 90 percent of the total benefits we were able to quantify include the
following:

    (1) Inhalation of fine particles is causally associated with premature death at concentrations
       near those experienced by most Americans on a daily basis. Although biological
       mechanisms for this effect have not yet been definitively established, the weight of the
       available epidemiological evidence supports an assumption of causality.
    (2) All fine particles, regardless  of their chemical composition,  are equally potent in causing
       premature mortality. This is an important assumption, because PM produced via
       transported precursors emitted from EGUs may differ significantly from direct PM
       released from diesel engines and  other industrial sources, but no clear scientific grounds
       exist for supporting differential effects estimates by particle type.
    (3) The impact function for fine particles is approximately linear within the range of ambient
       concentrations under consideration. Thus, the estimates include health benefits from
       reducing fine particles in areas with varied concentrations of PM, including both regions
       that are in attainment with fine particle standard and those that do not meet the standard.
    (4) The forecasts for future emissions and associated air quality modeling are valid.
       Although recognizing the difficulties, assumptions,  and inherent uncertainties in the
       overall enterprise, these analyses are based on peer-reviewed scientific literature and
       up-to-date assessment tools,  and we believe the results are highly useful in assessing this
       rule.

    In addition, we provide sensitivity analyses to illustrate the effects of uncertainty about key
analytical assumptions.  Our analysis of the preliminary control options did not include formal
integrated probabilistic uncertainty analyses, although we have conducted several sensitivity
tests based on changes to several key model parameters.  The recent NAS report on estimating
public health benefits of air pollution regulations recommended that EPA begin to move the

                                          9-28

-------
                                                                 Cost-Benefit Analysis
assessment of uncertainties from its ancillary analyses into its primary analyses by conducting
probabilistic, multiple-source uncertainty analyses. We are working to implement these
recommendations.

   In Appendix 9B, we present two types of probabilistic approaches designed to illustrate how
some aspects of the uncertainty in the C-R function could be handled in a PM benefits analysis.
The first approach generates a probabilistic estimate of statistical uncertainty based on standard
errors reported in the underlying studies used in the benefit modeling framework. In the second
illustrative approach, EPA, in collaboration with OMB, conducted a pilot expert  elicitation to
characterize uncertainties in the relationship between ambient PM2 5  and premature mortality
(TEc 2004).  This pilot was designed to improve our understanding of the design  and application
of expert elicitation methods to economic benefits analysis. For instance, the pilot was designed
to provide feedback on the efficacy of the protocol developed and the analytic challenges, as
well as to provide insight regarding potential implications of the results on the degree of
uncertainty surrounding the C-R function for PM2 5 mortality. The scope of the pilot was limited
in that we focused the elicitation on the C-R function of PM mass rather than on  individual
issues surrounding an estimate of the change in premature mortality due to PM exposure. In
Appendix 9B we present sensitivity analyses for illustrative purposes.

9.3.7 Model Results

   We summarize our preliminary control option modeling as background for calculating the
scaling factors.  The scaling factors are then used to estimate the PM-related benefits of the final
rule. Insights into ozone impacts can also be discerned.  As discussed in Table 9-5 above and
Table 9A-4 below, full implementation of the modeled preliminary control  options is projected
in 2020 to reduce 48-state emissions of land-based nonroad NOx by  663,600 tons (58 percent of
base case), SO2 by 305,000 tons (98.9 percent), VOC by 23,200 tons (24 percent) and directly
emitted PM25 by 91,300 tons (71 percent). In 2030, the modeled preliminary control option is
expected to reduce 48-state emissions of NOx by 1 million tons (82 percent), SO2 by 359,800
tons (99.7 percent), VOC by 34,000 tons (35 percent) and direct PM by 138,000  tons (90
percent).

   Based on these projected emission changes, REMSAD modeling results indicate the
pollution controls generate greater absolute air quality improvements in more populated, urban
areas. The rule will reduce average annual mean concentrations of PM25 across the U.S. by
roughly 2.5 percent (or 0.2 |ig/m3) and 3.4 percent (or 0.28 |ig/m3) in 2020 and 2030,
respectively.  The population-weighted average mean concentration declined by  3.3 percent (or
0.42 |ig/m3) in 2020 and 4.5 percent (or 0.59 |ig/m3) in 2030, which is much larger in absolute
terms than the spatial average for both years. Table 9-9 presents information on  the distribution
of modeled reductions in ambient PM concentrations across populations in the U.S.  By 2030,
slightly over 50 percent of U.S. populations  will  live in areas with reductions of greater than 0.5
|ig/m3. This information indicates how widespread the improvements in PM air quality are
expected to be.
                                          9-29

-------
Final Regulatory Impact Analysis
   Applying the health impact functions described in Table 9-5 to the estimated changes in
PM25 and ozone from the preliminary modeling yields estimates of the number of avoided
incidences for each health outcome.  These estimates are presented in Appendix A Table 9A-30
for the 2020 and 2030 model analysis years. To provide estimates of the monetized benefits of
the reductions in PM-related health outcomes described in Table 9A-30, we multiply the point
estimates of avoided incidences by unit values.  Values for welfare effects are based on
application of the economic models described above. The estimated total monetized health and
welfare benefits for the preliminary modeled scenario are also presented in Appendix A in Table
9A-31.

   The largest monetized health benefit is associated with reductions in the risk of premature
mortality, which accounts for 90 percent of total monetized health benefits.   The next largest
benefit is for chronic illness reductions (chronic bronchitis and nonfatal heart attacks), although
this value is more than an order of magnitude lower than for premature mortality. Minor
restricted activity days, work loss days, and hospital admissions account for the majority of the
remaining benefits. While the other categories account for less than $100 million each, they
represent a  large number of avoided incidences affecting many individuals.

   Ozone benefits arising from this rule are in aggregate positive for the nation.  However, due
to ozone increases occurring during certain hours of the day in some urban areas, in 2020 the net
effect is an  increase in ozone-related minor restricted activity days (MRAD), which are related to
changes in daily average ozone (which includes hours during which ozone levels are low, but are
increased relative to the baseline based on the preliminary modeling). However, by 2030, there
is a net decrease in ozone-related MRAD consistent with widespread reductions in ozone
concentrations from the increased NOx emissions reductions.  Note that in both years, the overall
impact of changes in both PM and ozone is a large decrease in the number of MRAD.  Overall,
ozone benefits are low relative to PM benefits for similar endpoint categories because of the
increases in ozone concentrations during  some hours of some days in certain urban areas. For  a
more complete discussion of this issue, see Chapter 2.

   Monetized and quantified welfare benefits are far outweighed by health benefits.  However,
we have not been able to quantify some important welfare categories, including the value of
changes in ecosystems from reduced deposition of nitrogen and sulfur and climate impacts.  The
welfare benefits we are able to quantify are dominated by the value of improved visibility.
Visibility benefits just in the limited set of parks included in the monetized total benefit estimate
are over $1.6 billion in 2030. Agricultural benefits, while small relative to visibility benefits, are
significant relative to ozone-related health benefits, representing the largest single benefit
category for ozone.
                                          9-30

-------
                                                    Cost-Benefit Analysis
                           Table 9-6
Endpoints and Studies Used to Calculate Total Monetized Health Benefits
Endpoint
Pollutant
Applied
Population
Source of Effect Estimate(s)
Source of Baseline
Incidence
Premature Mortality
Adults - Long-term
exposure
Infants
PM25
PM25
>29 years
<1
Pope, et al. (2002)
Woodruff etal. (1997)
CDC Wonder (1996-1998)
CDC Wonder (1996-1998)
Chronic Illness
Chronic Bronchitis
Non-fatal Heart
Attacks
PM25
PM25
> 26 years
Adults
Abbey, etal. ( 1995)
Peters etal. (2001)
1999 HIS (American Lung
Association, 2002b, Table
4); Abbey etal. (1993,
Table 3)
1999 NHDS public use
data files; adjusted by 0.93
for prob. of surviving after
28 days (Rosamond et al.,
1999)
Hospital Admissions
Respiratory
Cardiovascular
03
03
PM25
PM25
PM25
PM25
PM25
> 64 years
< 2 years
>64 years
20-64 years
> 64 years
< 65 years
> 64 years
Pooled estimate:
Schwartz (1995) - ICD 460-519
(all resp)
Schwartz (1994a, 1994b) - ICD
480-486 (pneumonia)
Moolgavkar et al. (1997) - ICD
480-487 (pneumonia)
Schwartz (1994b) - ICD 491-
492, 494-496 (COPD)
Moolgavkar et al. (1997) - ICD
490-496 (COPD)
Burnett etal. (2001)
Pooled estimate:
Moolgavkar (2003) - ICD 490-
496 (COPD)
Ito (2003) - ICD 490-496
(COPD)
Moolgavkar (2000) - ICD 490-
496 (COPD)
Ito (2003) - ICD 480-486
(pneumonia)
Sheppard, et al. (2003) - ICD
493 (asthma)
Pooled estimate:
1999 NHDS public use
data files
1999 NHDS public use
data files
1999 NHDS public use
data files
1999 NHDS public use
data files
1999 NHDS public use
data files
1999 NHDS public use
data files
1999 NHDS public use
                             9-31

-------
                                              Table 9-6
               Endpoints and Studies Used to Calculate Total Monetized Health Benefits
Endpoint

Asthma-Related ER
Visits
Pollutant
PM25
03
PM25
Applied
Population
20-64 years
All ages
0-18 years
Source of Effect Estimate(s)
Moolgavkar (2000) - ICD 390-
429 (all cardiovascular)
Pooled estimate: Weisel et al.
(1995), Cody etal. (1992),
Stiebetal. (1996)
Norrisetal. (1999)
Source of Baseline
Incidence
1999 NHDS public use
data files
2000 NHAMCS public use
data files3; 1999 NHDS
public use data files
2000 NHAMCS public use
data files; 1999 NHDS
public use data files
Other Health Endpoints
Acute Bronchitis
Asthma Exacerbations
Upper Respiratory
Symptoms
Lower Respiratory
Symptoms
Work Loss Days
School Absence Days
Worker Productivity
Minor Restricted
Activity Days
PM25
PM25
PM10
PM25
PM25
03
03
PM25, 03
8-12 years
6-18yearsA
Asthmatics, 9-
1 1 years
7-14 years
18-65 years
9-10 years
6-11 years
Outdoor
workers, 18-65
18-65 years
Dockery etal. (1996)
Pooled estimate:
Ostro etal. (2001) Cough
Ostro et al. (2001) Wheeze
Ostro et al. (2001) Shortness of
breath
Vedal etal. (1998) Cough
Pope etal. (1991)
Schwartz and Neas (2000)
Ostro (1987)
Pooled estimate:
Gilliland etal. (2001)
Chen et al. (2000)
Crocker and Horst (1981) and
U.S. EPA (1984)
Ostro and Rothschild (1989)
American Lung
Association (2002a, Table
11)
Ostro etal. (2001)
Vedal etal. (1998)
Pope etal. (1991, Table 2)
Schwartz (1994, Table 2)
1996 HIS (Adams etal.,
1999, Table 41); U.S.
Bureau of the Census
(2000)
National Center for
Education Statistics (1996)
NA
Ostro and Rothschild
(1989, p. 243)
A The original study populations were 8-13 for the Ostro et al. (2001) study and 6-13 for the Vedal et al. (1998)
study. Based on advice from the SAB-HES and NRC, we have extended the applied population to 6-18, reflecting
the common biological basis for the effect in children in the broader age group.

-------
o
o
o
 VI
•^
^a

°o
 a
•a
 O

 =
 o
 =
 o
•a
 v
 VI
 S

13
r-

a\












OI
s
o
3
3
H
08
55
£

cw
3

C3
>
h- 1
O
i

1
—
g
a
OI
J



























a
c
u
c
"e
1
OI
E
cj "a
C >
^ 4
O H
S
OI
o _
CJ 0
a -
HH 0
O H
(S
o

OI

O _
u a
4 -
fl S
1 H





— c

i- c
w "^
HH ^
f*






U I ^

| || 2l'u
? p] qj '55 2^
§! ^ii-s
^ § j^>=s
o 13 U S ^ §
i-H 1) ^ 1 "^ H
^ ^ ^ _H
in ' ci c o f+-*
,-,, f/3 si C .LH n
Point estimate is the mean of a normal distribution with a 5
interval between $1 and $10 million. Confidence interval i
analyses of the wage-risk VSL literature. $1 million repress
the interquartile range from the Mrozek and Taylor (2000)
million represents the upper end of the interquartile range f
Aldy (2003) meta-analysis. The VSL represents the value <
mortality risk aggregated over the affected population.

o
o^
I
a?
0
o
o"
o

*o"
a*5


0
o
0
o"
o
a§



ra
o
•i

-— '
-S"1
• ^
£-
0 &
s3
11
§ «
a •&
ft! OO
, -^ p_j







Vi  w ^ w
_ O ^ -M -^ 5P £2
-2 T3 M 'o ^3.5 o
PH -8 ^ '-§ 1 § -2

o^ ^
c^ £H
C^ °
cw
"^ H
"S ^
Point estimate is the mean of a generated distribution of W
pollution-related CB. WTP to avoid a case of pollution-reL
adjusting WTP (as described in Viscusi et al., 1991) to avoi
for the difference in severity and taking into account the el;
respect to severity of CB.
Age specific cost-of-illness values reflecting lost earnings i
costs over a 5 year period following a non-fatal MI. Lost e
based on Cropper and Krupnick (1990). Direct medical co:
average of estimates from Russell et al. (1998) and Wittels
Lost earnings:
Cropper and Krupnick (1990). Present discounted value of
earnings:

o
§
o"
CO
3

O
o
S

^T




o
r,
°
S

-e
ra
a
0
1
m a
U l~~l o
^T' ^3 ta
• i-H *T3
s ^ §
't—' O O
2 f .2
P3  m ^ o
*o ^r oo £2 *o
*o t^^ t^^ ^^ *o
— *„

 <^
^o" ^" oo 2 ^o"

^ft ^5 Sft ^ tft




(N vo ^T S  m ^A O
(^ *O OO 
O
^r ^r ^ 1?
T^- T^J- iy-j i,Q S
§ jn ^ jjij ^g
OD &0 &0 &0 OD
 jn
C*l ' — ' OO ^
in ro *o P?

a^ a^ a^ r^




m ON ,-H 2
ON ^r i> jn
(S) i — i OO <*
m" ro" >o" P^
aeaeae^







H
2

"S
o ^r ^ in ^o
O r^l i i i
M 1 1 m m m
^ o 
O
•3
erf
S
^ 00 00 00 00 00
i^ -^ -^ -^ -^


<;


u

.3
"S
0
o











































-------
•a
 O)
 a
 o
o
o
o
 K
H-^
^a

'o
 a
•a
 O

 =
 o
 =
 o
 S

13
i^

a\
 a>












CD
~03
a
"&»
Derivation of E:





0>
CJ
g
•a
=
HH

(J
•c
"03
55

1
^
"3

t|
o
"03
S


w
—
1

o
U
o>
S
11

f)
o
(S


o>
E
o —
hH J
(S
o
(S


o>
i

S "33
L^ ^
J
ON
ON






- "S
s '3

O "O

w






























































•5

i
1
1
'E.
o
w
<^ CJ
Q^|
1— 1 o tn
OX) P
^ r^H ^ O
o 1) 2o &0
a of S ff
^ (D . i '^
' 	 N > ^H .

M •> C71 S
O i • o ^

, — O (U ^_-
.!=) oj S o
"^ c3 Q 0
a J^ PH
tes (lost earnings plus direct
mation (e.g., average hospita
id weighted share of total CC
Ithcare Research and Quality
I £ § 8
'-£ rt >-, "
Q cd t-H
D i— ( +3 O
OJ cw (H_^
ill!


00
m
^"
a^






00
r^
a?






oo

f^
r^
a'j


ca
Q
^ NO
EH ON
o ^r
J 4
£ Tj-

i) rs

•-S ^r
CH <3\
ifi *^
O jg
o Sb Q
5 R O
0 <_-»
ON « ON s
Q o Q 13
Q ft Q i)
^ 0 ^ M o "g
. (3)2 w •> ox) }i^
^^u§j&p So^'^o
_Ql)'^H o D ^ &C
| f|U § HI*
' — ^ ^* ?n < ' — ^ >• ' ^ •


O -£ <3 ^ O -£ O ?*
O^OO OfflOOj^
•^§.30 "38"^'~'
.y oj rt S .y ; "S^S^
a^H^-^H a^H_s
tes (lost earnings plus direct
mation (e.g., average hospita
id weighted share of total pm
[ealthcare Research and Qua!
tes (lost earnings plus direct
mation (e.g., average hospita
id weighted share of total astl
Ithcare Research and Quality
ra i-H H LC Co i-H H Co
G O Co G O Co qj
"^ d >^ (_2 "-^ 3 "^
§J !— i -3 -^ §J !— i 3 Q

I II! 111!


s S
2 ^o"
^ft






§ S
^ ^o"
s ^^







c^ ^
^ ^0
•^f ^
s ^










|L_ CW
OO ¥*
7 1
00 I
^ 'a
.3 S -6
I's a
3 Q ^
go |
PH 0 <
ON ^
Q ,_, ^
O f+-* M ^
^ ° " ^
§•32^
"8 ^^ 1
" ^H ^ 1
_o i) £r*
u ^ ^o"
3 u <£ §
-^ S ^ fN
^ cs o
M ^ tH >^
o -J2 ca .-H

"3 8 3 3
.y i) c§ CX
13 c3 o "^
H ^^ 13 ra
tes (lost earnings plus direct
mation (e.g., average hospita
id weighted share of total car
ncy for Healthcare Research
a <° « OD
'-& s >^ ^C
§J !— i -3 CH
0) w ---1
O> i— i T3
1) rf 0)
lit!
H o ^ c


00
oo"
a^






00
m
oo"
a?







00
fYl.
00
a?










ON

"3 ^
^ ON
O 1)
11

ml o
< 0
















; of two unit COI values:
m Smith etal., 1997, and
m Stanford et al., 1999.
!<§<§
1-4 r, r.
> ^ ^n

"—i O
^^^


00
a4







NO
00
a4








NO
00
a4




a
1
c3


-------
•a
 v
 =
 o
o
o
o
 K
•^
^a

'o
 a
•a
 O

 =
 o
 =
 o
 W
UJ
 s

13
i^

a\
 a>


O)
•
o
o>
£
03
S
!fl
S
t
"a
o
U















Derivation of Estimates
O)
S
11
O H-J
f)
O


o
S
O i 	 i
CJ O>
s %
s
o


O)
s
O —
u 13
Jj
ON

^H





•*^
Si
03 C.
>OH T3




























fl
• 2
cS
N
ft
o
ffi
too

(S
o



'^
E?
.2

1

Combinations of the 3 symptoms for which WTP estimates are available that
closely match those listed by Pope, et al. result in 7 different "symptom
clusters," each describing a "type" of URS. A dollar value was derived for each
type of URS, using mid-range estimates of WTP (lEc, 1994) to avoid each
symptom in the cluster and assuming additivity of WTPs. The dollar value for
URS is the average of the dollar values for the 7 different types of URS.


CS









£






(N
ae




f 	 ^
GO
g
t/3
1
O-
a
GO
&>
O
2
8*
&
ft
ft

Combinations of the 4 symptoms for which WTP estimates are available that
closely match those listed by Schwartz, et al. result in 1 1 different "symptom
clusters," each describing a "type" of LRS. A dollar value was derived for each
type of LRS, using mid-range estimates of WTP (lEc, 1994) to avoid each
symptom in the cluster and assuming additivity of WTPs. The dollar value for
LRS is the average of the dollar values for the 1 1 different types of LRS.


£









£>






vo
ae




^ 	 ^
GO
Pi
1
Q
ft
a
GO
E?
o
2
'ft

-------
•a
 a>
 =
 o
o
o
o
 K
•^
^a

°o
 a
•a
 O

 =
 o
 =
 o
 w
UJ
 s

13
i^

a\
 a>
Derivation of Estimates


u o
§ 1-
•a o 13
'3 fi «
g O hj

;j  o
o
03

•g g
- 8-3
L ^^ ^
S3 ^^
a> cy\
U ^








sl
03 e





3
.2
>





i)
i
•c
>









Q
g
£_-
>>
Q
M
o
0

Based on expected lost wages from parent staying home with child. Estimated
daily lost wage (if a mother must stay at home with a sick child) is based on the
median weekly wage among women age 25 and older in 2000 (U.S. Census
Bureau, Statistical Abstract of the United States: 2001, Section 12: Labor
Force, Employment, and Earnings, Table No. 621). Thismedian wage is $551.
Dividing by 5 gives an estimated median daily wage of $103.
The expected loss in wages due to a day of school absence in which the mother
would have to stay home with her child is estimated as the probability that the
mother is in the workforce times the daily wage she would lose if she missed a
day = 72.85% of $103, or $75.



iC






iC






m
rf*.










m
~&
Q
i)
o

o
<
"o
O
o
GO
Based on $68 - median daily earnings of workers in farming, forestry and
fishing - from Table 621, Statistical Abstract of the United States ("Full-Time
Wage and Salary Workers - Number and Earnings: 1985 to 2000") (Source of
data in table: U.S. Bureau of Labor Statistics, Bulletin 2307 and Employment
and Earnings, monthly).
v° W
0s fl
^ ° o
p.i!

o jg bo CL,
o j|
^ o
x9 flj
^ §
S3 ' — N >-,

0 J£ W> ft
o S
:> r^H
I> O

^ S
^00
!ilt
°^ l-c u ^
O _^ 00 ft
0 ^
Z* o









>,
•^
o
•3
o
^H
1
0

Median WTP estimate to avoid one MRAD from Tolley, et al. (1986) .



ft






s






^H
^y-.








1
Q
'f
"5
^!
1
.0
-P
1) "M
<& Q
§ ^
11

-------
                                                                                 Cost-Benefit Analysis
                                                 Table 9-8
                       Primary Sources of Uncertainty in the Benefit Analysis
1.  Uncertainties Associated With Health Impact Functions
    The value of the ozone or PM effect estimate in each health impact function.
    Application of a single effect estimate to pollutant changes and populations in all locations.
    Similarity of future year effect estimates to current effect estimates.
    Correct functional form of each impact function.
    Extrapolation of effect estimates beyond the range of ozone or PM concentrations observed in the study.
    Application of effect estimates only to those subpopulations matching the original study population.
2.  Uncertainties Associated With Ozone and PM Concentrations
    Responsiveness of the models to changes in precursor emissions resulting from the control policy.
    Projections of future levels of precursor emissions, especially ammonia and crustal materials.
    Model chemistry for the formation of ambient nitrate concentrations.
    Lack of ozone monitors in rural areas requires extrapolation of observed ozone data from urban to rural areas.
    Use of separate air quality models for ozone and PM does not allow for a fully integrated analysis of pollutants and
         their interactions.
    Full ozone season air quality distributions are extrapolated from a limited number of simulation days.
    Comparison of model predictions of particulate nitrate with observed rural monitored nitrate levels indicates that
    REMSAD overpredicts nitrate in some parts of the Eastern US and underpredicts nitrate in parts of the Western
    US.
3.  Uncertainties Associated with PM Premature mortality Risk
    No scientific literature supporting a direct biological mechanism for observed epidemiological evidence.
    Direct causal agents within the complex mixture of PM have not been identified.
    The extent to which adverse health effects are associated with low level exposures that occur many times in the
    year versus peak exposures.
    The extent to which effects reported in the long-term exposure studies are associated with historically higher levels
    of PM rather than the levels occurring during the period of study.
    Reliability of the limited ambient PM25 monitoring data in reflecting actual PM25 exposures.
4.  Uncertainties Associated With Possible Lagged Effects
—   The portion of the PM-related long-term exposure mortality effects associated with changes in annual PM levels
    would occur in a single year is uncertain as well as the portion that might occur in subsequent years.
5.  Uncertainties Associated With Baseline Incidence Rates
—   Some baseline incidence rates are not location-specific (e.g., those taken from studies) and may therefore not
    accurately represent the actual location-specific rates.
—   Current baseline incidence rates may not approximate well baseline incidence rates in 2030.
—   Projected population and demographics may not represent well future-year population and demographics.
6.  Uncertainties Associated With Economic Valuation
    Unit dollar values associated with health and welfare endpoints are only estimates of mean WTP and therefore
    have uncertainty surrounding them.
    Mean WTP (in constant dollars) for each type of risk reduction may differ from current estimates due to
    differences in income or other factors.
    Future markets for agricultural products are uncertain.
7.  Uncertainties Associated With Aggregation of Monetized Benefits
—   Health and welfare benefits estimates are limited to the available effect estimates.  Thus, unquantified or
    unmonetized benefits are not included.
                                                    9-37

-------
Final Regulatory Impact Analysis
                                         Table 9-9
                 Distribution of PM2 5 Air Quality Improvements Over Population
                 Due to Nonroad Engine/Diesel Fuel Standards a in 2020 and 2030
Change in Annual Mean PM2 5
Concentrations (fig/m3)
0 < A PM25 Cone < 0.25
0.25 < A PM25 Cone < 0.5
0.51.75
2020 Population
Number (millions) Percent (%)
65.11
184.52
56.66
14.60
5.29
3.51
0
0
19.75%
55.97%
17.19%
4.43%
1.60%
1.06%
0.00%
0.00%
2030 Population
Number (millions) Percent (%)
28.60
147.09
107.47
38.50
8.82
15.52
5.70
4.19
8.04%
41.33%
30.20%
10.82%
2.48%
4.36%
1.60%
1.18%
 1 The change is defined as the control case value minus the base case value. The results reflect the modeling for the preliminary
 control option, not the final rule.
9.3.8 Apportionment of Benefits to NOx, SO2, and Direct PM Emissions Reductions

   As noted in the introduction to this chapter, the standards we are finalizing in this rule differ
from those that we used in modeling air quality and economic benefits.  As such, it is necessary
for us to scale the modeled benefits to reflect the difference in emissions reductions between the
final and preliminary modeled standards.  In order to do so, however, we must first apportion
total benefits to the NOx, SO2, and direct PM reductions for the modeled preliminary control
options.  This apportionment is necessary due to the differential contribution of each emission
species to the total change in ambient PM and total benefits. We do not attempt to develop
scaling factors for ozone benefits because of the difficulty in separating the contribution of NOx
and NMHC/VOC reductions to the change in ozone concentrations.

   As discussed in detail in Chapter 2, PM is a complex mixture of particles of varying species,
including nitrates, sulfates, and primary particles, including organic and elemental carbon.
These particles are formed in complex chemical reactions from emissions of precursor
pollutants, including NOx, S02, ammonia,  hydrocarbons, and directly emitted particles.  Different
emissions species contribute to the formation of PM in different amounts, so that a ton of
emissions of NOx contributes to total ambient PM mass differently than a ton of SO2 or directly
emitted PM. As such, it is inappropriate to scale benefits  by simply scaling the sum of all
precursor emissions. A more appropriate scaling method  is to first apportion total PM benefits to
the changes in underlying emission species and then scale the apportioned benefits.

   PM formation relative to any particular reduction in an emission species is a highly nonlinear
process, depending on meteorological conditions and baseline conditions, including the amount
                                           9-38

-------
                                                                 Cost-Benefit Analysis
of available ammonia to form ammonium nitrate and ammonium sulfate. Given the limited air
quality modeling conducted for this analysis, we make several simplifying assumptions about the
contributions of emissions reductions for specific species to changes in particulate species. For
this exercise, we assume that changes in sulfate particles are attributable to changes in SO2
emissions, changes in nitrate particles are attributable to changes in NOx emissions, and changes
in primary PM are attributable to changes in direct PM emissions.  These assumptions essentially
assume independence between SO2, NOx, and direct PM in the formation of ambient PM. This
is a reasonable assumption for direct PM, as it is generally not reactive in the atmosphere.
However, SO2 and NOx emissions interact with other compounds in the atmosphere to form
PM2 5.  For example, ammonia reacts with SO2 first to form ammonium sulfate. If there is
remaining ammonia, it reacts with NOx to form ammonium nitrate. When S02 alone is reduced,
ammonia is freed to react with any NOx that has not been used in forming ammonium nitrate. If
NOx is also reduced,  then there will be less available NOx to form ammonium nitrate from the
newly available ammonia. Thus, reducing SO2 can potentially lead to decreased ammonium
sulfate and increased nitrate, so that overall ambient PM benefits are less than the reduction in
sulfate particles. If NOx alone is reduced, there will be a direct reduction in ammonium nitrate,
although the amount of reduction depends on whether an area is ammonia limited.  If there is not
enough ammonia in an area to use up all of the available NOx, then NOx reductions will only
have an impact if they reduce emissions to the point where ammonium nitrate formation will  be
affected.  NOx reductions will not result in any offsetting increases in ambient PM under most
conditions. The implications of this for apportioning benefits between NOx, SO2, and direct PM
is that some of the sulfate-related benefits will be offset by reductions in nitrate benefits, so
benefits from SO2 reductions will be overstated, while NOx benefits will be understated. It is
not immediately apparent the size of this bias.

    The measure of change in ambient particulate mass that is most related to health benefits is
the population-weighted change in PM2 5 i-ig/m3, because health benefits are driven both by the
size of the change in PM25 and the populations exposed to that  change.  We calculate the
proportional share of total change in mass accounted for by nitrate, sulfate, and primary
particles. Results of these calculations for the 2020  and 2030 REMSAD modeling analysis are
presented in Table 9-10.  The sulfate percentage of total change is used to represent the S02
contribution to health benefits, the nitrate percentage is used to represent the NOx contribution to
health benefits, and the primary PM percentage is used to represent the direct PM contribution to
health benefits. These percentages will be applied to the PM-related health benefits estimates in
Appendix A in Tables 9A-30 and 9A-31 and combined with the emission scaling factors
developed in section 9.2 to estimate benefits for the  final set of standards.
                                          9-39

-------
Final Regulatory Impact Analysis
 Table 9-10.  Apportionment of Modeled Preliminary Control Option Population-weighted
           Change in Ambient PM2_s to Nitrate, Sulfate, and Primary Particles
2020 2030

Total PM2 5
Sulfate
Nitrate
Primary PM
Population- Percent of Total Population-
weighted Change weighted
Change (|ig/m3) Change (|ig/m3)
0.316 -- 0.438
0.071 22.5% 0.090
0.041 13.1% 0.073
0.203 64.4% 0.274
Percent of Total
Change
—
20.5%
16.8%
62.7%
   Visibility benefits are highly specific to the parks at which visibility improvement occur,
rather than where populations live.  As such, it is necessary to scale benefits at each individual
park and then aggregate to total scaled visibility benefits.  We apportion benefits at each park
using the contribution of changes in sulfates, nitrates, and primary particles to changes in light
extinction. The change in light extinction at each park is determined by the following equation
(Sisler, 1996):

AjSar = [3F(rh) * 1.375 * MSO4\ + [3F(rh) * 1.29 * A/WO3] + 10 * &PEC + 4 * MOA + kPMFINE + 0.6 * t^PMCOARSE

where rh is relative humidity, ATSO4 is the change in particulate sulfate,  APNO3 is the change
in particulate nitrate,  APEC is the change in primary elemental carbon, ATOA is the change in
total organic aerosols, APMFINE is the change in primary fine particles, and APMCOARSE is
the change in primary coarse particles.

The proportion of the total change in light extinction associated with changes in sulfate particles
is [3F(rh)  * 1.375 * A7SO4]/A/3sa-r .  The proportion of the total change in light extinction

associated with changes in nitrate particles is \^>F(rh) * 1.29 * APNO3\/ApEXT   Finally, the
proportion of the total change in light extinction associated with the change in directly emitted
particles is [l 0 * APEC + 4* A TOA + APMFINE + 0.6 * APMCOARSE]/A fi^ .

We calculate these proportions for each park to apportion park specific benefits  between S02,
NOx, and  PM. The apportioned benefits are then scaled using the emission ratios in Table 9-5.
Park specific apportionment of benefits is detailed in Appendix 9D.
                                          9-40

-------
                                                                Cost-Benefit Analysis
9.4 Estimated Benefits of Final Nonroad Diesel Engine Standards in 2020
and 2030

   To estimate the benefits of the NOx, SO2, and direct PM emission reductions from the
nonroad diesel engine standards in 2020 and 2030, we apply the emissions scaling factors
derived in section 9.2 and the apportionment factors described in section 9.3 to the benefits
estimates for 2020 and 2030 listed in Tables 9A-30 and 9A-3 1 . Note that we apply scaling and
apportionment factors only to PM and visibility related endpoints. Ozone related health and
welfare benefits are not estimated for the emissions reductions associated with the final
standards for reasons noted in the introduction to this chapter.

   The scaled avoided incidence estimate for any particular health endpoint is calculated using
the following equation:
Scaled Incidence = Modeled Incidence * 2_,
where Ri is the emissions ratio for emission species i from Table 9-4, and Aj is the health
benefits apportionment factor for emission species i, from Table 9-10. Essentially, benefits are
scaled using a weighted average of the species specific emissions ratios.  For example, the
calculation of the avoided incidence of premature mortality for the base estimate in 2020 is:

Scaled Premature Mortality Incidence = 7,821 * (0.759*0.131 + 0.800*0.225 + 0.869*0.644) =
6,562 (rounded to 6,600)

The monetized value for each endpoint is then obtained simply by multiplying the scaled
incidence estimate by the appropriate unit value in Table 9-6. The estimated changes in
incidence of health effects in 2020 and 2030 for the final rule based on application of the
weighted scaling factors are presented in Table 9-11. The estimated monetized benefits for both
PM health and visibility benefits are presented in Table 9-12. The visibility benefits are based
on application of the weighted scaling factors for visibility at each Class I area in the Chestnut
and Rowe study regions, aggregated to a national total for each year.
                                          9-41

-------
Final  Regulatory Impact Analysis
                                         Table 9-11.
      Reductions in Incidence of PM-related Adverse Health Effects Associated with
           the Final Full Program of Nonroad Diesel Engine and Fuel Standards
Endpoint
Premature mortality8: Long-term exposure (adults, 30 and over)
Infant mortality (infants under one year)
Chronic bronchitis (adults, 26 and over)
Non-fatal myocardial infarctions (adults, 18 and older)
Hospital admissions - Respiratory (adults, 20 and older)0
Hospital admissions - Cardiovascular (adults, 20 and older)D
Emergency Room Visits for Asthma (18 and younger)
Acute bronchitis (children, 8-12)
Asthma exacerbations (asthmatic children, 6-18)
Lower respiratory symptoms (children, 7-14)
Upper respiratory symptoms (asthmatic children, 9-11)
Work loss days (adults, 18-65)
Minor restricted activity days (adults, age 18-65)
Avoided IncidenceA
(cases/year)
2020
6,400
15
3,500
8,700
2,800
2,300
3,800
8,400
120,000
99,000
76,000
670,000
3,900,000
2030
12,000
22
5,600
15,000
5,100
3,800
6,000
13,000
200,000
160,000
120,000
1,000,000
5,900,000
A Incidences are rounded to two significant digits.
B Premature mortality associated with ozone is not separately included in this analysis
c Respiratory hospital admissions for PM includes admissions for COPD, pneumonia, and asthma.
D Cardiovascular hospital admissions for PM includes total cardiovascular and subcategories for ischemic heart
disease, dysrhythmias, and heart failure.
                                             9-42

-------
                                                                                       Cost-Benefit Analysis
                  Table 9-12. Results of PM Human Health and Welfare Benefits
     Valuation for the Final Full Program of Nonroad Diesel Engine and Fuel Standards
Endpoint
Premature mortality0: (adults, 30 and over)
3% discount rate
7% discount rate
Infant mortality (infants under one year)
Chronic bronchitis (adults, 26 and over)
Non-fatal myocardial infarctions0
3% discount rate
7% discount rate
Hospital Admissions from Respiratory CausesE
Hospital Admissions from Cardiovascular CausesF
Emergency Room Visits for Asthma
Acute bronchitis (children, 8-12)
Asthma exacerbations (asthmatic children, 6-18)
Lower respiratory symptoms (children, 7-14)
Upper respiratory symptoms (asthmatic children, 9-11)
Work loss days (adults, 18-65)
Minor restricted activity days (adults, age 18-65)
Recreational visibility (86 Class I Areas)
Monetized Total0
3% discount rate
7% discount rate
Monetary BenefitsA>B
(millions 2000$, Adjusted for Income
Growth)
2020

$40,000
$38,000
$960
$1,500

$740
$720
$49
$50
$1.0
$3.2
$5.7
$1.7
$2.0
$91
$210
$1,000
$44,000+B
$42,000+B
2030

$77,000
$72,000
$150
$2,400

$1,200
$1,200
$92
$83
$1.7
$5.1
$9.2
$2.7
$3.2
$130
$320
$1,700
$83,000+B
$78,000+B
 Monetary benefits are rounded to two significant digits.
B Monetary benefits are adjusted to account for growth in real GDP per capita between 1990 and the analysis year (2020 or 2030).
c Valuation of base estimate assumes discounting over the distributed lag structure described earlier. Results reflect the use of 3% and 7%
discount rates consistent with EPA and OMB's guidelines for preparing economic analyses (US EPA, 2000c, OMB Circular A-4).
D Estimates assume costs of illness and lost earnings in later life years are discounted using either 3 or 7 percent
E Respiratory hospital admissions for PM includes admissions for COPD, pneumonia, and asthma.
F Cardiovascular hospital admissions for PM includes total cardiovascular and subcategories for ischemic heart disease, dysrhythmias, and heart
failure.
0 B represents the monetary value of the unmonetized health and welfare benefits. A detailed listing of unquantified PM, ozone, CO, and NMHC
related health effects is provided in Table 9-1. These estimates do not include the benefits of reduced sulfur in home heating oil or benefits in Alaska or
Hawaii.
                                                        9-43

-------
Final Regulatory Impact Analysis
   We also evaluated the benefits of the NOx, SO2, and direct PM emission reductions from the
nonroad diesel engine standards in 2020 and 2030 of the fuel-only portions of the program.
Accordingly, we applied the benefits transfer methods to calculate similar results for the fuel
only portion of the program and the 500 ppm NRLM program. Because there would be no NOx
or NMHC reductions for the fuel-only components of the rule, the benefits transfer technique
may have more uncertainty in this application compared to the full program.  As discussed
above, we apply  scaling and apportionment factors only to PM health and visibility related
endpoints.  Toxics and ozone-related health and welfare benefits are not estimated for the
emissions reductions associated with the final standards for reasons noted in the introduction to
this chapter.

   The estimated changes in incidence of health effects in 2020 and 2030 for the fuel-only
components of the final rule based on application of the weighted scaling factors are presented in
Table 9-13. The estimated monetized benefits for both PM health and visibility benefits are
presented in Table 9-14. As described above, the visibility benefits are based on application of
the weighted scaling factors for visibility at each Class I area in the Chestnut and Rowe study
regions, aggregated to a national total for each year.
                                          9-44

-------
                                          Table 9-13.
                       Reductions in Incidence of PM-related Adverse
   Health Effects Associated with the Final Fuel-Related Components of Nonroad Diesel
                                            Standards
Endpoint
Premature mortality8: Long-term exposure (adults, 30
and over)
Infant mortality (infants under one year)
Chronic bronchitis (adults, 26 and over)
Non-fatal myocardial infarctions (adults, 18 and older)
Hospital admissions - Respiratory (adults, 20 and older)0
Hospital admissions - Cardiovascular (adults, 20 and
older)D
Emergency Room Visits for Asthma (18 and younger)
Acute bronchitis (children, 8-12)
Asthma exacerbations (asthmatic children, 6-18)
Lower respiratory symptoms (children, 7-14)
Upper respiratory symptoms (asthmatic children, 9-11)
Work loss days (adults, 18-65)
Minor restricted activity days (adults, age 18-65)
Avoided IncidenceA
(cases/year)
Fuel Only Program
2020
2,700
<10
1,500
3,600
1,200
900
1,600
3,500
51,000
41,000
31,000
280,000
1,600,000
2030
4,000
<10
1,900
5,200
1,700
1,300
2,000
4,600
68,000
54,000
41,000
340,000
2,000,000
500 ppm NRLM Fuel
2020
2,400
<10
1,300
3,200
1,000
900
1,400
3,100
46,000
37,000
28,000
250,000
1,500,000
2030
3,600
<10
1,700
4,700
1,600
1,100
1,800
4,100
61,000
49,000
37,000
300,000
1,800,000
A Incidences are rounded to two significant digits or nearest ten.  The estimates do not include the benefits of
reduced sulfur in home heating oil or benefits in Alaska or Hawaii.
B Premature mortality associated with ozone is not separately included in this analysis
c Respiratory hospital admissions for PM includes admissions for COPD, pneumonia, and asthma.
D Cardiovascular hospital admissions for PM includes total cardiovascular and subcategories for ischemic heart
disease, dysrhythmias, and heart failure.

-------
Final  Regulatory Impact Analysis
                  Table 9-14. Results of PM Human Health and Welfare Benefits Valuation
                   for the Final Fuel-Related Components of the Nonroad Diesel Standards
Endpoint
Premature mortality0: (adults, 30 and over)
3% discount rate
7% discount rate
Infant mortality (infants under one year)
Chronic bronchitis (adults, 26 and over)
Non- fatal myocardial infarctions13
3% discount rate
7% discount rate
Hospital Admissions from Respiratory CausesE
Hospital Admissions from Cardiovascular CausesF
Emergency Room Visits for Asthma
Acute bronchitis (children, 8-12)
Asthma exacerbations (asthmatic children, 6-18)
Lower respiratory symptoms (children, 7-14)
Upper respiratory symptoms (asthmatic children, 9-11)
Work loss days (adults, 18-65)
Minor restricted activity days (adults, age 18-65)
Recreational visibility (86 Class I Areas)
Monetized Total0
3% discount rate
7% discount rate
Monetary BenefitsA>B
(millions 2000$, Adjusted for Income Growth)
Fuel Only Program
2020

$17,000
$16,000
$40
$610

$310
$300
$20
$21
$0.4
$1.3
$2.3
$0.7
$0.8
$38
$90
$400
$18,000+B
$17,000+B
2030

$26,000
$24,000
$52
$820

$420
$410
$31
$28
$0.6
$1.7
$3.1
$0.9
$1.1
$43
$110
$550
$28,000+B
$26,000+B
500 ppm NRLM Fuel
2020

$15,000
$14,000
$36
$550

$280
$270
$18
$19
$0.4
$1.2
$2.1
$0.6
$0.7
$34
$80
$360
$16,000+B
$15,000+B
2030

$23,000
$22,000
$47
$740

$380
$370
$28
$25
$0.5
$1.6
$2.8
$0.8
$1.0
$39
$100
$500
$25,000+B
$24,000+B
 Monetary benefits are rounded to two significant digits
B Monetary benefits are adjusted to account for growth in real GDP per capita between 1990 and the analysis year (2020 or 2030).
c Valuation of base estimate assumes discounting over the distributed lag structure described earlier. Results reflect the use of 3% and 7% discount rates consistent
with EPA and OMB's guidelines for preparing economic analyses (US EPA, 2000c, OMB Circular A-4).
D Estimates assume costs of illness and lost earnings in later life years are discounted using either 3 or 7 percent
E Respiratory hospital admissions for PM includes admissions for COPD, pneumonia, and asthma.
F Cardiovascular hospital admissions for PM includes total cardiovascular and subcategories for ischemic heart disease, dysrhythmias, and heart failure.
G B represents the monetary value of the unmonetized health and welfare benefits. A detailed listing of unquantified PM, ozone, CO, and NMHC related health
effects is provided in Table 9-1. The estimates do not include the benefits of reduced sulfur in home heating oil or benefits in Alaska or Hawaii.
                                                              9-46

-------
                                                                Cost-Benefit Analysis
9.5  Development of Intertemporal Scaling Factors and Calculation of
Benefits Over Time

   To estimate the health and visibility benefits of the NOx, SO2, and direct PM emission
reductions from the final standards occurring in years other than 2020 and 2030, it is necessary
to develop factors to scale the modeled benefits in 2020 and 2030. In addition to scaling based
on the relative reductions in NOx, SO2, and direct PM, intertemporal  scaling requires additional
adjustments to reflect population growth, changes in the age composition of the population, and
per capita income levels.

   Two separate sets of scaling factors are required, one for PM related health benefits, and one
for visibility benefits.  For the first of these, PM health benefits, we need scaling factors based on
ambient PM25. Because of the nonproportional relationship between  precursor emissions and
ambient concentrations of PM25, it is necessary to first develop estimates of the marginal
contribution of reductions in each emission species to reductions in PM25 in each year. Because
we have only two points (2020 and 2030), we assume a very simple linear function for each
species over time (assuming that the marginal contribution of each emission species to PM25 is
independent of the  other emission species) again assuming that sulfate changes are primarily
associated with SO2 emission reductions, nitrate changes are primarily associated with NOx
emission reductions, and primary PM changes are associated with direct PM emission
reductions.

   Using the linear relationship, we estimate the marginal contribution of SO2 to sulfate, NOx to
nitrate, and direct PM to primary PM in each year. These marginal contribution estimates are
presented in Table 9-15. Note that these projections do not take into account differences in
overall baseline proportions of NOx,  SO2, and PM. They assume that the change in the relative
effectiveness of each emission species in reducing ambient PM that is observed between 2020
and 2030 can be extrapolated to other years. Because baseline emissions of NOx, SO2, and PM,
as well as ammonia and VOCs are changing between years, the relative effectiveness of NOx
and S02 emission reductions may change in a non-linear fashion.  It is  not clear what overall
biases these nonlinearities will introduce into the scaling exercise. However, without these
assumptions, it is not possible to develop year by year benefits estimates.

   Multiplying the year-specific marginal contribution estimates by the appropriate emissions
reductions in each year yields estimates of the population-weighted changes in PM25 constituent
species, which are summed to obtain  year specific population-weighted changes in total PM2 5.
Total benefits in each specific year are then developed by scaling total benefits in a base year
using the ratio of the change in PM2 5 in the target year to the base year, with additional scaling
factors to account for growth in total  population, age composition of the population, and growth
in per capita income.
                                          9-47

-------
Final Regulatory Impact Analysis
                                      Table 9-15.
                    Projected Marginal Contribution of Reductions
                   in Emission Species to Reductions in Ambient PM2-5
Change in PM9 , species (population-weighted i-ig/m3 per million tons reduced)
Year Sulfate/SO2 Nitrate/NOx Primary PM/direct PM
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
0.153
0.154
0.156
0.157
0.159
0.160
0.161
0.163
0.164
0.166
0.167
0.169
0.170
0.171
0.173
0.174
0.176
0.177
0.179
0.180
0.181
0.183
0.184
0.186
0.049
0.050
0.051
0.052
0.053
0.054
0.055
0.056
0.057
0.058
0.059
0.060
0.061
0.062
0.063
0.064
0.065
0.066
0.067
0.069
0.070
0.071
0.072
0.073
2.130
2.123
2.117
2.111
2.105
2.098
2.092
2.086
2.080
2.073
2.067
2.061
2.054
2.048
2.042
2.036
2.029
2.023
2.017
2.011
2.004
1.998
1.992
1.985
   Growth in population and changes in age composition are accounted for by apportioning
total benefits into benefits accruing to three different age groups, 0 to 18, 19 to 64, and 65 and
older. Benefits for each age group are then adjusted by the ratio of the age group population in
the target year to the age group population in the base year.  Age composition adjusted estimates
are then reaggregated to obtain total population and age composition adjusted benefits for each
year.  Growth in per capita income is accounted for by multiplying the target year estimate by
the ratio of the income adjustment factors in the target year to those in the base year.

   For example, for the targetyear of 2015, there are 193,431 tons of NOx reductions, 297,513
tons of SO2 reductions, and 53,072 tons of direct PM25 reductions. These are associated with a
                                         9-48

-------
                                                                 Cost-Benefit Analysis
populated weighted change in total PM25 of 0.17, calculated from Table 9-15. The ratio of this
change to the change in the 2030 base year is 0.392.  The age group apportionment factors
(based on using a 3% discount rate for 2030) are 0.2% for 0 to  18, 19.2% for 19 to 64, and
80.6% for 65 and older.  The age group population growth ratios for 2015 relative to 2030 are
0.891 for 0 to 18, 0.986 for 19 to 64, and 0.639 for 65 and older.  The income growth adjustment
ratios for 2015 are 0.936 for premature mortality endpoints and 0.928 for morbidity endpoints.
Premature mortality accounts for 93 percent of total health benefits and morbidity accounts for 7
percent of health benefits.  Combining these elements with the  total estimate of PM health
benefits in 2030 of $94.2 billion , total PM health benefits in 2015 for the final standards are
calculated as:

Total PM health benefits (2015) =

[$94.2 billion * 0.392*(0.002*0.891+0.192*0.986+0.806*0.639)*(0.93*0.936+0.07*0.928)]

= $24.2 billion

   In order to develop the time stream of visibility benefits, we need to develop  scaling factors
based on the contribution of each emission species to light extinction.  Similar to  ambient PM2 5,
because we have only two  estimates of the change in light extinction (2020 and 2030), we
assume a very simple linear function for each species over time (assuming that the marginal
contribution of each emission species to light extinction is independent of the other  emission
species) assuming that changes in the sulfate component of light extinction are associated with
S02 emission reductions, changes in the nitrate component of light extinction are primarily
associated with NOx emission reductions, and changes in the primary PM components of light
extinction are associated with direct PM emission reductions. Linear relationships (slope and
intercept) are calculated for each Class I area.

   Using the linear relationships, we estimate the marginal contribution of SO2, NOx, and direct
PM to the change in light extinction at each Class I area in each year. Again, note that these
estimates assume that the change in the relative effectiveness of each emission species in
reducing light extinction that is observed between 2020 and 2030 can be extrapolated to other
years.

   Multiplying the year specific marginal contribution estimates by the appropriate emissions
reductions in each year yields estimates of the changes  in light  extinction components, which are
summed to obtain year specific changes in total light extinction. Benefits for each park in each
specific year are then developed by scaling total benefits in a base year using the ratio of the
change in light extinction in the target year to the base year, with additional  scaling  factors to
account for growth in total population, and growth in per capita income.  Total national visibility
benefits for each year are obtained by summing the scaled benefits across Class I  areas.
                                          9-49

-------
Final Regulatory Impact Analysis
    Table 9-16 provides undiscounted estimates of the time stream of benefits for the final
standards using 3 and 7 percent concurrent discount rates.K Figure 9-1 shows the undiscounted
time stream of benefits using a 3 percent concurrent discount rate. Because of the assumptions
we made about the linearity of benefits for each emission species, overall benefits are also linear,
reflecting the relatively linear emissions reductions over time for each emission type. The
exception is during the early years of the program, where there is little NOx emission reduction,
so that benefits are dominated by SO2 and direct PM2 5 reductions.

    Using a 3 percent intertemporal discount rate, the present value in 2004 of the benefits of the
final standards is approximately $805 billion for the time period 2007 to 2036, using a matching
3 percent concurrent discount rate. Using a 7 percent intertemporal discount rate, the present
value in 2004 of the benefits of the final standards for the base estimate is approximately $352
billion using a matching 7 percent concurrent discount rate.

    Annualized benefits using 3 percent intertemporal and  concurrent discount rates are
approximately $39 billion.  Annualized benefits using 7 percent intertemporal and concurrent
discount rates are approximately $28 billion.
    KWe refer to discounting that occurs during the calculation of benefits for individual years as concurrent
discounting. This is distinct from discounting that occurs over the time stream of benefits, which is referred to as
intertemporal discounting.

                                           9-50

-------
                                                                                Cost-Benefit Analysis
            Table 9-16.  Time Stream of Benefits for Final Nonroad Diesel Engine StandardsA'B
Year
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2031
2032
2033
2034
2035
2036
Monetized PM-Health and Visbility Benefits
(Million 2000$)
3% Concurrent Discount Rate
$5,000
$9,100
$9,700
$11,000
$12,000
$15,000
$18,000
$21,000
$25,000
$28,000
$32,000
$36,000
$40,000
$44,000
$48,000
$52,000
$56,000
$61,000
$64,000
$68,000
$72,000
$76,000
$79,000
$83,000
$87,000
$90,000
$94,000
$98,000
$100,000
$110,000
Present Value in 2004
3% Intertemporal Discount Rate
7% Intertemporal Discount Rate
$805,000
-
7% Concurrent Discount Rate
$4,700
$8,600
$9,100
$10,000
$12,000
$14,000
$17,000
$20,000
$23,000
$27,000
$31,000
$34,000
$38,000
$42,000
$46,000
$49,000
$53,000
$57,000
$61,000
$64,000
$68,000
$71,000
$75,000
$78,000
$82,000
$85,000
$89,000
$92,000
$96,000
$100,000

—
$350,000
A All dollar estimates rounded to two significant digits.
B Results reflect the use of 3% and 7% discount rates consistent
2000c, OMB Circular A-4).
with EPA and OMB's guidelines for preparing economic analyses (US EPA,
                                                    9-51

-------
Final Regulatory Impact Analysis
    $120,000
    $100,000
                                        Figure 9-1.
             Base Estimate of the Stream of Annual Benefits for the Final Nonroad Diesel Engine
                                   Standards: 2007 to 2036
        $o
                         ./b  vt* >>  ,b
9.6 Comparison of Costs and Benefits

   The estimated social cost (measured as changes in consumer and producer surplus) in 2030
to implement the final rule, as described in Chapter 8 is $2.0 billion (here, converted to 2000$).
Thus, the net benefit (social benefits minus social costs) of the program at full implementation is
approximately $81 + B billion, where B represents the sum of all unquantified benefits and
disbenefits.  In 2020, partial implementation of the program yields net benefits of $42 + B
billion.  Therefore, implementation of the final rule is expected, based purely on economic
efficiency criteria, to provide society with a significant net gain in social welfare.  Table 9-17
presents a summary of the benefits, costs, and net benefits of the final rule. Figure 9-2 displays
the stream of benefits, costs, and net benefits of the Nonroad Diesel Engine and Fuel Standards
from 2007 to 2036.  In addition, Table 9-18  presents the present value  of the stream of benefits,
costs, and net benefits associated with the rule for this 30 year period.  The total present value of
the stream of monetized net benefits (benefits minus costs) is $750 billion (using a three percent
discount rate).
                                          9-52

-------
                                                                                 Cost-Benefit Analysis
                                                 Table 9-17.
                 Summary of Monetized Benefits, Costs, and Net Benefits of the
                Final Full Program Nonroad Diesel Engine and Fuel StandardsA


Social Costsc
Social Benefits13 E:
CO, VOC, Air Toxic-related benefits
Ozone-related benefits
PM-related Welfare benefits
PM-related Health benefits (3% discount rate)
PM-related Health benefits (7% discount rate)
Net Benefits (Benefits-Costs)1312 (3% discount rate)
Net Benefits (Benefits-Costs)1312 (7% discount rate)
Base Estimate8
2020
(Billions of 2000
dollars)
$1.8

Not monetized
Not monetized
$1.0
$43
$41
$42 +B
$41 +B
2030
(Billions of 2000
dollars)
$2.0

Not monetized
Not monetized
$1.7
$81
$78
$81 +B
$78 +B
A All costs and benefits are rounded to two significant digits.
B Base Estimate reflects premature mortality based on application of concentration-response function derived from long-term
exposure to PM2 5, valuation using the value of statistical lives saved apporach, and a willingness-to-pay approach for valuing
chronic bronchitis incidence.
c Note that costs are the total costs of reducing all pollutants, including CO, VOCs and air toxics, as well as NOx and PM.
Benefits in this table are associated only with PM, NOx and S02 reductions. These estimates do not include the benefits of
reduced sulfur in home heating oil or benefits  in Alaska or Hawaii. Costs are converted from 2002$ to 2000$ in this table using
the PPI for Total Manufacturing Industries.
D Not all possible benefits or disbenefits are quantified and monetized in this analysis. Potential benefit categories that have not
been quantified and monetized are listed in Table 9-1. These estimates do not include the benefits of reduced sulfur in home
heating oil or benefits in Alaska or Hawaii. B  is the sum of all unquantified benefits and disbenefits.
E Monetized benefits are presented using two different discount rates. Results reflect the use of 3% and 7% discount rates
consistent with EPA and OMB's guidelines for preparing  economic analyses (US EPA, 2000c, OMB Circular A-4).
                                                     9-53

-------
Final  Regulatory Impact Analysis
                                           Figure 9-2.
                       Stream of Benefits, Costs, and Net Benefits of the
                        Final Nonroad Diesel Engine and Fuel Standards
     $120,000
     $100,000
     $80,000
     $60,000
     $40,000
     $20,000
            2005
     $(20,000)
                                                                     	
                        2010
                                     2015
                                                  2020
                                                               2025
                                                                            2030
                                                                                         2035
                                  -Total Social Benefits	Total Social Costs -*— Net Benefits
                                              9-54

-------
                                                     Cost-Benefit Analysis
                         Table 9-18.
     Present Value in 2004 of the Stream of 30 Years of
Benefits, Costs, and Net Benefits for the Final Full Program
         Nonroad Diesel Engine and Fuel Standards
                     (Billions of 2000$)ab

Social Costs
Social Benefits
Net Benefits a
Billions of 2000$
3% Discount Rate
$27
$805
$780
Billions of 2000$
7% Discount Rate
$ 14
$352
$340
        a Rounded to two significant digits
 b Benefits represent 48-state benefits and exclude home heating oil sulfur reduction
 benefits, whereas costs include 50-state estimates. Costs were converted from 2002$
 to 2000$ using the PPI for Total Manufacturing Industries.
                             9-55

-------
Final Regulatory Impact Analysis
                                                Table 9-19.
                  Summary of Monetized Benefits, Costs, and Net Benefits of the
   Final Fuel Only Components of the  Nonroad Diesel Standards (Billions of 2000 dollars)A


Costs B'c
Fuel Only
Program
2020
$0.62
2030
$0.72
500 ppm
NRLM Fuel
2020
($0.28)
2030
($0.36)
Social Benefits0 DE:
CO, VOC, Air
Toxic-related benefits
Ozone-related
benefits
PM-related Welfare
benefits
PM-related Health
benefits
(3 % discount rate)
PM-related Health
benefits
(7% discount rate)
Net Benefits (3% discount
rate) = (Benefits-Costs)0 D E
Net Benefits (7% discount
rate) = (Benefits-Costs)0 D E
Not monetized
Not monetized
$0.4
$18
$17
$18 + B
$17 + B
Not monetized
Not monetized
$0.6
$28
$26
$28 + B
Not monetized
Not monetized
$0.4
$16
$15
$ 16 + B
$26 + B $16 + B
Not monetized
Not monetized
$0.5
$25
$23
$25 + B
$24 + B
A All costs and benefits are rounded to two significant digits.
B Engineering costs are presented instead of social costs. As discussed in previous chapters, total engineering costs include fuel
costs (refining, distribution, lubricity) and other operating costs (oil change maintenance savings). All engine and equipment
fixed cost expenditures are amortized using a seven percent capital cost to reflect the time value of money. The annual costs
presented here are the costs in the indicated year and are not the net present values.
c Note that costs are the total costs of reducing all pollutants, including CO, VOCs and air toxics, as well as NOx and PM.
Benefits in this table are associated only with PM, NOx and SO2 reductions. The estimates do not include the benefits of reduced
sulfur in home heating oil or benefits in Alaska or Hawaii. Costs were converted from 2002$ to  2000$ using the PPI for
Total Manufacturing Industries.
D Not all possible benefits or disbenefits are quantified and monetized in this analysis. Potential benefit categories that
have not been quantified and monetized are listed in Table 9-1.  B is the sum of all unquantified benefits and
disbenefits.
E Monetized costs and benefits are presented using two different discount rates.  Results reflect the use of 3% and 7%
discount rates consistent with EPA and OMB's guidelines for preparing economic analyses (US EPA, 2000c, OMB Circular A-
4).
                                                    9-56

-------
                                                                           Cost-Benefit Analysis
                                            Table 9-20.
                     Present Value in 2004 of the Stream of 30 Years of
                           Benefits, Costs, and Net Benefits for the
              Final Fuel Only Components of the Nonroad Diesel Standards
                                    (Billions of 2000$)ABCD

Fuel Only
Program
500 ppm
NRLM Fuel
3 % discount rate
Costs
Social Benefits
Net Benefits
$9.2
$340
$330
($0.54)
$310
$310
7 % discount rate
Costs
Social Benefits
Net Benefits
$4.6 1 ($0.3)
$160 1 $140
$160 1 $140
A Results are rounded to two significant digits.  Sums may differ because of rounding.
B Engineering costs are presented instead of social costs. As discussed in previous chapters, total engineering costs
include fuel costs (refining, distribution, lubricity) and other operating costs (oil change maintenance savings).
c Note that costs are the total costs of reducing all pollutants, including CO, VOCs and air toxics, as well as NOx and PM.
Benefits in this table are associated only with PM, NOx and SO2 reductions. The estimates do not include the benefits of
reduced sulfur in home heating oil or benefits in Alaska or Hawaii.
D Not all possible benefits or disbenefits  are quantified and monetized in this analysis.  Potential benefit
categories that have not been quantified and monetized are listed in Table 9-1.  B is the sum of all
unqualified benefits and disbenefits.
E Monetized costs and benefits are presented using two different discount rates. Results reflect the use of 3% and
7% discount rates consistent with EPA and OMB's guidelines for preparing economic analyses (US EPA, 2000c, OMB
Circular A-4).
                                               9-57

-------
Final Regulatory Impact Analysis
   A key input to our benefit-cost analysis is the social costs and emission reductions associated
with the final program.  Each of these elements also has associated uncertainty which contributes
to the overall uncertainty in our analysis of benefit-cost.

   EPA engineering cost estimates are based upon considerable expertise and experience within
the Agency. At the same time, any estimate of the future cost of control technology for engines
or the cost of removing sulfur from diesel fuel is inherently uncertain to some degree. At the
start is the question of what technology will actually be used to meet future standards, and what
such technology will cost at the time of implementation.  Our estimates of control costs are based
upon current technology plus newer technology already "in the pipeline."  New technology not
currently anticipated is by its nature not specifically included.  Potential new production
techniques which might lower costs are also not included in these estimates (although they are
partially included among factors contributing to learning curve effects).  On the other side of the
equation are unforseen technical hurdles that may act to increase control system costs.

   There is also uncertainty in our social cost estimates. Our Economic Impact Assessment
presented in Chapter 10 includes sensitivity analyses examining the effect of varying
assumptions surrounding the following key factors (Chapter 10, Appendix 10-1):

       market supply and demand elasticity parameters
   -   alternative assumptions about the fuel market supply shifts and fuel maintenance savings
   -   alternative assumptions about the engine and equipment market supply shifts

   For all of these factors, the change in social cost was estimated to be very small, with a
maximum impact of less than one percent.  These results are not surprising given the small  share
of total production costs of diesel engines, equipment, and fuel affected by the rule. See Chapter
10 for a more detailed discussion.

   Overall, we have limited means available  to develop quantitative estimates of total
uncertainty in costs.  Some of the factors identified above can act to either increase or decrease
actual cost compared to our estimates.  Some, such  as new technology developments and new
production techniques, will act to lower costs compared to our estimates.

   One source of a useful information about the overall uncertainty we might expect  to see in
cost is literature comparing historical rulemaking cost estimates with actual price increases when
                                          9-58

-------
                                                                    Cost-Benefit Analysis
new standards went into effect.L  Perhaps the most relevant of such studies is the paper by
Anderson and Sherwood analyzing these effects for those mobile source rules adopted since the
Clean Air Act Amendments of 1990. That paper reviewed six fuel quality  rules and ten light-
duty vehicle control rules that had been required by those amendments. It found that EPA
estimates of the costs for future standards tended to be similar to or higher than actual price
changes observed in the market place. Table 9-21 presents the results for some of the fuel and
vehicle rules reviewed in the paper.

                                        Table 9-21.
         Comparison of Historical EPA Cost Estimates with Actual Price Changes
EPA Rule
Phase 2 RVP control
Reformulated
Gasoline Phase 1
Reformulated
Gasoline Phase 2
SOOppm Sulfur
Highway Diesel Fuel
1994-2001 LDV
Regulations
EPA Mid-point
Estimate
1.1 c/gal
4.1 c/gal
5.7 c/gal
2.2 c/gal
$446/vehicle
Actual Price
Change
0.5 c/gal
2.2 c/gal
5.1 c/gal
2.2 c/gal
$347
Percent Difference
for Price vs EPA
-54%
-46%
-10%
0%
-22%
    The data in Table 9-21 would lead us to believe that cost uncertainty is largely a risk of
overestimation by EPA.  However, given the uncertainty in estimating costs, we believe it is
appropriate to consider the potential for both overestimation and underestimation. As a
sensitivity factor for social cost variability we have chosen to evaluate a range of possible errors
in social cost of from twenty percent higher to twenty percent lower than the EPA estimate.  The
resulting social cost range is shown in Table 9 -22. This uncertainty has virtually no impact on
    LFor this analysis, we based our cost estimates on information received from industry and technical reports
relevant to the US market.  We are also aware of two studies done to support nonroad standards development in
Europe, namely the VTT report and the EMA/Euromot report (Euromot 2002, Docket A-2001-28 Document number
II-B-12). We are not utilizing the cost information in these reports because neither one has sufficient information to
allow us to understand or derive the relevant cost figures and therefore provide us insufficient information that could
be used in trying to estimate cost uncertainty for nonroad diesel engine technologies.
                                            9-59

-------
Final Regulatory Impact Analysis
our estimates of the net benefits of the final rule, given the large magnitude by which benefits
exceed costs.
                                       Table 9-22.
                  Estimated Uncertainty for Cost of Final Full Program
Year
2010
2020
2030
Engineering Cost
Estimate
$0.30 billion
$1.8 billion
$2.1 billion
Uncertainty Range (-20 to +20 percent)
$0.24 - $0.36 billion
$1.5 -$2.2 billion
$1.7 -$2.6 billion
   Turning to the question of emissions uncertainty, the Agency does not at this time have
useful quantitative information to bring to bear on this question. For our estimates, we rely on
the best information that is available to us.  However, there is uncertainty involved in many
aspects of emissions estimations. Uncertainty exists in the estimates of emissions from the
nonroad sources affected by this final rule, as well as in the universe of other sources included in
the emission inventories used for our air quality modeling. To the extent that these other sources
are unchanged between our baseline and control case, the impact of uncertainty in those
estimates is lessened.  Similarly, since the key driver of the benefits of our final rule is the
changes produced by the new standards, the effect of uncertainty in the overall estimates of
nonroad emissions on our benefits estimates may be lessened.

   As discussed in Chapter 3 and  our summary and analysis of comments, the main sources of
uncertainty in our estimates of nonroad emissions fall in the three areas of population size
estimates, equipment usage rates (activity) and engine emission factors. Since nonroad
equipment is not subject to state registration and licensing requirements like those applying to
highway vehicles, it is difficult to develop precise equipment counts for in-use nonroad
equipment. Our modeled equipment populations are derived from related data about sales and
scrappage rates.  Similarly, annual amount of usage  and related load factor information is
estimated with some degree of uncertainty. We have access to extensive bodies of data on these
areas, but are also aware of the need for improvement.  Finally, the emission rates of engines in
actual field operation cannot readily be measured at the present time, but are estimated from
laboratory testing under a variety of typical operating cycles. While laboratory estimates are a
reliable  source of emissions data, they cannot fully capture all of the impacts of real in-use
operation on emissions, leading to some uncertainty about the results.  For further details on our
                                          9-60

-------
                                                                 Cost-Benefit Analysis
modeling of nonroad emissions, please refer to the discussions in Chapters 3 and Appendix 8A
ofthisRIA.

   We have ongoing efforts in all three of these areas designed to improve their accuracy.  Since
the opportunity to gather better data exists, we have chosen to focus our main efforts on
developing improved estimates rather than on developing elaborate techniques to estimate the
uncertainty of current estimates. In the long run, better estimates are the most desired outcome.

   One of the most important new tools we are developing is the use of portable emission
measurement devices to gather detailed  data on actual engines and equipment in daily use.
These devices have recently become practical due to advances in computing and sensor
technology, and will allow us to generate intensive data defining both activity-related factors
(e.g., hours of use, load factors, patterns of use) and in-use emissions data specific to the
measured activity and including effects from such things as age and emissions related
deterioration. The Agency is pursuing this equipment for improving both its highway and
nonroad engine emissions models. Because of the multiplicity of factors involved, we cannot
make a quantitative estimate of the uncertainty in our emissions estimates.
                                          9-61

-------
Final Regulatory Impact Analysis
Chapter 9 References

Abbey, D.E., B.L. Hwang, RJ. Burchette, T. Vancuren, and P.K. Mills. 1995. Estimated Long-
   Term Ambient Concentrations of PM(10) and Development of Respiratory Symptoms in a
   Nonsmoking Population. Archives of Environmental Health 50(2): 139-152.
Abbey, D.E., F. Petersen, P. K. Mills, and W. L. Beeson. 1993. Long-Term Ambient
   Concentrations of Total Suspended Particulates, Ozone, and Sulfur Dioxide and Respiratory
   Symptoms in a Nonsmoking Population. Archives of Environmental Health 48(1): 33-46.
Abbey, D.E., S.D. Colome, P.K. Mills, R. Burchette, W.L. Beeson and Y. Tian. 1993. Chronic
   Disease Associated With Long-Term Concentrations of Nitrogen Dioxide.  Journal of
   Exposure Analysis and Environmental Epidemiology. Vol. 3(2): 181-202.
Abbey, D.E., N. Nishino, W.F. McDonnell, RJ. Burchette, S.F. Knutsen, W. Lawrence Beeson
   and J.X. Yang.  1999. Long-term inhalable particles and other air pollutants related to
   mortality in nonsmokers [see comments]. Am J Respir Crit Care Med. Vol. 159(2): 373-82.
Abt Associates, Inc. 2003. ProposedNonroad Landbased Diesel Engine Rule: Air Quality
   Estimation, Selected Healthand Welfare Benefits Methods, and Benefit Analysis Results.
   Prepared for Office of Air Quality Planning and Standards, U.S. EPA. April, 2003.
Adams, P.F., G.E. Hendershot and M.A. Marano. 1999. Current Estimates from the National
   Health Interview Survey, 1996. Vital Health Stat. Vol. 10(200): 1-212.
Agency for Healthcare Research and Quality. 2000. HCUPnet, Healthcare Cost and Utilization
   Project.
American Lung Association, 1999. Chronic Bronchitis.  Web site available at:
   http://www.lungusa.org/diseases/lungchronic.html.
Anderson,!; Sherwood,T; Comparison of EPA and Other Estimates of Mobile Source Rule Costs
   to Actual Price Changes; Society of Automotive Engineers; SAE 2002-01-1980; May 14,
   2002.
Alberini, A., M. Cropper, A. Krupnick, N. Simon, (forthcoming). Does the Value of a Statistical
   Life Vary with Age and Health Status? Evidence from the U.S. and Canada, Journal of
   Environmental Economics and Management.
Alberini, A., M. Cropper, T.Fu, A. Krupnick, J. Liu, D.  Shaw, and W. Harrington. 1997. Valuing
   Health Effects of Air Pollution in Developing Countries: The Case of Taiwan. Journal of
   Environmental Economics and Management. 34: 107-126.
American Lung Association.  2002a.  Trends in Morbidity and Mortality: Pneumonia, Influenza,
   and Acute Respiratory Conditions. American Lung  Association, Best Practices and Program
   Services, Epidemiology and Statistics Unit.
American Lung Association.  2002b.  Trends in Chronic Bronchitis and Emphysema: Morbidity
   and Mortality. American Lung Association, Best Practices and Program Services,
   Epidemiology and Statistics Unit.
                                         9-62

-------
                                                               Cost-Benefit Analysis
American Lung Association.  2002c.  Trends in Asthma Morbidity and Mortality. American
   Lung Association, Best Practices and Program Services, Epidemiology and Statistics Unit.
Banzhaf, S., D. Burtraw, and K. Palmer. 2002. Efficient Emission Fees in the U.S. Electricity
   Sector. Resources for the Future Discussion Paper 02-45, October.
Berger, M.C., G.C. Blomquist, D. Kenkel, and G.S. Tolley. 1987. Valuing Changes in Health
   Risks: A Comparison of Alternative Measures. The Southern Economic Journal 53: 977-984.
Bricker, S. B., C. G.  Clement, D. E. Pirhalla, S. P. Orlando and D. R. G. Farrow. 1999. National
   Estuarine Eutrophi cation Assessment: Effects of Nutrient Enrichment in the Nation's
   Estuaries.  National Oceanic and Atmospheric Administration, National Ocean Service,
   Special  Projects Office and the National Centers for Coastal Ocean Science. Silver Spring,
   Maryland. 71p
Burnett RT, Smith-Doiron M, Stieb D, Raizenne ME, Brook JR, Dales RE, Leech JA, Cakmak
   S, Krewski D. 2001. Association between ozone and hospitalization for acute respiratory
   diseases in children less than 2 years of age. Am JEpidemiol 153:444-52
Carnethon MR, Liao D, Evans GW, Cascio WE, Chambless LE, Rosamond WD, Heiss G. 2002.
   Does the cardiac autonomic response to postural change predict incident coronary heart
   disease  and mortality? The Atherosclerosis Risk in Communities Study. American Journal of
   Epidemiology, 155(l):48-56
Chen, L., B.L. Jennison, W. Yang and S.T. Omaye. 2000.  Elementary school absenteeism and
   air pollution. Inhal Toxicol. Vol.  12(11): 997-1016.
Chestnut, L.G.  1997. Draft Memorandum:  Methodology for Estimating  Values for Changes in
   Visibility at National Parks. April 15.
Chestnut, L.G. and R.L. Dennis.  1997. Economic Benefits of Improvements in Visibility: Acid
   Rain Provisions of the 1990 Clean Air Act Amendments. Journal of Air and Waste
   Management Association 47:395-402.
Chestnut, L.G. and R.D. Rowe.  1990a. Preservation Values for Visibility Protection at the
   National Parks: Draft Final Report. Prepared for Office of Air Quality Planning and
   Standards, US Environmental Protection Agency, Research Triangle Park, NC and Air
   Quality  Management Division, National Park Service, Denver, CO.
Chestnut, L.G.,  and R.D. Rowe. 1990b. A New National Park Visibility Value Estimates. In
   Visibility and Fine Particles, Transactions of an AWMA/EPA International Specialty
   Conference, C. V. Mathai, ed. Air and Waste Management Association, Pittsburgh.
CMS (2002).  Centers for Medicare and Medicaid Services. Conditions of Participation:
   Immunization Standards for Hospitals, Long-Term Care Facilities, and Home Health
   Agencies.  67 FR 61808, October 2, 2002.
Cody, R.P., C.P. Weisel, G. Birnbaum and PJ. Lioy.  1992. The effect of ozone associated with
   summertime photochemical smog on the frequency of asthma visits to hospital emergency
   departments. Environ Res. Vol. 58(2): 184-94.
                                         9-63

-------
Final Regulatory Impact Analysis
Crocker, T.D. and R.L. Horst, Jr. 1981. Hours of Work, Labor Productivity, and Environmental
   Conditions: A Case Study.  The Review of Economics and Statistics. Vol. 63: 361-368.
Cropper, M.L. and AJ. Krupnick.  1990.  The Social Costs of Chronic Heart and Lung Disease.
   Resources for the Future. Washington, DC.  Discussion Paper QE 89-16-REV.
Daniels MJ, Dominici F, Samet JM, Zeger SL. 2000. Estimating particulate matter-mortality
   dose-response curves and threshold levels: an analysis of daily time-series for the 20 largest
   US cities. Am JEpidemiol  152(5):397-406
Dockery, D.W., C.A. Pope, X.P. Xu, J.D. Spengler, J.H. Ware, M.E. Fay, E.G. Ferris and F.E.
   Speizer. 1993. An association between air pollution and mortality in six U.S. cities. New
   England Journal of Medicine 329(24): 1753-1759.
Dockery, D.W., J. Cunningham, A.I. Damokosh, L.M. Neas, J.D. Spengler, P. Koutrakis, J.H.
   Ware, M. Raizenne and F.E. Speizer.  1996. "Health Effects of Acid Aerosols On North
   American Children-Respiratory Symptoms." Environmental Health Perspectives. 104(5):
   500-505.
Dominici F, McDermott A, Zeger SL, Samet JM. 2002. On the use of generalized additive
   models in time-series studies of air pollution and health. Am J Epidemiol 156(3): 193-203
Dekker J.M., R.S. Crow, A.R. Folsom, P.J. Hannan, D. Liao, C.A. Swenne, and E. G. Schouten.
   2000.  Low Heart Rate Variability in a 2-Minute Rhythm Strip Predicts Risk of Coronary
   Heart Disease and Mortality From Several Causes : The ARIC Study. Circulation 2000 102:
   1239-1244.
Eisenstein, E.L., L.K.  Shaw, K.J. Anstrom, C.L. Nelson, Z. Hakim, V. Hasselblad and D.B.
   Mark.  2001.  Assessing the clinical and economic burden of coronary artery disease:
   1986-1998. Med Care. Vol. 39(8): 824-35.
EPA-SAB-COUNCIL-ADV-99-05, 1999. An SAB Advisory on the Health and Ecological
   Effects Initial Studies of the Section 812 Prospective Study: Report to Congress: Advisory
   by the Health and Ecological Effects Subcommittee, February.
EPA-SAB-COUNCIL-ADV-98-003, 1998. Advisory Council on Clean Air Compliance  Analysis
   Advisory on the Clean Air Act Amendments (CAAA) of 1990 Section 812 Prospective
   Study: Overview of Air Quality and Emissions Estimates: Modeling, Health and Ecological
   Valuation Issues Initial Studies.
EPA-SAB-COUNCIL-ADV-99-012, 1999. The Clean Air Act Amendments (CAAA) Section
   812 Prospective Study of Costs and Benefits (1999): Advisory by the  Health and Ecological
   Effects Subcommittee on Initial Assessments of Health and Ecological Effects: Part 1. July.
EPA-SAB-COUNCIL-ADV-00-001, 1999. The Clean Air Act Amendments (CAAA) Section
   812 Prospective Study of Costs and Benefits (1999): Advisory by the  Health and Ecological
   Effects Subcommittee on Initial Assessments of Health and Ecological Effects: Part 2.
   October, 1999.
EPA-SAB-COUNCIL-ADV-00-002, 1999. The Clean Air Act Amendments (CAAA) Section
   812 Prospective Study of Costs and Benefits (1999): Advisory by the  Advisory Council on
                                         9-64

-------
                                                               Cost-Benefit Analysis
   Clean Air Compliance Analysis: Costs and Benefits of the CAAA. Effects Subcommittee on
   Initial Assessments of Health and Ecological Effects: Part 2. October, 1999.
EPA-SAB-EEAC-00-013, 2000.  An SAB Report on EPA's White Paper Valuing the Benefits of
   Fatal Cancer Risk Reduction. July.
EPA-SAB-COUNCIL-ADV-01-004. 2001. Review of the Draft Analytical Plan for EPA's
   Second Prospective Analysis - Benefits and Costs of the Clean Air Act 1990-2020: An
   Advisory by a Special Panel of the Advisory Council on Clean Air Compliance Analysis.
   September.
Evans, William N., and W. Kip Viscusi. 1993. Income Effects and the Value of Health.  Journal
   of Human Resources 28(3):497-518.
Euromot and EMA, 2002. Investigation of the Feasibility of PM Filterns for NRMM.   The
   European Association of Internal Combustion Engine Manufacturers and the Engine
   Manufacturers Association (USA). Revised July 12, 2002.  Docket A-2001-28. Document
   Number II-B-12.
Fox, S.,  and R.A. Mickler, 1995. Impact of Air Pollutants on Southern Pine Forests Ecological
   Studies 118. Springer Verlag: New York.
Freeman, A. M. III. 1993. The Measurement of Environmental and Resource Values: Theory and
   Methods. Resources for the Future, Washington, D.C.
Garcia, P., Dixon, B. and Mjelde, J. (1986): Measuring the benefits of environmental change
   using a duality approach: The case of Ozone and Illinios cash grain farms. Journal of
   Environmental Economics and Management.
Gilliland, F.D., K. Berhane, E.B. Rappaport, D.C. Thomas, E. Avol, W.J. Gauderman, S.J.
   London, H.G. Margolis, R. McConnell, K.T. Islam and J.M. Peters.  2001.  The effects of
   ambient air pollution on school absenteeism due to respiratory illnesses.  Epidemiology. Vol.
   12(1): 43-54.
Gold DR, Litonjua A, Schwartz J, Lovett E, Larson A, Nearing B, Allen G, Verrier M, Cherry R,
   Verrier R. 2000. Ambient pollution and heart rate variability. Circulation 101(11): 1267-73
Greenbaum, D. 2002a. Letter. Health Effects Institute. May 30.  Available online at:
   http://www.healtheffects.org/Pubs/NMMAPSletter.pdf.  Accessed March 20, 2003.
Grosclaude, P.  and N.C. Soguel.  1994. "Valuing Damage to Historic Buildings Using a
   Contingent Market: A Case Study of Road Traffic Externalities." Journal of Environmental
   Planning and Management 37: 279-287.
Guo, Y.L., Y.C. Lin, F.C. Sung, S.L. Huang, Y.C. Ko, J.S. Lai, H.J. Su, C.K. Shaw, R.S. Lin,
   D.W. Dockery.  1999. Climate, Traffic-Related  Air Pollutants, and Asthma Prevalence in
   Middle-School Children in Taiwan. Environmental Health Perspectives 107: 1001-1006.
Harrington, W. and P. R. Portney. 1987. Valuing the Benefits of Health and Safety Regulation.
   Journal of Urban Economics 22:101-112.
Hollman, F.W., T.J. Mulder, and J.E. Kalian. 2000.  Methodology and Assumptions for the
   Population Projections of the United States: 1999 to 2100. Population Division Working
                                         9-65

-------
Final Regulatory Impact Analysis
   Paper No. 38, Population Projections Branch, Population Division, U.S. Census Bureau,
   Department of Commerce.  January.
HRSA (1998). Health Resources and Services Administration: Procurement and Transplantation
   Network; Final Rule. 63 FR 16295, April 2, 1998.
Ibald-Mulli, A., J. Stieber, H.-E. Wichmann, W. Koenig, and A. Peters. 2001. Effects of Air
   Pollution on Blood Pressure: A Population-Based Approach.  American Journal of Public
   Health. 91: 571-577.
Industrial Economics, Incorporated (lEc).  1994.  Memorandum to Jim DeMocker, Office of Air
   and Radiation, Office of Policy Analysis and Review, US  Environmental Protection Agency,
   March 31.
Ito, K. and G.D. Thurston.  1996. Daily PMlO/mortality associations: an investigations of at-risk
   subpopulations. Journal of Exposure Analysis and Environmental Epidemiology. Vol. 6(1):
   79-95.
Jones-Lee, M.W., M. Hammerton and P.R. Philips. 1985.  The Value of Safety: Result of a
   National Sample Survey. Economic Journal. 95(March): 49-72.
Jones-Lee, M.W.  1989. The Economics of Safety and Physical Risk. Oxford: Basil Blackwell.
Jones-Lee, M.W., G. Loonies, D. O'Reilly, and P.R. Phillips.  1993. The Value of Preventing
   Non-fatal Road Injuries: Findings of a Willingness-to-pay National Sample Survey.  TRY
   Working Paper, WP SRC2.
Kleckner, N. and  J. Neumann. 1999.  "Recommended Approach to Adjusting WTP Estimates to
   Reflect Changes in Real Income. Memorandum to Jim Democker, US EPA/OPAR, June 3.
Krewski D, Burnett RT, Goldbert MS, Hoover K, Siemiatycki J, Jerrett M, Abrahamowicz M,
   White WH. 2000. Reanalysis of the Harvard Six Cities Study and the American Cancer
   Society Study of Parti culate Air Pollution and Mortality. Special Report to the Health
   Effects Institute, Cambridge MA, July 2000
Krupnick,  A.J. and M.L. Cropper. 1992. "The Effect of Information on Health Risk
   Valuations." Journal of Risk and Uncertainty 5(2): 29-48.
Krupnick,  A., M.  Cropper., A. Alberini, N. Simon, B. O'Brien, R. Goeree, and M. Heintzelman.
   2002. Age, Health and the Willingness to Pay for Mortality Risk Reductions: A Contingent
   Valuation Study of Ontario Residents, Journal of Risk and Uncertainty, 24, 161-186.
Kunzli, N., R. Kaiser, S. Medina, M. Studnicka, O. Chanel, P. Filliger, M. Kerry, F. Horak Jr.,
   V. Puybonnieux-Texier, P. Quenel, J. Schneider, R.  Seethaler, J-C Vergnaud, and H.
   Sommer. 2000. Public-health Impact of Outdoor and Traffic-related Air Pollution: A
   European Assessment. The Lancet, 356: 795-801.
Kunzli N, Medina S, Kaiser R, Quenel P, Horak F Jr,  Studnicka M. 2001. Assessment of deaths
   attributable to air pollution: should we use risk estimates based on time series or on cohort
   studies? Am JEpidemiol 153(11): 1050-5
Lareau, T.J. and D.A.  Rae.  1989. Valuing WTP for Diesel Odor Reductions: An Application of
   Contingent Ranking Techniques, Southern Economic Journal, 55: 728- 742.
                                         9-66

-------
                                                               Cost-Benefit Analysis
Lave, L.B. and E.P. Seskin.  1977.  Air Pollution and Human Health. Johns Hopkins University
   Press for Resources for the Future: Baltimore.
Levy, J.I., J.K. Hammitt, Y. Yanagisawa, and J.D. Spengler. 1999. Development of a New
   Damage Function Model for Power Plants: Methodology and Applications. Environmental
   Science and Technology, 33: 4364-4372.
Levy, J.I., TJ. Carrothers, J.T. Tuomisto, J.K. Hammitt, and J.S. Evans. 2001. Assessing the
   Public Health Benefits of Reduced Ozone Concentrations. Environmental Health
   Perspectives. 109: 1215-1226.
Liao D, Cai J, Rosamond WD, Barnes RW, Hutchinson RG, Whitsel EA, Rautaharju P, Heiss G.
   1997. Cardiac autonomic function and incident coronary heart disease: a population-based
   case-cohort study. The ARIC Study. Atherosclerosis Risk in Communities Study. American
   Journal  of Epidemiology, 145(8):696-706.
Liao D, Creason J, Shy C, Williams R, Watts R, Zweidinger R.  1999. Daily  variation of
   particulate air pollution and poor cardiac autonomic control in the elderly.  Environ Health
   Perspect 107:521-5
Lipfert, F.W., S.C. Morris and R.E. Wyzga.  1989.  Acid Aerosols - the Next Criteria Air
   Pollutant. Environmental Science & Technology. Vol. 23(11): 1316-1322.
Lipfert, F.W.  ; H. Mitchell Perry Jr ; J. Philip Miller ; Jack D. Baty ; Ronald E. Wyzga ; Sharon
   E. Carmody 2000. The Washington University-EPRI Veterans' Cohort Mortality Study:
   Preliminary Results, Inhalation Toxicology, 12: 41-74
Lippmann, M., K. Ito, A. Nadas, and R.T. Burnett. 2000. Association of Particulate Matter
   Components with Daily Mortality and Morbidity in Urban Populations. Health Effects
   Institute Research Report Number 95, August.
Magari SR, Hauser R, Schwartz J, Williams PL, Smith TJ, Christiani DC. 2001. Association of
   heart rate variability with occupational and environmental exposure to particulate air
   pollution. Circulation 104(9):986-91
McClelland, G., W. Schulze, D. Waldman, J. Irwin, D. Schenk, T. Stewart, L.  Deck, and M.
   Thayer. 1993. Valuing Eastern Visibility: A Field Test of the Contingent Valuation Method.
   Prepared for Office of Policy, Planning and Evaluation, US  Environmental Protection
   Agency. September.
McConnell, R., K. Berhane, F. Gilliland, S.J. London, H. Vora, E. Avol, W.J. Gauderman, H.G.
   Margolis, F. Lurmann, D.C. Thomas, and J.M. Peters. 1999. Air Pollution and Bronchitic
   Symptoms in Southern California Children with Asthma.  Environmental Health
   Perspectives, 107(9): 757-760.
McConnell  R, Berhane K, Gilliland F, London SJ, Islam T, Gauderman WJ, Avol E, Margolis
   HG, Peters JM. 2002. Asthma in exercising children exposed to ozone: a cohort study.
   Lancet 359(9309):896.
                                         9-67

-------
Final Regulatory Impact Analysis
McDonnell, W.F., D.E. Abbey, N. Nishino and M.D. Lebowitz. 1999. Long-term ambient ozone
   concentration and the incidence of asthma in nonsmoking adults: the ahsmog study.
   Environmental Research. 80(2 Pt 1): 110-21.
Miller, T.R. 2000. Variations between Countries in Values of Statistical Life. Journal of
   Transport Economics and Policy. 34: 169-188.
Moolgavkar SH, Luebeck EG, Anderson EL. 1997. Air pollution and hospital admissions for
   respiratory causes in Minneapolis-St. Paul and Birmingham. Epidemiology 8:364-70
Moolgavkar, S.H. 2000. Air pollution and hospital admissions for diseases of the circulatory
   system in three U.S. metropolitan areas. J Air Waste Manag Assoc 50:1199-206
Moore and  Viscusi (1988).  The Quality-Adjusted Value of Life. Economic Inquiry. 26(3): 369-
   388.
Mrozek, JR and Taylor, LO (2002).  What Determines the Value of Life? A Meta-Analysis.
   Journal  of Policy Analysis and Management, Vol 21, No.2, 253-270.
National Center for Education Statistics. 1996 The Condition of Education 1996, Indicator 42:
   Student Absenteeism and Tardiness.  U.S. Department of Education National Center for
   Education Statistics. Washington DC.
National Research Council  (NRC). 1998. Research Priorities for Airborne Particulate Matter: I.
   Immediate Priorities and a Long-Range Research Portfolio. The National Academies Press:
   Washington, D.C.
National Research Council  (NRC). 2002. Estimating the Public Health Benefits of Proposed Air
   Pollution Regulations. The National Academies Press: Washington, D.C.
NCLAN. 1988. Assessment of Crop Loss from Air Pollutants. (Eds. Walter W. Heck, O. Clifton
   Taylor and David T. Tingey) Elsevier Science Publishing Co.: New York, Pp. 1-5. (ERL,GB
   639).
Neumann, I.E., M.T. Dickie, and R.E. Unsworth. 1994. Linkage Between Health Effects
   Estimation and Morbidity Valuation in the Section 812 Analysis — Draft Valuation
   Document.  Industrial Economics Incorporated (ffic) Memorandum to Jim DeMocker, U.S.
   Environmental Protection Agency, Office of Air and Radiation, Office of Policy Analysis
   and Review. March 31.
Norris, G., S.N. YoungPong, J.Q. Koenig, T.V. Larson, L. Sheppard and J.W. Stout.  1999.  An
   association between fine particles and asthma emergency department visits for children in
   Seattle.  Environ Health Perspect. Vol.  107(6): 489-93.
Ostro, B.D. 1987.  Air Pollution and Morbidity Revisited: a Specification Test. Journal of
   Environmental Economics Management. 14:  87-98.
Ostro, B. and L. Chestnut. 1998. Assessing the Health Benefits of Reducing Particulate Matter
   Air Pollution in the United States. Environmental Research, Section A, 76: 94-106.
Ostro B.D.  and S. Rothschild.  1989.  Air Pollution and Acute Respiratory Morbidity: An
   Observational Study of Multiple Pollutants. Environmental Research 50:238-247.
                                         9-68

-------
                                                                Cost-Benefit Analysis
Ostro, B.D., MJ. Lipsett, M.B. Wiener and J.C. Seiner.  1991.  Asthmatic Responses to Airborne
   Acid Aerosols.  Am J Public Health. Vol. 81(6): 694-702.
Ostro, B., M. Lipsett, J. Mann, H. Braxton-Owens and M. White. 2001. Air pollution and
   exacerbation of asthma in African-American children in Los Angeles. Epidemiology. Vol.
   12(2): 200-8.
Ozkaynak, H. and G.D. Thurston. 1987. Associations between 1980 U.S. mortality rates and
   alternative measures of airborne particle concentration.  Risk Anal. Vol. 7(4): 449-61.
Peters A, Dockery DW, Muller JE, Mittleman MA. 2001. Increased particulate air pollution and
   the triggering of myocardial infarction. Circulation. 103:2810-2815.
Poloniecki JD, Atkinson RW, de Leon AP, Anderson HR. 1997. Daily time series for
   cardiovascular hospital admissions and previous day's air pollution in London, UK. Occup
   Environ M?d54(8):535-40.
Pope, C.A.  2000. Invited Commentary: Particulate Matter-Mortality Expsoure-Response
   Relations and Thresholds. American Journal of Epidemiology., 152: 407-412.
Pope, C.A., III, R.T. Burnett, MJ. Thun, E.E. Calle, D. Krewski, K. Ito, G.D. Thurston. 2002.
   Lung Cancer, Cardiopulmonary Mortality, and Long-term Exposure to Fine Particulate  Air
   Pollution. Journal of the American Medical Association. 287: 1132-1141.
Pope, C.A., III, MJ. Thun, M.M. Namboodiri, D.W. Dockery, J.S. Evans, F.E. Speizer, and
   C.W. Heath, Jr. 1995. Particulate Air Pollution as a Predictor of Mortality in a Prospective
   Study of U.S. Adults. American Journal of Respiratory Critical Care Medicine 151: 669-
   674.
Pope, C.A., III, D.W. Dockery, J.D. Spengler, and M.E. Raizenne. 1991. Respiratory Health
   and PM10 Pollution: a Daily Time Series Analysis American Review of Respiratory Diseases
   144: 668-674.
Ransom, M.R. and C.A. Pope. 1992. Elementary School Absences and PM(10) Pollution in
   Utah Valley. Environmental Research. Vol. 58(2): 204-219.
Rosamond, W.,  G. Broda, E. Kawalec, S. Rywik, A. Pajak, L. Cooper and L. Chambless. 1999.
   Comparison of medical care and survival of hospitalized patients with acute myocardial
   infarction in Poland and the United States. American Journal of Cardiology. 83:1180-5.
Rossi G, Vigotti MA, Zanobetti A, Repetto F, Gianelle V, Schwartz J. 1999.  Air pollution and
   cause-specific mortality  in Milan, Italy, 1980-1989. Arch Environ Health  54(3): 158-64
Rowlatt et al.  1998. Valuation of Deaths from Air Pollution. NERA and CASPAR for DETR.
Russell, M.W., D.M. Huse, S. Drowns, E.C. Hamel and S.C. Hartz. 1998. Direct medical costs
   of coronary artery disease in the United States. Am J Cardiol. Vol. 81(9): 1110-5.
Samet, J.M., S.L. Zeger, I.E. Kelsall, J. Xu and L.S. Kalkstein.   1997.  Air Pollution, Weather,
   and Mortality in Philadelphia 1973-1988. Health Effects Institute.  Cambridge, MA.  March.
Samet JM, Zeger SL, Dominici F, Curriero F, Coursac I, Dockery DW, Schwartz J, Zanobetti A.
   2000. The National Morbidity, Mortality and Air Pollution Study: Part II: Morbidity,
                                         9-69

-------
Final Regulatory Impact Analysis
   Mortality and Air Pollution in the United States.  Research Report No. 94, Part II. Health
   Effects Institute, Cambridge MA, June 2000.
Schwartz, I, Dockery, D.W., Neas, L.M., Wypij, D., Ware, J.H., Spengler, J.D., Koutrakis, P.,
   Speizer, F.E., and Ferris, Jr., E.G.   1994. Acute Effects of Summer Air Pollution on
   Respiratory Symptom Reporting in Children American Journal of Respiratory Critical Care
   Medicine 150: 1234-1242.
Schwartz J, Laden F, Zanobetti A. 2002. The concentration-response relation between PM(2.5)
   and daily deaths. Environmental Health Perspectives 110:1025-9
Schwartz J. 2000. The distributed lag between air pollution and daily deaths. Epidemiology.
   2000 May;ll(3):320-6.
Schwartz, J. 2000.  Assessing  confounding, effect modification, and thresholds in the
   association between ambient particles and daily deaths. Environmental Health Perspectives
   108(6):  563-8.
Schwartz, J. 1995. Short term fluctuations in air pollution and hospital admissions of the elderly
   for respiratory disease. Thorax 50(5):531-8
Schwartz, J. 1993.  Particulate Air Pollution and Chronic Respiratory Disease Environmental
   Research 62: 7-13.
Schwartz J, Dockery DW, Neas LM. 1996. Is daily mortality associated specifically with fine
   particles? J Air  Waste ManagAssoc. 46:927-39.
Schwartz J and Zanobetti A. 2000. Using meta-smoothing to estimate dose-response trends
   across multiple  studies, with  application to air pollution and daily death.
   Epidemiology. 11:666-72.
Schwartz J, Neas LM. 2000. Fine particles are more strongly associated than coarse particles
   with acute respiratory health  effects in schoolchildren. Epidemiology 11:6-10.
Seigneur, C., G. Hidy, I. Tombach, J. Vimont, and P. Amar. 1999. Scientific Peer Review of the
   Regulatory Modeling System for Aerosols and Deposition (REMSAD). Prepared for the
   KEVRIC Company, Inc.
Sheppard, L., D. Levy, G. Norris, T.V. Larson and J.Q. Koenig. 1999. Effects of ambient air
   pollution on nonelderly asthma hospital admissions in  Seattle, Washington,  1987-1994.
   Epidemiology. Vol.  10: 23-30.
Shogren, J.  and T. Stamland. 2002.  Skill and the Value of Life.  Journal of Political Economy.
   110: 1168-1197.
Sisler, J.F. 1996. Spatial and Seasonal Patterns and Long Term Variability of the Composition of
   the Haze in the United States: An Analysis of Data from the IMPROVE Network.
   Cooperative Institute for Research in the Atmosphere,  Colorado State University; Fort
   Collins, CO July.
Smith, D.H., D.C. Malone, K.A.  Lawson, L.J. Okamoto, C. Battista and W.B. Saunders.  1997.
   A national estimate of the economic costs of asthma. Am JRespir Crit Care Med. 156(3 Pt
   1): 787-93.
                                         9-70

-------
                                                               Cost-Benefit Analysis
Smith, V. K., G.Van Houtven, and S.K. Pattanayak. 2002. Benefit Transfer via Preference
   Calibration. Land Economics. 78: 132-152.
Stanford, R., T. McLaughlin and LJ. Okamoto. 1999. The cost of asthma in the emergency
   department and hospital. Am J Respir Crit Care Med. Vol. 160(1): 211-5.
Stieb, D.M., R.T. Burnett, R.C. Beveridge and J.R. Brook.  1996.  Association between ozone
   and asthma emergency department visits in Saint John, New Brunswick, Canada.
   Environmental Health Perspectives. Vol. 104(12): 1354-1360.
Stieb DM, Judek S, Burnett RT. 2002.  Meta-analysis of time-series studies of air pollution and
   mortality: effects of gases and particles and the influence of cause of death, age, and season.
   J Air Waste Manag Assoc 52(4):470-84
Taylor, C.R., K.H. Reichelderfer, and S.R. Johnson.  1993. Agricultural Sector Models for the
   United States: Descriptions and Selected Policy Applications.  Iowa State University Press:
   Ames, IA.
Thurston, G.D. and K. Ito. 2001. Epidemiological studies of acute ozone exposures and
   mortality. J Expo Anal Environ Epidemiol. Vol. 11(4): 286-94.
Tolley, G.S. et al. 1986. Valuation of Reductions in Human Health Symptoms and Risks.
   University of Chicago. Final Report for the US Environmental Protection Agency. January.
Tsuji H, Larson MG, Venditti FJ Jr, Manders ES, Evans JC, Feldman CL, Levy D.  1996. Impact
   of reduced heart rate variability on risk for cardiac events. The Framingham Heart Study.
   Circulation 94(ll):2850-5
US Bureau of the Census. 2002. Statistical Abstract of the United States: 2001. Washington
   DC.
US Department of Commerce, Bureau of Economic Analysis. BEA Regional Projections to
   2045: Vol.  1, States.  Washington, DC US  Govt. Printing Office, July 1995.
US Department of Health and Human Services, Centers for Disease Control and Prevention,
   National Center for Health Statistics.  1999. National Vital Statistics Reports, 47(19).
US Environmental Protection Agency. 2002. Third External Review Draft of Air Quality Criteria
   for Paniculate Matter (April, 2002): Volume II. EPA/600/P-99/002aC
US Environmental Protection Agency. 2003a. Emissions Inventory Technical Support Document
   for the Proposed Nonroad Diesel Engines Rule.
US Environmental Protection Agency. 2003b. Air Quality Technical Support Document for the
   Proposed Nonroad Diesel Engines Rule.
US Environmental Protection Agency, 1996a. Review of the National Ambient Air Quality
   Standards for Ozone: Assessment of Scientific and Technical Information. Office of Air
   Quality Planning and Standards, Research Triangle Park, NC EPA report no. EP A/4521R-
   96-007.
US Environmental Protection Agency, 1996b. Review of the National Ambient Air Quality
   Standards for Particulate Matter: Assessment of Scientific and Technical Information.
                                         9-71

-------
Final Regulatory Impact Analysis
   Office of Air Quality Planning and Standards, Research Triangle Park, NC EPA report no.
   EPA/4521R-96-013.
US Environmental Protection Agency, 1999. The Benefits and Costs of the Clean Air Act, 1990-
   2010. Prepared for US  Congress by US EPA, Office of Air and Radiation/Office of Policy
   Analysis and Review, Washington, DC, November; EPA report no. EPA-410-R-99-001.
US Environmental Protection Agency, 1993. External Draft, Air Quality Criteria for Ozone and
   Related Photochemical Oxidants. Volume II. US EPA, Office of Health and Environmental
   Assessment. Research Triangle Park, NC, EPA/600/AP-93/004b.3v.
US Environmental Protection Agency, 2000a. Regulatory Impact Analysis: Heavy-Duty Engine
   and Vehicle Standards and Highway Diesel Fuel Sulfur Control Requirements. Prepared by:
   Office of Air and Radiation. Available at http://www.epa.gov/otaq/diesel.htm Accessed
   March 20, 2003.
US Environmental Protection Agency, 2000b. Valuing Fatal Cancer Risk Reductions. White
   Paper for Review by the EPA Science Advisory Board.
US Environmental Protection Agency 2000c. Guidelines for Preparing Economic Analyses.
   EPA240-R-00-003. September.
US Environmental Protection Agency, 1997. The Benefits and Costs of the Clean Air Act, 1970
   to 1990. Prepared for US Congress by US EPA, Office of Air and Radiation/Office of
   Policy Analysis and Review, Washington, DC
U.S. EPA (1997).  Regulatory Impact Analysis for Particulate Matter and Ozone National
   Ambient Air Quality Standards and Proposed Regional Haze Rule.  July 1997.
US Environemental Protection Agency, 2002. Technical Addendum: Methodologies for the
   Benefit Analysis of the Clear Skies Initiative. September. Available online at
   http://www.epa.gov/air/clearskies/tech_adden.pdf Accessed March 20, 2003.
U.S. FDA (1995).  U.S. Food and Drug Administration: Procedures for the Safe and Sanitary
   Processing and Importing of Fish and Fishery Products; Final Rule.  60 FR 65095, December
   18, 1995.
U.S. FDA (1996).  U.S. Food and Drug Administration: Regulations Restricting the Sale and
   Distribution of Cigarettes and Smokeless Tobacco to Protect Children and Adolescents, Final
   Rule. 61 FR 44395, August 28, 1996.
U.S. FDA (1997).  U.S. Food and Drug Administration: Quality Mammography Standards, Final
   Rule. 62 FR 55851, October 28, 1997.
U.S. FDA (1998).  U.S. Food and Drug Administration: Food Labeling, Warning and Notice
   Statement, Labeling of Juice Products, Final Rule.  63 FR 37029, July 1998.
U.S. FDA (1999).  U.S. Food and Drug Administration: Food Labeling, Trans Fatty Acids in
   Nutrition Laveling, Nutrient Content Claims, and Health Claims, Proposed Rule. 64 FR
   62746, November 17, 1999.
                                        9-72

-------
                                                               Cost-Benefit Analysis
U.S. FDA (2000). U.S. Food and Drug Administration: Food Labeling, Safe Handling
    Statements, Laveling of Shell Eggs, Refrigeration of Shell Eggs Held for Retail Distribution,
    Final Rule. 65 FR 76091, December 5, 2000.
U.S. FDA (2001). U.S. Food and Drug Administration: Hazard Analysis and Critical Control
    Point, Procedures for the Safe and Sanitary Processing and Importing of Juice, Final Rule.
    66 FR 6137, January  19, 2001

US Office of Management and Budget. 1992. Guidelines and Discount Rates for Benefit-Cost
    Analysis of Federal Programs. Circular No. A-94. October.
Vedal,  S.,  J. Petkau, R. White and J. Blair.  1998.  Acute effects of ambient inhalable particles in
    asthmatic and nonasthmatic children. American Journal of Respiratory and Critical Care
    Medicine. Vol.  157(4): 1034-1043.
Viscusi, W.K. 1992. Fatal Tradeoffs: Public and Private Responsibilities for Risk. (New York:
    Oxford University Press).
Viscusi W.K. and J.E. Aldy. 2003 forthcoming.  "The Value of A Statistical Life: A Critical
    Review of Market Estimates Throughout the World."  Journal of Risk and Uncertainty.
Viscusi, W.K., W.A. Magat, and J. Huber.  1991. "Pricing Environmental Health Risks: Survey
    Assessments  of Risk-Risk and Risk-Dollar Trade-Offs for Chronic Bronchitis" Journal of
    Environmental Economics and Management, 21: 32-51.
VTT Processes for the European Commission, "Feasibility Study on a Third Stage of Emission
    Limits for Compression Ignition Engines with a Power Output Between 18 and 560 kW"
Weisel, C.P., R.P. Cody and P.J. Lioy. 1995. Relationship between summertime ambient ozone
    levels and emergency department visits for asthma in central New Jersey.  Environ Health
    Perspect. Vol. 103 Suppl 2:  97-102.
Whittemore, A.S. and E.L. Korn. 1980.  Asthma and Air Pollution in the Los Angeles Area.
    American Journal of Public Health. 70: 687-696.
Wittels, E.H., J.W. Hay and A.M. Gotto, Jr. 1990. Medical costs of coronary artery disease in
    the United States.  Am J Cardiol. Vol. 65(7): 432-40.
Woodruff, T.J., J. Grillo and K.C. Schoendorf.  1997.  The relationship between selected causes
    of postneonatal  infant mortality and particulate air pollution in the United States.
    Environmental Health Perspectives. Vol. 105(6): 608-612.
Woods & Poole Economics Inc. 2001. Population by Single Year of Age CD. Woods & Poole
    Economics, Inc.
World Health Organization. 2002.  "Global Burden of Disease Study." World Health
    Organization.
Yu, O., L. Sheppard, T. Lumley, J.Q. Koenig and G.G. Shapiro. 2000.  Effects of Ambient Air
    Pollution on Symptoms of Asthma in Seattle-Area Children Enrolled in the CAMP Study.
    Environ Health  Perspect. Vol. 108(12): 1209-1214.
                                         9-73

-------
Final Regulatory Impact Analysis
Zanobetti, A., J. Schwartz, E. Samoli, A. Gryparis, G. Touloumi, R. Atkinson, A. Le Tertre, J.
   Bobros, M. Celko, A. Goren, B. Forsberg, P. Michelozzi, D. Rabczenko, E. Aranguez Ruiz
   and K. Katsouyanni.  2002.  The temporal pattern of mortality responses to air pollution: a
   multicity assessment of mortality displacement. Epidemiology. Vol. 13(1): 87-93.
                                         9-74

-------
                                                                  Cost-Benefit Analysis
APPENDIX 9A: Benefits Analysis of Modeled Preliminary Control Option

   9A.1 Summary of Emissions Inventories and Modeled Changes in Emissions from Nonroad Engii§te89
   9A.2 Air Quality Impacts  	9-92
      9A.2.1 PM Air Quality Estimates	9-93
          9A.2.1.1 Modeling Domain  	9-95
          9A.2.1.2 Simulation Periods	9-95
          9A.2.1.3 Model Inputs  	9-98
          9A.2.1.4 Converting REMSAD Outputs to Benefits Inputs	9-98
          9A.2.1.5 PM Air Quality Results	9-99
      9A.2.2 Ozone Air Quality Estimates  	9-102
          9A.2.2.1 Modeling Domain  	9-103
          9A.2.2.2 Simulation Periods	9-106
          9A.2.2.3 Converting CAMx Outputs to Full-Season Profiles for Benefits Analysis ....  9-106
          9A.2.2.4 Ozone Air Quality Results	9-107
      9A.2.3 Visibility Degradation Estimates 	9-111
          9A.2.3.1 Residential Visibility Improvements	9-113
          9A.2.3.2 Recreational Visibility Improvements	9-114
   9A.3 Benefit Analysis- Data and Methods 	9-116
      9A.3.1 Valuation Concepts	9-118
      9A.3.2 Growth in WTP Reflecting National Income Growth Over Time  	9-120
      9A.3.3 Methods for Describing Uncertainty	9-123
      9A.3.4 Demographic Projections  	9-128
      9A.3.5 Health Benefits Assessment Methods	9-129
      9A.3.5.1 Selecting Health Endpoints and Epidemiological Effect Estimates	9-131
      9A.3.5.2 Uncertainties Associated with Health Impact Functions	9-147
      9A.3.5.3 Baseline Health Effect Incidence Rates	9-151
      9A.3.6 Human Welfare Impact Assessment	9-173
          9A.3.6.1 Visibility Benefits  	9-173
          9A.3.6.2 Agricultural, Forestry and other Vegetation Related Benefits  	9-176
             9A.3.6.2.1 Agricultural Benefits 	9-177
             9A.3.6.2.2 Forestry Benefits  	9-178
             9A.3.6.2.3 Other Vegetation Effects  	9-178
          9A.3.6.3 Benefits  from Reductions in Materials Damage and Odor 	9-179
          9A.3.6.4 Benefits  from Reduced Ecosystem Damage 	9-180
   9A.4 Benefits Analysis—Results  	9-181
                                           9-75

-------
Final Regulatory Impact Analysis
    This appendix details the models and methods used to generate the benefits estimates from
which the benefits of the final standards presented in Chapter IX are derived.  This analysis uses
a methodology generally consistent with benefits analyses performed for the recent analysis of
the Heavy Duty Engines/Diesel Fuel rulemaking (U.S. EPA, 2000a) and the proposed Interstate
Air Quality Rule (U.S. EPA, 2004).  The benefits analysis relies on three major modeling
components:

    1) Calculation of the impact that a set of preliminary fuel and engine standards would have
       on the nationwide inventories for NOx, non-methane hydrocarbons (NMHC), SO2, and
       PM emissions in 2020 and 2030;

    2) Air quality modeling for 2020 and 2030 to determine changes in ambient concentrations
       of ozone and particulate matter, reflecting baseline and post-control emissions
       inventories.

    3) A benefits analysis to determine the changes in human health and welfare,  both in terms
       of physical effects and monetary value, that result from the projected changes in ambient
       concentrations of various pollutants for the modeled standards.

    Potential human health effects linked to PM2 5 range from premature mortality  linked to long-
term exposure to PM, to a range of morbidity effects linked to long-term (chronic) and shorter-
term (acute) exposures (e.g., respiratory and cardiovascular symptoms resulting in hospital
admissions,  asthma exacerbations, and acute and chronic bronchitis). Exposure to ozone has
also been linked to a variety of respiratory effects including hospital admissions and illnesses
resulting in school absences.3 Welfare effects potentially linked to PM include materials damage
and visibility impacts, while ozone can adversely affect the agricultural and forestry sectors by
decreasing yields of crops and forests.  Although methods exist for quantifying the benefits
associated with many of these human health and welfare categories, not all can be  evaluated at
this time due to limitations in methods and/or data. Table 4-1 lists the full complement of human
health and welfare effects associated with PM and ozone and identifies those effects that are
quantified for the primary estimate,  are quantified as part of the sensitivity analysis (to be
completed for the supplemental analysis), and remain unquantified because of to current
limitations in methods or available data.
   AShort-term exposure to ambient ozone has also been linked to premature death. The EPA is currently
evaluating the epidemiological literature examining the relationship between ozone and premature mortality,
sponsoring three independent meta-analyses of the literature. Once this evaluation has been completed and peer-
reviewed, the EPA will consider including ozone-related premature mortality in the primary benefits analysis for
future rules.

                                           9-76

-------
                                                                 Cost-Benefit Analysis
   Figure 9A.1 illustrates the major steps in the analysis. Given baseline and post-control
emissions inventories for the emission species expected to impact ambient air quality, we use
sophisticated photochemical air quality models to estimate baseline and post-control ambient
concentrations of ozone and PM, and deposition of nitrogen and sulfur for each year. The
estimated changes in ambient concentrations are then combined with monitoring data to estimate
population level exposures to changes in ambient concentrations for use in estimating health
effects.  Modeled changes in ambient data are also used to estimate changes in visibility, and
changes in other air quality  statistics that are necessary to estimate welfare effects. Changes in
population exposure to ambient air pollution are then input to concentration-response functions
to generate changes in incidence of health effects, or, changes in other exposure metrics are input
to dose-response functions to generate changes in welfare effects.  The resulting effects changes
are then assigned monetary  values, taking into account adjustments to values  for growth in real
income out to the year of analysis (values for health and welfare effects are in general positively
related to real income levels). Finally, values for individual health and welfare effects are
summed to obtain an estimate of the total monetary value of the changes in emissions.

   On September 26, 2002, the National Academy of Sciences (NAS) released a report on its
review of the Agency's methodology for analyzing the health benefits of measures taken to
reduce air pollution.  The report focused on the EPA's approach for estimating the health
benefits of regulations designed to reduce concentrations of airborne PM.

   In its report, the NAS said that the EPA has generally used a reasonable framework for
analyzing the health benefits of PM-control measures. It recommended, however, that the
Agency take a number of steps to improve its benefits analysis. In particular, the NAS stated
that the Agency should

      •  include benefits  estimates for a range of regulatory options;

      •  estimate benefits for intervals, such as every 5 years, rather than a single year;

      •  clearly state the projected baseline statistics used in estimating health benefits,
          including those for air emissions, air quality, and health outcomes;

      •  examine whether implementation of regulations might cause unintended impacts on
          human health or the environment;

      •  when appropriate, use data from non-U.S. studies to broaden age ranges to which
          current estimates apply and to include more types of relevant health outcomes; and

      •  begin to move the assessment of uncertainties from its ancillary analyses into its base
          analyses by  conducting probabilistic, multiple-source uncertainty analyses. This
          assessment should be based on available data and expert judgment.

                                          9-77

-------
Final Regulatory Impact Analysis
   Although the NAS made a number of recommendations for improvement in the EPA's
approach, it found that the studies selected by the Agency for use in its benefits analysis were
generally reasonable choices.  In particular, the NAS agreed with the EPA's decision to use
cohort studies to derive benefits estimates. It also concluded that the Agency's selection of the
American Cancer Society (ACS) study for the evaluation of PM-related premature mortality was
reasonable, although it noted the publication  of new cohort studies that the Agency should
evaluate. Since the publication of the NAS report, the EPA has reviewed new cohort studies,
including reanalyses of the ACS  study data and has carefully considered these new study data in
developing the analytical approach for the final rule (see below).

   In addition to the NAS report, the EPA has also received technical guidance and input
regarding its methodology for conducting PM- and ozone-related benefits analysis from two
additional sources, including the Health Effects Subgroup (HES) of the  SAB Council reviewing
the 812 blueprint (SAB-HES, 2003) and the Office of Management and Budget (OMB) through
ongoing discussions regarding methods used in conducting regulatory impact analyses (RIAs)
(e.g., see OMB Circular A-4).  The SAB HES recommendations include the following (SAB-
HES, 2003):

      •   use of the updated ACS Pope et al. (2002) study rather than the ACS Krewski et al.
          study to estimate premature mortality for the primary analysis;

      •   dropping the alternative estimate used in the proposal  RIA and instead including a
          primary estimate that incorporates consideration of uncertainly in key effects
          categories such as premature mortality directly into the estimates (e.g., use of the
          standard errors from the Pope et al. [2002] study in deriving  confidence bounds for
          the adult mortality estimates);

      •   addition of infant mortality (children under the age of one) into the primary estimate,
          based on supporting evidence from the World Health Organization Global Burden of
          Disease study and other published studies that strengthen the evidence for a
          relationship between PM exposure and respiratory inflamation and infection in
          children leading to death;

      •   inclusion of asthma exacerbations for children in the primary estimate;

      •   expansion of the age groups evaluated for a range of morbidity effects beyond the
          narrow band of the studies to the broader (total) age group (e.g., expanding a study
          population for 7 to 11 year olds to cover the entire child age  range of 6 to 18 years).

      •   inclusion of new endpoints (school absences [ozone], nonfatal heart attacks in adults
          [PM], hospital admissions for children under two [ozone]), and suggestion of a new
                                          9-78

-------
                                                                   Cost-Benefit Analysis
          meta-analysis of hospital admissions (PM10) rather than using a few PM25 studies;13
          and

       •  updating of populations and baseline incidences.

   Recommendations from OMB regarding RIA methods have focused on the approach used to
characterize uncertainty in the benefits estimates generated for RIAs, as well as the approach
used to value premature mortality estimates. The EPA is currently in the process of developing a
comprehensive integrated strategy for characterizing the impact of uncertainty in key elements of
the benefits modeling process (e.g., emissions modeling, air quality modeling, health effects
incidence estimation, valuation) on the results that are generated.

   We are also altering the value of a statistical life (VSL) used in the analysis to reflect new
information in the ongoing academic debate over the appropriate characterization of the value of
reducing the risk of premature mortality. In previous analyses, we used a distribution of VSL
based on 26 VSL estimates from the economics literature. For this analysis, we are
characterizing the VSL distribution in a  more general fashion, based on two recent meta-analyses
of the wage-risk-based VSL literature. The new distribution is assumed to be normal, with a
mean of $5.5 million and a 95 percent confidence interval between $1 and $10 million.  The
EPA welcomes public comment on the appropriate methodology for valuing reductions in the
risk of premature death.

   The EPA has addressed the comments received from the public, the NAS, the  SAB-HES, and
OMB in developing the analytical approach for the final rule. We have also reflected advances in
data and methods in air quality modeling, epidemiology, and economics that have occurred  since
the proposal analysis. Updates to the assumptions and methods used in estimating PM2 5-related
and ozone-related benefits since completion of the Proposed Nonroad Diesel Rule include the
following:

Health Endpoints

       •  The primary analysis incorporates updated impact functions to reflect updated time-
          series studies of hospital admissions to correct for errors in application of the
    BNote that the SAB-HES comments were made in the context of a review of the methods for the Section 812
analysis of the costs and benefits of the Clean Air Act. This context is pertinent to our interpretation of the SAB-
HES comments on the selection of effect estimates for hospital admissions associated with PM (SAB-HES, 2003).
The Section 812 analysis is focused on a broad set of air quality changes, including both the coarse and fine fractions
of PM10.  As such, impact functions that focus on the full impact of PM10 are appropriate. However, for the Nonroad
Diesel Engines rule, which is expected to affect primarily the fine fraction (PM25) of PM10, impact functions that
focus primarily on PM2 5 are more appropriate.

                                           9-79

-------
Final Regulatory Impact Analysis
          generalized additive model (GAM) functions in S-plus. More information on this
          issue is available at http://www.healtheffects.org.

       •   The primary analysis uses an all cause mortality effect estimate based on the Pope et
          al. (2002) reanalysis of the ACS study data.  In addition, we provide a breakout for
          two major cause of death categories—cardiopulmonary and lung cancer.

       •   Infant mortality is included in the primary analysis (infants age 0-1 years).

       •   Asthma exacerbations are incorporated into the primary analysis. Although the
          analysis of the proposed rule included asthma exacerbations as a separate endpoint
          outside of the base case analysis, for the final rule, we will include asthma
          exacerbations in children 6 to 18 years of age as part of the primary analysis.

Valuation
       •   In generating the monetized benefits for premature mortality in the primary analysis,
          the VSL is entered as a mean (best estimate) of 5.5 million.  Unlike the analysis of the
          proposed rule, the analysis of the final rule does not include a value of statistical life
          year (VSLY) estimate.

   The analysis of the proposed rule included an alternative estimate in addition to the primary
estimate that was intended to evaluate the impact of several key assumptions on the estimated
reductions in premature premature mortality and chronic bronchitis.  However, reflecting
comments from the public, the SAB-HES as well as the NAS panel, rather than including an
alternative estimate in the analysis, the EPA will investigate the impact of key assumptions on
mortality and morbidity estimates through a series of sensitivity analyses.

   The benefits estimates generated for the final Nonroad Diesel Engine rule are subject to a
number of assumptions and uncertainties, which are discussed throughout the document. For
example, key assumptions underlying the primary estimate for the premature mortality category
include the following:

   (1) Inhalation of fine particles is causally associated with premature death at concentrations
       near those experienced by most Americans on a daily basis.  Although biological
       mechanisms for this effect have not yet been definitively established, the weight of the
       available epidemiological evidence supports an assumption of causality.
   (2) All fine particles, regardless of their chemical composition, are equally potent in causing
       premature mortality. This is an important assumption, because PM produced via
       transported precursors emitted from EGUs may differ significantly from direct PM
       released from diesel engines and other industrial sources, but no clear scientific grounds
       exist for supporting differential effects estimates by particle type.
                                          9-80

-------
                                                                 Cost-Benefit Analysis
    (3) The impact function for fine particles is approximately linear within the range of ambient
       concentrations under consideration.  Thus, the estimates include health benefits from
       reducing fine particles in areas with varied concentrations of PM, including both regions
       that are in attainment with fine particle standard and those that do not meet the standard.
    (4) The forecasts for future emissions and associated air quality modeling are valid.
       Although recognizing the difficulties, assumptions, and inherent uncertainties in the
       overall enterprise, these analyses are based on peer-reviewed scientific literature and
       up-to-date assessment tools, and we believe the results are highly useful in assessing this
       rule.

    In addition to the quantified and monetized benefits summarized above, a number of
additional categories are not currently amenable to quantification or valuation.  These include
reduced acid and particulate deposition damage to cultural monuments and other materials,
reduced odor, reduced ozone effects on forested ecosystems, and environmental benefits due to
reductions of impacts of acidification in lakes and streams and eutrophication in coastal areas.
Additionally, we have not quantified a number of known or suspected health effects linked with
PM and ozone for which appropriate health impact functions are not available or which do not
provide easily  interpretable outcomes (i.e., changes in forced expiratory volume [FEV1]).  As a
result, monetized benefits generated for the primary estimate may underestimate the total
benefits attributable to the final regulatory option.

    Benefits estimates for the final Nonroad Diesel Engines rule were generated using BenMAP,
which is a computer program developed by the EPA that integrates a number of the modeling
elements used  in previous RIAs (e.g., interpolation functions, population projections, health
impact functions, valuation functions, analysis and pooling methods) to translate modeled air
concentration estimates into health effects incidence estimates and monetized benefits estimates.
BenMAP  provides estimates of both the mean impacts and the distribution of impacts.

    In general, the chapter is organized around the steps illustrated in Figure 9A. 1. In section A,
we describe and summarize the emissions inventories and modeled reductions in emissions of
NOx, VOC, SO2, and directly emitted diesel PM for the set of preliminary control options.  In
section B, we describe and summarize the air quality models and results, including both baseline
and post-control conditions, and discuss the way modeled air quality changes are used in the
benefits analysis.  In Section C, we provide and overview of the data and methods that  are used
to quantify and value health and welfare  endpoints, and provide a discussion of how we
incorporate uncertainty into our analysis. In Section D, we report the results of the analysis for
human health and welfare effects. Additional sensitivity analyses are provided in Appendix 9B
and 9C.
                                          9-81

-------
Final Regulatory Impact Analysis
                     Table 9A.1.  Summary of Results: Estimated Benefits
                          of the Modeled Preliminary Control Option
Discount Rate
3% discount rate
7% discount rate
Total BenefitsA B
(Billions 2000$)
2020
$52+B
$49+B
2030
$92+B
$87+B
A Benefits of CO and HAP emission reductions are not quantified in this analysis and, therefore, are not presented in
this table. The quantifiable benefits are from emission reductions of NOX, NMHC, SO2 and PM only. For
notational purposes, unqualified benefits are indicated with a "B" to represent the sum of additional monetary
benefits and disbenefits. A detailed listing of unqualified health and welfare effects is provided in Table 9A-2.
B Results reflect the use of 3% and 7% discount rates consistent with EPA and OMB's guidelines for preparing economic
analyses (US EPA, 2000c, OMB Circular A-4). Results are rounded to two significant digits.
                                                9-82

-------
                                                                      Cost-Benefit Analysis
Figure 9A. 1.  Key Steps in Air Quality Modeling Based Benefits Analysis

 INPUTS                              PROCESSES
                                                                                   INPUTS
  Emission Inventories
  (1996 NET, Mobile
  5b, NONROAD)
   Air Quality
   Monitor Data
   (AIRS)
 .	.
   Concentration-
   Response Functions
   Incidence and
   Prevalence Rates
   Population and
   Demographic Data
  Valuation Functions
r
GDP Projections
i
  Income Elasticities   !
                                         Model baseline and post-
                                         control ambient air quality
                                           (REMSAD, CAM-X)
                          Model Population Exposure
                            to Changes in Ambient
                               Concentrations
                                 (BenMAP)
                                  T
                           Estimate Expected
                           Changes in Human
                           Health Outcomes
                              (BenljlAP)
                      Estimate Expected
                      Changes inWelfare
                          Effects
                            Estimate Monetary
                             Value of Changes
                             in Human Health
                                Outcomes
                                (BenMAP)
                                                 Estimate Monetary
                                                Value of Changes in
                                                  Welfare Effects
Adjust Monetary Values for Growth in Real
       Income to Year of Analysis
                             Sum Health and Welfare Monetary Values to
                                  Obtain Total Monetary Benefits
                                                                              Dose-response
                                                                                Functions
                                               Sector Models
                                               (AGSDVI)
                                                                            Valuation Functions
                                             9-83

-------
                                                                  Table 9A.2.
               Human Health and Welfare Effects of Pollutants Affected by the Final Nonroad Diesel Engine Rule
Pollutant/Effect
Quantified and Monetized Effects in
         Primary Analysis
Quantified and/or Monetized Effects
      in Sensitivity Analyses
Unquantified Effects
PM/Health
Premature mortality in adults - long
    term exposures
Infant mortality
Bronchitis - chronic and acute
Hospital admissions - respiratory
    and cardiovascular
Emergency room visits for asthma
Non-fatal heart attacks (myocardial
    infarction)
Asthma exacerbations
Lower and upper respiratory illness
Minor restricted activity days
Work loss days
                                   Low birth weight
                                   Changes in pulmonary function
                                   Chronic respiratory diseases other than chronic bronchitis
                                   Morphological changes
                                   Altered host defense mechanisms
                                   Cancer
                                   Non-asthma respiratory emergency room visits
                                   Changes in cardiac function (e.g. heart rate variability)
                                   Allergic responses (to diesel exhaust)
PM/Welfare
Visibility in California,
Southwestern, and Southeastern
Class I areas
Visibility in Northeastern,
    Northwestern, and Midwestern
    Class I areas
Visibility in residential and non-
    Class I areas
Household soiling

-------
Pollutant/Effect
 Quantified and Monetized Effects in
	Primary Analysis	
 Quantified and/or Monetized Effects
	in Sensitivity Analyses	
Unquantified Effects
Ozone/Health
                                                                         Increased airway responsiveness to stimuli
                                                                         Inflammation in the lung
                                                                         Chronic respiratory damage
                                                                         Premature aging of the lungs
                                                                         Acute inflammation and respiratory cell damage
                                                                         Increased susceptibility to respiratory infection
                                                                         Non-asthma respiratory emergency room visits
                                                                         Hospital admissions - respiratory
                                                                         Emergency room visits for asthma
                                                                         Minor restricted activity days
                                                                         School loss days
                                                                         Chronic Asthma3
                                                                         Asthma attacks
                                                                         Cardiovascular emergency room visits
                                                                         Premature mortality - acute exposures'1
                                                                         Acute respiratory symptoms
Ozone/Welfare
                                                                         Decreased commercial forest productivity
                                                                         Decreased yields for fruits and vegetables
                                                                         Decreased yields for commercial and non-commercial crops
                                                                         Damage to urban ornamental plants
                                                                         Impacts on recreational demand from damaged forest aesthetics
                                                                         Damage to ecosystem functions
                                                                         Decreased outdoor worker productivity
Nitrogen and
Sulfate
Deposition/
Welfare
                                     Costs of nitrogen controls to reduce
                                         eutrophication in selected
                                         eastern estuaries
                                     Impacts of acidic sulfate and nitrate deposition on commercial
                                         forests
                                     Impacts of acidic deposition on commercial freshwater fishing
                                     Impacts of acidic deposition on recreation in terrestrial
                                         ecosystems
                                     Impacts of nitrogen deposition on commercial fishing,
                                         agriculture, and forests
                                     Impacts of nitrogen deposition on recreation in estuarine
                                         ecosystems
                                     Reduced existence values for currently healthy ecosystems

-------
 Pollutant/Effect
 Quantified and Monetized Effects in
	Primary Analysis	
 Quantified and/or Monetized Effects
	in Sensitivity Analyses	
Unquantified Effects
  SCX/Health
                                                                         Hospital admissions for respiratory and cardiac diseases
                                                                         Respiratory symptoms in asthmatics
 NOx/Health
                                                                         Lung irritation
                                                                         Lowered resistance to respiratory infection
                                                                         Hospital Admissions for respiratory and cardiac diseases
 CO/Health
                                                                         Premature mortality
                                                                         Behavioral effects
                                                                         Hospital admissions - respiratory, cardiovascular, and other
                                                                         Other cardiovascular effects
                                                                         Developmental effects
                                                                         Decreased time to onset of angina
 NMHCsc
 Health
                                                                         Cancer (diesel PM, benzene, 1,3-butadiene, formaldehyde,
                                                                             acetaldehyde)
                                                                         Anemia (benzene)
                                                                         Disruption of production of blood components (benzene)
                                                                         Reduction in the number of blood platelets (benzene)
                                                                         Excessive bone marrow formation (benzene)
                                                                         Depression of lymphocyte counts (benzene)
                                                                         Reproductive and developmental effects (1,3-butadiene)
                                                                         Irritation of eyes and mucous membranes (formaldehyde)
                                                                         Respiratory and respiratory tract
                                                                         Asthma attacks in asthmatics (formaldehyde)
                                                                         Asthma-like symptoms in non-asthmatics (formaldehyde)
                                                                         Irritation of the eyes, skin, and respiratory tract (acetaldehyde)
                                                                         Upper respiratory tract irritation & congestion (acrolein)
 NMHCsc
 Welfare
                                                                         Direct toxic effects to animals
                                                                         Bioaccumlation in the food chain
                                                                         Reduced odors
a While no causal mechanism has been identified linking new incidences of chronic asthma to ozone exposure, two epidemiological studies shows a statistical
association between long-term exposure to ozone and incidences of chronic asthma in exercising children and some non-smoking men (McConnell, 2002; McDonnell,
etal, 1999).
b Premature mortality associated with ozone is not separately included in the calculation of total monetized benefits.
c All non-methane hydrocarbons (NMHCs) listed in the table are also hazardous air pollutants listed in Section 112(b) of the Clean Air Act.

-------
                                                               Cost-Benefit Analysis
9A.1 Summary of Emissions Inventories and Modeled Changes in Emissions
from Nonroad Engines

   For the preliminary control options we modeled, implementation will occur in two ways:
reduction in sulfur content of nonroad diesel fuel and adoption of controls on new engines.
Because full turnover of the fleet of nonroad diesel engines will not occur for many years, the
emission reduction benefits of the final standards will not be fully realized until decades after the
initial reduction in fuel sulfur content.  Based on the projected time paths for emissions
reductions, EPA chose to focus detailed emissions and air quality modeling on two future years,
2020 and 2030, which reflect partial and close to complete turnover of the fleet of nonroad diesel
engines to models meeting the preliminary control options. Tables 9A-3 and 9A-4 summarize
the baseline emissions of NOX, SO2, VOC, and direct diesel PM2 5 and the change in the
emissions from nonroad engines used in modeling air quality changes.

   Emissions and air quality modeling decisions are made early in the analytical process.  Since
the preliminary control scenario was developed, EPA has gathered more information and
received public comment regarding the technical feasibility of the standards, and EPA has
revised the control scenario accordingly. Section 3.6 of the RIA describes the changes in the
inputs and resulting emission inventories between the preliminary baseline and  control scenarios
used for the air quality modeling and the baseline and control scenarios.

   Chapter 3 discussed the development of the 1996, 2020 and 2030 baseline emissions
inventories for the nonroad sector and for the sectors not affected by this rule. The emission
sources and the basis for current and future-year inventories are listed in Table 9A-5.
                                         9-87

-------
Final Regulatory Impact Analysis
                                  Table 9A-3
    Summary of Baseline Emissions for Preliminary Nonroad Engine Control Options

Source
Pollutant Emissions (tons)
NOX
SO2
voc
PM25
1996 Baseline
Nonroad Engines
All Other Sources
Total, All Sources
1,583,641
22,974,945
24,558,586
172,175
18,251,679
18,423,854
221,398
18,377,795
18,599,193
178,500
2,038,726
2,217,226
2020 Base Case
Nonroad Engines
All Other Sources
Total, All Sources
1,144,686
14,394,399
15,539,085
308,075
14,882,962
15,191,037
97,113
13,812,619
13,909,732
127,755
1,940,307
2,068,062
2030 Base Case
Nonroad Engines
All Other Sources
Total, All Sources
1,231,981
14,316,841
15,548,822
360,933
15,190,439
15,551,372
97,345
15,310,670
15,408,015
143,185
2,066,918
2,210,103

-------
                                                                  Cost-Benefit Analysis
                                       Table 9A-4
      Summary of Emissions Changes for the Preliminary Nonroad Control Options*

Item
Pollutant
NOX
SO2
voc
PM25
2020 Nationwide Emission Changes
Absolute Tons
Percent Reduction from Landbased
Nonroad Emissions
Percentage Reduction from All
Manmade Sources
663,618
58.0%
4.5%
304,735
98.9%
2.1%
23,172
23.9%
0.2%
91,278
71.4%
4.6%
2030 Emission Changes
Absolute Tons
Percent Reduction from Landbased
Nonroad Emissions
Percentage Reduction from All
Manmade Sources
1,009,744
82.0%
6.3%
359,774
99.7%
2.1%
34,060
35.0%
0.2%
129,073
90.0%
5.5%
* Does not include SO2 and PM2 5 reductions from recreational marine diesel engines, commercial marine diesel
engines, and locomotives due to control of diesel fuel sulfur levels.
                                           9-89

-------
Final Regulatory Impact Analysis
                                      Table 9A-5
          Emissions Sources and Basis for Current and Future-Year Inventories
 Emissions Source
1996 Base year
Future-year Base Case Projections
 Utilities
1996 NEI Version 3.12
(CEM data)
Integrated Planning Model (IPM)
 Non-Utility Point and Area
 sources
1996 NEI
Version 3.12 (point)
Version 3.11 (area)
BEA growth projections
 Highway vehicles
MOBILESb model with
MOBILE6 adjustment
factors for VOC and
NOX;
PARTS model for PM
VMT projection data
 Nonroad engines (except
 locomotives, commercial
 marine vessels, and
 aircraft)
NONROAD2002 model
BEA and Nonroad equipment
growth projections
Note: Full description of data, models, and methods applied for emissions inventory development and modeling are
provided in Emissions Inventory TSD (EPA, 2003a).
9A.2 Air Quality Impacts

   This section summarizes the methods for and results of estimating air quality for the 2020
and 2030 base cases and control scenarios for the purposes of benefit-cost analyses. EPA has
focused on the health, welfare, and ecological effects that have been linked to air quality
changes.  These air quality changes include the following:

   -  Ambient paniculate matter (PM10 and PM2 5)-as estimated using a national-scale version
       of the REgional Modeling System for Aerosols and Deposition (REMSAD);

   -  Ambient ozone-as estimated using regional-scale applications of the Comprehensive Air
       Quality Model with Extensions (CAMx); and

   -  Visibility degradation (i.e., regional haze), as developed using empirical estimates of
       light extinction coefficients and efficiencies in combination with REMSAD modeled
       reductions in pollutant concentrations.
                                         9-90

-------
                                                                  Cost-Benefit Analysis
Although we expect reductions in airborne sulfur and nitrogen deposition, these air quality
impacts have not been quantified for this rule nor have the associated benefits been estimated.

    The air quality estimates in this section are based on the emission changes for the modeled
preliminary control program discussed in Chapter 3. These air quality results are in turn
associated with human populations and ecosystems to estimate changes in health and welfare
effects. In Section B-l, we describe the estimation of PM air quality using REMSAD, and in
Section B-2, we cover the estimation of ozone air quality using CAMx. Lastly, in Section B-3,
we discuss the  estimation of visibility degradation.

9A.2.1 PM Air Quality Estimates

    We use the emissions inputs summarized above with a national-scale version of the REgional
Model System for Aerosols and Deposition (REMSAD) to estimate PM air quality in the
contiguous U.S. REMSAD is  a three-dimensional grid-based Eulerian air quality model
designed to estimate annual paniculate concentrations and deposition over large spatial scales
(e.g., over the contiguous U.S.). Consideration of the different processes that affect primary
(directly emitted) and secondary (formed by atmospheric processes) PM at the regional scale in
different locations is fundamental to understanding and assessing the effects of pollution control
measures that affect ozone, PM and deposition of pollutants to the surface.0  Because it accounts
for spatial and temporal variations as well as differences in the reactivity  of emissions,
REMSAD is useful for evaluating the impacts of the rule on U.S. PM concentrations.

    REMSAD was peer-reviewed in 1999 for EPA as reported in "Scientific Peer-Review of the
Regulatory Modeling System for Aerosols and Deposition" (Seigneur et al., 1999).  Earlier
versions of REMSAD have been employed for the EPA's Prospective 812 Report to Congress,
EPA's HD Engine/Diesel Fuel rule, and EPA's air quality assessment of the Clear Skies
Initiative.  Version 7 of REMSAD was employed for this analysis and is fully described in the
air quality modeling technical  support document (US EPA, 2003b). This version reflects
updates in the following areas  to improve performance and address comments from the 1999
peer-review:

    -  Gas phase chemistry updates to "micro-CB4" mechanism including new treatment for the
       NO3 and N2O5 species and the addition of several reactions to better account for the
    c Given the potential impact of the Nonroad Engine/Diesel Fuel rule on secondarily formed particles it is
important to employ a Eulerian model such as REMSAD. The impact of secondarily formed pollutants typically
involves primary precursor emissions from a multitude of widely dispersed sources, and chemical and physical
processes of pollutants that are best addressed using an air quality model that employs an Eulerian grid model
design.

                                          9-91

-------
Final Regulatory Impact Analysis
       wide ranges in temperature, pressure, and concentrations that are encountered for
       regional and national applications.

   -  PM chemistry updates to calculate particulate nitrate concentrations through use of the
       MARS-A equilibrium algorithm and internal calculation of secondary organic aerosols
       from both biogenic (terpene) and anthropogenic (estimated aromatic) VOC emissions.

   -  Aqueous phase chemistry updates to incorporate the oxidation of SO2 by O3 and O2 and
       to include the cloud and rain liquid water content from MM5 meteorological data directly
       in sulfate production and deposition calculations.

   As discussed earlier in Chapter 2, the model tends to underestimate observed PM2 5
concentrations nationwide, especially over the western U.S.

   Our analysis applies the modeling system to the entire U.S. for the five emissions scenarios:
a 1996 baseline projection, a 2020 baseline projection and a 2020 projection with nonroad
controls, a 2030 baseline projection and a 2030 projection with nonroad controls. As discussed
in the Benefits Analysis TSD, we use the relative predictions from the model by combining the
1996 base-year and each future-year scenario with ambient  air quality observations to determine
the expected change in 2020 or 2030 ozone concentrations due to the rule (Abt Associates,
2003). These results are used solely in the benefits analysis.

   REMSAD simulates every hour of every day of the year and, thus, requires a variety of input
files that contain information pertaining to the modeling domain and simulation period.  These
include gridded, 1-hour average emissions estimates and meteorological fields, initial and
boundary conditions, and land-use information. As applied to the contiguous U.S., the model
segments the area within the region into square blocks called grids (roughly  equal in size to
counties), each of which has several layers of air conditions. Using this data, REMSAD
generates predictions of 1-hour average PM concentrations for every grid. We then calibrate the
modeling results to develop 2020 and 2030 PM estimates at monitor sites by normalizing the
observations to the observed 1996 concentrations at each monitor site. For areas (grids) without
PM monitoring data, we interpolated concentration values using data from monitors surrounding
the area. After completing this process, we then calculated  daily and seasonal PM air quality
metrics as inputs to the health and welfare C-R functions of the benefits analysis. The following
sections provide a more detailed discussion of each of the steps in this evaluation and a summary
of the results.
                                          9-92

-------
                                                                Cost-Benefit Analysis
   9A.2.1.1 Modeling Domain

    The PM air quality analyses employed the modeling domain used previously in support of
Clear Skies air quality assessment. As shown in Figure 9A-2, the modeling domain
encompasses the lower 48 States and extends from 126 degrees to 66 degrees west longitude and
from 24 degrees north latitude to 52 degrees north latitude.  The model contains horizontal grid-
cells across the model domain of roughly 36 km by 36 km.  There are 12 vertical layers of
atmospheric conditions with the top of the modeling domain at 16,200 meters.  The 36 by 36 km
horizontal grid results in a 120 by 84 grid (or 10,080 grid-cells) for each vertical layer. Figure
9A-3 illustrates the horizontal grid-cells for Maryland and surrounding areas.

   9A.2.1.2 Simulation Periods

   For use in this benefits analysis, the simulation periods modeled by REMSAD included
separate full-year application for each of the five emissions scenarios as described in Chapter 3,
i.e., 1996 baseline and the 2020 and 2030 base cases and control scenarios.
                                         9-93

-------
Final Regulatory Impact Analysis
                                   Figure 9A-2
              REMSAD Modeling Domain for Continental United States
Note: Gray markings define individual grid-cells in the REMSAD model.
                                          9-94

-------
                                                      Cost-Benefit Analysis
49
42
    93
104
      Figure 9A-3. Example of REMSAD 36 x 36km Grid-cells for Maryland Area
                                   9-95

-------
Final Regulatory Impact Analysis
   9A.2.1.3 Model Inputs

   REMSAD requires a variety of input files that contain information pertaining to the
modeling domain and simulation period. These include gridded, 1-hour average emissions
estimates and meteorological fields, initial and boundary conditions, and land-use information.
Separate emissions inventories were prepared for the 1996 baseline and each of the future-year
base cases and control scenarios. All other inputs were specified for the 1996 baseline model
application and remained unchanged for each future-year modeling scenario.

   Similar to CAMx, REMSAD requires detailed emissions inventories containing temporally
allocated emissions for each grid-cell in the modeling domain for each  species being simulated.
The previously described annual emission inventories were preprocessed into model-ready
inputs through the SMOKE emissions preprocessing system. Details of the preprocessing of
emissions through SMOKE as provided in the emissions modeling TSD. Meteorological inputs
reflecting 1996 conditions  across the contiguous U.S. were derived from Version 5 of the
Mesoscale Model (MM5).  These inputs included horizontal wind components (i.e., speed and
direction), temperature, moisture, vertical diffusion rates, and rainfall rates for each grid cell in
each vertical layer. Details of the annual 1996 MM5 modeling are provided in Olerud (2000).

   Initial species concentrations and lateral boundary conditions were  specified to approximate
background concentrations of the species; for the lateral boundaries the concentrations varied
(decreased parabolically) with height.  These background concentrations are provided in the air
quality modeling TSD (U.S. EPA, 2003b). Land use information was obtained from the U.S.
Geological Survey  database at 10 km resolution and aggregated to the -36 KM horizontal
resolution used for this REMSAD application.

   9A.2.1.4 Converting REMSAD Outputs to Benefits Inputs

   REMSAD generates predictions of hourly PM concentrations for every grid.  The particulate
matter species modeled by REMSAD include a primary coarse fraction (corresponding to PM in
the 2.5 to 10 micron size range), a primary fine fraction (corresponding to PM less than 2.5
microns in diameter), and several secondary particles (e.g., sulfates, nitrates, and organics).
PM25 is calculated  as the sum of the primary fine fraction and all of the secondarily-formed
particles. These hourly predictions for each REMSAD grid-cell are aggregated to daily averages
and used in conjunction with observed PM concentrations from AIRS to generate the predicted
changes in the daily and annual PM air quality metrics (i.e.,  annual mean PM  concentration)
from the future-year base case to future-year control scenario as inputs  to the health and welfare
                                          9-96

-------
                                                                  Cost-Benefit Analysis
C-R functions of the benefits analysis.d In addition, the speciated predictions from REMSAD are
employed as inputs to a post-processing module that estimates atmospheric visibility, as
discussed later in Section 9A.3.

    In order to estimate PM-related health and welfare effects for the contiguous U.S., daily and
annual average PM concentrations are required for every location. Given available PM
monitoring data, we generated an annual profile for each location in the contiguous 48 States in
two steps: (1) we combine monitored observations and modeled PM predictions to interpolate
forecasted daily PM concentrations for each REMSAD grid-cell, and (2) we compute the daily
and annual PM measures of interest based on the annual PM profiles.e These methods are
described in detail in the benefits analysis technical support document (Abt Associates, 2003).

    9A.2.1.5 PM Air Quality Results

    Table 9A-5 provides a summary of the predicted ambient PM2 5 concentrations for the 2020
and 2030 base cases and changes associated with Nonroad Engine/Diesel Fuel control scenarios.
The REMSAD results indicate that the  predicted change in PM concentrations is composed
almost entirely of reductions in fine particulates (PM25) with little or no reduction in coarse
particles (PM10less PM25).  Therefore, the observed changes in PM10are composed primarily of
changes in PM25. In addition to the standard frequency statistics (e.g., minimum, maximum,
average, median), Table 9A-5 provides the population-weighted average which better reflects the
baseline levels and predicted changes for more populated areas of the nation. This measure,
therefore, will better reflect the potential benefits of these predicted changes through exposure
changes to these populations. As shown, the average annual mean concentrations of PM2 5
across all U.S. grid-cells declines by roughly 2.5 percent (or 0.2 |ig/m3) and 3.4 percent (or 0.28
|ig/m3) in 2020 and 2030, respectively.  The population-weighted average mean concentration
declined by 3.3 percent (or 0.42 |ig/m3) in 2020  and 4.5 percent (or 0.59 |ig/m3) in  2030, which
is much larger in absolute terms than the spatial  average for both years.  This indicates the rule
may generate greater absolute air quality improvements in more populated, urban areas.
   DBased on AIRS, there were 1,071 FRM PM monitors with valid data as defined as more than 11 observations
per season.

   EThis approach is a generalization of planar interpolation that is technically referred to as enhanced Voronoi
Neighbor Averaging  (EVNA) spatial interpolation (See Abt Associates (2003) for a more detailed description).

                                           9-97

-------
Final Regulatory Impact Analysis
                                         Table 9A-6.
                            Summary of Base Case PM Air Quality
  and Changes Due to Preliminary Control Option for Nonroad Diesel Standards: 2020 and 2030
Statistic
2020
Base Case
Change"
Percent
Change
PM25 (ng/m3)
Minimum Annual Mean b
Maximum Annual Mean b
Average Annual Mean
Median Annual Mean
Pop- Weighted Average Annual Mean c
2.18
29.85
8.10
7.50
12.42
-0.02
-1.36
-0.20
-0.18
-0.42
-0.78%
-4.56%
-2.49%
-2.68%
-3.34%
2050
Base Case
Change"
Percent
Change

2.33
32.85
8.37
7.71
13.07
-0.02
-2.03
-0.28
-0.22
-0.59
-1.01%
-6.18%
-3.38%
-2.80%
-4.48%
 1 The change is defined as the control case value minus the base case value.

 b The base case minimum (maximum) is the value for the populated grid-cell with the lowest (highest) annual average. The
 change relative to the base case is the observed change for the populated grid-cell with the lowest (highest) annual average in the
 base case.

 c Calculated by summing the product of the projected REMSAD grid-cell population and the estimated PM concentration, for that
 grid-cell and then dividing by the total population in the 48 contiguous States.
    Table 9A-6 provides information on the populations in 2020 and 2030 that will experience
improved PM air quality.  There are significant populations that live in areas with meaningful
potential reductions in annual mean PM2 5 concentrations resulting from the rule. As shown,
almost 10 percent of the 2030 U.S.  population are predicted to experience reductions of greater
than 1 |ig/m3.  This is  an increase from the 2.7 percent of the U.S. population that are expected to
experience such reductions in 2020. Furthermore, just over 20 percent of the 2030 U.S.
population will benefit from reductions in annual mean PM25 concentrations of greater than 0.75
|ig/m3 and slightly over  50 percent will live in areas with reductions of greater than 0.5 |ig/m3.
This information indicates how widespread the improvements in PM air quality  are expected to
be and the large populations that will benefit from these improvements.
                                             9-98

-------
                                                                  Cost-Benefit Analysis
                                        Table 9A-7
   Distribution of PM2 s Air Quality Improvements Over Population Due to Preliminary Control
               Option for Nonroad Engine/Diesel Fuel Standards: 2020 and 2030
Change in Annual Mean PM2 s
Concentrations (/ug/m3)
0 < A PM25 Cone < 0.25
0.25 < A PM25 Cone < 0.5
0.5 < A PM25 Cone < 0.75
0.75 1.75
2020 Population
Number (millions) Percent (%)
65.11
184.52
56.66
14.60
5.29
3.51
0
0
19.75%
55.97%
17.19%
4.43%
1.60%
1.06%
0.00%
0.00%
2050 Population
Number (millions) Percent (%)
28.60
147.09
107.47
38.50
88.22
15.52
5.70
4.19
8.04%
41.33%
30.20%
10.82%
2.48%
4.36%
1.60%
1.18%
 1 The change is defined as the control case value minus the base case value.
    Table 9A-7 provides additional insights on the potential changes in PM air quality resulting
from the standards.  The information presented previously in Table 9A-5 illustrated the absolute
and relative changes for different points along the distribution of baseline 2020 and 2030 PM2 5
concentration levels, e.g., the change reflects the lowering of the minimum predicted baseline
concentration rather than the minimum predicted change for 2020 and 2030. The latter is the
focus of Table 9A-7 as it presents the distribution of predicted changes in both absolute terms
(i.e., |ig/m3) and relative terms (i.e., percent) across individual REMSAD grid-cells. Therefore,
it provide more information on the  range of predicted changes associated with the rule. As
shown for 2020, the absolute reduction in annual mean PM2 5 concentration ranged from a low of
0.02 |ig/m3 to a high of 1.36 |ig/m3, while the relative reduction ranged from a low of 0.3 percent
to a high of 12.2 percent. Alternatively, for 2030, the absolute reduction ranged from 0.02 to
2.03 |ig/m3, while the relative reduction ranged from 0.4 to 15.5 percent.
                                           9-99

-------
Final Regulatory Impact Analysis
                                        Table 9A-8.
Summary of Absolute and Relative Changes in PM Air Quality Due to Preliminary Control
             Option for Nonroad Engine/Diesel Fuel Standards: 2020 and 2030
Statistic
Absolute Change from Base Case fag/m3)"
Minimum
Maximum
Average
Median
Population-Weighted Average c
Relative Change from Base Case (%)b
Minimum
Maximum
Average
Median
Population-Weighted Average c
2020
PM2 5 Annual Mean

-0.02
-1.36
-0.20
-0.19
-0.42

-0.33%
-12.24%
-2.44%
-2.33%
-3.28%
2030
PM2 5 Annual Mean

-0.02
-2.03
-0.28
-0.26
-0.59

-0.44%
-15.52%
-3.32%
-3.13%
-4.38%
 1 The absolute change is defined as the control case value minus the base case value for each REMSAD grid-cell.

 b The relative change is defined as the absolute change divided by the base case value, or the percentage change, for each gridcell.
 The information reported in this section does not necessarily reflect the same gridcell as is portrayed in the absolute change
 section.

 c Calculated by summing the product of the projected gridcell population and the estimated gridcell PM absolute/relative measure
 of change, and then dividing by the total population in the 48 contiguous states.
9A.2.2 Ozone Air Quality Estimates

    We use the emissions inputs summarized in Section 9A. 1 with a regional-scale version of
CAMx to estimate ozone air quality in the Eastern and Western U.S. CAMx is an Eulerian
three-dimensional photochemical grid air quality model designed to calculate the concentrations
of both inert and chemically reactive pollutants by simulating the physical and chemical
processes in the atmosphere that affect ozone formation. Because it accounts for spatial and
temporal variations as well as differences in the reactivity of emissions, the CAMx is useful for
evaluating the impacts of the rule on U.S. ozone concentrations.  As discussed earlier in Chapter
2, although the model tends to underestimate observed ozone, especially over the western U.S., it
exhibits less bias and error than any past regional ozone modeling application conducted by EPA
(i.e., Ozone Transport Assessment Group (OTAG), On-highway  Tier-2, and HD Engine/Diesel
Fuel).
                                            9-100

-------
                                                                Cost-Benefit Analysis
   Our analysis applies the modeling system separately to the Eastern and Western U.S. for five
emissions scenarios: a 1996 baseline projection, a 2020 baseline projection and a 2020 projection
with preliminary nonroad controls, a 2030 baseline projection and a 2030 projection with
preliminary nonroad controls.  As discussed in the Benefits Analysis TSD, we use the relative
predictions from the model by combining the 1996 base-year and each future-year scenario with
ambient air quality observations to determine the expected change in 2020 or 2030 ozone
concentrations due to the rule (Abt Associates, 2003).  These results are used solely in the
benefits analysis.

   The CAMx modeling system requires a variety of input files that contain information
pertaining to the modeling domain and simulation period.  These include gridded, day-specific
emissions estimates and meteorological fields, initial and boundary conditions, and land-use
information. The model divides the continental United States into two regions: East and West.
As applied to each  region, the model segments the area within the subject region into square
blocks called grids (roughly equal in size to counties), each of which has several layers of air
conditions that are  considered in the analysis. Using this data, the CAMx model generates
predictions of hourly ozone concentrations for every grid. We then calibrate the results of this
process to develop  2020 and 2030 ozone profiles at monitor sites by normalizing the
observations to the observed ozone concentrations at each monitor site. For areas (grids) without
ozone monitoring data, we interpolated ozone values using data from monitors surrounding the
area. After completing this process, we calculated daily and seasonal ozone metrics to be used
as inputs to the health and welfare C-R functions of the benefits analysis.  The following sections
provide a more detailed discussion of each of the steps in this evaluation and a summary of the
results.

   9A.2.2.1 Modeling Domain

   The modeling domain representing the Eastern U.S. is the same as that used previously for
OTAG and the On-highway Tier-2 rulemaking.  As shown in Figure 9A-4, this domain
encompasses most  of the Eastern U.S. from the East coast to mid-Texas and consists of two grids
with differing resolutions. The modeling domain extends from 99 degrees to 67 degrees west
longitude and from 26 degrees to 47 degrees north latitude. The inner portion of the modeling
domain shown in Figure 9A-4 uses a relatively fine grid of 12 km consisting of nine vertical
layers. The outer area has less horizontal resolution, as it uses a 36 km grid with the same nine
vertical layers.  The vertical height of the modeling domain is 4,000 meters above ground level
for both areas.

   The modeling domain representing the Western U.S. is the same as that used previously for
the On-highway Tier-2 rulemaking. As shown in Figure 9A-5, this domain encompasses the
area west of the 99th degree longitude (which runs through North and South Dakota, Nebraska,

                                         9-101

-------
Final Regulatory Impact Analysis
Kansas, Oklahoma, and Texas) and consists of two grids with differing resolutions. The domain
extends from 127 degrees to 99 degrees west longitude and from 26 degrees to 52 degrees north
latitude. The inner portion of the modeling domain shown in Figure 9A-5 uses a relatively fine
grid of 12 km consisting of eleven vertical layers.  The outer area has less horizontal resolution,
as it uses a 36 km grid with the same eleven vertical layers. The vertical height of the modeling
domain is 4,800 meters above ground level.
                                        9-102

-------
                                                                       Cost-Benefit Analysis
                 Figure 9A-4 CAMx Eastern U.S. Modeling Domain
                                                  1
Note: The inner area represents fine grid modeling at 12 km resolution, while the outer area represents the coarse grid
modeling at 36 km resolution.
             Figure 9A-5 CAMx Western U.S. Modeling Domain
Note: The inner area represents fine grid modeling at 12 km resolution, while the outer area represents the coarse grid
modeling at 36 km resolution.
                                            9-103

-------
Final Regulatory Impact Analysis
   9A.2.2.2 Simulation Periods

   For use in this benefits analysis, the simulation periods modeled by CAMx included several
multi-day periods when ambient measurements recorded high ozone concentrations. A
simulation period, or episode, consists of meteorological data characterized over a block of days
that are used as inputs to the air quality model.  A simulation period is selected to characterize a
variety of ozone conditions including some days with high ozone concentrations in one or more
portions of the U.S. and observed exceedances of the  1-hour NAAQS for ozone  being recorded
at monitors.  We focused on the summer of 1995 for selecting the episodes to model in the East
and the summer of 1996 for selecting the episodes to model in the West because each is a recent
time period for which we had model-ready meteorological inputs and this timeframe contained
several periods of elevated ozone over the Eastern and Western U.S., respectively.  As detailed
in the air quality modeling TSD, this analysis used three multi-day meteorological scenarios
during the summer of 1995 for the model simulations over the eastern U.S.: June 12-24, July 5-
15, and August 7-21. Two multi-day meteorological scenarios during the summer of 1996 were
used in the model simulations over the western U.S.: July 5-15 and July 18-31.  Each of the five
emissions scenarios (1996 base year, 2020 base, 2020 control, 2030 baseline, 2030 control) were
simulated for the selected episodes. These episodes include a three day "ramp-up" period to
initialize the model, but the results for these days are not used in this analysis.

   9A.2.2.3 Converting CAMx Outputs to Full-Season Profiles for  Benefits Analysis

   This study extracted hourly, surface-layer ozone concentrations for each grid-cell from the
standard CAMx output file containing hourly average ozone values. These model  predictions
are used in conjunction with the observed concentrations obtained from the Aerometric
Information Retrieval System (AIRS) to generate ozone concentrations for the entire ozone
season.f>g  The predicted changes in ozone concentrations from  the future-year base case to
future-year control scenario serve as inputs to the health and welfare C-R functions of the
benefits analysis, i.e., BENMAP.

   In order to estimate ozone-related health and welfare effects for the contiguous U.S., full-
season ozone data are required for every CAPMS grid-cell.  Given available ozone monitoring
data, we generated full-season ozone profiles for each location in the contiguous 48 States in two
steps: (1) we combine monitored observations and modeled ozone predictions to interpolate
   F The ozone season for this analysis is defined as the 5-month period from May to September; however, to
estimate certain crop yield benefits, the modeling results were extended to include months outside the 5-month
ozone season.

   GBased on AIRS, there were 961 ozone monitors with sufficient data, i.e., 50 percent or more days reporting at
least 9 hourly observations per day (8 am to 8 pm) during the ozone season.

                                          9-104

-------
                                                                  Cost-Benefit Analysis
hourly ozone concentrations to a grid of 8 km by 8 km population grid-cells, and (2) we
converted these full-season hourly ozone profiles to an ozone measure of interest, such as the
daily average. hjl For the analysis of ozone impacts on agriculture and commercial forestry, we
use a similar approach except air quality is interpolated to county centroids as opposed to
population grid-cells.  We report ozone concentrations as a cumulative index called the SUM06.
The SUM06 is the sum of the ozone concentrations for every hour that exceeds 0.06 parts per
million (ppm) within a 12-hour period from 8 am to 8 pm in the months of May to September.
These methods are described in detail in the benefits analysis technical support document (Abt
Associates, 2003).

    9A.2.2.4 Ozone Air Quality Results

    This section provides a summary the predicted ambient ozone concentrations from the
CAMx model for the 2020 and 2030 base cases and changes associated with the Nonroad
Engine/Diesel Fuel control scenario.  In Tables 9A-8 and 9A-9, we provide those ozone metrics
for grid-cells in the Eastern and Western U.S. respectively, that enter the concentration response
functions for health benefits endpoints. In addition to the standard frequency statistics (e.g.,
minimum, maximum, average, median), we provide the population-weighted average which
better reflects the baseline levels and predicted changes for more populated areas  of the nation.
This measure, therefore, will better reflect the potential benefits of these predicted changes
through exposure changes to these populations.

    As shown in Table 9A-8, for the 2020 ozone season, the rule results in average reductions of
roughly 2 percent, or between 0.57 to 0.85 ppb, in the daily average ozone concentration metrics
across the Eastern U.S. population grid-cells.  For the 2030 ozone season,  the average reductions
in the daily average ozone concentration are between 3 and 3.5 percent, or between 0.91 to 1.35
ppb. A slightly lower relative decline is predicted for the population-weighted average, which
reflects the observed increases in ozone concentrations for certain hours during the year in
highly populated urban areas associated with  NOx emissions reductions (see more detailed
discussion in Chapter 2). Additionally, the daily 1-hour maximum ozone concentrations are
predicted to decline between 2.3 and 3.6 percent in 2020 and 2030 respectively, i.e., between
1.05 and 1.66 ppb.

    As shown in Table 9A-9, for the 2020 ozone season, the rule results in average reductions of
roughly 1.5  percent, or between 0.57 to 0.52 ppb, in the daily average ozone concentration
   HThe 8 km grid squares contain the population data used in the health benefits analysis model, CAPMS. See
Section C of this appendix for a discussion of this model.

   'This approach is a generalization of planar interpolation that is technically referred to as enhanced Voronoi
Neighbor Averaging (EVNA) spatial interpolation (See Abt Associates (2003) for a more detailed description).

                                          9-105

-------
Final Regulatory Impact Analysis
metrics across the Western U.S. population grid-cells. For the 2030 ozone season, the average
reductions in the daily average ozone concentration are roughly 2 percent, or between 0.61 to
0.82 ppb.  Additionally, the daily 1-hour maximum ozone concentrations are predicted to decline
between 1.3 and 2.1 percent in 2020 and 2030 respectively, i.e., between 0.62 and 0.97 ppb.

    As discussed in more detail in Chapter 2, our ozone air quality modeling showed that the
NOx emissions reductions from the preliminary modeled standards are projected to result in
increases in ozone concentrations for certain hours during the year, especially in urban, NOx
limited areas. These increases are often observed within the highly populated urban areas in
California. As a result, the population-weighted metrics for ozone shown in Table 9A-9 indicate
increases in concentrations. Most of these increases are expected to occur during hours where
ozone levels are low (and often below the one-hour ozone standard). These increase are
accounted for in the benefits analysis because it relies on the changes in ozone concentrations
across the entire distribution of baseline levels.  However, as detailed in Chapter 2 and illustrated
by the results from Tables 9A-8 and 9A-9, most of the country experiences decreases in ozone
concentrations for most hours in the year.

    In Table 9A-10, we provide the seasonal SUM06 ozone metric for counties in the Eastern
and Western U.S. that enters the concentration response function for agriculture benefit end-
points. This metric is a cumulative threshold measure so that the increase in baseline NOx
emissions from Tier 2 post-control to this rulemaking have resulted in a larger number of rural
counties exceeding the hourly 0.06 ppm threshold. As a result, changes in ozone concentrations
for these counties are contributing to greater impacts of the Nonroad Diesel Engine rule on the
seasonal SUM06 ozone metric. As shown, the average across all Eastern U.S. counties declined
by 78 percent, or almost 17 ppb.  Similarly high percentage reductions are observed across the
other points on the distribution with the maximum declining by almost 30 ppb, or 55 percent,
and the median declining by almost 20 ppb, or 83  percent.
                                          9-106

-------
                                                          Table 9A-9.
Summary of CAMx Derived Ozone Air Quality Metrics Due to Preliminary Control Option for Nonroad
                    Engine/Diesel Fuel Standards for Health Benefits EndPoints: Eastern U.S.
Statistic °
2020
Base Case
Change b
Percent Change
2030
Base Case
Change b
Percent Change b
Daily 1-Hour Maximum Concentration (ppb)
Minimum c
Maximum '
Average
Median
Population-Weighted Average d
28.85
93.94
45.54
45.45
51.34
-0.81
-0.85
-1.05
-1.23
-0.67
-2.80%
-0.90%
-2.30%
-2.71%
-1.31%
28.81
94.70
45.65
45.52
51.47
-1.24
-1.61
-1.66
-1.73
-1.16
-4.31%
-1.70%
-3.64%
-3.80%
-2.25%
Daily 5-Hour Average Concentration (ppb)
Minimum c
Maximum '
Average
Median
Population-Weighted Average d
24.90
68.69
38.99
38.94
42.77
-0.67
-0.20
-0.85
-0.92
-0.47
-2.68%
-0.29%
-2.17%
-2.39%
-1.10%
24.87
69.11
39.08
39.00
42.90
-1.03
-0.44
-1.35
-1.40
-0.84
-4.13%
-0.64%
-3.45%
-3.58%
-1.96%
Daily 8-Hour Average Concentration (ppb)
Minimum '
Maximum c
Average
Median
Population-Weighted Average d
24.15
68.30
38.46
38.44
42.07
-0.64
-0.21
-0.83
-0.89
-0.46
-2.64%
-0.31%
-2.16%
-2.33%
-1.08%
24.12
68.72
38.55
38.50
42.19
-0.98
-0.46
-1.33
-1.45
-0.82
-4.07%
-0.67%
-3.44%
-3.76%
-1.93%
Daily 12-Hour Average Concentration (ppb)
Minimum '
Maximum c
Average
Median
Population-Weighted Average d
22.42
66.06
36.59
36.61
39.65
-0.58
-0.17
-0.78
-0.84
-0.40
-2.57%
-0.25%
-2.13%
-2.30%
-1.00
22.40
66.46
36.66
36.66
39.75
-0.89
-0.38
-1.25
-1.43
-0.72
-3.96%
-0.58%
-3.40%
-3.89%
-1.80%
Daily 24-Hour Average Concentration (ppb)
Minimum c
Maximum '
Average
Median
Population-Weighted Average d
15.20
55.95
28.93
28.92
30.24
-0.35
0.10
-0.57
-0.63
-0.18
-2.28%
0.18%
-1.96%
-2.15%
-0.60%
15.19
56.23
28.98
28.98
30.29
-0.54
0.04
-0.91
-1.01
-0.37
-3.52%
0.07%
-3.14%
-3.48%
-1.23%
a These ozone metrics are calculated at the CAMX grid-cell level for use in health effects estimates based on the results of spatial and temporal Voronoi Neighbor Averaging.
Except for the daily 24-hour average, these ozone metrics are calculated over relevant time periods during the daylight hours of the "ozone season," i.e., May through
September. For the 5-hour average, the relevant time period is 10 am to 3 pm; for the 8-hr average, it is 9 am to 5 pm; and, for the 12-hr average it is 8 am to 8 pm.

b  The change is defined as the control case value minus the base case value. The percent change is the "Change" divided by the "Base Case," and then multiplied by 100 to
convert the value to a percentage.

c The base case minimum (maximum) is the value for the CAMX grid cell with the lowest (highest) value.
d Calculated by summing the product of the projected CAMX grid-cell population and the estimated CAMX gnd-cell seasonal ozone concentration, and then dividing by the
total population.

-------
                                                           Table 9A-10.
Summary of CAMx Derived Ozone Air Quality Metrics Due to Preliminary Control  Option for Nonroad
                   Engine/Diesel Fuel Standards for Health Benefits EndPoints: Western U.S.
Statistic a
2020
Base Case
Change b
Percent Change b
2030
Base Case
Change b
Percent Change b
Daily 1-Hour Maximum Concentration (ppb)
Minimum c
Maximum c
Average
Median
Population- Weighted Average d
27 .48
201.28
47.02
46.10
63.80
-0.01
4.87
-0.62
-0.56
0.34
-0.03%
2.42%
-1.31%
-1.19%
0.54%
27 .48
208.02
47.04
46.06
64.23
-0.01
6.26
-0.97
-0.66
0.38
-0.05%
3.01%
-2.07%
-1.43%
0.58%
Daily 5-Hour Average Concentration (ppb)
Minimum c
Maximum c
Average
Median
Population- Weighted Average d
24.20
163.41
41.11
40.48
53.56
-0.01
2.55
-0.52
-0.40
0.45
-0.04%
1.56%
-1.26%
-1.04%
0.84%
24.21
168.89
41.13
40.46
53.89
-0.01
6.04
-0.82
-0.69
0.55
-0.05%
3.57%
-2.00%
-1.70%
1.03%
Daily 8-Hour Average Concentration (ppb)
Minimum c
Maximum c
Average
Median
Population- Weighted Average d
23.77
157.49
40.68
40.11
51.96
-0.01
1.33
-0.51
-0.36
0.46
-0.04%
0.84%
-1.25%
-1.03%
0.88%
23.77
161.92
40.69
40.09
52.29
-0.01
5.94
-0.81
-0.72
0.57
-0.05%
3.67%
-1.99%
-1.79%
1.10%
Daily 12-Hour Average Concentration (ppb)
Minimum c
Maximum c
Average
Median
Population- Weighted Average d
22.13
140.48
39.30
38.85
47.68
0.31
1.65
-0.48
-0.38
0.49
1.39%
1.18%
-1.23%
-0.97%
1.02%
22.09
143.59
39.31
38.82
47.99
0.44
1.78
-0.77
-0.58
0.63
2.01%
1.24%
-1.95%
-1.50%
1.32%
Daily 24-Hour Average Concentration (ppb)
Minimum c
Maximum c
Average
Median
Population- Weighted Average d
14.08
95.27
33.42
32.97
35.53
0.22
0.41
-0.38
-0.30
0.47
1.60%
0.43%
-1.14%
-0.89%
1.31%
14.03
96.59
33.42
32.95
35.74
0.32
0.29
-0.61
-0.61
0.63
2.30%
0.30%
-1.82%
-1.85%
1.77%
3 These ozone metrics are calculated at the CAMX grid-cell level for use in health effects estimates based on the results of spatial and temporal Voronoi Neighbor Averaging.
Except for the daily 24-hour average, these ozone metrics are calculated over relevant time periods during the daylight hours of the "ozone season," i.e., May through
September. For the 5-hour average, the relevant time period is 10 am to 3 pm; for the 8-hr average, it is 9 am to 5 pm; and, for the 12-hr average it is 8 am to 8 pm.

b  The change is defined as the control case value minus the base case value. The percent change is the "Change" divided by the "Base Case," and then multiplied by 100 to
convert the value to a percentage.

c The base case minimum (maximum) is the value for the CAMX grid cell with the lowest (highest) value.
d Calculated by summing the product of the projected CAMX grid-cell population and the estimated CAMX grid-cell seasonal ozone concentration, and then dividing by the
total population.

-------
                                                                        Cost-Benefit Analysis
                                          Table 9A-11.
 Summary of CAMx Derived Ozone Air Quality Metrics Due to Preliminary Control Option for Nonroad
             Engine/Diesel Fuel Standards for Welfare Benefits Endpoints: 2020 and 2030
Statistic "
Sum06 (ppm)
Minimum c
Maximum c
Average
Median
2020
Base Case
Change b
Percent
Change b
2050
Base Case
Change b
Percent
Change b
Eastern U.S.
0.00
67.24
4.74
2.18
0.00
-3.30
-0.72
-0.76
-
-4.91
-15.10
-35.02
0.00
68.63
4.88
2.21
0.00
-5.54
-1.09
-0.77
-
-8.07%
-22.43%
-34.84%
Western U.S.
Sum06 (ppm)
Minimum c
Maximum c
Average
Median
0.00
132.73
2.78
0.00
0.00
6.09
-0.22
0.00
-
4.59
-7.85
-
0.00
137.71
2.83
0.00
0.00
8.45
-0.33
0.00
-
6.14%
-11.72%
-
 1 SUM06 is defined as the cumulative sum of hourly ozone concentrations over 0.06 ppm (or 60 ppb) that occur during daylight
 hours (from Sam to 8pm) in the months of May through September. It is calculated at the county level for use in agricultural
 benefits based on the results of temporal and spatial Voronoi Neighbor Averaging.

 b The change is defined as the control case value minus the base case value. The percent change is the "Change" divided by the
 "Base Case," which is then multiplied by 100 to convert the value to a percentage.

 c The base case minimum (maximum) is the value for the county level observation with the lowest (highest) concentration.
9A.2.3 Visibility Degradation Estimates

    Visibility degradation is often directly proportional to decreases in light transmittal in the
atmosphere.  Scattering and absorption by both gases and particles decrease light transmittance.
To quantify changes in visibility, our analysis computes a light-extinction coefficient, based on
the work of Sisler (1996), which shows the total fraction of light that is decreased per unit
distance. This coefficient accounts for the scattering and absorption of light by both particles
and gases, and accounts for the higher extinction efficiency of fine particles compared to coarse
particles. Fine particles with significant light-extinction efficiencies include sulfates, nitrates,
organic carbon, elemental carbon (soot), and soil (Sisler, 1996).

    Based upon the light-extinction coefficient, we also calculated a unitless visibility index,
called a "deciview," which is used in the valuation of visibility. The deciview metric provides a
                                              9-109

-------
Final Regulatory Impact Analysis
linear scale for perceived visual changes over the entire range of conditions, from clear to hazy.
Under many scenic conditions, the average person can generally perceive a change of one
deciview. The higher the deciview value, the worse the visibility.  Thus, an improvement in
visibility is a decrease in deciview value.

    Table 9A-11 provides the distribution of visibility improvements across 2020 and 2030
populations resulting from the Nonroad Engine/Diesel Fuel rule. The majority of the 2030 U.S.
population live in areas with predicted improvement in annual average visibility of between 0.4
to 0.6 deciviews resulting from the rule.  As shown, almost 20 percent of the 2030 U.S.
population are predicted to experience improved annual average visibility of greater than 0.6
deciviews.  Furthermore, roughly 70 percent of the 2030 U.S. population will benefit from
reductions in annual average visibility of greater than 0.4 deciviews. The information provided
in Table 9A-11 indicates how widespread the improvements in visibility are expected to be and
the share of populations that will benefit from these improvements.

    Because the visibility benefits analysis distinguishes between general regional visibility
degradation and that particular to Federally-designated Class I areas (i.e., national parks, forests,
recreation areas, wilderness areas, etc.), we separated estimates of visibility degradation into
"residential" and "recreational"  categories.  The estimates of visibility degradation for the
"recreational" category apply to Federally-designated Class I areas, while estimates for the
"residential" category apply to non-Class I areas. Deciview estimates are estimated using
outputs  from REMSAD for the 2020 and 2030 base cases and control scenarios.

                                      Table 9A-12.
  Distribution of Populations Experiencing Visibility Improvements Due to Preliminary Control
                 Option for Nonroad Diesel Engine Standards: 2020 and 2030
Improvements in Visibility "
(annual average deciviews)
0 < A Deciview < 0.2
0.2 < A Deciview < 0.4
0.4 < A Deciview < 0.6
0.6 < A Deciview < 0.8
0.8 < A Deciview < 1.0
A Deciview > 1.0
2020 Population
Number (millions) Percent (%)
52.0
115.5
81.3
62.0
13.2
5.6
15.8%
35.0%
24.7%
18.8%
4.0%
1.7%
2030 Population
Number (millions) Percent (%)
11.6
179.7
90.5
49.1
16.4
8.5
3.3%
50.5%
25.4%
13.8%
4.6%
2.4%
 1 The change is defined as the control case deciview level minus the base case deciview level.
                                          9-110

-------
                                                                  Cost-Benefit Analysis
    9A.2.3.1 Residential Visibility Improvements

    Air quality modeling results predict that the Nonroad Engine/Diesel Fuel rule will create
improvements in visibility through the country.  In Table 9A-12, we summarize residential
visibility improvements across the Eastern and Western U.S. in 2020 and 2030. The baseline
annual average visibility for all U.S. counties is 14.8 deciviews. The mean improvement across
all U.S. counties is 0.28 deciviews, or almost 2 percent. In urban areas with a population of
250,000 or more (i.e., 1,209 out of 5,147 counties), the mean improvement in annual visibility
was 0.39 deciviews and ranged from 0.05 to  1.08 deciviews. In rural areas (i.e., 3,938 counties),
the mean improvement in visibility was 0.25 deciviews in 2030 and ranged from 0.02 to 0.94
deciviews.

    On average, the Eastern U.S. experienced slightly larger absolute but smaller relative
improvements in visibility than the Western U.S. from the Nonroad Engine/Diesel Fuel
reductions. In Eastern U.S., the mean improvement was 0.34 deciviews from an average
baseline of 19.32 deciviews. Western counties experienced a mean improvement of 0.21
deciviews from an average baseline of 9.75 deciviews projected in 2030. Overall, the data
suggest that the Nonroad Engine/Diesel Fuel rule has the potential to provide widespread
improvements in visibility for 2020 and 2030.
                                      Table9A-13.
      Summary of Baseline Residential Visibility and Changes by Region: 2020 and 2030
                                (Annual Average Deciviews)
Regions"
Eastern U.S.
Urban
Rural
Western U.S.
Urban
Rural
National, all counties
Urban
Rural
2020
Base Case
20.27
21.61
19.73
8.69
9.55
8.50
14.77
17.21
14.02
Change11
0.24
0.24
0.24
0.18
0.25
0.17
0.21
0.24
0.20
Percent
Change
1.3%
1.2%
1.3%
2.1%
2.7%
2.0%
1.7%
1.7%
1.6%
2030
Base Case
20.54
21.94
19.98
8.83
9.78
8.61
14.98
17.51
14.20
Change11
0.33
0.33
0.33
0.25
0.35
0.23
0.29
0.34
0.28
Percent
Change
1.7%
1.6%
1.8%
2.8%
3.6%
2.7%
2.3%
2.3%
2.2%
 a Eastern and Western regions are separated by 100 degrees north longitude. Background visibility conditions differ by
 region.
 b An improvement in visibility is a decrease in deciview value.  The change is defined as the Nonroad Engine/Diesel
 Fuel control case deciview level minus the basecase deciview level.
                                          9-111

-------
Final Regulatory Impact Analysis
   9A.2.3.2 Recreational Visibility Improvements

   In Table 9A-13, we summarize recreational visibility improvements by region in 2020 and
2030 in Federal Class I areas.  These recreational visibility regions are shown in Figure 9A-6.
As shown, the national improvement in visibility for these areas increases from 1.5  percent, or
0.18 deciviews, in 2020 to 2.1 percent, or 0.24 deciviews, in 2030. Predicted relative visibility
improvements are the largest in the Western U.S. as shown for California (3.2% in 2030), and
the Southwest (2.9%) and the Rocky Mountain (2.5%).  Federal Class I areas in the Eastern U.S.
are predicted to have an absolute improvement of 0.24 deciviews in 2030, which reflects a 1.1 to
1.3 percent change from 2030 baseline visibility of 20.01 deciviews.
                                         9-112

-------
                                                                 Cost-Benefit Analysis
Figure 9A-6. Recreational Visibility Regions for Continental U.S.
               Study Region
               Transfer Region
   Note: Study regions were represented in the Chestnut and Rowe (1990a, 1990b) studies used
   in evaluating the benefits of visibility improvements, while transfer regions used extrapolated
   study results.
                                          9-113

-------
                                      Table 9A-14.
      Summary of Baseline Recreational Visibility and Changes by Region: 2020 and 2030
                               (Annual Average Deciviews)
Class I Visibility Regions"
Eastern U.S.
Southeast
Northeast/Midwest
Western U.S.
California
Southwest
Rocky Mountain
Northwest
National Average (unweighted)
2020
Base Case
19.72
21.31
18.30
8.80
9.33
6.87
8.46
12.05
11.61
Change11
0.18
0.18
0.18
0.17
0.21
0.16
0.15
0.18
0.18
Percent
Change
0.9%
0.9%
1.0%
2.0%
2.3%
2.3%
1.8%
1.5%
1.5%
2030
Base Case
20.01
21.62
18.56
8.96
9.56
7.03
8.55
12.18
11.80
Change11
0.24
0.24
0.24
0.24
0.30
0.21
0.21
0.24
0.24
Percent
Change
1.2%
1.1%
1.3%
2.7%
3.2%
2.9%
2.5%
2.0%
2.1%
 a Regions are pictured in Figure VI-5 and are defined in the technical support document (see Abt Associates, 2003).
 b An improvement in visibility is a decrease in deciview value. The change is defined as the Nonroad Engine/Diesel
 Fuel control case deciview level minus the basecase deciview level.
9A.3 Benefit Analysis- Data and Methods

   Environmental and health economists have a number of methods for estimating the economic
value of improvements in (or deterioration of) environmental quality. The method used in any
given situation depends on the nature of the effect and the kinds of data, time, and resources that
are available for investigation and analysis. This section provides an overview of the methods
we selected to quantify and monetize the benefits included in this RIA.

   Given changes in environmental quality (ambient air quality, visibility, nitrogen and sulfate
deposition, odor), the next step is to determine the economic value of those changes.  We follow
a "damage-function" approach in calculating total benefits of the modeled changes in
environmental quality. This  approach estimates changes in individual health and welfare
endpoints (specific effects that can be associated with changes in air quality) and assigns values
to those changes assuming independence of the individual values.  Total benefits are calculated
simply as the sum of the values for  all non-overlapping health and welfare endpoints.  This
imposes no overall preference structure, and does not account for potential income or
substitution effects, i.e. adding a new endpoint will not reduce the value of changes in other
endpoints. The "damage-function"  approach is the standard approach for most cost-benefit
analyses  of environmental quality programs, and has been used in several recent published
analyses  (Banzhaf et al., 2002; Levy et al., 2001; Levy et al., 1999; Ostro and Chestnut, 1998).

-------
                                                                    Cost-Benefit Analysis
    In order to assess economic value in a damage-function framework, the changes in
environmental quality must be translated into effects on people or on the things that people
value.  In some cases, the changes in environmental quality can be directly valued, as is the case
for changes in visibility. In other cases, such as for changes in ozone and PM, a health and
welfare impact analysis must first be conducted to convert air quality changes into effects that
can be assigned dollar values.

    For the purposes of this RIA, the health impacts analysis is limited to those health effects that
are directly linked to ambient levels of air pollution, and specifically to those linked to ozone and
particulate matter. There are known health effects associated with other emissions expected to
be reduced by these  standards, however, due to limitations in air quality models, we are unable
to quantify the changes in the ambient levels of CO, SO2, and air toxics such as benzene.j There
may be other, indirect health impacts associated with implementation of controls to meet the
preliminary control options, such as occupational health impacts for equipment operators. These
impacts may be positive or negative, but in general, for this set of preliminary control options,
are expected to be small relative to the direct air pollution related impacts.

    The welfare impacts analysis is limited to changes in the environment that have a direct
impact on human welfare.  For this analysis, we are limited by the available data to examining
impacts of changes in visibility and agricultural yields. We also  provide qualitative discussions
of the impact of changes in other environmental and ecological effects, for example,  changes in
deposition of nitrogen and sulfur to terrestrial and aquatic ecosystems and odor, but we are
unable to place an economic value  on these changes.

    We note at the outset that EPA  rarely has the time or resources to perform extensive new
research to measure  either the health outcomes or their values for this analysis.  Thus, similar to
Kunzli et al.  (2000) and other recent health impact analyses, our  estimates are based  on the best
available methods of benefits transfer. Benefits transfer is the science and art of adapting
primary research from similar contexts to obtain the most accurate measure of benefits for the
environmental quality change under analysis. Where appropriate, adjustments are made for the
level of environmental quality change, the sociodemographic and economic characteristics of the
    J Several commentators from the public and from public interest groups noted that occupational studies have
shown diesel exhaust, as a mixture, to be carcinogenic. In addition, several of these commentors also noted that
diesel exhaust contains carcinogenic hazardous air pollutants (HAPs). For these reasons, it was suggested that EPA
should include modeling of cancer incidence associated with exposure to the carcinogenic components of diesel
exhaust. Diesel particles producing lung cancer mortality may be included in the lung cancer mortality estimates for
PM2 5. We also acknowledge both that diesel exhaust as a mixture is likely to be carcinogenic and that it contains
specific carcinogenic HAPs which represent a cancer risk. However, at this time, as discussed in Chapter 2, we do
not believe that the data support the determination of a unit risk for diesel exhaust as a mixture and therefore,
lifetime mortality attributable to diesel exhaust exposure cannot be quantified for purposes of benefits analysis.

                                            9-115

-------
Final Regulatory Impact Analysis
affected population, and other factors in order to improve the accuracy and robustness of benefits
estimates.

9A.3.1 Valuation Concepts

   In valuing health impacts, we note that reductions in ambient concentrations of air pollution
generally lower the risk of future adverse health affects by a fairly small amount for a large
population. The appropriate economic measure is therefore willingness-to-pay for changes in
risk prior to the regulation (Freeman, 1993). In general, economists tend to view an individual's
willingness-to-pay (WTP) for a improvement in environmental quality as the appropriate
measure of the value of a risk reduction.  An individual's willingness-to-accept (WTA)
compensation for not receiving the improvement is also a valid measure. However, WTP is
generally considered to be a more readily available and conservative measure of benefits.
Adoption of WTP as the measure of value implies that the value of environmental quality
improvements is dependent on the individual preferences of the affected population and that the
existing distribution of income (ability to pay) is appropriate.  For some health effects, such as
hospital admissions, WTP estimates are generally not available. In these cases, we use the cost
of treating or mitigating the effect as a primary  estimate.  These costs of illness (COI) estimates
generally understate the true value of reductions in risk of a health effect, reflecting the direct
expenditures related to treatment but not the value of avoided pain and suffering from the health
effect (Harrrington and Portnoy, 1987; Berger,  1987).

   For many goods, WTP can be observed by examining actual market transactions. For
example, if a  gallon of bottled drinking water sells for one dollar, it can be observed that at least
some persons are willing to pay one dollar for such water. For goods not exchanged in the
market, such as most environmental "goods," valuation is not as straightforward. Nevertheless,
a value may be inferred from observed behavior, such as sales and prices of products that result
in similar effects or risk reductions, (e.g., non-toxic cleaners or bike helmets).  Alternatively,
surveys may be used in an attempt to directly elicit WTP for an environmental improvement.

   One distinction in environmental benefits estimation is between use values and non-use
values. Although no general agreement exists among economists on a precise distinction
between the two (see Freeman, 1993), the general nature of the difference is clear.  Use values
are those aspects of environmental quality that affect an individual's welfare more or less
directly.  These effects include changes in product prices, quality, and availability, changes in
the quality of outdoor recreation and outdoor aesthetics, changes in health or life expectancy, and
the costs of actions taken to avoid negative effects of environmental quality changes.

   Non-use values are those for which an individual is willing to pay for reasons that do not
relate to the direct use or enjoyment of any environmental benefit, but might relate to existence

                                          9-116

-------
                                                                    Cost-Benefit Analysis
values and bequest values. Non-use values are not traded, directly or indirectly, in markets. For
this reason, the measurement of non-use values has proved to be significantly more difficult than
the measurement of use values.  The air quality changes produced by the Nonroad Diesel Engine
rule cause changes in both use and non-use values, but the monetary benefit estimates are almost
exclusively for use values.

   More frequently than not,  the economic benefits from environmental quality changes are not
traded in markets, so direct measurement techniques can not be used.  There are three main non-
market valuation methods used to develop values for endpoints considered in this analysis.
These include stated preference (or contingent valuation), indirect market  (e.g. hedonic wage),
and avoided cost methods.

   The stated preference or CV method values endpoints by using carefully structured surveys
to ask a sample of people what amount of compensation is equivalent to a  given change in
environmental quality.  There is an extensive scientific literature and body of practice on both
the theory and technique of stated preference based valuation.  EPA believes that well-designed
and well-executed stated preference studies are valid for estimating the benefits of air quality
regulation.15 Stated preference valuation studies form the basis for valuing a number of health
and welfare endpoints, including the value of premature mortality risk reductions, chronic
bronchitis risk reductions, minor illness risk reductions,  and visibility improvements.

   Indirect market methods can also be used to infer the benefits of pollution reduction. The
most important application of this technique for our analysis is the calculation of the value of a
statistical life for use in the estimate of benefits from premature mortality  risk reductions. There
exists no market where changes in the probability of death are directly exchanged. However,
people make decisions about occupation, precautionary behavior, and other activities associated
with changes in the risk of death.  By examining these risk changes and the other characteristics
of people's choices, it is possible to infer information about the monetary  values associated with
changes  in premature mortality risk (see Section 9A.3.5.5.1).

   Avoided cost methods are ways to estimate the costs of pollution by using the expenditures
made necessary by pollution damage.  For example, if buildings must be cleaned or painted more
    KConcerns about the reliability of value estimates from CV studies arose because research has shown that bias
can be introduced easily into these studies if they are not carefully conducted. Accurately measuring WTP for
avoided health and welfare losses depends on the reliability and validity of the data collected.  There are several
issues to consider when evaluating study quality, including but not limited to 1) whether the sample estimates of
WTP are representative of the population WTP; 2) whether the good to be valued is comprehended and accepted by
the respondent; 3) whether the WTP elicitation format is designed to minimize strategic responses; 4) whether WTP
is sensitive to respondent familiarity with the good, to the size of the change in the good, and to income; 5) whether
the estimates of WTP are broadly consistent with other estimates of WTP for similar goods; and 6) the extent to
which WTP responses are consistent with established economic principles.

                                           9-117

-------
Final Regulatory Impact Analysis
frequently as levels of PM increase, then the appropriately calculated increment of these costs is
a reasonable lower bound estimate (under most conditions) of true economic benefits when PM
levels are reduced. Avoided costs methods are also used to estimate some of the health-related
benefits related to morbidity, such as hospital admissions (see section 9A.3.5).

   The most direct way to measure the economic value of air quality changes is in cases where
the endpoints have market prices. For the final rule, this can only be done for effects on
commercial agriculture. Well-established economic modeling approaches are used to predict
price changes that result from predicted changes in agricultural outputs. Consumer and producer
surplus measures can then be developed to give reliable indications of the benefits of changes  in
ambient air quality for this category (see Section 9A.3.6.2).

9A.3.2 Growth in WTP Reflecting National Income Growth Over Time

   Our analysis accounts for expected growth in real income over time.  Economic theory
argues that WTP for most goods (such as environmental protection) will increase if real incomes
increase. There is substantial empirical evidence that the income elasticity1 of WTP for health
risk reductions is positive, although there is uncertainty about its exact value. Thus, as real
income increases the WTP for environmental improvements also increases. While many
analyses assume that the income elasticity of WTP is unit elastic (i.e., ten percent higher real
income level implies a ten percent higher WTP to reduce risk changes), empirical evidence
suggests that income elasticity is substantially less than one and thus relatively inelastic. As real
income rises, the WTP value also rises but at a slower rate than real income.

   The effects of real income changes on WTP estimates can influence benefit estimates in two
different ways: (1) through real income growth between the year a WTP study was conducted
and the year for which benefits are estimated, and (2) through differences in income between
study populations  and the affected populations at a particular time.  Empirical evidence of the
effect of real income on WTP gathered to date is based on studies examining the former. The
Environmental Economics Advisory Committee (EEAC) of the SAB advised EPA to adjust
WTP for increases in real income over time, but not to adjust WTP to account for cross-sectional
income differences "because of the sensitivity of making such distinctions, and because of
insufficient evidence available at present" (EPA-SAB-EEAC-00-013).

   Based  on a review of the available income elasticity literature, we adjust the valuation of
human health benefits upward to account for projected growth in real U.S. income. Faced with a
dearth of estimates of income elasticities derived from time-series studies, we applied estimates
   LIncome elasticity is a common economic measure equal to the percentage change in WTP for a one percent
change in income.

                                         9-118

-------
                                                                     Cost-Benefit Analysis
derived from cross-sectional studies in our analysis.  Details of the procedure can be found in
Kleckner and Neumann (1999).  An abbreviated description of the procedure we used to account
for WTP for real income growth between  1990 and 2030 is  presented below."1

   Reported income elasticities suggest that the severity of a health effect is a primary
determinant of the strength of the relationship between changes in real income and WTP.  As
such, we use different elasticity estimates  to adjust the WTP for minor health effects, severe and
chronic health effects, and premature mortality.  We also expect that the WTP for improved
visibility in Class I areas would increase with growth in real income. The elasticity values used
to adjust estimates of benefits in 2020 and 2030 are presented in Table 9A-11.

   Table 9A-15. Elasticity Values Used to Account for Projected Real Income GrowthA
Benefit Category
Minor Health Effect
Severe and Chronic Health Effects
Premature Mortality
Visibility8
Central Elasticity Estimate
0.14
0.45
0.40
0.90
A Derivation of estimates can be found in Kleckner and Neumann (1999) and Chestnut (1997). Cost of Illness (COI) estimates
are assigned an adjustment factor of 1.0.
B No range was applied for visibility because no ranges were available in the current published literature.
    In addition to elasticity estimates, projections of real GDP and populations from 1990 to
2020 and 2030 are needed to adjust benefits to reflect real per capita income growth. For
consistency with the emissions and benefits modeling, we use national population estimates for
the years 1990 to 1999 based on U.S. Census Bureau estimates (Hollman, Mulder and Kalian,
2000). These population estimates are based on application of a cohort-component model
applied to 1990 U.S. Census data projections".  For the years between 2000 and 2030, we applied
growth rates based on the U.S. Census Bureau projections to the U.S. Census estimate of
national  population in 2000. We use projections of real GDP provided in Kleckner and
    M Industry commentors suggest that the income elasticity values used to adjust willingness to pay (WTP) values
for avoidance of adverse health effects are based on incorrect methodology. Specifically, they assert that EPA
values are based on cross-sectional data when they should be based on time series data. The method we used to
derive income adjustment factors, which is detailed here, is consistent with advice from the SAB-EEAC and reflect
modest increases in WTP over time. Some recent evidence from published meta-analyses (see Viscusi and Aldy,
2003) suggest that we should be using a larger income adjustment factor for premature mortality.

    NU.S. Bureau of Census. Annual Projections of the Total Resident Population, Middle Series, 1999-2100.
(Available on the internet at http://www.census.gov/population/www/projections/natsum-Tl.html)

                                            9-119

-------
Final Regulatory Impact Analysis
Neumann (1999) for the years 1990 to 2010°. We use projections of real GDP (in chained 1996
dollars) provided by Standard and Poor'sp for the years 2010 to 2024q.  The Standard and Poor's
database only provides estimates of real GDP between 1990 and 2024.  We were unable to find
reliable projections of GDP past 2024. As such, we assume that per capita GDP remains
constant between 2024 and 2030.

    Using the method outlined in Kleckner and Neumann (1999), and the population and income
data described above, we calculate WTP adjustment factors for each of the elasticity estimates
listed in Table 1. Benefits for each of the  categories (minor health effects, severe and chronic
health effects, premature mortality, and visibility) will be adjusted by multiplying the unadjusted
benefits by the appropriate adjustment factor. Table 2 lists the estimated adjustment factors.
Note that for premature mortality, we apply the income adjustment factor ex post to the present
discounted value of the stream of avoided mortalities occurring over the lag period.  Also note
that no adjustments will be made to benefits based on the cost-of-illness approach or to work loss
days and worker productivity. This assumption will also lead us to under predict benefits in
future years since it is likely that increases in real U.S.  income would also result in increased
cost-of-illness (due, for example, to increases in wages paid to medical workers) and increased
cost of work loss days and lost worker productivity (reflecting that if worker incomes are higher,
the losses resulting from reduced worker production would also be higher). No adjustments are
needed  for agricultural benefits,  as the model is based on projections of supply and demand in
future years and should already incorporate future changes in real income.
    °U.S. Bureau of Economic Analysis, Table 2A (1992$). (Available on the internet at
http://www.bea.doc.goWbea/dn/0897nip2/tab2a.htni) and U.S. Bureau of Economic Analysis, Economics and
Budget Outlook.  Note that projections for 2007 to 2010 are based on average GDP growth rates between 1999 and
2007.

    Standard and Poor's. 2000. "The U.S. Economy: The 25 Year Focus." Winter.

    QIn previous analyses, we used the Standard and Poor's projections of GDP directly. This led to an apparent
discontinuity in the adjustment factors between 2010 and 2011. We refined the method by applying the relative
growth rates for GDP derived from the Standard and Poor's projections to the 2010 projected GDP based on the
Bureau of Economic Analysis projections.

                                           9-120

-------
                                                                     Cost-Benefit Analysis
  Table 9A-16.  Adjustment Factors Used to Account for Projected Real Income GrowthA>B
Benefit Category
Minor Health Effect
Severe and Chronic Health Effects
Premature Mortality
Visibility
2020
1.066
1.229
1.201
1.516
2030C
1.076
1.266
1.233
1.613
A Based on elasticity values reported in Table 9A-11, US Census population projections, and projections of real
gross domestic product per capita.
B Note that these factors have been modified from the proposal analysis to refelect relative growth rates for GDP
derived from the Standard and Poor's projections rather than absolute growth rates.
c Income growth adjustment factor for 2030 is based on an assumption that there is no growth in per capita income
between 2024 and 2030, based on a lack of available GDP projections beyond 2024.

9A.3.3 Methods for Describing Uncertainty

    In any complex analysis using estimated parameters and inputs from numerous models, there
are likely to be many sources of uncertainty/ This analysis is no exception.  As outlined both in
this and preceding chapters, many inputs are used to derive the final estimate of benefits,
including emission inventories, air quality models (with their associated parameters and inputs),
epidemiological health effect estimates, estimates of values (both from WTP and cost-of-illness
studies), population estimates, income estimates, and estimates of the future state of the world
(i.e., regulations, technology, and human behavior).  Each of these inputs may be uncertain, and
depending on their location in the benefits analysis, may have a disproportionately large impact
on final estimates of total benefits.  For example, emissions estimates are used in the first stage
of the analysis. As such, any uncertainty in emissions estimates will be propagated through the
entire analysis. When compounded with uncertainty in later stages, small uncertainties in
emission levels can lead to much larger impacts on total benefits. A more thorough discussion of
uncertainty can be found in the benefits technical support document (TSD) (Abt Associates,
2003).

    Some key sources of uncertainty in each stage of the benefits analysis are:

    -   Gaps in scientific data and inquiry;
    R  It should be recognized that in addition to uncertainty, the annual benefit estimates for the Nonroad Diesel
Engines rulemaking presented in this analysis are also inherently variable, due to the truly random processes that
govern pollutant emissions and ambient air quality in a given year. Factors such as engine hours and weather
display constant variability regardless of our ability to accurately measure them. As such, the estimates of annual
benefits should be viewed as representative of the types of benefits that will be realized, rather than the actual
benefits that would occur every year.
                                            9-121

-------
Final Regulatory Impact Analysis
    -   Variability in estimated relationships, such as epidemiological effect estimates,
       introduced through differences in study design and statistical modeling;
       Errors in measurement and projection for variables such as population growth rates;
       Errors due to misspecification of model structures, including the use of surrogate
       variables, such as using PM10 when PM2 5 is not available, excluded variables, and
       simplification of complex functions; and
    -   Biases due to omissions or other research limitations.

    Some of the key uncertainties in the benefits analysis are presented in Table 9A-13. Given
the wide variety of sources for uncertainty and the potentially large  degree of uncertainty about
any primary estimate, it is necessary for us to address this issue in several ways, based on the
following types of uncertainty:

a.   Quantifiable uncertainty in benefits estimates.   For some parameters or inputs it may be
    possible to provide a statistical representation of the underlying  uncertainty distribution.
    Quantitative uncertainty may include measurement uncertainty or variation in estimates
    across or within studies.  For example, the variation in VSL results across available meta-
    analyses provides a quantifiable basis for representing some uncertainty that can be
    calculated for monetized benefits.  Methods typically used to evaluate the impact of these
    quantifiable sources of uncertainty on benefits and incidence estimates center on Monte
    Carlo-based probabilistic simulation.  This technique allows uncertainty in key inputs to be
    propagated through the model to generate a single distribution of results reflecting the
    combined impact of multiple sources of uncertainty.  Variability can also be considered
    along with uncertainty using nested two-stage Monte Carlo simulation.

b.   Uncertainty in the basis for quantified estimates. Often it is possible to identify a source of
    uncertainty (for example, an ongoing debate over the proper method to estimate premature
    mortality) that is not readily addressed through traditional uncertainty analysis. In these
    cases, it is possible to characterize the potential impact of this uncertainty on the overall
    benefits estimates through sensitivity analyses.

c.   Nonquantifiable uncertainty.  Uncertainties may also result from omissions of known effects
    from the benefits calculation, perhaps owing to a lack of data or modeling capability. For
    example, in this analysis we were unable to quantify the benefits of avoided airborne
    nitrogen deposition on aquatic and terrestrial ecosystems, diesel  odor, or avoided health and
    environmental effects associated with reductions in CO emissions.

It should be noted that, even for individual endpoints, there is usually more than one source of
uncertainty. This makes it difficult to provide an overall quantified  uncertainty estimate for
individual endpoints or for total benefits, without conducting a comprehensive uncertainty

                                          9-122

-------
                                                                 Cost-Benefit Analysis
analysis that considers the aggregate impact of multiple sources of uncertainty on benefits
estimates.

   The NAS report on the EPA's benefits analysis methodology highlighted the need for the
EPA to conduct rigorous quantitative analysis of uncertainty in its benefits estimates.  In
response to these comments, the EPA has initiated the development of a comprehensive
methodology for characterizing the aggregate impact of uncertainty in key modeling elements on
both health incidence and benefits estimates. This methodology will begin by identifying those
modeling elements that have a significant impact on benefits due to either the magnitude of their
uncertainty or other factors such as nonlinearity within the modeling framework. A combination
of influence analysis and sensitivity analysis methods may be used to focus the analysis of
uncertainty on these key sources of uncertainty. A probabilistic simulation approach based on
Monte Carlo methods will be developed for propagating the impact of these sources of
uncertainty through the modeling framework.  Issues such as correlation between input
parameters and the identification of reasonable upper and lower bounds for input distributions
characterizing uncertainty will be addressed in developing the approach.

   For this analysis of the final rule, EPA has addressed key sources of uncertainty through a
series of sensitivity analyses examining the impact of alternate assumptions on the benefits
estimates that are generated.  Sensitivity estimates are presented in Appendix  9C.  We also
present information related to an expert elicitation pilot in Appendix 9B.

   Our estimate of total benefits should be viewed as an approximate result because of the
sources of uncertainty discussed above (see Table 9A-13). Uncertainty about specific aspects of
the health and welfare estimation models are discussed in greater detail in the following sections
and in the benefits  TSD (Abt Associates, 2003). The total benefits estimate may understate or
overstate actual benefits of the rule.

   In considering the monetized benefits estimates, the reader should remain aware of the many
limitations of conducting these analyses mentioned throughout this RIA. One significant
limitation of both the health and welfare benefits analyses is the inability to quantify many of the
serious effects listed in Table 9A-1.  For many health and welfare effects, such as changes in
ecosystem functions and PM-related materials damage, reliable C-R functions and/or valuation
functions are not currently available.  In general, if it were possible to monetize these benefits
categories, the benefits estimates presented in this analysis would increase.  Unquantified
benefits are qualitatively discussed in the health and welfare effects sections. In addition to
unquantified benefits, there may also be environmental costs that we are unable to quantify.
Several of these environmental cost categories are related to nitrogen deposition, while one
category is related to the issue of ultraviolet light. These endpoints are qualitatively discussed in
                                          9-123

-------
the health and welfare effects sections as well. The net effect of excluding benefit and disbenefit
categories from the estimate of total benefits depends on the relative magnitude of the effects.

-------
                                                                                 Cost-Benefit Analysis
             Table 9A-17.  Primary Sources of Uncertainty in the Benefit Analysis
1.  Uncertainties Associated With Health Impact Functions
    The value of the ozone or PM effect estimate in each health impact function.
    Application of a single effect estimate to pollutant changes and populations in all locations.
    Similarity of future year effect estimates to current effect estimates.
    Correct functional form of each impact function.
    Extrapolation of effect estimates beyond the range of ozone or PM concentrations observed in the study.
    Application of effect estimates only to those subpopulations matching the original study population.
2.  Uncertainties Associated With Ozone and PM Concentrations
    Responsiveness of the models to changes in precursor emissions resulting from the control policy.
    Projections of future levels of precursor emissions, especially ammonia and crustal materials.
    Model chemistry for the formation of ambient nitrate concentrations.
    Lack of ozone monitors in rural areas requires extrapolation of observed ozone data from urban to rural areas.
    Use of separate air quality models for ozone and PM does not allow for a fully integrated analysis of pollutants and
        their interactions.
    Full ozone season air quality distributions are extrapolated from a limited number of simulation days.
    Comparison of model predictions of particulate nitrate with observed rural monitored nitrate levels indicates that
    REMSAD overpredicts nitrate in some parts of the Eastern US and underpredicts nitrate in parts of the Western
    US.
3.  Uncertainties Associated with PM Premature mortality Risk
    No scientific literature supporting a direct biological mechanism for observed epidemiological evidence.
    Direct causal agents within the complex mixture of PM have not been identified.
    The extent to which adverse health effects are associated with low level exposures that occur many times in the
    year versus peak exposures.
    The extent to which effects reported in the long-term exposure studies are associated with historically higher levels
    of PM rather than the  levels occurring during the period of study.
    Reliability of the limited ambient PM25 monitoring data in reflecting actual PM25 exposures.
4.  Uncertainties Associated With Possible Lagged Effects
—   The portion of the PM-related long-term exposure mortality effects associated with changes in annual PM levels
    would occur in a single year is uncertain as well as the portion that might occur in subsequent years.
5.  Uncertainties Associated With Baseline Incidence Rates
—   Some baseline incidence rates are not location-specific (e.g., those taken from studies) and may therefore not
    accurately represent the actual location-specific rates.
—   Current baseline incidence rates may not approximate well baseline incidence rates in 2030.
-   Projected population and demographics may not represent well future-year population and demographics.
6.  Uncertainties Associated With Economic Valuation
    Unit dollar values associated with health and welfare endpoints are only estimates of mean WTP and therefore
    have uncertainty surrounding them.
    Mean WTP (in constant dollars) for each type of risk reduction may differ from current estimates due to
    differences in income or other factors.
    Future markets for agricultural products are uncertain.
7.  Uncertainties Associated With Aggregation of Monetized Benefits
—   Health and welfare benefits estimates are limited to the available effect estimates. Thus, unquantified or
    unmonetized benefits are not included.
                                                   9-125

-------
Final Regulatory Impact Analysis
9A.3.4 Demographic Projections

   Quantified and monetized human health impacts depend critically on the demographic
characteristics of the population, including age, location, and income. In previous analyses, we
have used simple projections of total population that did not take into account changes in
demographic composition over time. In the current analysis, we use more sophisticated
projections based on economic forecasting models developed by Woods and Poole, Inc. The
Woods and Poole (WP) database contains county level projections of population by age, sex, and
race out to 2025. Projections in each county are determined simultaneously with every other
county in the U.S. to take into account patterns of economic growth and migration. The sum of
growth in  county level populations is constrained to equal a previously determined national
population growth, based on Bureau of Census estimates (Hollman, Mulder and Kalian, 2000).
According to WP, linking county level growth projections together and constraining to a national
level total growth avoids potential errors introduced by forecasting each county independently.
County projections are developed in a four stage process.  First, national level variables such as
income, employment, populations, etc. are forecasted.  Second, employment projections are
made for 172 economic areas defined by the Bureau of Economic Analysis, using an "export-
base" approach, which relies on linking industrial sector production of non-locally consumed
production items, such as outputs from mining, agriculture, and manufacturing with the national
economy.  The export-base approach requires estimation of demand equations or calculation of
historical growth rates for output and employment by sector. Third, population is projected for
each economic area  based on net migration rates derived from employment opportunities, and
following  a cohort-component method based on fertility and mortality in each area. Fourth,
employment and population projections are repeated for counties, using the economic region
totals as bounds. The age, sex, and race distributions for each region or county are determined
by aging the population by single year of age by sex and race for each year through 2025 based
on historical rates of mortality, fertility,  and migration.

   The WP projections of county level population are based on historical population data from
1969-1999, and do not include the 2000 Census results. Given the availability of detailed 2000
Census data, we constructed adjusted county level population projections for each future year
using a two stage process. First, we constructed ratios of the projected WP populations in a
future year to the projected WP population in 2000 for each future year by age, sex, and race.
Second, we multiplied the block level 2000 Census population data by the appropriate age,  sex,
and race specific WP ratio for the county containing the census block, for each future year.  This
results in a set of future population projections that is consistent with the most recent detailed
census data. The WP projections extend only through 2025.  To calculate populations for 2030,
we applied the growth rate from 2024 to 2025 to each year between 2025 and 2030.
                                         9-126

-------
                                                                 Cost-Benefit Analysis
   Figure 9 A-7 shows the projected trends in total U.S. population and the percentage of total
population aged zero to eighteen and over 65.  This figure illustrates that total populations are
projected increase from 281 million in 2000 to 345 million in 2025.  The percent of the
population 18 and under is expected to decrease slightly, from 27 to 25 percent, and the percent
of the population over 65 is expected to increase from 12  percent to 18 percent.

populations. For consistency with the emissions and benefits modeling, we use national
population estimates based on the U.S. Census Bureau projections.  We use projections of real
GDP provided in Kleckner and Neumann (1999) for the years 1990 to 2010.s We use projections
of real GDP (in chained  1996 dollars) provided by Standard and Poor's for the years 2010 to
2024.' The Standard and Poor's database only provides estimates of real GDP between 1990 and
2024. We were unable to find reliable projections of GDP beyond 2024. As such,  we assume
that per capita GDP remains constant between 2024 and 2030.  This assumption will lead us to
under-predict benefits because at least some level  of income growth would be projected to occur
between the years 2024 and 2030.

9A.3.5 Health Benefits Assessment Methods

       The most significant monetized benefits of reducing ambient concentrations of PM and
ozone are attributable to reductions in health risks associated with air pollution. The EPA's
Criteria Documents for ozone and PM list numerous health effects known to be linked to
ambient concentrations of these pollutants (EPA,  1996a and 1996b). As illustrated in Figure 9A-
1, quantification of health impacts requires several inputs, including epidemiological effect
estimates,  baseline incidence and prevalence rates, potentially affected populations, and
estimates of changes in ambient concentrations of air pollution.  Previous sections have
described the population and air quality inputs. This section  describes the effect estimates and
baseline incidence  and prevalence inputs and the methods used to quantify and monetize changes
in the expected number of incidences of various health  effects.
   s US Bureau of Economic Analysis, Table 2A (1992$). (Available on the internet at
http://www.bea.doc.goWbea/dn/0897nip2/tab2a.htm) and US Bureau of Economic Analysis, Economics and Budget
Outlook. Note that projections for 2007 to 2010 are based on average GDP growth rates between 1999 and 2007.

   T Standard and Poor's. 2000. "The U.S. Economy: The 25 Year Focus." Winter 2000.

                                          9-127

-------
      400.0
      350.0
      300.0
(/)
o  250.0

1
c
|  200.0

Q.
O
Q.
S  150.0
o
      100.0
       50.0
        0.0
                                                      Figure 9A-7.
                                       Projections of U.S. Population, 2000-2025
                                                                                            25.0%
                                                                                         30.0%
                                                                                            20.0% =
                                                                                                  o
                                                                                                  Q.
                                                                                            15.0% -g
                                                                                                  i-
                                                                                                  "5
                                                                                                  +J
                                                                                                  §
                                                                                            10.0% £
                                                                                            5.0%
                                                                                                     Population
                                                                                                    -% 18 and under
                                                                                                    -% 65 and over
                                                                                         0.0%
            2000 2002 2004 2005  2006  2008  2010  2012  2014  2015 2016  2018 2020 2022 2024 2025
   As noted above, values for environmental quality improvements are expected to increase with growth in real per capita income.
Accounting for real income growth over time requires projections of both real gross domestic product (GDP) and total U.S.

-------
                                                                     Cost-Benefit Analysis
    9A.3.5.1   Selecting Health Endpoints and Epidemiological Effect Estimates

    Quantifiable health benefits of the rule may be related to ozone only, PM only, or both
pollutants. Decreased worker productivity, respiratory hospital admissions for children under
two, and school absences are related to ozone but not PM. PM-only health effects include
premature mortality, nonfatal heart attacks, chronic bronchitis, acute bronchitis, upper and lower
respiratory symptoms, asthma exacerbations, and work loss days." Health effects related to both
PM and ozone include hospital admissions, emergency room visits for asthma, and minor
restricted activity days.

    We relied on the available published scientific literature to ascertain the relationship between
PM and ozone exposure and adverse human health effects.  We evaluated studies using the
selection criteria summarized in Table 9A-18.  These criteria include consideration of whether
the study was peer reviewed, the match between the pollutant studied and the pollutant of
interest, the study design and location, and characteristics of the study population, among other
considerations. The selection of C-R functions for the benefits analysis is guided by the goal of
achieving a balance between comprehensiveness and scientific defensibility.

    The Health Effects Institute (HEI) reported findings by health researchers at Johns Hopkins
University and others that have raised concerns about aspects of the statistical methods used in a
number of recent time-series studies of short-term exposures to air pollution and health effects
(Greenbaum, 2002). The estimates derived from the long-term exposure studies, which account
for a major share of the economic benefits described in this chapter, are not affected. Similarly,
the time-series studies employing generalized linear models (GLMs) or other parametric
methods, as well as case-crossover studies, are not affected.  As discussed in HEI materials
provided to the EPA and to CASAC (Greenbaum, 2002), researchers working on the National
Morbidity, Mortality, and Air Pollution Study (NMMAPS) found problems in the default
    "Evidence has been found linking ozone exposures with premature mortality independent of PM exposures. A
recent analysis by Thurston and Ito (2001) reviewed previously published time-series studies of the effect of daily
ozone levels on daily mortality and found that previous EPA estimates of the short-term exposure mortality benefits
of the ozone NAAQS (EPA, 1997) may have been underestimated by up to a factor of two, even when PM is
controlled for in the models. In its September 2001 advisory on the draft analytical blueprint for the second Section
812 prospective analysis, the SAB cited the Thurston and Ito study as a significant advance in understanding the
effects of ozone on daily mortality and recommended re-evaluation of the ozone mortality endpoint for inclusion in
the next prospective study (EPA-SAB-COUNCIL-ADV-01-004, 2001).  In addition, a recent World Health
Organization (WHO) report found that "recent epidemiological studies have strengthened the evidence that there are
short-term O3 effects on premature mortality and respiratory morbidity and provided further information on
exposure-response relationships and effect modification." (WHO, 2003).  Based on these new analyses and
recommendations, the EPA is currently reevaluating ozone-related mortality for inclusion in the primary benefits
analysis. The EPA is sponsoring three independent meta-analyses of the ozone-mortality epidemiology literature to
inform a determination on inclusion of this important health endpoint. Upon completion and peer review of the
meta-analyses, the EPA will make its determination on whether benefits of reductions in ozone-related mortality will
be included in the future benefits analyses.

                                             9-129

-------
Final Regulatory Impact Analysis
"convergence criteria" used in Generalized Additive Models (GAM) and a separate issue first
identified by Canadian investigators about the potential to underestimate standard errors in the
same statistical package. Following identification of the GAM issue, a number of time-series
studies were reanalyzed using alternative methods, typically GAM with more stringent
convergence criteria and an alternative model such as generalized linear models (GLM) with
natural smoothing splines, and the results of the reanalyses have been compiled and reviewed in
a recent HEI publication (HEI, 2003a). In most, but not all, of the reanalyzed studies, it was
found that risk estimates were reduced and confidence intervals increased with the use of GAM
with more stringent convergence criteria or GLM analyses; however, the reanalyses generally
did not substantially change the findings of the original studies, and the changes in risk estimates
with alternative analysis methods were much  smaller than the variation in effects across  studies.
The HEI review committee concluded the following:

    -  Although the number of studies showing an association of PM with premature mortality
       was slightly smaller, the PM association persisted in the majority of studies.
    -  In some of the large number of studies in which the PM association persisted, the
       estimates of PM effect were substantially smaller.
    -  In the few studies in which investigators performed further sensitivity analyses, some
       showed marked sensitivity of the PM effect estimate to the degree of smoothing and/or
       the specification of weather (HEI, 2003b, p. 269)

    Examination of the original studies used in our benefits analysis found that the health
endpoints that  are potentially affected by the GAM issues include reduced hospital admissions
and reduced lower respiratory symptoms.  For the analysis of the final rule, we have
incorporated a number of studies that  have been updated to correct for the GAM issue, including
Ito et al. (2003) for respiratory-related hospital admissions (COPD and pneumonia),  Shepard et
al. (2003) for respiratory-related hospital admissions (asthma), Moolgavkar (2003) for
cardiovascular-related hospital admissions (ICD codes 390-429), and Ito et al. (2003) for
cardiovascular-related hospital admissions (ischemic heart disease, dysrhythmia, and heart
failure).  Several additional hospital admissions-related studies have not yet been formally
updated to correct for the GAM issue.  These  include the lower respiratory symptoms study and
hospital admissions for respiratory and cardiovascular causes in populations aged 20 to 64.
However, as discussed above, available evidence suggests that the errors introduced into effect
estimates due to the GAM issue  should not significantly affect incidence results.
                                         9-130

-------
                                                           Cost-Benefit Analysis
Table 9A-18. Summary of Considerations Used in Selecting C-R Functions
Consideration
Peer reviewed
research
Study type
Study period
Population
attributes
Study size
Study location
Pollutants included
in model
Measure of
pollutant
Economically
valuable health
effects
Non-overlapping
endpoints
Comments
Peer reviewed research is preferred to research that has not undergone the peer review process.
Among studies that consider chronic exposure (e.g., over a year or longer) prospective cohort studies
are preferred over cross-sectional studies because they control for important individual-level
confounding variables that cannot be controlled for in cross-sectional studies.
Studies examining a relatively longer period of time (and therefore having more data) are preferred,
because they have greater statistical power to detect effects. More recent studies are also preferred
because of possible changes in pollution mixes, medical care, and life style over time. However, when
there are only a few studies available, studies from all years will be included.
The most technically appropriate measures of benefits would be based on impact functions that cover
the entire sensitive population, but allow for heterogeneity across age or other relevant demographic
factors. In the absence of effect estimates specific to age, sex, preexisting condition status, or other
relevant factors, it may be appropriate to select effect estimates that cover the broadest population, to
match with the desired outcome of the analysis, which is total national- level health impacts.
Studies examining a relatively large sample are preferred because they generally have more power to
detect small magnitude effects. A large sample can be obtained in several ways, either through a large
population, or through repeated observations on a smaller population, i.e. through a symptom diary
recorded for a panel of asthmatic children.
U.S. studies are more desirable than non-U.S. studies because of potential differences in pollution
characteristics, exposure patterns, medical care system, population behavior and life style.
When modeling the effects of ozone and PM (or other pollutant combinations) jointly, it is important
to use properly specified impact functions that include both pollutants. Use of single pollutant models
in cases where both pollutants are expected to affect a health outcome can lead to double-counting
when pollutants are correlated.
For this analysis for PM-related effects, impact functions based on PM25 are preferred to PM10 because
the Nonroad Diesel Engine rule will regulate emissions of PM25 precursors and air quality modeling
was conducted for this size fraction of PM. Where PM25 functions are not available, PM10 functions
are used as surrogates, recognizing that there will be potential downward (upward) biases if the fine
fraction of PM10 is more (less) toxic than the coarse fraction. Adequacy of ozone exposure metrics in
studies was also considered.
Some health effects, such as forced expiratory volume and other technical measurements of lung
function, are difficult to value in monetary terms. These health effects are not quantified in this
analysis.
Although the benefits associated with each individual health endpoint may be analyzed separately,
care must be exercised in selecting health endpoints to include in the overall benefits analysis because
of the possibility of double counting of benefits.
                                      9-131

-------
Final Regulatory Impact Analysis
   It is important to reiterate that the estimates derived from the long-term exposure studies,
which account for a major share of the economic benefits described in this chapter, are not
affected by the GAM issue.  Similarly, the time-series studies employing GLMs or other
parametric methods, as well as case-crossover studies, are not affected.

   Although a broad range of serious health effects has been associated with exposure to
elevated ozone and PM levels (as noted for example in Table 9A-1 and described more fully in
the ozone and PM Criteria Documents (EPA, 1996a, 1996b)), we include only a subset of health
effects in this quantified benefit analysis. Health effects are excluded from this analysis for three
reasons: the possibility of double counting (such as hospital admissions for specific respiratory
diseases); uncertainties in applying effect relationships based on clinical studies to the affected
population; or a lack of an established relationship between the health effect and pollutant in the
published epidemiological literature.

   In general, the use of results from more than a single study can provide a more robust
estimate of the relationship between a pollutant and a given health effect.  However, there are
often differences between studies examining the same endpoint, making it difficult to pool the
results in a consistent manner. For example, studies may examine different pollutants or
different age groups. For this reason, we consider very carefully the set of studies available
examining each endpoint and select a consistent subset that provides a good balance of
population coverage and match with the pollutant of interest. In many cases, either because of a
lack of multiple studies, consistency problems, or clear superiority in the quality or
comprehensiveness of one study over others, a single published study is selected as the basis of
the effect estimate.

   When several effect estimates for a pollutant and a given health endpoint have been selected,
they are quantitatively combined or pooled to derive a more robust estimate of the relationship.
The benefits Technical Support Document (TSD) completed for the nonroad diesel rulemaking
provides details of the procedures used to combine  multiple impact functions (Abt Associates,
2003).  In general, we use fixed or random effects models to pool estimates from different
studies of the same endpoint. Fixed effects pooling simply weights each study's estimate by the
inverse variance, giving more weight to studies with greater statistical power (lower variance).
Random effects pooling accounts for both within-study variance and between-study variability,
due, for example, to differences in population susceptibility. We use the fixed effects model  as
our null hypothesis and then determine whether the data suggest that we should reject this null
hypothesis, in which case we would use the random effects model.v Pooled impact functions are
   vThe fixed effects model assumes that there is only one pollutant coefficient for the entire modeled area. The
random effects model assumes that different studies are estimating different parameters; therefore, there may be a
number of different underlying pollutant coefficients.

                                           9-132

-------
                                                                 Cost-Benefit Analysis
used to estimate hospital admissions (PM), school absence days (ozone), lower respiratory
symptoms (PM), asthma exacerbations (PM), and asthma-related emergency room visits (ozone).
For more details on methods used to pool incidence estimates, see the benefits TSD for the
nonroad diesel rulemaking (Abt Associates, 2003).

   Effect estimates for a pollutant and a given health endpoint are applied consistently across all
locations nationwide.  This applies to both impact functions defined by a single effect estimate
and those defined by a pooling of multiple effect estimates. Although the effect estimate may, in
fact, vary from one location to another (e.g., due to differences in population susceptibilities or
differences in the  composition of PM), location-specific effect estimates are generally not
available.

   The specific studies from which effect estimates for the primary analysis are drawn are
included  in Table 9A-19.

   Premature Mortality. Both long- and short-term exposures to ambient levels of air pollution
have been associated with increased risk of premature mortality.  The size of the premature
mortality risk estimates from these epidemiological studies, the serious nature of the effect itself,
and the high monetary value ascribed to prolonging life make premature mortality risk reduction
the most important health endpoint quantified in this analysis.

   Epidemiological analyses have consistently linked air pollution, especially PM, with excess
mortality. Although a number of uncertainties remain to be addressed by continued research
(NRC, 1998), a substantial body of published scientific literature documents the correlation
between elevated  PM  concentrations and increased mortality rates. Community epidemiological
studies that have used both short-term and long-term exposures and response have been used to
estimate PM/ mortality relationships. Short-term studies use a time-series approach to relate
short-term (often day-to-day) changes in PM concentrations and changes in daily mortality rates
up to several days after a period of elevated PM concentrations. Long-term studies examine the
potential  relationship between community-level PM exposures over multiple years and
community-level annual mortality rates.

   Researchers have found statistically significant associations between PM and premature
mortality using both types  of studies. In general, the risk estimates based on the long-term
exposure studies are larger than those derived from short-term studies. Cohort analyses are
better able to capture the full public health impact of exposure to air pollution  over time (Kunzli,
2001; NRC, 2002). This section discusses some of the issues surrounding the  estimation of
premature mortality.
                                          9-133

-------
Final Regulatory Impact Analysis
    Over a dozen studies have found significant associations between various measures of
long-term exposure to PM and elevated rates of annual mortality, beginning with Lave and
Seskin (1977). Most of the published studies found positive (but not always statistically
significant) associations with available PM indices such as total suspended particles (TSP).
Particles of different fine particles components (i.e., sulfates), and fine particles, as well as
exploration of alternative model specifications sometimes found inconsistencies (e.g., Lipfert,
[1989]).  These early "cross-sectional" studies (e.g., Lave and Seskin [1977]; Ozkaynak and
Thurston [1987]) were criticized for a number of methodological limitations, particularly for
inadequate control at the individual level for variables that are potentially important in causing
mortality, such as wealth, smoking, and diet.

    More recently, several  long-term studies have been published that use improved approaches
and appear to be consistent with the earlier body of literature. These new "prospective cohort"
studies reflect a significant improvement over the earlier work because they include individual-
level information with respect to health status and residence.  The most extensive study and
analyses has been based on data from two prospective cohort groups, often referred to as the
Harvard "Six-City Study"  (Dockery et al., 1993) and the "American Cancer Society or ACS
study" (Pope et  al., 1995 and Pope et al., 2002);  these studies have found consistent
relationships between fine particle indicators and premature mortality across multiple locations
in the United States.  A third major data set comes from the California based 7th Day Adventist
Study (e.g., Abbey et al., 1999), which reported associations between long-term PM exposure
and premature mortality in men. Results from this cohort, however, have been inconsistent and
the air quality results are not geographically representative of most of the United States.  The
Veterans Study was originally designed as a means of assessing the efficacy of anti-hypertensive
drugs in reducing morbidity and mortality in a population with pre-existing high blood pressure
(in this case, male veterans) (Lipfert et al., 2000). Unlike previous long-term analyses, this study
found some associations between premature mortality and ozone but found inconsistent results
for PM indicators.  A variety of issues associated with the study design, including sample
representativeness and loss to follow up, make this cohort a poor choice for extrapolating to the
general public. Furthermore, because of the selective nature of the population in the veteran's
cohort and methodological weaknesses, which may have resulted in estimates of relative risk that
are biased relative to a relative risk for the general population, we have chosen not to include any
effect estimates from the Lipfert et al. (2000) study in our benefits assessment. We note that,
while the PM analyses considering segmented (shorter) time periods such as Lipfert et al. gave
differing results (including significantly negative mortality coefficients for some PM metrics),
when methods consistent with the past studies were used (i.e., many- year average PM
concentrations),  similar results were reported:  the authors found that "(t)he single-mortality-
period responses without ecological variables are qualitatively similar to what has been reported
before (SO4 > PM2 5 > PM15)."
                                          9-134

-------
                                                            Cost-Benefit Analysis
Table 9A-19. Endpoints and Studies Used to Calculate Total Monetized Health Benefits
Endpoint
Pollutant | Study
Study Population
Premature Mortality
Premature Mortality — Long-
term exposure, all-cause
Premature Mortality — Long-
term exposure, all-cause
PM25
PM25
Pope et al. (2002)
Woodruff etal., 1997
>29 years
Infant (<1 yr)
Chronic Illness
Chronic Bronchitis
Non-fatal Heart Attacks
PM25
PM25
Abbey, et al. (1995)
Peters etal. (2001)
> 26 years
Adults
Hospital Admissions
Respiratory
Cardiovascular
Asthma- Related ER Visits
Ozone
Ozone
PM2.5
PM2.5
PM2.5
PM2.5
PM2.5
PM2.5
Ozone
PM25
Pooled estimate:
Schwartz (1995) - ICD 460-5 19 (all resp)
Schwartz (1994a, 1994b) - ICD 480-486 (pneumonia)
Moolgavkar et al. (1997) - ICD 480-487 (pneumonia)
Schwartz (1994b) - ICD 491-492, 494-496 (COPD)
Moolgavkar et al. (1997) - ICD 490-496 (COPD)
Burnett etal. (2001)
Pooled estimate:
Moolgavkar (2003) - ICD 490-496 (COPD)
Ito (2003) - ICD 490-496 (COPD)
Moolgavkar (2000) - ICD 490-496 (COPD)
Ito (2003) - ICD 480-486 (pneumonia)
Sheppard, et al. (2003) - ICD 493 (asthma)
Pooled estimate:
Moolgavkar (2003) - ICD 390-429 (all cardiovascular)
Ito (2003) - ICD 410-414, 427-428 (ischemic heart disease,
dysrhythmia, heart failure)
Moolgavkar (2000) - ICD 390-429 (all cardiovascular)
Pooled estimate: Weisel et al. (1995), Cody et al. (1992),
Stiebetal. (1996)
Norrisetal. (1999)
> 64 years
< 2 years
> 64 years
20-64 years
> 64 years
< 65 years
> 64 years
20-64 years
All ages
0-1 8 years
                                                                         (continued)
                                      9-135

-------
Final Regulatory Impact Analysis
Table 9A-19.  Endpoints and Studies Used to Calculate Total Monetized Health Benefits
(continued)
Endpoint
Pollutant | Study
Study Population
Other Health Endpoints
Acute Bronchitis
Upper Respiratory Symptoms
Lower Respiratory
Symptoms
Asthma Exacerbations
Work Loss Days
School Absence Days
Worker Productivity
Minor Restricted Activity
Days
PM25
PM10
PM25
PM25
PM25
Ozone
Ozone
PM25,
Ozone
Dockeryetal.(1996)
Popeetal. (1991)
Schwartz and Neas (2000)
Pooled estimate:
Ostro et al. (2001) (cough, wheeze and shortness of breath)
Vedaletal. (1998) Cough
Ostro (1987)
Pooled estimate:
Gilliland et al. (2001)
Chen et al. (2000)
Crocker and Horst (1981)
Ostro and Rothschild (1989)
8- 12 years
Asthmatics, 9-11
years
7-14 years
6- 18 years1
18-65 years
9- 10 years
6-11 years
Outdoor workers,
18-65
18-65 years
   The original study populations were 8 to 13 for the Ostro et al. (2001) study and 6 to 13 for the Vedal et al.
   (1998) study. Based on advice from the SAB-HES, we have extended the applied population to 6 to 18,
   reflecting the common biological basis for the effect in children in the broader age group.
   Given their consistent results and broad geographic coverage, the Six-City and ACS data
have been particularly important in benefits analyses.  The credibility of these two studies is
further enhanced by the fact that they were subject to extensive reexamination and reanalysis by
an independent team of scientific experts commissioned by HEI (Krewski et al., 2000). The
final results of the reanalysis were then independently peer reviewed by a Special Panel of the
HEI Health Review Committee. The results of these reanalyses confirmed and expanded those
of the original investigators.  This intensive independent reanalysis effort was occasioned both
by the importance of the original findings as well as concerns that the underlying individual
health effects information has never been made publicly available.

   The HEI re-examination lends credibility to the original studies and highlights sensitivities
concerning the relative impact of various pollutants, the potential role of education in mediating
the association between pollution and premature  mortality, and the influence of spatial
                                          9-136

-------
                                                                      Cost-Benefit Analysis
correlation modeling.™ Further confirmation and extension of the overall findings using more
recent air quality and a longer follow-up period for the ACS cohort was recently published in the
Journal of the American Medical Association (Pope et al., 2002).

    In developing and improving the methods for estimating and valuing the potential reductions
in premature mortality risk over the years, the EPA has consulted  with the SAB-HES.  That
panel recommended use of long-term prospective cohort studies in estimating premature
mortality risk reduction (EPA-SAB-COUNCIL-ADV-99-005, 1999).  This recommendation has
been confirmed by a recent report from the National Research Council, which stated that "it is
essential to use the cohort studies in benefits analysis to capture all important effects from air
pollution exposure" (NAS, 2002, p. 108).  In the NRC's view, compared with the time-series
studies, cohort studies give a more complete assessment of the long-term, cumulative effects of
air pollution.  The overall effect estimates may be a combination of effects  from long-term
exposure plus some fraction from short-term exposure, but the amount of overlap is unknown.
Additionally, the SAB recommended emphasis on the ACS study  because it includes a much
larger sample size and longer exposure interval and covers more locations (e.g., 50 cities
compared to the Six Cities Study) than other studies of its kind. As explained in the regulatory
impact analysis for the Heavy-Duty Engine/Diesel Fuel rule (EPA, 2000a), more recent EPA
benefits analyses have relied on an improved specification of the ACS cohort data that was
developed in the HEI reanalysis (Krewski et al., 2000).  The latest reanalysis of the ACS cohort
data (Pope et al., 2002), provides additional refinements to the analysis of PM-related mortality
by (a) extending the follow-up period for the ACS study subjects to 16 years, which triples the
size of the mortality data set; (b) substantially increasing exposure data, including consideration
    w Regarding potential confounding by copollutants, commentors noted that the HEI reanalysis of the ACS study
data for long-term exposure mortality found an association between SO2 and premature mortality and did not find a
strong association between PM2 5 and premature mortality. These commentors suggest that these findings regarding
potential confounding compromise the accuracy of the ACS study. While recognizing the need for research into the
issue of copollutants, including SO2, we disagree with the commentor's interpretation of the HEI reanalysis. While
this study did find an association between premature mortality and SO2, such an association was also reported for
fine particles and sulfate. In addition, the HEI reanalysis, as well as other studies examining the copollutant issue
(Samet et al., 2000, 2001) have suggested that SO2 might represent a surrogate for ambient PM2 5 concentrations and
is likely associated with sulfate concentrations since it is a precursor. This could partially explain the association
between SO2 and premature mortality found in the HEI reanalysis. Finally, we have updated our methods for
characterizing premature mortality and are now using the Pope et al. 2002 reanalysis of the ACS study data. While
this study continues to find and association between SO2 and cardiovascular mortality, it also finds the strongest
association yet between long term PM2 5 exposure and premature mortality.
    Commentors have also suggested that both the ACS and Six Cities studies provide evidence for confounding by
socio-economic factors in the chronic exposure mortality endpoint. Following recommendations by the SAB-HES,
we have updated our analytical framework to use the Pope et al. 2002 reanalysis of the ACS study data in estimating
long-term exposure mortality. This study incorporates consideration for a variety of potential risk factors including
smoking, educational status and age. With the exception of smoking status, none of the socio-economic factors
examined in the Pope et al. 2002 reanalysis had a significant effect on the association between premature mortality
and PM25 exposure. Rather than representing confounders, several of these socio-economic factors, including
educational status, were identified as potential effects modifiers.

                                             9-137

-------
Final  Regulatory Impact Analysis
for cohort exposure to PM25 following implementation of PM25 standard in 1999; (c) controlling
for a variety of personal risk factors including occupational exposure and diet; and (d) using
advanced statistical methods to evaluate specific issues that can adversely affect risk estimates
including the possibility of spatial autocorrelation of survival times in communities located near
each other.  Because of these refinements, the SAB- HES recommends using the Pope et al.
(2002) study as the basis for the primary mortality estimate for adults and suggests that alternate
estimates of premature mortality generated using other cohort and time series studies could be
included as part of the sensitivity analysis (SAB-HES, 2003).

    The SAB-HES also recommended using the estimated relative risks from the Pope et al.
(2002) study based on the average exposure to PM2 5, measured by the average of two PM25
measurements, over the periods 1979-1983, and 1999-2000. In addition to relative risks for all-
cause mortality, the Pope et al. (2002) study provides relative risks for cardiopulmonary, lung
cancer, and all other cause mortality.x Because of concerns regarding the statistical reliability of
the all-other cause mortality relative risk estimates, we calculate premature mortality impacts for
the primary analysis based on the all-cause relative risk.  However, we provide separate
estimates of cardiopulmonary and lung cancer deaths to show how these important causes of
death are affected by reductions in PM2 5.

    In previous RIAs, infant mortality has not been evaluated as part of the primary analysis
because of uncertainty  in the strength of the association between exposure to PM and
postneonatal mortality. Instead, benefits estimates related to reduced infant mortality have been
included as part of the sensitivity analysis for RIAs. However, recently published studies have
strengthened the case for an association between PM exposure and respiratory inflamation and
infection leading to premature mortality in children under 5 years of age. Specifically, the SAB-
HES noted the release of the World Health Organization Global Burden of Disease Study
focusing on ambient air, which cites several recently published time-series studies relating daily
PM exposure to mortality  in children (SAB-HES, 2003). The SAB-HES also cites the study by
Belanger et al. (2003) as corroborating findings linking PM exposure to increased respiratory
    x Commentors pointed out that both cardiovascular disease and cancer have latency periods of from 15 to 20
years.  Therefore, given that PM concentrations were four times higher in the 1960's compared with the 1980's, we
may be overestimating mortality incidence by using effects estimates, based on the original ACS study data, that do
not sufficiently correct for these higher PM concentrations during earlier segments of the exposure period for target
populations. We recognize that uncertainty is introduced into benefits estimates as a result of both latency and lag
issues. As the SAB-HES pointed out, the lack of detailed temporal exposure data for long term prospective cohort
studies makes it difficult to characterize latency and lag periods and evaluate the importance of temporal variation in
exposure levels. The Pope et al. 2002 reanalysis of the ACS study data, which includes additional years of follow-up
data for the original study population, does suggest that lung cancer may have a longer latency period. However,
inclusion of additional years of exposure data, in the case of lung cancer has served to strengthen, rather than
weaken the association between PM2 5 and premature mortality. By contrast, inclusion of additional follow-on data
for cardiovascular effects has suggested that this endpoint may have a shorter latency/lag period in that the effects
estimate has been reduced and not strengthened with the inclusion of the additional data.

                                            9-138

-------
                                                                 Cost-Benefit Analysis
inflamation and infections in children. Recently, a study by Chay and Greenstone (2003) found
that reductions in TSP caused by the recession of 1981-1982 were related to reductions in infant
mortality at the county level.  With regard to the cohort study conducted by Woodruff et al.
(1997), the SAB- HES notes several strengths of the study, including the use of a larger cohort
drawn from a large number of metropolitan areas and efforts to control for a variety of individual
risk factors in infants (e.g., maternal educational level, maternal ethnicity, parental marital status,
and maternal smoking status). Based on these findings, the SAB-HES recommends that the EPA
incorporate infant mortality into the primary benefits estimate and that infant mortality be
evaluated using a impact function developed from the Woodruff et al. (1997) study (SAB-HES,
2003).

   Chronic Bronchitis.  Chronic bronchitis is characterized by mucus in the lungs and a
persistent wet cough for at least 3 months a year for several years in a row. Chronic bronchitis
affects an estimated 5 percent of the U.S. population (American Lung Association, 1999). A
limited number of studies  have estimated the impact of air pollution on new incidences of
chronic bronchitis.  Schwartz (1993) and Abbey et al.(1995) provide evidence that long-term PM
exposure gives rise to the development of chronic bronchitis in the United States. Because the
Nonroad Diesel regulations are expected to reduce primarily PM2 5, this analysis uses only the
Abbey et al.  (1995)  study, because it is the only study focusing on the relationship between PM25
and new incidences  of chronic bronchitis.

   NonfatalMyocardial Infarctions (heart attacks). Nonfatal heart attacks have been linked
with short-term exposures to PM2 5 in the United States (Peters et al., 2001) and other countries
(Poloniecki et al.  ,1997). We use a recent study by Peters  et al. (2001) as the basis for the impact
function estimating the relationship between PM25 and nonfatal heart attacks.  Peters et al. is the
only available U.S. study to provide a specific estimate for heart attacks. Other studies, such as
Samet et al. (2000) and Moolgavkar et al. (2000), show a consistent relationship between all
cardiovascular hospital admissions, including for nonfatal  heart attacks, and PM. Given the
lasting impact of a heart attack on longer-term health costs and earnings, we choose to provide a
separate estimate for nonfatal heart attacks based on the single available U.S. effect estimate.
The finding of a  specific impact on heart attacks is consistent with hospital admission and other
studies showing relationships between fine particles and cardiovascular effects both within and
outside the United States.  These studies provide a weight  of evidence for this type of effect, as
discussed in the Criteria Document.  Several epidemiologic studies (Liao et al., 1999; Gold et al.,
2000; Magari et al.,  2001) have shown that heart rate variability (an indicator of how much the
heart is able to speed up or slow down in response to momentary stresses) is negatively related to
PM levels. Heart rate variability is a risk factor for heart attacks and other coronary heart
diseases (Carthenon et a.l, 2002; Dekker et al., 2000; Liao et al., 1997, Tsuji et al., 1996).  As
such, significant impacts of PM on heart rate variability are consistent with an increased risk of
heart attacks.

                                          9-139

-------
Final Regulatory Impact Analysis
   Hospital and Emergency Room Admissions. Because of the availability of detailed hospital
admission and discharge records, there is an extensive body of literature examining the
relationship between hospital admissions and air pollution. Because of this, many of the hospital
admission endpoints use pooled impact functions based on the results of a number of studies.  In
addition, some studies have examined the relationship between air pollution and emergency
room (ER) visits. Because most ER visits do not result in an admission to the hospital (the
majority of people going to the ER are treated and return home), we treat hospital admissions
and ER visits separately, taking account of the  fraction of ER visits that are admitted to the
hospital.

   Hospital admissions require the patient to be examined by a physician and, on average, may
represent more serious incidents than ER visits. The two main groups of hospital admissions
estimated in this analysis are respiratory admissions and cardiovascular admissions. There is not
much evidence linking ozone or PM with other types  of hospital admissions.  The only type of
ER visits that have been consistently linked to ozone and PM in the United States are asthma-
related visits.

   To estimate avoided incidences of cardiovascular hospital admissions associated with PM25,
we use studies by Moolgavkar (2003) and Ito et al. (2003). There are additional published
studies showing a statistically significant relationship between PM10 and cardiovascular hospital
admissions. However, given that the preliminary control options we are analyzing are expected
to reduce primarily PM2 5, we have chosen to focus on the two studies focusing on PM2 5.  Both
of these studies provide  an effect estimate for populations over 65, allowing us to pool the
impact functions for this age group. Only Moolgavkar (2000) provided a separate effect estimate
for populations 20 to 647  Total cardiovascular hospital admissions are thus the sum of the
pooled estimate for populations over 65  and the single study  estimate for populations 20 to 64.
Cardiovascular hospital  admissions include admissions for myocardial infarctions. To avoid
double counting benefits from reductions in myocardial infarctions when applying the impact
function for cardiovascular hospital admissions, we first adjusted the baseline cardiovascular
hospital admissions to remove admissions for myocardial infarctions.

   To estimate total avoided incidences of respiratory hospital admissions, we use impact
functions for several respiratory causes,  including chronic obstructive pulmonary disease
(COPD), pneumonia, and asthma. As with cardiovascular admissions, there are additional
published studies showing a statistically significant relationship between PM10 and respiratory
   YNote that the Moolgavkar (2000) study has not been updated to reflect the more stringent GAM convergence
criteria. However, given that no other estimates are available for this age group, we have chosen to use the existing
study.  Given the very small (<5 percent) difference in the effect estimates for 65 and older cardiovascular hospital
admissions between the original and reanalyzed results, we do not expect there to be much bias introduced by this
choice.

                                          9-140

-------
                                                                 Cost-Benefit Analysis
hospital admissions. We use only those focusing on PM2 5. Both Moolgavkar (2000) and Ito et
al. (2003) provide effect estimates for COPD in populations over 65, allowing us to pool the
impact functions for this group.  Only Moolgavkar (2000) provided a separate effect estimate for
populations 20 to 64.z  Total COPD hospital admissions are thus the sum of the pooled estimate
for populations over 65 and the single study estimate for populations 20 to 64.  Only Ito et al.
(2003) estimated pneumonia, and only for the population 65 and older. In addition, Sheppard et
al. (2003) provided an effect estimate for asthma hospital admissions for populations under age
65. Total avoided incidences of PM-related respiratory-related hospital admissions is the sum of
COPD, pneumonia, and asthma admissions.

   To estimate the effects of PM air pollution reductions on asthma-related ER visits, we use the
effect estimate from a study of children  18 and under by Norris et al. (1999). As noted earlier,
there is another study by Schwartz examining a broader age group (less than 65), but the
Schwartz study focused on PM10 rather than PM25. We selected the Norris et al. (1999) effect
estimate because it better matched the pollutant of interest. Because children tend to have higher
rates of hospitalization  for asthma relative to adults under 65, we will likely capture the majority
of the impact of PM25 on asthma ER visits in populations under 65, although there may still be
significant impacts in the adult population under 65.

   To estimate avoided incidences of respiratory hospital admissions associated with ozone, we
use a number of studies examining hospital admissions for a range of respiratory illnesses,
including pneumonia and COPD. Two age groups, adults over 65  and children under 2, are
examined. For adults over 65, Schwartz (1995) provides effect estimates for two different cities
relating ozone and hospital admissions for all respiratory causes (defined as ICD codes 460-519).
Impact functions based on these studies are pooled first before being pooled with other studies.
Two studies (Moolgavkar et al.,  1997; Schwartz, 1994a) examined ozone and pneumonia
hospital admissions in Minneapolis. One additional study (Schwartz, 1994b) examined ozone
and pneumonia hospital admissions in Detroit.  The impact functions for Minneapolis are pooled
together first, and the resulting impact function  is then pooled with the impact function for
Detroit. This avoids assigning too much weight to the information coming from one city. For
COPD hospital admissions, there are two available studies, Moolgavkar et al. (1997), conducted
in Minneapolis, and Schwartz (1994b), conducted in Detroit. These two studies are pooled
together. To estimate total respiratory hospital admissions for adults over 65, COPD admissions
are added to pneumonia admissions, and the result is pooled with the Schwartz (1995) estimate
of total respiratory admissions. Burnett et al. (2001) is the only study providing an effect
estimate for respiratory hospital  admissions in children under 2.
   zAgain, given the very small (<10 percent) difference in the effect estimates for 65 and older COPD hospital
admissions between the original and reanalyzed results, we do not expect there to be much bias introduced by this
choice.

                                          9-141

-------
Final Regulatory Impact Analysis
   Acute Health Events and School/Work Loss Days.  As indicated in Table 9A-1, in addition to
mortality, chronic illness, and hospital admissions, a number of acute health effects not requiring
hospitalization are associated with exposure to ambient levels of ozone and PM.  The sources for
the effect estimates used to quantify these effects are described below.

   Around 4 percent of U.S. children between ages 5 and 17 experience episodes of acute
bronchitis annually (American Lung Association,  2002). Acute bronchitis is characterized by
coughing, chest discomfort, slight fever, and extreme tiredness, lasting for a number of days.
According to the MedlinePlus medical encyclopedia,aa with the exception of cough, most acute
bronchitis symptoms abate within 7 to 10 days.  Incidence of episodes of acute bronchitis in
children between the ages of 5 and 17 are estimated using an effect estimate developed from
Dockery etal. (1996).

   Incidences of lower respiratory symptoms (e.g., wheezing, deep cough) in children aged 7 to
14 are estimated using an effect estimate from Schwartz and Neas (2000).

   Because asthmatics have greater sensitivity to stimuli (including air pollution), children with
asthma can be more susceptible to a variety of upper respiratory symptoms (e.g., runny or stuffy
nose; wet cough; and burning, aching, or red eyes). Research on the effects of air pollution on
upper respiratory symptoms has thus focused on effects in asthmatics. Incidences of upper
respiratory  symptoms in asthmatic children aged 9 to 11 are estimated using an effect estimate
developed from Pope et al. (1991).

   Health effects from air pollution can also result in missed days of work (either from personal
symptoms or from caring for a sick family member). Work loss days due to PM2 5 are  estimated
using an effect estimate developed from Ostro (1987).  Children may also be absent from school
due to respiratory or other diseases caused by exposure to air pollution.  Most studies examining
school absence rates have found little or no association with PM2 5, but several studies  have
found a significant association between ozone levels and school absence rates. We use two
recent studies, Gilliland et al. (2001) and Chen et al. (2000), to  estimate changes in absences
(school loss days) due to changes in ozone levels.  The Gilliland et al. study  estimated the
incidence of new periods of absence, while the Chen et al. study examined absence on a given
day.  We convert the Gilliland estimate to days of absence by multiplying the absence  periods by
the average duration of an absence. We estimate an average duration of school absence of 1.6
days by dividing the average daily school absence rate from Chen et al. (2000) and Ransom and
Pope (1992) by the episodic absence rate from Gilliland et al. (2001). This provides estimates
from Chen et al. (2000) and Gilliland et al. (2000), which can be pooled to provide an  overall
estimate.
        http://www.nlm.nih.gov/medlineplus/ency/article/000124.htm, accessed January 2002.

                                         9-142

-------
                                                                Cost-Benefit Analysis
   Minor restricted activity days (MRAD) result when individuals reduce most usual daily
activities and replace them with less strenuous activities or rest, yet not to the point of missing
work or school. For example, a mechanic who would usually be doing physical work most of
the day will instead spend the day at a desk doing paper and phone work due to difficulty
breathing or chest pain. The effect of PM2 5 and ozone on MRAD  is estimated using an effect
estimate derived from Ostro and Rothschild (1989).

   In previous RIAs, we have not included estimates of asthma exacerbations in the asthmatic
population in the primary analysis because of concerns over double counting of benefits and
difficulties in differentiating asthma symptoms for purposes of first developing impact functions
that cover distinct endpoints and then establishing the baseline incidence estimates required for
predicting incidence reductions.  Concerns over double counting stem from the fact that studies
of the general population also include asthmatics, so estimates based solely on the asthmatic
population cannot be directly added to the general population numbers without double counting.
In one specific case (upper respiratory symptoms in children), the  only study available was
limited to asthmatic children, so this endpoint can be readily included in the calculation of total
benefits.  However, other endpoints, such as lower respiratory symptoms and MRADs, are
estimated for the total population that includes asthmatics. Therefore, to simply add predictions
of asthma-related symptoms generated for the population of asthmatics to these total population-
based estimates could result in double counting, especially if they  evaluate similar endpoints.

   The SAB-HES, in commenting on the analytical blueprint for 812 acknowledged these
challenges in evaluating asthmatic symptoms and appropriately adding them into the primary
analysis (SAB-HES, 2003).  However, despite these challenges, the SAB-HES recommends the
addition of asthma-related symptoms (i.e., asthma exacerbations) to the primary analysis,
provided that the studies use the panel study approach and that they have comparable design and
baseline frequencies in both asthma prevalence and exacerbation rates.  Note also, that the SAB-
HES, while supporting the incorporation of asthma exacerbation estimates,  does not believe that
the association between ambient air pollution, including ozone and PM, and the new onset of
asthma is sufficiently strong to support inclusion of this asthma-related endpoint in the primary
estimate. For this analysis, we have followed the SAB-HES  recommendations regarding asthma
exacerbations in developing the primary estimate.  To prevent double counting, we are focusing
the estimation on asthma exacerbations occurring in children and are excluding adults from the
calculation.  Asthma exacerbations occurring in adults are assumed to be captured in the general
population endpoints such as work loss days and MRADs. Consequently, if we had included an
adult-specific asthma exacerbation estimate, we would likely double count incidence for this
endpoint.  However, because the general population endpoints do not cover children (with regard
to asthmatic effects), an analysis focused specifically on asthma exacerbations for children (6 to
18 years of age) could be conducted without concern for double counting.
                                         9-143

-------
Final Regulatory Impact Analysis
   To characterize asthma exacerbations in children, we selected two studies (Ostro et al., 2001
and Vedal et al., 1998) that followed panels of asthmatic children.  Ostro et al. (2001) followed a
group of 138 African-American children in Los Angeles for 13 weeks, recording daily
occurrences of respiratory symptoms associated with asthma exacerbations (e.g., shortness of
breath, wheeze, and cough).  This study found a statistically significant association between
PM25, measured as a 12-hour average, and the daily prevalence of shortness of breath and
wheeze endpoints.  Although the association was not statistically significant for cough, the
results were still positive and close to significance; consequently, we decided to include this
endpoint, along with shortness of breath and wheeze, in generating incidence  estimates (see
below). Vedal et al. (1998) followed a group of elementary school children, including 74
asthmatics, located on the west coast of Vancouver Island for 18 months including
measurements of daily peak expiratory flow (PEF) and the tracking of respiratory symptoms
(e.g., cough, phlegm, wheeze, chest tightness) through the use of daily diaries. Association
between PM10 and respiratory symptoms for the asthmatic population was only reported for two
endpoints:  cough and PEF.  Because it is difficult to translate PEF measures into clearly defined
health endpoints that can be monetized, we only included the cough-related effect estimate from
this study in quantifying asthma exacerbations.  We employed the following pooling approach in
combining estimates generated using effect estimates from the two studies to produce a single
asthma exacerbation incidence estimate. First, we pooled the separate incidence estimates for
shortness of breath, wheeze, and cough generated using effect estimates from the Ostro et al.
study, because each of these endpoints is aimed at capturing the same overall  endpoint (asthma
exacerbations) and there could be overlap in their predictions. The pooled estimate from the
Ostro et al. study is then pooled with the cough-related estimate generated using the Vedal study.
The rationale for this second pooling step is similar to the first; both studies are  attempting to
quantify the same overall endpoint (asthma exacerbations).

   Additional epidemiological studies are available for characterizing asthma-related health
endpoints (the full list of epidemiological studies considered for modeling asthma-related
incidence are presented in Table 9A-20). However, based on recommendations from the SAB-
FIES, we decided not to use these additional studies in generating the primary estimate.  In
particular, the Yu et al. (2000) estimates show a much higher baseline incidence rate than other
studies, which may lead to an overstatement of the expected impacts in the overall asthmatic
population. The Whittemore and Korn (1980) study did not use a well-defined endpoint, instead
focusing on a respondent-defined "asthma attack." Other studies looked at respiratory symptoms
in asthmatics but did not focus on specific exacerbations of asthma.

   9A.3.5.2  Uncertainties Associated with Health Impact Functions

   Within-Study Variation. Within-study variation refers to the precision with which a given
study estimates the relationship between air quality changes and health effects. Health effects

                                          9-144

-------
                                                                   Cost-Benefit Analysis
studies provide both a "best estimate" of this relationship plus a measure of the statistical
uncertainty of the relationship.  This size of this uncertainty depends on factors such as the
number of subjects studied and the size of the effect being measured. The results of even the
most well-designed epidemiological studies are characterized by this type of uncertainty, though
well-designed studies typically report narrower uncertainty bounds around the best estimate than
do studies of lesser quality. In selecting health endpoints, we generally focus on endpoints
where a statistically significant relationship has been observed in at least some studies, although
we may pool together results from studies with both statistically  significant and  insignificant
estimates to avoid selection bias.

   Across-Study Variation. Across-study variation refers to the fact that different published
studies of the same pollutant/health effect relationship typically do not report identical findings;
in some instances the differences are substantial. These differences can exist even between
equally reputable studies and may result in health effect estimates that vary considerably.
Across-study variation can result from two possible causes. One possibility is that  studies report
different estimates of the single true relationship between a given pollutant and a health effect
due to differences in study design, random chance, or other factors.  For example, a hypothetical
study conducted in New York and one conducted in Seattle may report  different C-R functions
for the relationship between PM and mortality, in part because of differences between these two
study populations (e.g., demographics, activity patterns). Alternatively, study results may differ
because these two studies are in fact estimating different relationships; that is, the same
reduction in PM in New York and Seattle may result in different reductions in premature
mortality. This may result from a number of factors, such as differences in the relative
sensitivity of these two populations to PM pollution and differences in the composition of PM in
these two locations.  In either case, where we identified multiple studies that are appropriate for
estimating a given health effect, we generated a pooled estimate  of results from each of those
studies.
                                           9-145

-------
Final Regulatory Impact Analysis
Table 9A-20. Studies Examining Health Impacts in the Asthmatic Population Evaluated
for Use in the Benefits Analysis
Endpoint
Definition
Pollutant (Study
Asthma Attack Indicators1
Shortness of
breath
Cough
Wheeze
Asthma
exacerbation
Cough
Prevalence of shortness of
breath; incidence of
shortness of breath
Prevalence of cough;
incidence of cough
Prevalence of wheeze;
incidence of wheeze
> 1 mild asthma symptom:
wheeze, cough, chest
tightness, shortness of
breath)
Prevalence of cough
PM25
PM25
PM25
PM10,
PM,0
PM10
Ostroetal. (2001)
Ostroetal. (2001)
Ostroetal. (2001)
Yu et al. (2000)
Vedaletal. (1998)
Study Population

African-American
asthmatics, 8-13
African-American
asthmatics, 8-13
African-American
asthmatics, 8-13
Asthmatics, 5-13
Asthmatics, 6-13
Other symptoms/illness endpoints
Upper respiratory
symptoms
Moderate or worse
asthma
Acute bronchitis
Phlegm
Asthma attacks
> 1 of the following: runny
or stuffy nose; wet cough;
burning, aching, or red
eyes
Probability of moderate (or
worse) rating of overall
asthma status
> 1 episodes of bronchitis
in the past 12 months
"Other than with colds,
does this child usually
seem congested in the
chest or bring up phlegm?"
Respondent-defined
asthma attack
PM10
PM25
PM25
PM25
PM25,
ozone
Pope etal. (1991)
Ostroetal. (1991)
McConnell et al.
(1999)
McConnell et al.
(1999)
Whittemore and
Korn (1980)
Asthmatics 9-11
Asthmatics, all ages
Asthmatics, 9-15*
Asthmatics, 9-15*
Asthmatics, all ages
                                      9-146

-------
                                                                  Cost-Benefit Analysis
   Application ofC-R Relationship Nationwide. Regardless of the use of impact functions
based on effect estimates from a single epidemiological study or multiple studies, each impact
function was applied uniformly throughout the United States to generate health benefit estimates.
However, to the extent that pollutant/health effect relationships are region-specific, applying a
location-specific impact function at all locations in the United States may result in overestimates
of health effect changes in some locations and underestimates of health effect changes in other
locations. It is not possible, however, to know the extent or direction of the overall effect on
health benefit estimates introduced by application of a single impact function to the entire United
States. This may be a significant uncertainty in the analysis, but the current state of the scientific
literature does not allow for a region-specific estimation of health benefits.bb

   Extrapolation of Impact Functions Across Populations. Epidemiological  studies  often focus
on specific age ranges, either due to data availability limitations (e.g., most hospital admission
data come from Medicare records, which are limited to populations 65 and older), or to simplify
data collection (e.g., some asthma symptom studies focus on children at summer camps, which
usually have a limited age range). We have assumed for the primary analysis that most impact
functions should be applied only to those populations with ages that strictly match the
populations in the underlying epidemiological studies. However, in many cases, there is no
biological reason why the observed health effect would not also occur in other populations
within a reasonable range of the studied population.  For example, Dockery et al. (1996)
examined acute bronchitis in children aged 8 to 12.  There  is no biological reason to expect a
very different response in children aged 6 or 14. By excluding populations outside the range in
the studies, we may be underestimating the health impact in the overall population. In  response
to recommendations from the SAB-HES, where there appears to be a reasonable physiological
basis for expanding the age group associated with a specific effect estimate beyond the study
population to cover the full age group (e.g., expanding from a study population of 7 to  11 year
olds to the full 6to 18 year child age group), we have done so and used those expanded incidence
estimates in the primary analysis.

    Uncertainties in the PMMortality Relationship.  Health researchers have consistently linked
air pollution, especially PM, with excess mortality. A substantial body of published scientific
literature recognizes a correlation between elevated PM concentrations and increased premature
mortality rates. However, much about this relationship is still uncertain.  These uncertainties
include the following:
   BBAlthough we are not able to use region-specific effect estimates, we use region-specific baseline incidence
rates where available.  This allows us to take into account regional differences in health status, which can have a
significant impact on estimated health benefits.

                                          9-147

-------
Final Regulatory Impact Analysis
       •   Causality:  A substantial number of published epidemiological studies recognize an
          association between elevated PM concentrations and increased premature mortality
          rates; however, these epidemiological studies are not designed to and cannot
          definitively prove causation. For the analysis of the final Nonroad Diesel Engines
          rulemaking, we assumed a causal relationship between exposure to elevated PM and
          premature mortality, based on the consistent evidence of a correlation between PM
          and premature mortality reported in the substantial body of published scientific
          literature.

       •   Other Pollutants:  PM concentrations are correlated with the concentrations of other
          criteria pollutants, such as ozone and CO, and it is unclear how much each of these
          pollutants may influence mortality rates. Recent studies (see Thurston and Ito
          [2001]) have explored whether ozone may have premature mortality effects
          independent of PM, but we do not view the  evidence as conclusive at this time.  The
          EPA is currently evaluating the epidemiological literature on the relationship between
          ozone and premature mortality and in future regulatory analyses may include ozone
          mortality as a separate impact in  the primary analysis.  To the extent that the effect
          estimates we use to evaluate the preliminary control options in fact capture premature
          mortality effects of other criteria pollutants besides  PM, we may be overestimating
          the benefits of reductions in PM. However, we are  not providing separate estimates
          of the premature mortality benefits from the ozone and CO reductions likely to occur
          due to the preliminary control options.

       •   Shape of the C-R Function:  The shape of the true PM premature mortality C-R
          function is uncertain, but this analysis assumes the C-R function to have a log-linear
          form (as derived from the literature) throughout the relevant range of exposures. If
          this is not the correct form of the C-R function, or if certain scenarios predict
          concentrations well above the range of values for which the C-R function was fitted,
          avoided premature mortality may be mis-estimated.

       •   Regional Differences: As discussed above,  significant variability exists in the results
          of different PM/mortality studies. This variability may reflect regionally specific C-R
          functions resulting from regional differences in factors such as the physical and
          chemical composition of PM. If true regional differences exist, applying the
          PM/mortality C-R function to regions outside the study location could result in
          mis-estimation of effects in these regions.

       •   Exposure/Mortality Lags:  There is a potential time lag between changes in PM
          exposures and changes in premature mortality rates. For the chronic PM/mortality
          relationship, the length of the lag is unknown and may be dependent on the kind of
          exposure. The existence of such a lag is important for the valuation of premature
          mortality incidence because economic theory suggests that benefits occurring in the
          future should be discounted.  There is no specific scientific evidence of the existence
          or structure of a PM effects lag.  However, current scientific literature on adverse
          health effects similar to those associated with PM (e.g., smoking-related disease) and
          the difference in the effect size between chronic exposure studies and daily mortality

                                          9-148

-------
                                                                 Cost-Benefit Analysis
          studies suggests that all incidences of premature mortality reduction associated with a
          given incremental change in PM exposure probably would not occur in the same year
          as the exposure reduction.  The smoking-related literature also implies that lags of up
          to a few years or longer are plausible. Adopting the lag structure used in the Tier
          2/Gasoline Sulfur and Heavy-Duty Engine/Diesel Fuel RIAs and endorsed by the
          SAB (EPA-SAB-COUNCIL-ADV-00-001, 1999), we assume a 5-year lag structure.
          This approach assumes that 25 percent of PM-related premature deaths occur in each
          of the first 2 years after the exposure and the rest occur in equal parts (approximately
          17 percent) in each of the ensuing 3 years.

          Cumulative Effects: As a general point, we attribute the PM/mortality relationship in
          the underlying epidemiological studies to cumulative exposure to PM. However, the
          relative roles of PM exposure duration and PM exposure level in inducing premature
          mortality remain unknown at this time.
9A.3.5.3 Baseline Health Effect Incidence Rates

   The epidemiological studies of the association between pollution levels and adverse health
effects generally provide a direct estimate of the relationship of air quality changes to the relative
risk of a health effect, rather than an estimate of the absolute number of avoided cases. For
example, a typical result might be that a 10 |ig/m3 decrease in daily PM25 levels might decrease
hospital admissions by 3 percent. The baseline incidence of the health effect is necessary to
convert this relative change into a number of cases. The baseline incidence rate provides an
estimate of the incidence rate (number of cases of the health effect per year, usually per 10,000
or 100,000 general population) in the  assessment location corresponding to baseline pollutant
levels in that location. To derive the total baseline incidence per year, this rate must be
multiplied by the corresponding population number (e.g., if the baseline incidence rate is number
of cases per year per  100,000 population, it must be multiplied by the number of 100,000s in the
population).

   Some epidemiological studies examine the association between pollution levels and adverse
health effects in a specific subpopulation, such as asthmatics or diabetics.  In these cases, it is
necessary to develop  not only baseline incidence rates, but also prevalence rates for the defining
condition (e.g., asthma). For both baseline incidence and prevalence data, we use age-specific
rates where available. Impact functions are applied to individual age groups and then summed
over the relevant age  range to provide an estimate of total population benefits.

   In most cases, because of a lack of data or methods, we have not attempted to project
incidence rates to future years,  instead assuming that the most recent data on incidence rates are
the best prediction of future incidence rates. In recent years, better data on trends in incidence
and prevalence rates for some endpoints, such  as asthma, have become available. We are

                                         9-149

-------
Final Regulatory Impact Analysis
working to develop methods to use these data to project future incidence rates.  However, for our
primary benefits analysis of the final rule, we will continue to use current incidence rates.

   Table 9A-21 summarizes the baseline incidence data and sources used in the benefits
analysis. In most cases, a single national incidence rate is used, due to a lack of more spatially
disaggregated data. We used national incidence rates whenever possible, because these data are
most applicable to a national assessment of benefits.  However, for some studies, the only
available incidence information comes from the studies themselves; in these cases, incidence in
the study population is assumed to represent typical incidence at the national level.  However,
for hospital admissions, regional rates are available, and for premature mortality, county-level
data are available.

   Age-, cause-, and county-specific mortality rates were obtained from the U.S. Centers for
Disease Control  (CDC) for the years 1996 through 1998.  CDC maintains an online data
repository of health statistics, CDC Wonder,  accessible at http://wonder.cdc.gov/. The mortality
rates provided are derived from U.S. death records and U.S. Census Bureau postcensal
population estimates.  Mortality rates were averaged across 3 years (1996 through 1998) to
provide more stable estimates.  When estimating rates for age groups that differed from the CDC
Wonder groupings, we assumed that rates were uniform across all ages in the reported age group.
For example, to estimate mortality rates for individuals ages 30 and up, we scaled the 25- to 34-
year old death count and population by one-half and then generated a population-weighted
mortality rate using data for the older age groups. Note that we have not projected any changes
in mortality rates over time.  We are aware that the U.S. Census projections of total  and age-
specific mortality rates used in our population projections are based on projections of declines in
national level mortality rates for younger populations and increases in mortality rates for older
populations over time.  We are evaluating the most appropriate way to incorporate these
projections of changes in overall national  mortality rates into our database of county-level cause-
specific mortality rates. In the interim, we have not attempted to adjust future mortality rates.
This will lead to an overestimate of premature mortality benefits in future years, with the
overestimation bias increasing the further benefits are projected into the future. We do not at
this time have a quantified estimate of the magnitude of the potential bias in the years analyzed
for this rule (2010 and 2015).
                                          9-150

-------
Table 9A-21. Baseline Incidence Rates and Population Prevalence Rates for Use in Impact
Functions, General Population
Endpoint
Premature
mortality
Hospitalizations
Asthma ER visits
Chronic
Bronchitis
Nonfatal MI
(heart attacks)
Asthma
Exacerbations
Acute Bronchitis
Parameter
Daily or annual mortality
rate
Daily hospitalization rate
Daily asthma ER visit rate
Annual prevalence rate per
person
Age 18-44
Age 45-64
Age 65 and older
Annual incidence rate per
person
Daily nonfatal myocardial
infarction incidence rate
per person, 18+
Northeast
Midwest
South
West
Incidence (and prevalence)
among asthmatic African
American children
- daily wheeze
- daily cough
- daily dyspnea
Prevalence among
asthmatic children
- daily wheeze
- daily cough
- daily dyspnea
Annual bronchitis
incidence rate, children
Rates
Value
Age, cause, and county-
specific rate
Age, region, cause-specific
rate
Age, Region specific visit
rate
0.0367
0.0505
0.0587
0.00378
0.0000159
0.0000135
0.0000111
0.0000100
0.076(0.173)
0.067(0.145)
0.037 (0.074)
0.038
0.086
0.045
0.043
Source3
CDC Wonder (1996-1998)
1999 NHDS public use data
files"
2000 NHAMCS public use data
files0; 1999 NHDS public use
data filesb
1999 HIS (American Lung
Association, 2002b, Table 4)
Abbey etal. (1993, Table 3)
1999 NHDS public use data
filesb; adjusted by 0.93 for
prob. of surviving after 28 days
(Rosamond etal., 1999)
Ostro etal. (2001)
Vedal etal. (1998)
American Lung Association
(2002a, Table 11)
                                                                           (continued)

-------
Table 9A-21. Baseline Incidence Rates and Population Prevalence Rates for Use in Impact
Functions, General Population (continued)
Endpoint
Lower
Respiratory
Symptoms
Upper
Respiratory
Symptoms
Work Loss Days
Minor Restricted
Activity Days
School Loss
Days6

Parameter
Daily lower respiratory
symptom incidence among
childrend
Daily upper respiratory
symptom incidence among
asthmatic children
Daily WLD incidence rate
per person (18-65)
Age 18-24
Age 25-44
Age 45-64
Daily MRAD incidence
rate per person
Daily school absence rate
per person
Daily illness-related school
absence rate per person6
Northeast
Midwest
South
Southwest
Daily respiratory illness-
related school absence rate
per person
Northeast
Midwest
South
West
Rates
Value
0.0012
0.3419
0.00540
0.00678
0.00492
0.02137
0.055
0.0136
0.0146
0.0142
0.0206
0.0073
0.0092
0.0061
0.0124
Source3
Schwartz (1994, Table 2)
Popeetal. (1991, Table 2)
1996 HIS (Adams et al., 1999,
Table 41); U.S. Bureau of the
Census (2000)
Ostro and Rothschild (1989, p.
243)
National Center for Education
Statistics (1996)
1996 HIS (Adams et al., 1999,
Table 47); estimate of 180
school days per year
1996 HIS (Adams et al., 1999,
Table 47); estimate of 180
school days per year
        The following abbreviations are used to describe the national surveys conducted by the National Center for
        Health Statistics: HIS refers to the National Health Interview Survey; NHDS—National Hospital
        Discharge Survey; NHAMCS—National Hospital Ambulatory Medical Care Survey.
        Seeftp://ftp.cdc.gov/pub/Health_Statistics/NCHS/Datasets/NHDS/
        Seeftp://ftp.cdc.gov/pub/Health_Statistics/NCHS/Datasets/NHAMCS/
        Lower Respiratory Symptoms  are defined as > 2 of the following:  cough, chest pain, phlegm, wheeze
        The estimate of daily illness-related school absences excludes school loss days associated with injuries to
        match the definition in the Gilliland et al. (2001) study.

-------
Final Regulatory Impact Analysis
   For the set of endpoints affecting the asthmatic population, in addition to baseline incidence
rates, prevalence rates of asthma in the population are needed to define the applicable
population. Table 9A-22 lists the prevalence rates used to determine the applicable population
for asthma symptom endpoints. Note that these reflect current asthma prevalence and assume no
change in prevalence rates in future years. As noted above, we are investigating methods for
projecting asthma prevalence rates in future years.

   9A.3.5.4 Accounting for Potential Health Effect Thresholds

   When conducting clinical (chamber) and epidemiological studies, functions may be
estimated with or without explicit thresholds. Air pollution levels below the threshold are
assumed to have no associated adverse health effects. When a threshold is not assumed, as is
often the case in epidemiological studies, any exposure level is assumed to pose a nonzero risk
of response to at least one segment of the population.

   The possible existence of an effect  threshold is a very important scientific question and issue
for policy analyses such as this one. The EPA SAB Advisory Council for Clean Air
Compliance, which provides advice and review of the EPA's methods for assessing the benefits
and costs of the Clean Air Act under Section 812 of the Clean Air Act, has advised the EPA that
there is currently no scientific basis for selecting a threshold of 15 |ig/m3 or any other specific
threshold for the PM-related health effects considered in typical benefits analyses (EPA-SAB-
Council-ADV-99-012, 1999). This is supported by the recent literature on health effects of PM
exposure (Daniels et al., 2000; Pope, 2000; Rossi et al., 1999; Schwartz, 2000) that finds in most
cases no evidence of a nonlinear relationship between PM and health effects and certainly does
not find a distinct threshold. The most recent draft of the EPA Air Quality Criteria for
Particulate Matter (EPA, 2004) reports only one study, analyzing data from Phoenix, AZ, that
reported even limited evidence suggestive of a possible threshold for PM25 (Smith et al., 2000).
                                         9-152

-------
                                                                 Cost-Benefit Analysis
Table 9A-22. Asthma Prevalence Rates Used to Estimate Asthmatic Populations in Impact
Functions
Population Group
All Ages
<18
5-17
18-44
45-64
65+
Male, 27+
African- American, 5 to 17
African- American, <18
Asthma Prevalence Rates
Value
0.0386
0.0527
0.0567
0.0371
0.0333
0.0221
0.021
0.0726
0.0735
Source
American Lung Association (2002c, Table 7) — based on 1999 HIS
American Lung Association (2002c, Table 7)— based on 1999 HIS
American Lung Association (2002c, Table 7) — based on 1999 HIS
American Lung Association (2002c, Table 7) — based on 1999 HIS
American Lung Association (2002c, Table 7)— based on 1999 HIS
American Lung Association (2002c, Table 7) — based on 1999 HIS
2000 HIS public use data files8
American Lung Association (2002c, Table 9) — based on 1999 HIS
American Lung Association (2002c, Table 9) — based on 1999 HIS
   Seeftp://ftp.cdc.gov/pub/Health_Statistics/NCHS/Datasets/HIS/2000/
   Recent cohort analyses by HEI (Krewski et al., 2000) and Pope et al. (2002) provide
additional evidence of a quasi-linear relationship between long-term exposures to PM2 5 and
premature mortality.  According to the latest draft PM criteria document, Krewski et al. (2000)
found a "visually near-linear relationship between all-cause and cardiopulmonary mortality
residuals and mean sulfate concentrations, near-linear between cardiopulmonary mortality and
mean PM2 5, but a somewhat nonlinear relationship between all-cause mortality residuals and
mean PM2 5 concentrations that flattens above about 20 |ig/m3. The confidence bands around the
fitted curves are very wide, however, neither requiring a linear relationship nor precluding a
nonlinear relationship if suggested by reanalyses."

   The Pope et al. (2002) analysis, which represented an extension to the Krewski et al.
analysis,  found that the functions relating PM2 5 and premature mortality "were not significantly
different  from linear associations."

   Daniels et al. (2000) examined the presence of thresholds in PM10 C-R relationships for daily
mortality using the largest 20 U.S. cities for 1987-1994. The results of their models suggest that
the linear model was preferred over spline and threshold models. Thus, these results suggest that
linear models without a threshold may well be appropriate for estimating the effects of PM10 on
the types of premature mortality of main interest. Schwartz and Zanobetti (2000) investigated
                                         9-153

-------
Final Regulatory Impact Analysis
the presence of threshold by simulation and actual data analysis of 10 U.S. cities.  In the analysis
of data from 10 cities, the combined C-R curve did not show evidence of a threshold in the PM10-
mortality associations.  Schwartz, Laden, and Zanobetti (2002) investigated thresholds by
combining data on the PM2 5-mortality relationships for six cities and found an essentially linear
relationship down to 2 i-ig/m3, which is at or below anthropogenic background in most areas.
They also examined just traffic-related particles and again found  no evidence of a threshold. The
Smith et al. (2000) study of associations between daily total mortality and PM25 and PM10_25 in
Phoenix, AZ, (during 1995-1997) also investigated the possibility of a threshold using a
piecewise linear model and a cubic spline model. For both the piecewise linear and cubic spline
models, the analysis suggested a threshold of around 20 to 25 |ig/m3.  However, the C-R curve
for PM25 presented in this publication suggests more of a U- or V-shaped relationship than the
usual "hockey stick" threshold relationship.

   Based on the recent literature and  advice from the  SAB, we assume there are no thresholds
for modeling health  effects. Although not included in the primary analysis, the potential impact
of a health effects threshold on avoided incidences of PM-related premature mortality is
explored as a key sensitivity analysis  and is presented  in Appendix 9-B.

   Our assumptions regarding thresholds are supported by the National Research Council in its
recent review of methods for estimating the public health benefits of air pollution regulations. In
their review, the National Research Council concluded that there is no evidence for any
departure from linearity in the observed range of exposure to PM10 or PM25, nor any indication
of a threshold. They cite the weight of evidence available from both short- and long-term
exposure models and the similar effects found in cities with low and high ambient concentrations
ofPM.

9A.3.5.5  Selecting Unit Values for Monetizing Health Endpoints

   The appropriate  economic value of a change in a health effect depends on whether the health
effect is viewed ex ante (before the effect has occurred) or ex post (after the effect has occurred).
Reductions in ambient concentrations of air pollution generally lower the risk of future adverse
health affects by a fairly small amount for a large population. The appropriate economic
measure is therefore ex ante WTP for changes in risk.  However,  epidemiological studies
generally provide estimates of the relative risks of a particular health effect avoided due to a
reduction in air pollution. A convenient way to use this data in a consistent framework is to
convert probabilities to units of avoided statistical incidences.  This measure is calculated by
dividing individual WTP for a risk reduction by the related observed change in risk. For
example, suppose a measure is able to reduce the risk of premature mortality from 2 in 10,000 to
1 in 10,000 (a reduction of 1  in 10,000).  If individual  WTP for this risk reduction is $100, then
the WTP for an avoided statistical premature mortality amounts to $1 million ($100/0.0001

                                          9-154

-------
                                                                 Cost-Benefit Analysis
change in risk). Using this approach, the size of the affected population is automatically taken
into account by the number of incidences predicted by epidemiological studies applied to the
relevant population. The same type of calculation can produce values for statistical incidences
of other health endpoints.

   For some health effects, such as hospital admissions, WTP estimates are generally not
available. In these cases, we use the cost of treating or mitigating the effect as a primary
estimate. For example, for the valuation of hospital admissions we use the  avoided medical costs
as an estimate of the value of avoiding the health effects causing the admission. These COI
estimates generally understate the true value of reductions in risk of a health effect. They tend to
reflect the direct expenditures related to treatment but not the value of avoided pain and suffering
from the health effect. Table 9A-23  summarizes the value estimates per health effect that we
used in this analysis. Values are presented both for a 1990 base income level and adjusted for
income growth in the two future analysis years, 2010 and 2015. Note that the unit values for
hospital admissions are the weighted averages of the ICD-9 code-specific values for the group of
ICD-9 codes included in the hospital admission categories.  A discussion of the valuation
methods for premature mortality and chronic bronchitis is provided here because of the relative
importance of these effects. Discussions of the methods used to value nonfatal myocardial
infarctions (heart attacks) and school absence days are provided because these endpoints have
only recently been added to the analysis and the valuation methods are still under development.
In the following discussions, unit values are presented at 1990 levels of income for consistency
with previous analyses. Equivalent future year values can be obtained from Table 9A-23.
                                          9-155

-------
 o
 a
•a
 =
 o

 =
 S

13
 S
 o
 =
 o
a\
a
c
u
s
c
1
0)
CJ
S2
0)
s
5
istical J
C5
^
^
0)
3

«

0
Ol
C5
E

^
1
=
Ol
J













Ol

o _
u a
5 z
o
Ol





o
g
o"
m



8
a
0
1
m 5
\^/ "—j -4-J
ra crt
•5 ^ -
TH S-H TH
_3 <& 3
fi O 3
5 O O
I f a
P3  m
*O "^ OO

\ ^O ^\~
o t^ co
*o "^r oo
^^ ^^ ^^








^r ^r
•^ ^ in
o 
-------
•a
 a>
 S
 =
 o
a\
a\
 o
 a
•a
 =
 o

 =
 S

13
 S
 o
 =
 o
 w
UJ
a\












S
1
*•*
Derivation of E





a
tj
a
O)
3
CJ
"3
tj
03
.
PH

B
"3
>•
'o
"03
a

w
S
^
0
U
o>
E
0 3
o
N
o>
i
8-3

M 33
S
o


o
E
o "3
fi £
ON
C"\
i-H




x "a
s '3
Q) ^3
M a
UJ




















































issioni
1
1
'&
1
ON j-
Q 5
a ° °
O *"M £^
T3 Oj) g ^

S _u ^ >
0  M rf ^
^^j^.S ^ ^ t^ rp
III! 111!

i. S
5 ^




s ^
^O^ V_Q
^ \o'
a? ^^




ro
^N ^J
^ >0
2 !g
ae ^








| .I
oo .M
.3 S -1
CH ""O ™
is a
3 Q J3
^ 1
C^ ^
Q ^ S)
O (H— i t/3 ^
i— i O 1) H
O *^ §j .
-T-i &0 ri S
13 ri 3 &
« ^ ^ €

§ |tl
^ cd o
CW *~ ^H >^

•a 8 | 3
^3 H > ,^
W j O ^
H ^^ T3 ra
,tes (lost earnings plus direct
mation (e.g., average hospita
id weighted share of total car
ncy for Healthcare Research

. Jj <§ ^ ^
M ^ ri
^ ""55 w -^
O1""1 > i— i T3
1) rt 0)
111!

00
oo"




00
f^
00
a?




^
00
oo"
a^








||
GO ^
O 1)
11

^0















: of two unit COI values:
m Smith etal., 1997, and
m Stanford et al., 1999.
^ o o
oo. S ,K
^ f+^ f+^
g in" l>"
^ i/~l *O
i-; o

VO
00
S




vo
00
a4






VO
00




a
1
^

-------
•a
 v
 s
 =
 o
a\
a\
a\
 o
 a
•a
 a
 o

 a
 s
13
 S
 o
 a
 o
 w
UJ
a\

0)
CJ
=

"S
'3
fi
"3
CJ
03
L.
PH
01
"3
>•
'o
fll
i
03
a

w
£
1
u














Derivation of Estimates II
O)
S
O , 	 i
cj 13

0 hJ
o
N
O)
s
8-3
= >
S
o


o
S
S|
3
o\
ON






.a ."s
03 &
O* "O
M a
UJ


























q
"S
1
o
ffi
&0

&
"o


'^
E?
,2

CD
Pi
Combinations of the 3 symptoms for which WTP estimates are available that
closely match those listed by Pope, et al. result in 7 different "symptom
clusters," each describing a "type" of URS. A dollar value was derived for each
type of URS, using mid-range estimates of WTP (lEc, 1994) to avoid each
symptom in the cluster and assuming additivity of WTPs. The dollar value for
URS is the average of the dollar values for the 7 different types of URS.



t-~

ae





£4






•n
£4



f 	 ^
GO
g
W
I
1
GO
&>
O
2
8*
&
ft
ft

Combinations of the 4 symptoms for which WTP estimates are available that
closely match those listed by Schwartz, et al. result in 1 1 different "symptom
clusters," each describing a "type" of LRS. A dollar value was derived for each
type of LRS, using mid-range estimates of WTP (lEc, 1994) to avoid each
symptom in the cluster and assuming additivity of WTPs. The dollar value for
LRS is the average of the dollar values for the 1 1 different types of LRS.



t-~

ae





~






^o
£,



f 	 ^
GO
Pi
1
.2
ft
GO
E?
o
2
'ft
u
Pi
S3
o

Asthma exacerbations are valued at $42 per incidence, based on the mean of
average WTP estimates for the four severity definitions of a "bad asthma day,"
described in Rowe and Chestnut (1986). This study surveyed asthmatics to
estimate WTP for avoidance of a "bad asthma day," as defined by the subjects.
For purposes of valuation, an asthma attack is assumed to be equivalent to a day
in which asthma is moderate or worse as reported in the Rowe and Chestnut
(1986) study.



in

ae





3







-------
•a
 a>
 S
 =
 o
a\
a\
a\
 o
 a
•a
 a
 o

 a
 s
13
 S
 o
 a
 o
 w
UJ
a\




tj
S

3
jg
«M
C3
'1
"S
55
1

"3
>
*^
o
03
s
•c
w
2
^^
=
u






















Derivation of Estimates
i
o _i
(J ^
OJ3
g
™s

4>
S|

O hJ
®


O)
s
8^3
= >
ON
ON








js . S
S o
C3 &
fli ^3
W a
w






























|>
Q
0
J
"o
0
o
1
o
1
c3

'.£
Q

^
^
•c
w
&


<•
County-specific median annual wages divided by 50 (assuming 1 weeks of
vacation) and then by 5 - to get median daily wage. U.S. Year 2000 Census
compiled by Geolytics, Inc.


1
§
**



JH
.2
I





^
.2
>









Q
t^
^
>-,

Q
M
o
>^
0

(U • ^
•^ ^ ^ JB a *

Based on expected lost wages from parent staying home with child. Estima
daily lost wage (if a mother must stay at home with a sick child) is based 01
median weekly wage among women age 25 and older in 2000 (U.S. Census
Bureau, Statistical Abstract of the United States: 2001, Section 12: Labor
Force, Employment, and Earnings, Table No. 621). This median wage is $
Dividing by 5 gives an estimated median daily wage of $103.
The expected loss in wages due to a day of school absence in which the mo
would have to stay home with her child is estimated as the probability that
mother is in the workforce times the daily wage she would lose if she misse
day = 72.85% of $103, or $75.


in
£





[C
V)






s











"**-.
Q
o


o
<
"o
^
o
GO

-,

(^ J^ flj i-H
g ^ g1 ft
o §
^ o









>^
>
o


o
53

0




Median WTP estimate to avoid one MRAD from Tolley, et al. (1986) .


in
£





iri
v>






in








&
Q
•f
tj

1
0

w
u ^
Pi Q
o ^
rt PH
11

-------
Final Regulatory Impact Analysis
    9A. 3.5.5.1 Valuing Reductions in Premature Mortality Risk.

    We estimate the monetary benefit of reducing premature mortality risk using the "value of
statistical lives saved" (VSL) approach, which is a summary measure for the value of small
changes in premature mortality risk experienced by a large number of people. The VSL
approach applies information from several published value-of-life studies to determine a
reasonable benefit of preventing premature mortality.  The mean value of avoiding one statistical
death is assumed to be $5.5 million in 1999 dollars.  This represents  a central value consistent
with the range of values suggested by recent meta-analyses of the wage-risk VSL literature.  The
distribution of VSL is characterized by a confidence interval from $1 to $10 million, based on
two meta-analyses of the wage-risk VSL literature. The $1 million lower confidence limit
represents the lower end of the interquartile range from the Mrozek and Taylor (2000)
meta-analysis. The $10 million upper confidence limit represents the upper end of the
interquartile range from the Viscusi and Aldy (2003) meta-analysis.

    In previous analyses, we used an estimate of mean VSL equal to  $6.3 million, based on a
distribution fitted to the estimates from 26 value-of-life studies identified in the Section 812
reports as "applicable to policy analysis." cc

    As indicated in the previous section on quantification of premature mortality benefits, we
assume for this analysis that some of the incidences of premature mortality related to PM
exposures occur in a distributed fashion over the  5 years following exposure. To take this into
account in the valuation of reductions in premature mortality, we apply 3 percent and 7 percent
discount rates to the value of premature mortality occurring in future years.dd
    cc Commentors have suggested that the VSL used in the Draft RIA may not be appropriate for populations
impacted by the rule in that it may not reflect the risk preference of the of the target population. We recognize the
large amount of uncertainty in the VSL for application to environmental policy. Following SAB-EEAC guidance,
we used a wage-risk-based VSL in valuing premature mortality for the primary estimate in the final rule. In
response to concerns about the range of estimates included in the VSL distribution, we modified the value of life
distribution used for the final rule. As described above, the new mean value of avoiding one statistical death ($5.5
million in 1999 dollars) represents a central value consistent with the range of values suggested by recent meta-
analyses of the wage-risk VSL literature.  The distribution of VSL used in this  RIA is characterized by a confidence
interval from $1 to $10 million, based on two meta-analyses of the wage-risk VSL literature.  Following SAB-EEAC
guidance, we discount over the lag period between exposure and premature mortality in valuing reductions in
mortality incidence (see Section 9.A.3.5.2).

    DDThe choice of a discount rate, and its associated conceptual basis, is a topic of ongoing discussion within the
federal government. The EPA adopted a 3 percent discount rate for its base estimate in this case to reflect reliance
on a "social rate of time preference" discounting concept. We  have also calculated benefits and costs using a 7
percent rate consistent with an "opportunity cost of capital" concept to reflect the time value of resources directed to
meet regulatory requirements. In this case, the benefit and cost estimates were not significantly affected by the
choice of discount rate. Further discussion of this topic appears in the EPA's Guidelines for Preparing Economic
Analyses (EPA 2000c).

                                             9-160

-------
                                                                Cost-Benefit Analysis
   The economics literature concerning the appropriate method for valuing reductions in
premature mortality risk is still developing. The adoption of a value for the projected reduction
in the risk of premature mortality is the subject of continuing discussion within the economics
and public policy analysis community.  Regardless of the theoretical economic considerations,
the EPA prefers not to draw distinctions in the monetary value assigned to the lives saved even if
they differ in age, health status, socioeconomic status, gender, or other characteristic of the adult
population.

   Following the advice of the EEAC of the SAB, the EPA currently uses the VSL approach in
calculating the primary estimate  of premature mortality benefits, because we believe this
calculation provides the most reasonable single estimate of an individual's willingness to trade
off money for reductions in premature mortality risk (EPA-SAB-EEAC-00-013).  Although there
are several differences between the labor market studies the EPA uses to derive a VSL estimate
and the PM air pollution context addressed here, those differences in the affected populations
and the nature of the risks imply both upward and downward adjustments.  Table 9A-24 lists
some of these differences and the expected effect on the VSL estimate for air pollution-related
premature mortality.  In the absence of a comprehensive and balanced set of adjustment factors,
the EPA believes it is reasonable to continue to use the $5.5 million value while acknowledging
the significant limitations and uncertainties in the available literature.

Table 9A-24. Expected Impact on Estimated Benefits of Premature Mortality Reductions
of Differences Between Factors Used in Developing Applied VSL and Theoretically
Appropriate VSL
Attribute
Age
Life expectancy/health status
Attitudes toward risk
Income
Voluntary vs. Involuntary
Catastrophic vs. protracted death
Expected Direction of Bias
Uncertain, perhaps overestimate
Uncertain, perhaps overestimate
Underestimate
Uncertain
Uncertain, perhaps underestimate
Uncertain, perhaps underestimate
    Some economists emphasize that the VSL is not a single number relevant for all situations.
Indeed, the VSL estimate of $5.5 million (1999 dollars) is itself the central tendency of a number
of estimates of the VSL for some rather narrowly defined populations. When there are
significant differences between the population affected by a particular health risk and the
                                         9-161

-------
Final Regulatory Impact Analysis
populations used in the labor market studies, as is the case here, some economists prefer to adjust
the VSL estimate to reflect those differences.

   The SAB-EEAC has advised that the EPA "continue to use a wage-risk-based VSL as its
primary estimate, including appropriate sensitivity analyses to reflect the uncertainty of these
estimates," and that "the only risk characteristic for which adjustments to the VSL can be made
is the timing of the risk" (EPA-SAB-EEAC-00-013, EPA, 2000b).  In developing our primary
estimate of the benefits of premature mortality reductions, we have followed this advice and
discounted over the lag period between exposure and premature mortality.

   Uncertainties Specific to Premature Mortality Valuation.  The economic benefits associated
with premature mortality are the largest category of monetized benefits of this rule. In addition,
in prior analyses, the EPA has identified valuation of premature mortality benefits as the  largest
contributor to the range of uncertainty in monetized benefits (see EPA [1999]). Because  of the
uncertainty in estimates of the value of premature mortality avoidance, it is important to
adequately characterize and understand the various types of economic approaches  available for
premature mortality valuation. Such an assessment also requires an understanding of how
alternative valuation approaches reflect that some individuals may be more susceptible to air
pollution-induced premature mortality or reflect differences in the nature of the risk presented by
air pollution relative to the risks studied in the relevant economics literature.

   The health science literature on air pollution indicates that several human characteristics
affect the degree to which mortality risk affects an individual. For example, some age groups
appear to be more susceptible to air pollution than others (e.g., the elderly and children).  Health
status prior to exposure also affects susceptibility.  An ideal benefits estimate of mortality risk
reduction would reflect these human characteristics, in addition to an individual's WTP to
improve one's own chances of survival plus WTP to  improve other individuals'  survival  rates.
The ideal measure would also take into account the specific nature of the risk reduction
commodity that is provided to individuals, as well as the context in which risk is reduced. To
measure this value, it is important to assess how reductions in air pollution reduce  the risk of
dying from the time that reductions take effect onward, and how individuals value these changes.
Each individual's survival curve, or the probability of surviving beyond a given  age, should  shift
as a result of an environmental quality improvement. For example, changing the current
probability of survival for an individual also shifts future probabilities of that individual's
survival. This probability shift will differ across individuals because survival curves depend on
such characteristics as age, health state, and the current age to which the individual is likely to
survive.

   Although a survival curve approach provides a theoretically preferred method for valuing the
benefits of reduced risk of premature mortality associated with reducing air pollution, the

                                          9-162

-------
                                                                  Cost-Benefit Analysis
approach requires a great deal of data to implement. The economic valuation literature does not
yet include good estimates of the value of this risk reduction commodity. As a result, in this
study we value avoided premature mortality risk using the VSL approach.

   Other uncertainties specific to premature mortality valuation include the following:

       •  Across-study variation:  There is considerable uncertainty as to whether the available
          literature on VSL provides adequate estimates of the VSL saved by air pollution
          reduction.  Although there is considerable variation in the analytical designs and data
          used in the existing literature,  the majority of the studies involve the value of risks to
          a middle-aged working population.  Most of the studies examine differences in wages
          of risky occupations, using a wage-hedonic approach. Certain characteristics of both
          the population affected and the mortality risk facing that population are believed to
          affect the average WTP to reduce the risk.  The appropriateness of a distribution of
          WTP based on the current VSL literature for valuing the premature mortality-related
          benefits of reductions in air pollution concentrations therefore depends not only on
          the quality of the studies (i.e.,  how well  they measure what they are trying to
          measure), but also on the extent to which the risks being valued are similar and the
          extent to which the subjects in the studies are similar to the population affected by
          changes in pollution concentrations.

       •  Level of risk reduction:  The transferability of estimates of the VSL from the wage-
          risk studies to the context of the this rulemaking analysis rests on the assumption that,
          within a reasonable range, WTP for reductions in mortality risk is linear in risk
          reduction.  For example, suppose a study estimates that the average WTP for a
          reduction in mortality risk of 1/100,000  is $50, but that the actual mortality risk
          reduction resulting from a given pollutant reduction is 1/10,000.  If WTP for
          reductions in mortality risk is linear in risk reduction, then a WTP of $50 for a
          reduction of 1/100,000 implies a WTP of $500 for a risk reduction of 1/10,000
          (which is 10 times the risk reduction valued in the study). Under the assumption of
          linearity, the estimate of the VSL does not depend on the particular amount of risk
          reduction being valued.  This assumption has been  shown to be reasonable provided
          the change in the risk being valued is within the range of risks evaluated in the
          underlying studies (Rowlatt et al., 1998).

       •  Voluntariness of risks evaluated:  Although job-related mortality risks may differ in
          several ways from air pollution-related mortality risks, the most important difference
          may be that job-related risks are incurred voluntarily, or generally assumed to be,
          whereas  air pollution-related risks are incurred involuntarily.  Some evidence
          suggests that people will pay more to reduce involuntarily incurred risks than risks
          incurred voluntarily.  If this is the case, WTP estimates based on wage-risk studies
          may understate WTP to reduce involuntarily incurred air pollution-related mortality
          risks.
                                          9-163

-------
Final Regulatory Impact Analysis
       •  Sudden versus protracted death:  A final important difference related to the nature of
          the risk may be that some workplace mortality risks tend to involve sudden,
          catastrophic events, whereas air pollution-related risks tend to involve longer periods
          of disease and suffering prior to death.  Some evidence suggests that WTP to avoid a
          risk of a protracted death involving prolonged suffering and loss of dignity and
          personal control is greater than the WTP to avoid a risk (of identical magnitude) of
          sudden death. To the extent that the mortality risks addressed in this assessment are
          associated with longer periods of illness or greater pain and suffering than are the
          risks addressed in the valuation literature, the WTP measurements employed in the
          present analysis would reflect a downward bias.

       •  Self-selection and skill in avoiding risk. Recent research (Shogren et al., 2002)
          suggests that VSL estimates based on hedonic wage studies may overstate the
          average value of a risk reduction. This is based on the fact that the risk-wage tradeoff
          revealed in hedonic studies reflects the preferences of the marginal worker (i.e., that
          worker who demands the highest compensation for his risk reduction).  This worker
          must have either higher risk, lower risk tolerance,  or both.  However, the risk estimate
          used in hedonic studies is generally based on average risk, so the VSL may be
          upwardly biased because the wage differential and risk measures do not match.

   For more discussion, see Appendix 9B.

   9A. 3.5.5.2 Valuing Reductions in the Risk of Chronic Bronchitis.

   The best available estimate of WTP to avoid a case of chronic bronchitis comes from Viscusi
et al. (1991).  The Viscusi et al. study, however, describes a severe case of chronic bronchitis to
the survey respondents.  We therefore employ an estimate of WTP to avoid a pollution-related
case of chronic bronchitis, based on adjusting the Viscusi et al. (1991) estimate of the WTP to
avoid a severe case. This is done to account for the likelihood that an average case of pollution-
related chronic bronchitis is not as severe. The adjustment is  made by applying the elasticity of
WTP with respect to severity reported in the Krupnick and Cropper (1992) study.  Details of this
adjustment procedure are provided in the benefits TSD for the nonroad diesel rulemaking (Abt
Associates, 2003).

   We use the mean of a distribution of WTP estimates as the central tendency estimate of WTP
to avoid a pollution-related case of chronic bronchitis in this analysis.  The distribution
incorporates uncertainly from three sources:  the WTP to avoid a case of severe chronic
bronchitis, as described by Viscusi et al.; the severity level of an average pollution-related case
of chronic bronchitis (relative to that of the case described by Viscusi et al.); and the elasticity
of WTP with respect to severity of the illness. Based on assumptions about the distributions of
each of these three uncertain components, we derive a distribution of WTP to avoid a pollution-
related case of chronic bronchitis by statistical uncertainty analysis techniques. The expected
                                          9-164

-------
                                                                 Cost-Benefit Analysis
value (i.e., mean) of this distribution, which is about $331,000 (2000$), is taken as the central
tendency estimate of WTP to avoid a PM-related case of chronic bronchitis.

   9A3.5.5.3 Valuing Reductions in Non-Fatal Myocardial Infarctions (Heart Attacks).

The Agency has recently incorporated into its analyses the impact of air pollution on the
expected number of nonfatal heart attacks, although it has examined the impact of reductions in
other related cardiovascular endpoints. We were not  able to identify a suitable WTP value for
reductions in the risk of nonfatal heart attacks. Instead, we propose a COI unit value with two
components:  the direct medical costs and the opportunity cost (lost earnings) associated with the
illness event. Because the costs associated with an myocardial infarction extend beyond the
initial event itself, we consider costs incurred over several years.  Using age-specific annual lost
earnings estimated by Cropper and Krupnick (1990) and a 3 percent discount rate, we estimated
a present discounted value in lost earnings (in 2000$) over 5 years due to an  myocardial
infarction of $8,774 for someone between the ages of 25 and 44, $12,932 for someone between
the ages of 45 and 54, and $74,746 for someone between the ages of 55 and 65. The
corresponding age-specific estimates of lost earnings  (in 2000$) using a 7 percent discount rate
are $7,855, $11,578, and $66,920, respectively.  Cropper and Krupnick (1990) do not provide
lost earnings estimates for populations under 25 or over 65.  As such, we  do  not include lost
earnings in the cost estimates for these age groups.

   We found three possible sources in the literature of estimates of the direct medical  costs  of
myocardial infarction:

       •  Wittels et al.  (1990) estimated expected total medical costs of myocardial infarction
          over 5 years to be $51,211 (in 1986$) for people who were admitted to the hospital
          and survived  hospitalization. (There does not appear to be any discounting used.)
          Wittels et al.  was used to value coronary heart disease in the 812 Retrospective
          Analysis of the Clean Air Act.  Using the  CPI-U for medical care, the Wittels
          estimate is $109,474 in year 2000$. This  estimated cost is based  on a medical  cost
          model, which incorporated therapeutic options, projected outcomes, and prices (using
          "knowledgeable cardiologists" as consultants). The model used medical data and
          medical decision algorithms to estimate the probabilities of certain events and/or
          medical procedures being used.  The authors note that the average length of
          hospitalization for acute myocardial infarction has decreased over time (from an
          average of 12.9 days in 1980 to an average of 11 days  in 1983). Wittels et al. used  10
          days as the average in their study.  It is unclear how much further the length of stay
          for myocardial infarction may have decreased from 1983 to the present.  The average
          length of stay for ICD code 410 (myocardial infarction) in the year-2000 AHQR
          HCUP database is 5.5  days. However, this may include patients who died in the
          hospital (not included  among our nonfatal myocardial  infarction cases), whose length
          of stay was therefore substantially shorter  than it would be if they had not died.

                                         9-165

-------
Final Regulatory Impact Analysis
       •  Eisenstein et al. (2001) estimated 10-year costs of $44,663 in 1997$, or $49,651 in
          2000$ for myocardial infarction patients, using statistical prediction (regression)
          models to estimate inpatient costs.  Only inpatient costs (physician fees and hospital
          costs) were included.

       •  Russell et al. (1998) estimated first-year direct medical costs of treating nonfatal
          myocardial infarction of $15,540 (in 1995$) and $1,051 annually thereafter.
          Converting to year 2000$, that would be $23,353 for a 5-year period (without
          discounting) or $29,568 for a 10-year period.

   In summary, the three different studies provided significantly different values (see Table 9A-
25).

   As noted above, the estimates from these three studies are substantially different, and we
have not adequately resolved the sources of differences in the estimates.  Because the wage-
related opportunity cost estimates from Cropper and Krupnick (1990) cover a 5-year period, we
use estimates for medical costs that similarly cover a 5-year period (i.e., estimates from Wittels
et al. (1990) and Russell et al. (1998). We use a simple average of the two 5-year estimates, or
$65,902, and  add it to the 5-year opportunity cost estimate.  The resulting estimates are given in
Table 9A-26.
                                          9-166

-------
                                                                  Cost-Benefit Analysis
Table 9A-25.  Alternative Direct Medical Cost of Illness Estimates for Nonfatal Heart
Attacks
Study
Wittelsetal. (1990)
Russell etal. (1998)
Eisenstein et al. (2001)
Russell etal. (1998)
Direct Medical Costs (2000$)
$109,474a
$22,33 lb
$49,65 lb
$27,242b
Over an x-Year Period, for x =
5
5
10
10
a   Wittels et al. did not appear to discount costs incurred in future years.
b   Using a 3 percent discount rate.
Table 9A-26.  Estimated Costs Over a 5-Year Period (in 2000$) of a Nonfatal Myocardial
Infarction
Age Group
0-24
25-44
45-54
55-65
>65
Opportunity Cost
$0
$8,774b
$12,253b
$70,619b
$0
Medical Cost3
$65,902
$65,902
$65,902
$65,902
$65,902
Total Cost
$65,902
$74,676
$78,834
$140,649
$65,902
a   An average of the 5-year costs estimated by Wittels et al., 1990, and Russell et al., 1998.
b   From Cropper and Krupnick, 1990, using a 3 percent discount rate.
   9A. 3.5.5.4 Valuing Reductions in School Absence Days.

   School absences associated with exposure to ozone are likely to be due to respiratory-related
symptoms and illnesses. Because the respiratory symptom and illness endpoints we are
including are all PM-related rather than ozone-related, we do not have to be concerned about
double counting of benefits if we aggregate the benefits of avoiding ozone-related school
absences with the benefits of avoiding PM-related respiratory symptoms and illnesses.

   One possible approach to valuing a school absence is using a parental opportunity cost
approach.  This method requires two steps: estimate the probability that, if a school child stays
home from school, a parent will have to stay home from work to care for the child, and  value the
                                          9-167

-------
Final Regulatory Impact Analysis
lost productivity at the person's wage. Using this method, we would estimate the proportion of
families with school-age children in which both parents work, and value a school loss day as the
probability of a work loss day resulting from a school loss day (i.e., the proportion of households
with school-age children in which both parents work) times some measure of lost wages
(whatever measure we use to value work loss days).  There are three significant problems with
this method, however. First, it omits WTP to avoid the symptoms/illness that resulted in the
school absence. Second, it effectively gives zero value to school absences which do not result in
a work loss day (unless we derive an alternative estimate of the value of the parent's time for
those cases in which the parent is not in the labor force).  Third, it makes an assumption about
the gender of the parent that would miss work.  We are investigating approaches using WTP for
avoid the symptoms/illnesses causing the absence. In the interim, we will use the parental
opportunity cost approach.

   For the parental opportunity cost approach, we make  an explicit, lower assumption that in
married  households with two working parents, the female parent will stay home with a sick
child. From the U.S.  Census Bureau, Statistical Abstract of the United States:  2001, we
obtained (1) the numbers of single, married, and "other" (i.e., widowed, divorced, or  separated)
women with children in the workforce, and (2) the rates of participation in the workforce of
single, married, and "other" women with children. From these two sets of statistics, we inferred
the numbers of single, married, and "other" women with  children, and the corresponding
percentages. These percentages were used to calculate a  weighted average participation rate, as
shown in Table 9A-27. We do not take into account that many single and "other" women with
children may lose their jobs if they are repeatedly absent  due to their children's illnesses.

   Our  estimated daily lost wage (if a mother must stay at home with a sick child) is based on
the median weekly wage among women age 25 and older in 2000 (U.S. Census Bureau,
Statistical Abstract of the United States: 2001, Section 12: Labor Force, Employment, and
Earnings, Table No. 621). This median wage is $551. Dividing by 5 gives an estimated median
daily wage of $103.
                                         9-168

-------
                                                                 Cost-Benefit Analysis
Table 9A-27.  Women with Children: Number and Percent in the Labor Force, 2000, and
Weighted Average Participation Ratea






Single
Married
Other"
Total:
Number (in
millions) in
Labor Force


(1)
3.1
18.2
4.5

Participation
Rate



(2)
73.9%
70.6%
82.7%

Implied Total
Number in
Population (in
millions)

(3) = (l)/(2)
4.19
25.78
5.44
35.42
Implied
Percent in
Population


(4)
11.84%
72.79%
15.36%


Weighted
Average
Participation
Rate [=sum
(2)*(4) over
rows]




72.85%
a   Data in columns (1) and (2) are from U.S. Census Bureau, Statistical Abstract of the United States: 2001,
   Section 12: Labor Force, Employment, and Earnings, Table No. 577.
b   Widowed, divorced, or separated.
    The expected loss in wages due to a day of school absence in which the mother would have
to stay home with her child is estimated as the probability that the mother is in the workforce
times the daily wage she would lose if she missed a day = 72.85% of $103, or $75.ee

    9A.3.5.6 Unqualified Health Effects

    In addition to the health effects discussed above, there is emerging evidence that human
exposure to ozone may be associated with premature mortality (Ito and Thurston, 1996; Samet,
et al.  1997, Ito and Thurston, 2001), PM and ozone with increased emergency room visits for
non-asthma respiratory causes (US EPA, 1996a; 1996b), ozone with impaired airway
responsiveness (US EPA, 1996a), ozone with increased susceptibility to respiratory infection
(US EPA, 1996a), ozone with acute inflammation and respiratory cell damage (US EPA, 1996a),
ozone and PM with premature aging of the lungs and chronic respiratory damage (US EPA,
1996a; 1996b), ozone with onset of asthma in exercising children (McConnell et al. 2002), and
PM with reduced heart rate variability and other changes in cardiac function. An improvement
in ambient PM and ozone air quality may  reduce the number of incidences within each effect
category that the U.S. population would experience.  Although these health effects are believed
   EEIn a very recent article, Hall, Brajer, and Lurmann (2003) use a similar methodology to derive a mid-estimate
value per school absence day for California of between $70 and $81, depending on differences in incomes between
three counties in California. Our national average estimate of $75 per absence is consistent with these published
values.
                                          9-169

-------
Final Regulatory Impact Analysis
to be PM or ozone-induced, effect estimates are not available for quantifying the benefits
associated with reducing these effects. The inability to quantify these effects lends a downward
bias to the monetized benefits presented in this analysis.

9A.3.6 Human Welfare Impact Assessment

    PM and ozone have numerous documented effects on environmental quality that affect
human welfare. These welfare effects include direct damages to property, either through impacts
on material structures  or by soiling of surfaces, direct economic damages in the form of lost
productivity of crops and trees, indirect damages through alteration of ecosystem functions, and
indirect economic damages through the loss in value of recreational experiences or the existence
value of important resources.  EPA's Criteria Documents for PM and ozone list numerous
physical and ecological effects known to be linked to ambient concentrations of these pollutants
(US EPA,  1996a; 1996b). This section describes individual effects and how we quantify and
monetize them. These effects include changes in commercial crop and forest yields, visibility,
and nitrogen deposition to estuaries.

    9A.3.6.1 Visibility Benefits

    Changes in the level of ambient particulate matter caused by the reduction in emissions from
the preliminary control options will change the level of visibility in much of the U.S.  Visibility
directly affects people's enjoyment of a variety of daily activities. Individuals value visibility
both in the  places they live and work, in the places they travel to for recreational purposes, and at
sites of unique public value, such as  the Grand Canyon. This section discusses the measurement
of the economic benefits of visibility.

    It is difficult to quantitatively define a visibility endpoint that can be used for valuation.
Increases in PM concentrations cause increases in light extinction. Light extinction is a measure
of how much the components of the  atmosphere absorb light. More light absorption means that
the clarity of visual images and visual range is reduced, ceteris paribus. Light absorption is a
variable that can be accurately measured.  Sisler (1996) created a unitless measure of visibility
based directly on the degree of measured light absorption called the deciview.  Deciviews are
standardized for a reference distance in such a way that one deciview corresponds to a change of
about 10 percent in available light. Sisler characterized a change in light extinction of one
deciview as "a small but perceptible scenic change under many circumstances."  Air quality
                                          9-170

-------
                                                                     Cost-Benefit Analysis
models were used to predict the change in visibility, measured in deciviews, of the areas affected
by the preliminary control options.ff

    EPA considers benefits from two categories of visibility changes: residential visibility and
recreational visibility. In both cases economic benefits are believed to consist of both use values
and non-use values. Use values include the aesthetic benefits of better visibility, improved road
and air safety, and enhanced recreation in activities like hunting and birdwatching. Non-use
values are based on people's beliefs that the environment ought to exist free of human-induced
haze. Non-use values may be a more important component of value for recreational areas,
particularly national parks and monuments.

    Residential visibility benefits are those that occur from visibility changes in urban, suburban,
and rural areas,  and also in recreational areas not listed as federal Class I areas.88 For the
purposes of this analysis, recreational visibility improvements are defined as those that occur
specifically in federal Class I areas. A key distinction between recreational and residential
benefits is that only those people living in residential areas are assumed to receive benefits from
residential visibility, while all households in the U.S. are assumed to derive some benefit from
improvements in Class I areas. Values are assumed to be higher if the Class I area is located
close to their home.1111

    Only two existing studies provide defensible monetary estimates of the value of visibility
changes. One is a study on residential visibility conducted in 1990 (McClelland, et. al., 1993)
and the other is  a 1988 survey on recreational visibility value (Chestnut and Rowe, 1990a;
1990b). Both utilize the contingent valuation method. There has been a great deal of
controversy and significant development of both theoretical and empirical knowledge about how
to conduct CV surveys in the past decade. In EPA's judgment, the Chestnut and Rowe study
contains many of the elements of a valid CV study and is sufficiently reliable to serve as the
    FF A change of less than 10 percent in the light extinction budget represents a measurable improvement in
visibility, but may not be perceptible to the eye in many cases.  Some of the average regional changes invisibility
are less than one deciview (i.e. less than 10 percent of the light extinction budget), and thus less than perceptible.
However, this does not mean that these changes are not real or significant. Our assumption is then that individuals
can place values on changes in visibility that may not be perceptible.  This is quite plausible if individuals are aware
that many regulations lead to small improvements invisibility which when considered together amount to
perceptible changes invisibility.

    GG The Clean Air Act designates 156 national parks and wilderness areas as Class I areas for visibility
protection.

    HH For details of the visibility estimates discussed in this chapter, please refer to the benefits technical support
document for this RIA (Abt Associates 2003).

                                            9-171

-------
Final Regulatory Impact Analysis
basis for monetary estimates of the benefits of visibility changes in recreational areas." This
study serves as an essential input to our estimates of the benefits of recreational visibility
improvements in the primary benefits estimates.  Consistent with SAB advice, EPA has
designated the McClelland, et al. study as significantly less reliable for regulatory benefit-cost
analysis, although it does provide useful estimates on the order of magnitude of residential
visibility benefits (EPA-SAB-COlMCIL-ADV-00-002,  1999). Residential visibility benefits are
therefore only included as a sensitivity  estimate in Appendix 9-B.

    The Chestnut and Rowe study measured the demand for visibility in Class I areas managed
by the National Park Service (NFS) in three broad regions  of the country: California, the
Southwest, and the  Southeast.  Respondents in five states were asked about their willingness to
pay to protect national  parks or NPS-managed wilderness areas within a particular region.  The
survey used photographs reflecting different visibility levels in the  specified recreational areas.
The visibility levels in  these photographs were later converted to deciviews for the current
analysis. The survey data collected were used to estimate a WTP equation for improved
visibility.  In addition to the visibility change variable, the  estimating equation also included
household income as an explanatory variable.

    The Chestnut and Rowe study did not measure values for visibility improvement in Class I
areas outside the three  regions.  Their study covered 86 of  the 156 Class I areas in the U.S. We
can infer the value of visibility changes in the other Class I areas by transferring values of
visibility changes at Class I areas in the study regions. However, these values are not as
defensible and are thus presented only as an alternative calculation in Table 9A-25.  A complete
description of the benefits transfer method used to infer values for visibility changes in Class I
areas outside the study regions is provided in the benefits TSD for this RIA (Abt Associates,
2003).

    The estimated relationship from the Chestnut and Rowe study is only  directly applicable to
the populations represented by survey respondents. EPA used benefits transfer methodology to
extrapolate these results to the population affected by the Nonroad Diesel Engines rule.   A
general willingness to pay equation for improved visibility (measured in deciviews) was
developed as a function of the baseline level of visibility, the magnitude of the visibility
improvement, and household income. The behavioral parameters of this equation were taken
from analysis of the Chestnut and Rowe data. These parameters were used to calibrate WTP for
the visibility changes resulting from the Nonroad Diesel  Engines rule. The method for
   11 An SAB advisory letter indicates that"many members of the Council believe that the Chestnut and Rowe
study is the best available." (EPA-SAB-COUNCIL-ADV-00-002, 1999) However, the committee did not formally
approve use of these estimates because of concerns about the peer-reviewed status of the study. EPA believes the
study has received adequate review and has been cited in numerous peer-reviewed publications (Chestnut and
Dennis, 1997).

                                          9-172

-------
                                                                 Cost-Benefit Analysis
developing calibrated WTP functions is based on the approach developed by Smith, et al. (2002).
Available evidence indicates that households are willing to pay more for a given visibility
improvement as their income increases (Chestnut, 1997). The benefits estimates here
incorporate Chestnut's estimate that a 1 percent increase in income is associated with a 0.9
percent increase in WTP for a given change in visibility.

   Using the methodology outlined above, EPA estimates that the total WTP for the visibility
improvements in California, Southwestern, and Southeastern Class I areas brought about by the
Nonroad Diesel Engines rule is $2.2 billion. This value includes the value to households living
in the same state as the Class I area as well as values for all households in the U.S. living outside
the state containing the Class I area, and the value accounts for growth in real income. We
examine the impact of expanding the visibility benefits analysis  to other areas of the country in a
sensitivity analysis presented in Appendix 9-B.

   One major source of uncertainty for the visibility benefit estimate is the benefits transfer
process used. Judgments used to choose the functional form and key parameters of the
estimating equation for willingness to pay for the affected population could have significant
effects on the size of the estimates. Assumptions about how individuals respond to changes in
visibility that are either very small, or outside the range covered in the  Chestnut and Rowe study,
could also affect the results.

   9A.3.6.2 Agricultural, Forestry and other Vegetation Related Benefits

   The Ozone Criteria Document notes that "ozone affects vegetation  throughout the United
States, impairing crops, native vegetation, and ecosystems more than any other air pollutant"
(US  EPA, 1996). Changes in ground level ozone resulting from the preliminary control options
are expected to impact crop and forest yields throughout the affected area.

   Well-developed techniques exist to provide monetary estimates of these benefits to
agricultural producers and to consumers. These techniques use models  of planting decisions,
yield response functions, and agricultural products supply and demand. The resulting welfare
measures are based on predicted changes in market prices and production costs. Models also
exist to measure benefits to silvicultural producers and consumers. However, these models have
not been adapted for use in analyzing ozone related forest impacts. As such, our analysis
provides monetized estimates of agricultural benefits, and a discussion of the impact of ozone
changes on forest productivity, but does not monetize commercial forest related benefits.
                                          9-173

-------
Final Regulatory Impact Analysis
    9A. 3.6.2.1 Agricultural Benefits

    Laboratory and field experiments have shown reductions in yields for agronomic crops
exposed to ozone, including vegetables (e.g., lettuce) and field crops (e.g., cotton and wheat).
The most extensive field experiments, conducted under the National Crop Loss Assessment
Network (NCLAN) examined 15 species and numerous cultivars.  The NCLAN results show that
"several economically important crop species are  sensitive to ozone levels typical of those found
in the U.S." (US EPA, 1996).  In addition, economic studies have shown a relationship between
observed ozone levels and crop yields (Garcia, et  al., 1986). The economic value associated with
varying levels of yield loss for ozone-sensitive commodity crops is analyzed using the AGSIM®
agricultural benefits model (Taylor, et al., 1993).  AGSIM® is an econometric-simulation model
that is based on a large set of statistically estimated demand and supply equations for agricultural
commodities produced in the United States.  The model is capable of analyzing the effects of
changes in policies (in this case, the implementation of the Nonroad Diesel Engines rule) that
affect commodity crop yields or production costs.jj

    The measure of benefits  calculated by the model is the net change in consumer and producer
surplus from baseline ozone concentrations to the ozone concentrations resulting from
attainment of particular standards.  Using the baseline and post-control equilibria, the model
calculates the change in net consumer and producer surplus on a crop-by-crop basis.kk Dollar
values are aggregated across crops for each standard. The total dollar value represents a measure
of the change in social welfare associated with the Nonroad Diesel Engines rule.

    The model employs biological  exposure-response information derived from  controlled
experiments conducted by the NCLAN (NCLAN, 1996).  For the purpose of our analysis, we
analyze changes for the six most economically significant crops for which C-R functions are
available: corn, cotton, peanuts, sorghum, soybean, and winter wheat.11 For some crops there are
multiple C-R functions, some more sensitive to ozone and some less.  Our base estimate assumes
that crops are evenly mixed between relatively sensitive and  relatively insensitive varieties.
Sensitivity to this assumption is tested in Appendix 9-B.
    "AGSIM6 is designed to forecast agricultural supply and demand out to 2010. We were not able to adapt the
model to forecast out to 2030. Instead, we apply percentage increases in yields from decreased ambient ozone levels
in 2030 to 2010 yield levels, and input these into an agricultural sector model held at 2010 levels of demand and
supply.  It is uncertain what impact this assumption will have on net changes in surplus.

    KK Agricultural benefits differ from other health and welfare endpoints in the length of the assumed ozone
season.  For agriculture, the ozone season is assumed to extend from April to September. This assumption is made
to ensure proper calculation of the ozone statistic used in the exposure-response functions. The only crop affected
by changes in ozone during April is winter wheat.

    LL The total value for these crops in 1998 was $47 billion.

                                          9-174

-------
                                                                 Cost-Benefit Analysis
   9A. 3.6.2.2 Forestry Benefits

   Ozone also has been shown conclusively to cause discernible injury to forest trees (US EPA,
1996; Fox and Mickler, 1996). In our previous analysis of the HD Engine/Diesel Fuel rule, we
were able to quantify the effects of changes in ozone concentrations on tree growth for a limited
set of species. Due to data limitations, we were not able to quantify such impacts for this
analysis. We plan to assess both physical impacts on tree growth and the economic value of
those phyisical impacts in our analysis of the final rule. We will use econometric models of
forest product supply and demand to estimate changes in prices, producer profits and consumer
surplus.

   9A.3.6.2.3 Other Vegetation Effects

   An additional welfare benefit expected to accrue as a result of reductions in ambient ozone
concentrations in the U.S. is the economic value the public receives from reduced aesthetic
injury to forests. There is sufficient scientific information available to reliably establish that
ambient ozone levels cause visible injury to foliage and impair the growth of some sensitive
plant species (US EPA, 1996c, p. 5-521). However,  present analytic tools and resources
preclude EPA from quantifying the benefits of improved forest aesthetics.

   Urban ornamentals represent an additional vegetation category likely to experience some
degree of negative effects associated with exposure to ambient ozone levels and likely to impact
large economic  sectors. In the absence of adequate exposure-response functions and economic
damage functions for the potential range of effects relevant to these types of vegetation, no direct
quantitative economic benefits analysis has been conducted.  It is estimated that more than $20
billion (1990 dollars) are spent annually on landscaping using ornamentals (Abt Associates,
1995), both by private property owners/tenants  and by governmental units responsible for public
areas. This is therefore a potentially important welfare effects category. However, information
and valuation methods are not available to  allow for plausible estimates of the percentage of
these expenditures that may be related to impacts associated with ozone exposure.

   The nonroad diesel standards, by reducing NOX emissions, will also reduce nitrogen
deposition on agricultural land and forests. There is  some evidence that nitrogen deposition may
have positive effects on agricultural output through passive fertilization.  Holding all other
factors constant, farmers' use of purchased fertilizers or manure may increase as deposited
nitrogen is reduced. Estimates of the potential value of this possible increase in the use of
purchased fertilizers are not available, but it is likely that the overall value is very small relative
to other health and welfare effects. The share of nitrogen requirements provided by this
deposition is  small, and the marginal cost of providing this nitrogen from alternative sources is
quite low. In some areas, agricultural lands suffer from nitrogen over-saturation due to an

                                          9-175

-------
Final Regulatory Impact Analysis
abundance of on-farm nitrogen production, primarily from animal manure. In these areas,
reductions in atmospheric deposition of nitrogen from PM represent additional agricultural
benefits.

   Information on the effects of changes in passive nitrogen deposition on forests and other
terrestrial  ecosystems is very limited. The multiplicity of factors affecting forests, including
other potential stressors such as ozone, and limiting factors such as moisture and other nutrients,
confound  assessments of marginal changes in any one stressor or nutrient in forest ecosystems.
However,  reductions in deposition of nitrogen could have negative effects on forest and
vegetation growth in ecosystems where nitrogen is a limiting factor (US EPA, 1993).

   On the other hand, there is evidence that forest ecosystems in some areas of the United States
are nitrogen saturated (US EPA, 1993). Once saturation is reached, adverse effects of additional
nitrogen begin to occur such as soil acidification which can lead to leaching of nutrients needed
for plant growth and mobilization of harmful elements such as aluminum. Increased soil
acidification is also linked to higher amounts of acidic runoff to streams and lakes and leaching
of harmful elements into aquatic ecosystems.

   9A.3.6.3 Benefits from Reductions in Materials Damage and Odor

   The preliminary control options that we modeled are expected to produce economic benefits
in the form of reduced materials damage.  There are two important categories of these benefits.
Household soiling refers to the accumulation of dirt, dust, and ash on exposed surfaces. Criteria
pollutants  also have corrosive effects on commercial/industrial buildings and structures of
cultural and historical significance. The effects on historic buildings and outdoor works of art
are of particular concern because of the uniqueness and irreplaceability of many of these objects.

   Previous EPA benefit analyses have been  able to provide quantitative estimates of household
soiling damage. Consistent with SAB advice, we  determined that the existing data (based on
consumer  expenditures from the early 1970's) are too out of date to provide a reliable enough
estimate of current household soiling damages (EPA-SAB-Council-ADV-003,  1998) to include
in our base estimate. We calculate household soiling damages in a sensitivity estimate provided
in Appendix 9C.

   EPA is unable to estimate any benefits to commercial  and industrial entities from reduced
materials damage.  Nor is EPA able to estimate the benefits of reductions in PM-related damage
to historic buildings and outdoor works of art. Existing studies of damage to this latter category
in Sweden (Grosclaude and Soguel, 1994) indicate that these benefits could be an order of
magnitude larger than household soiling benefits.
                                         9-176

-------
                                                                 Cost-Benefit Analysis
   Reductions in emissions of diesel hydrocarbons that result in unpleasant odors may also lead
to improvements in public welfare.  The magnitude of this benefit is very uncertain, however,
Lareau and Rae (1989) found a significant and positive WTP to reduce the number of exposures
to diesel odors.  They found that households were on average willing to pay around $20 to $27
(2000$) per year for a reduction of one exposure to intense diesel odors per week (translating
this to a national level, for the approximately 125 million households in 2020, the total WTP
would be between $2.5 and $3.4 billion annually). Their results are not in a form that can be
transferred to the context of this analysis, but the general magnitude of their results suggests this
could be a significant welfare benefit of the rule.

   9A.3.6.4 Benefits from Reduced Ecosystem Damage

   The effects of air pollution on the health and stability of ecosystems are potentially very
important, but are at present poorly understood and difficult to measure.  The reductions in NOX
caused by the final rule could produce significant benefits. Excess nutrient loads, especially of
nitrogen, cause a variety of adverse consequences to the health of estuarine and coastal waters.
These effects include toxic and/or noxious algal blooms such as brown and red tides, low
(hypoxic) or zero (anoxic) concentrations of dissolved oxygen in bottom waters, the loss of
submerged aquatic vegetation due to the light-filtering effect of thick algal mats, and
fundamental shifts in phytoplankton community structure (Bricker et al.,  1999).

   Direct C-R functions relating changes in nitrogen loadings to changes in estuarine benefits
are not available. The preferred WTP based measure of benefits depends on the availability of
these C-R functions and on estimates of the value of environmental responses.  Because neither
appropriate C-R functions nor sufficient information to estimate the marginal value of changes in
water quality exist at present, calculation of a WTP measure is not possible.

   If better models of ecological effects can be defined,  EPA believes that progress can be made
in estimating WTP measures for ecosystem functions.  These estimates would be superior to
avoided cost estimates in placing economic values on the welfare changes associated with air
pollution damage to ecosystem health.  For example, if nitrogen or sulfate loadings can be linked
to measurable and definable changes in fish populations or definable indexes of biodiversity,
then CV studies can be designed to elicit individuals' WTP for changes in these effects. This is
an important area for further  research and analysis, and will require close collaboration among
air quality modelers, natural scientists, and economists.

9A.4 Benefits Analysis—Results

   Applying the C-R and valuation functions described in Section C to the estimated changes in
ozone and PM described in Section B yields estimates of the changes in physical damages (i.e.

                                         9-177

-------
Final Regulatory Impact Analysis
premature mortalities, cases, admissions, change in deciviews, increased crop yields, etc.) and
the associated monetary values for those changes. Estimates of physical health impacts are
presented in Table 9A.9. Monetized values for both health and welfare endpoints are presented
in Table 9 A. 10, along with total aggregate monetized benefits.  All of the monetary benefits are
in constant year 2000 dollars.

   Not all known PM- and ozone-related health and welfare effects could be quantified or
monetized. The monetized value of these unquantified effects is represented by adding an
unknown "B" to the aggregate total. The estimate of total monetized health benefits is thus
equal to the subset of monetized PM- and ozone-related health and welfare benefits plus B, the
sum  of the unmonetized health and welfare benefits.

   The total monetized estimates are dominated by benefits of premature mortality risk
reductions. Our benefits analysis projects that the modeled preliminary control options will
result in 7,800 avoided premature deaths in 2020 and 14,000 avoided premature deaths in 2030.
The increase in benefits from 2020 to 2030 reflects additional emission reductions from the
standards, as well as increases in total population and the average age (and thus baseline
mortality risk) of the population.

   Our primary estimate of total monetized benefits (including PM health, ozone health and
welfare, and visibility) in 2030 for the modeled nonroad preliminary control options is $96
billion using a 3 percent discount rate and $91 billion using a 7 percent discount rate. In 2020,
the monetized benefits are estimated at $54 billion using a 3 percent discount rate and $51 billion
using a 7 percent discount rate. Health benefits account for 97 percent of total benefits. The
monetized benefit associated with reductions in the risk of premature mortality, which accounts
for $89 billion in 2030 and $49 billion in 2020, is over 90 percent of total monetized health
benefits.  The next largest benefit is for reductions in chronic illness (chronic bronchitis and non-
fatal heart attacks), although this value is more than an order of magnitude lower than for
premature mortality.  Visibility, minor restricted activity days, work loss days, school absence
days, and worker productivity account for the majority of the remaining benefits. The remaining
categories account for less than $10 million each, however, they represent a large number of
avoided incidences affecting many  individuals.

   A comparison of the incidence table to the monetary benefits table reveals that there is not
always a close correspondence between the number of incidences avoided for a given endpoint
and the monetary value associated with that endpoint.  For example, there are 100 times more
work loss days than premature mortalities, yet work loss days account for only a very small
fraction of total monetized benefits. This reflects the fact that many of the less severe health
effects, while more common, are valued at a lower level than the more severe health effects.
Also, some effects, such as hospital admissions, are valued using a proxy measure known to

                                          9-178

-------
                                                                Cost-Benefit Analysis
underestimate WTP.  As such the true value of these effects may be higher than that reported in
Table 9A.9.

   Ozone benefits are in aggregate positive for the nation. However, due to ozone increases
occurring during certain hours of the day in some urban areas, in 2020 the net effect is an
increase in minor restricted activity days, which are related to changes in daily average ozone
(which includes hours during which ozone levels are low, but are increased relative to the
baseline). However, by 2030, there is a net decrease in MRAD consistent with widespread
reductions in ozone concentrations from the increased NOX emissions reductions. Overall,
ozone benefits are low relative to PM benefits for similar endpoint categories because of the
increases in ozone concentrations during some hours of some days in certain urban areas. For a
more complete discussion of this issue, see Chapter 3.
                                         9-179

-------
Final Regulatory Impact Analysis
                                         Table 9A.30.
      Reductions in Incidence of Adverse Health Effects Associated with Reductions in
  Particulate Matter and Ozone Associated with the Modeled Preliminary Control Option
Endpoint
PM-related Endpoints
Premature mortality: Long-term exposure (adults, 30 and over)8
Infant mortality (infants under one year)
Chronic bronchitis (adults, 26 and over)
Non-fatal myocardial infarctions (adults, 18 and older)
Hospital admissions — Respiratory (adults, 20 and older)0
Hospital admissions — Cardiovascular (adults, 20 and older)D
Emergency Room Visits for Asthma (18 and younger)
Acute bronchitis (children, 8-12)
Asthma exacerbations (asthmatic children, 6-18)
Lower respiratory symptoms (children, 7-14)
Upper respiratory symptoms (asthmatic children, 9-11)
Work loss days (adults, 18-65)
Minor restricted activity days (adults, age 18-65)
Ozone-related Endpoints
Hospital Admissions - Respiratory Causes (adults, 65 and older)E
Hospital Admissions - Respiratory Causes (children, under 2 years)
Emergency Room Visits for Asthma (all ages)
Minor restricted activity days (adults, age 18-65)
School absence days (children, age 6-11)
Avoided Incidence*
(cases/year)
2020

7,800
18
4,300
10,600
3,400
2,800
4,600
10,000
150,000
120,000
92,000
810,000
4,800,000

370
150
93
(2,400)
65,000
2030

13,800
26
6,500
17,700
6,000
4,400
6,900
16,000
230,000
190,000
141,000
1,160,000
6,800,000

1,100
280
200
96,000
96,000
A Incidences are rounded to two significant digits.
B Premature mortality associated with ozone is not separately included in this analysis
c Respiratory hospital admissions for PM includes admissions for COPD, pneumonia, and asthma.
D Cardiovascular hospital admissions for PM includes total cardiovascular and subcategories for ischemic heart
disease, dysrhythmias, and heart failure.
E Respiratory hospital admissions for ozone includes admissions for all respiratory causes and subcategories for
COPD and pneumonia.
                                             9-180

-------
                                                     Table 9A.31
   Results of Human Health and Welfare Benefits Valuation for the Modeled Preliminary
                                      Nonroad Diesel Engine Standards


Endpoint

Premature mortality0: (adults, 30 and over)
3% discount rate
7% discount rate
Infant mortality (infants under one year)
Chronic bronchitis (adults, 26 and over)
Non-fatal myocardial infarctions
3% discount rate
7% discount rate
Hospital Admissions from Respiratory CausesD>F

Hospital Admissions from Cardiovascular CausesE
Emergency Room Visits for Asthma

Acute bronchitis (children, 8-12)
Asthma exacerbations (asthmatic children, 6-18)
Lower respiratory symptoms (children, 7-14)
Upper respiratory symptoms (asthmatic children, 9-11)
Work loss days (adults, 18-65)
Minor restricted activity days (adults, age 18-65)

School absence days (children, age 6-11)
Worker productivity (outdoor workers, age 18-65)
Recreational visibility (86 Class I Areas)
Agricultural crop damage (6 crops)
Monetized TotalH
3% discount rate
7% discount rate


Pollutant

PM


PM
PM
PM


03
PM
PM
03
PM
PM
PM
PM
PM
PM
03
PM
03
03
PM
03
O3 and PM


Monetary BenefitsA'B
(millions 2000$, Adjusted for
Income Growth)
2020

$49,000
$46,000
$120
$1,800

$910
$880
$7.4
$60
$61
$0.03
$1.3
$3.9
$6.9
$2.0
$2.4
$110
($0.1)
$260
$4.8
$4.2
$1,300
$88

$54,000+B
$51,000+B
2030

$89,000
$84,000
$180
$2,800

$1,440
$1,400
$21
$110
$96
$0.06
$2.0
$6.0
$10.7
$3.1
$3.7
$150
$4.9
$370
$10
$6.9
$2,100
$137

$96,000+B
$91,000+B
  Monetary benefits are rounded to two significant digits.
B Monetary benefits are adjusted to account for growth in real GDP per capita between 1990 and the analysis year (2020 or 2030).
c Premature mortality associated with ozone is not separately included in this analysis. It is assumed that the C-R function for premature mortality
captures both PM mortality benefits and any mortality benefits associated with other air pollutants.  Also note that the valuation assumes the 5
year distributed lag structure described earlier.  Results reflect the use 3% and 7% discount rates consistent with EPA and OMB's guidelines for preparing
economic analyses (US EPA, 2000c, OMB Circular A-4).

D Respiratory hospital admissions for PM includes admissions for  COPD, pneumonia, and asthma.
E Cardiovascular hospital admissions for PM includes total cardiovascular and subcategories for ischemic heart disease, dysrhythmias, and heart
failure.
F Respiratory hospital admissions for ozone includes admissions for all respiratory causes and subcategories for COPD and pneumonia.
G B represents the monetary value of the unmonetized health and welfare benefits. A detailed listing of unquantified PM, ozone, CO, and NMHC
related health effects is provided in Table XI-B. 1.

-------
Final Regulatory Impact Analysis
9A.5  Discussion

   This analysis has estimated the health and welfare benefits of reductions in ambient
concentrations of particulate matter resulting from reduced emissions of NOx, SO2, VOC, and
diesel PM from nonroad diesel engines. The result suggests there will be significant health and
welfare benefits arising from the regulation of emissions from nonroad engines in the U.S. Our
estimate that 14,000 premature mortalities would be avoided in 2030, when emission reductions
from the regulation are fully realized, provides additional evidence of the important role that
pollution from the nonroad sector plays in the public health impacts of air pollution.

   We provide sensitivity analyses in Appendix 9C to examine key modeling assumptions.  In
addition, there are other uncertainties that we could not quantify, such as the importance of
unquantified effects and uncertainties in the modeling of ambient air quality. Inherent in any
analysis of future regulatory programs are uncertainties in projecting atmospheric conditions,
source-level emissions, and engine use hours, as well as population, health baselines, incomes,
technology, and other factors. The assumptions used to capture these elements are reasonable
based on the available evidence. However, data limitations prevent an overall quantitative
estimate of the uncertainty associated with estimates of total economic benefits. If one is
mindful of these limitations, the magnitude of the benefit estimates presented here can be useful
information in expanding the understanding of the public health impacts of reducing air pollution
from nonroad engines.

   The U.S. EPA will continue to evaluate new methods and models and select those most
appropriate for the estimation the health benefits of reductions in air pollution.  It is important to
continue improving benefits transfer methods in terms of transferring economic values and
transferring estimated C-R functions. The development of both better models of current health
outcomes and new models for additional health effects such as asthma and high blood pressure
will be essential to future improvements in the accuracy and reliability of benefits analyses (Guo
et al., 1999; Ibald-Mulli et al., 2001). Enhanced collaboration between air quality modelers,
epidemiologists, and economists should result in a more tightly integrated analytical framework
for measuring health benefits of air pollution policies. The Agency welcomes comments on how
we can improve the quantification and monetization of health and welfare effects and on
methods for characterizing uncertainty in our estimates.
                                         9-182

-------
                                                               Cost-Benefit Analysis
Appendix 9A References

    Abbey, D.E., F. Petersen, P.K. Mills, and W.L. Beeson.  1993. "Long-Term Ambient
Concentrations of Total Suspended Particulates, Ozone, and Sulfur Dioxide and Respiratory
Symptoms in a Nonsmoking Population." Archives of Environmental Health 48(1): 33-46.
    Abbey, D.E., S.D. Colome, P.K. Mills, R. Burchette, W.L. Beeson, and Y. Tian. 1993.
"Chronic Disease Associated With Long-Term Concentrations of Nitrogen Dioxide." Journal of
Exposure Analysis and Environmental Epidemiology 3 (2): 181 -202.
    Abbey, D.E., B.L. Hwang, RJ. Burchette, T. Vancuren, and P.K. Mills.  1995.  "Estimated
Long-Term Ambient Concentrations of PM(10) and Development of Respiratory Symptoms in a
Nonsmoking Population." Archives of Environmental Health 50(2): 139-152.
    Abbey, D.E., N. Nishino, W.F. McDonnell, RJ. Burchette, S.F. Knutsen, W. Lawrence
Beeson, and J.X. Yang. 1999. "Long-term inhalable particles and other air pollutants related to
mortality in nonsmokers [see comments]." Am JRespir Crit Care Med. 159(2):373-82.
    Abt Associates, Inc. 1995. Urban Ornamental Plants:  Sensitivity to Ozone and Potential
Economic Losses.  Prepared for the U.S. Environmental Protection Agency, Office of Air
Quality Planning and Standards; Research Triangle Park, NC.
    Abt Associates, Inc. April 2003. ProposedNonroad Land-based Diesel Engine Rule: Air
Quality Estimation, Selected Health and Welfare Benefits Methods, and Benefit Analysis Results.
Prepared for Office of Air Quality Planning and Standards, U.S. EPA.
    Adams, M.B., T.R. Angradi, and J.N. Kochenderfer.  1997.  "Stream Water and Soil Solution
Responses to 5 Years of Nitrogen and sSulfur Additions at the Fernow Experimental Forest,
West Virginia." Forest Ecology and Management 95:79-91.
    Adams, P.F., G.E. Hendershot, and M.A. Marano.  1999.  "Current Estimates from the
National Health Interview Survey, 1996." Vital Health Stat. 10(200): 1-212.
    Agency for Healthcare Research and Quality.  2000.  HCUPnet, Healthcare Cost and
Utilization Project.
    Alberini, A., M. Cropper, T. Fu, A. Krupnick,  J. Liu, D. Shaw, and W. Harrington. 1997.
"Valuing Health Effects of Air Pollution in Developing Countries: The Case of Taiwan."
Journal of Environmental Economics and Management 34:107-126.
    American Lung Association.  1999.  Chronic Bronchitis. Available at
http://www.lungusa.org/diseases/lungchronic.html.
    American Lung Association.  2002a. Trends in Morbidity and Mortality: Pneumonia,
Influenza,  and Acute Respiratory Conditions. American Lung Association, Best Practices and
Program Services, Epidemiology and Statistics Unit.
    American Lung Association.  2002b. Trends in Chronic Bronchitis and Emphysema:
Morbidity and Mortality.  American Lung Association, Best Practices and Program Services,
Epidemiology and Statistics Unit.
    American Lung Association.  2002c. Trends in Asthma Morbidity and Mortality.  American
Lung Association, Best Practices and Program Services, Epidemiology and Statistics Unit.
                                        9-183

-------
Final Regulatory Impact Analysis
   Banzhaf, S., D. Burtraw, and K. Palmer. October 2002. "Efficient Emission Fees in the U.S.
Electricity Sector." Resources for the Future Discussion Paper 02-45.
   Beier, C., H. Hultberg, F. Moldan, and Wright. 1995. "MAGIC Applied to Roof
Experiments (Risdalsheia, N.; Gardsjon, S.; Klosterhede, D.K.) to Evaluate the Rate of
Reversibility of Acidification Following Experimentally Reduced Acid Deposition." Water Air
SoilPollut. 85:1745-1751.
   Belanger, K., W. Beckett, E.  Triche, M.B. Bracken, T.  Holford, P. Ren, I.E. McSharry, D.R.
Gold, T.A. Platts-Mills, and B.P. Leaderer. 2003. "Symptoms of Wheeze and Persistent Cough
in the First Year of Life: Associations with Indoor Allergens, Air Contaminants, and Maternal
History of Asthma." Am JEpidemiol. 158:195-202.
   Berger, M.C., G.C. Blomquist, D. Kenkel, and G.S. Tolley. 1987. "Valuing Changes in
Health Risks: A Comparison of Alternative Measures." The Southern Economic Journal 53:
977-984.
   Booty, W.G.  and J.R. Kramer.  1984.  "Sensitivity Analysis of a Watershed Acidification
Model."  Philos.  Trans. R. Soc. London, Ser. B 305:441-449.
   Brenden, N., H. Rabbani, and M. Abedi-Valugerdi. 2001. "Analysis of Mercury-Induced
Immune Activation in Nonobese Diabetic (NOD) Mice."  Clinical and Experimental
Immunology 125(2):202-10.
   Bricker, S.B., C.G. Clement, D.E. Pirhalla, S.P. Orlando, and D.R.G. Farrow. 1999.
National Estuarine Eutrophication Assessment: Effects of Nutrient Enrichment in the Nation's
Estuaries. National Oceanic and Atmospheric Administration, National Ocean Service, Special
Projects Office and the National  Centers for Coastal Ocean Science. Silver Spring, Maryland.
   Bricker, S.B., C.G. Clements, D.E. Pirhalla, S.P. Orlando, and D.R.G. Farrow. 1999.
National Eutrophication Assessment: Effects of Nutrient Enrichment in the Nation's Estuaries.
NOAA, National Ocean Service, Special Projects Office and the National Centers for Coastal
Ocean Science. Silver Spring, MD.
   Burnett R.T., M. Smith-Doiron, D. Stieb, M.E. Raizenne, J.R. Brook, R.E Dales, J.A. Leech,
S. Cakmak, D. Krewski. 2001.  "Association between Ozone and Hospitalization for Acute
Respiratory Diseases in Children less than 2 Years of Age." Am JEpidemiol 153:444-52
   Carnethon M.R., D. Liao, G.W. Evans, W.E. Cascio, L.E. Chambless, W.D. Rosamond, and
G. Heiss. 2002.  "Does the Cardiac Autonomic Response to Postural Change Predict Incident
Coronary Heart Disease and Mortality?  The Atherosclerosis Risk in Communities Study."
American Journal of Epidemiology 155(l):48-56.
   Centers for Disease Control and Prevention.  2001.  National Report on Human Exposure to
Environmental Chemicals.  Atlanta, GA:  Department of Health and Human Services. 01-0379.
http://www.cdc.gov/nceh/dls/report.
   Chay, K.Y. and M. Greenstone. 2003. "The Impact of Air Pollution on Infant Mortality:
Evidence from Geographic Variation in Pollution Shocks Induced by a Recession." Quarterly
Journal of Economics 118(3).
   Chen, L., B.L. Jennison, W. Yang, and S.T. Omaye. 2000. "Elementary School
Absenteeism and Air Pollution." Inhal  Toxicol. 12(11):997-1016.
                                         9-184

-------
                                                               Cost-Benefit Analysis
    Chestnut, L.G.  April 15, 1997. Draft Memorandum: Methodology for Estimating Values
for Changes in Visibility at National Parks.
    Chestnut, L.G. and R.L. Dennis. 1997. "Economic Benefits of Improvements in Visibility:
Acid Rain Provisions of the 1990 Clean Air Act Amendments." Journal of Air and Waste
Management Association 47:395-402.
    Chestnut, L.G. and R.D. Rowe. 1990a. Preservation Values for Visibility Protection at the
National Parks: Draft Final Report. Prepared for Office of Air Quality Planning and Standards,
U.S. Environmental Protection Agency, Research Triangle Park, NC, and Air Quality
Management Division, National Park Service, Denver, CO.
    Chestnut, L.G. and R.D. Rowe. 1990b. "A New National Park Visibility Value Estimates."
In Visibility and Fine Particles, Transactions  of an AWMA/EPA International Specialty
Conference, C. V. Mathai, ed. Air and Waste Management Association, Pittsburgh.
    Christophersen, N. and R.F. Wright.  1981. "A Model for Stream Water Chemistry at
Birkenes, Norway." Water Re sour. Res. 17:377-389.
    Christophersen, N., H.M.  Seip, and R.F. Wright. 1982. "A Model for Streamwater
Chemistry at Birkenes, Norway."  Water Resour. Res. 18:977-997.
    Church, M.R., P.W. Shaffer, K.W. Thornton, D.L. Cassell, C.I. Liff, M.G Johnson, D.A.
Lammers, JJ. Lee, G.R. Holdren, J.S. Kern, L.H. Liegel, S.M. Pierson, D.L. Stevens, B.P.
Rochelle, and R.S. Turner.  1992.  Direct/Delayed Response Project:  Future Effects of Long-
Term Sulfur Deposition on Stream Chemistry in the Mid-Appalachian Region of the Eastern
United States.  U.S. Environmental Agency, EPA/600/R-92/186, Washington, DC. 384 pp.
    Cody, R.P., C.P. Weisel, G Birnbaum, and P.J. Lioy.  1992. "The Effect of Ozone
Associated with Summertime Photochemical  Smog on the Frequency of Asthma Visits to
Hospital Emergency Departments." Environ Res. 58(2): 184-94.
    Cosby, BJ. and R.F. Wright.  1998. "Modelling Regional Response of Lakewater Chemistry
to Changes in Acid Deposition: The MAGIC Model Applied to Lake Surveys in Southernmost
Norway 1974-1986-1995." Hydrol.Earth System Sci. 2:563-576.
    Cosby, B.J., R.F. Wright,  G.M. Hornberger, and J.N.  Galloway.  1985a.  Modelling the
Effects of Acid Deposition: Assessment of a  Lumped Parameter Model of Soil Water and
Streamwater Chemistry." Water Resour. Res.  21:51-63.
    Cosby, B.J., R.F. Wright,  G.M. Hornberger, and J.N.  Galloway.  1985b. "Modelling the
Effects of Acid Deposition: Estimation of Long-Term Water Quality Responses in a Small
Forested Catchment."  Water Resour. Res. 21:1591 -1601.
    Cosby, B.J., G.M. Hornberger, J.N. Galloway, and R.F. Wright.  1985c.  "Time Scales of
Catchment Acidification: A Quantitative Model for Estimating Freshwater Acidification."
Environ. Sci. Technol. 19:1144-1149.
    Cosby, B.J., A. Jenkins, J.D. Miller, R.C.  Ferrier, and T.A.B. Walker.  1990. "Modelling
Stream Acidification in Afforested Catchments: Long-Term Reconstructions at Two Sites in
Central  Scotland." J.Hydrol. 120:143-162.
                                        9-185

-------
Final Regulatory Impact Analysis
   Cosby, B.J., R.F. Wright, and E. Gjessing.  1995.  "An Acidification Model (MAGIC) with
Organic Acids Evaluated Using Whole-Catchment Manipulations in Norway." J.Hydrol.
170:101-122.
   Cosby, B.J., R.C. Ferrier, A. Jenkins, and R.F. Wright.  2001. "Modelling the Effects of
Acid Deposition: Refinements, Adjustments and Inclusion of Nitrogen Dynamics in the MAGIC
Model." Hydrol. Earth Syst. Sci. 5:499-517.
   Crocker, T.D. and R.L. Horst, Jr.  1981.  "Hours of Work, Labor Productivity, and
Environmental Conditions: A Case Study."  The Review of Economics and Statistics 63:361-368.
   Cropper, M.L. and A.J. Krupnick.  1990. "The Social Costs of Chronic Heart and Lung
Disease." Resources for the Future. Washington, DC. Discussion Paper QE 89-16-REV.
   Daniels M.J., F. Dominici, J.M. Samet, and S.L. Zeger. 2000. "Estimating Particulate
Matter-Mortality Dose-Response Curves and Threshold Levels: An Analysis of Daily
Time-Series for the 20 Largest U.S. Cities."  Am J Epidemiol 152(5):397-406.
   DeHayes, D.H., P.G.  Schaberg, G.J. Hawley, and G.R. Strimbeck.  1999.  "Acid Rain
Impacts cCalcium Nutrition and Forest Health: Alteration of Membrane-Associated Calcium
Leads to Membrane Destabilization and Foliar Injury in Red Spruce." BioScience 49:789-800.
   Dekker J.M., R.S. Crow, A.R. Folsom, P.J. Hannan, D. Liao, C.A.  Swenne, and E.G.
Schouten.  2000. "Low Heart Rate Variability in a 2-Minute Rhythm Strip Predicts Risk of
Coronary Heart Disease and Mortality From Several Causes:  The ARIC Study." Circulation
2000 102:1239-1244.
   Dockery, D.W., C.A.  Pope, X.P. Xu, J.D. Spengler, J.H. Ware, M.E. Fay, E.G.  Ferris, and
F.E. Speizer. 1993. "An Association between Air Pollution and Mortality in Six U.S. Cities."
New England Journal of Medicine 329(24): 1753-1759.
   Dockery, D.W., J. Cunningham, A.I. Damokosh, L.M. Neas, J.D. Spengler, P. Koutrakis,
J.H. Ware, M. Raizenne,  and F.E. Speizer. 1996.  "Health Effects of Acid Aerosols On North
American Children-Respiratory Symptoms." Environmental Health Perspectives 104(5): 500-
505.
   Dominici, F., A. McDermott, M. Daniels, et al. 2002. Report to the Health Effects Institute:
"Reanalyses of the NMMAPS Database." Available at
www.biostat.jhsph.edu/~fominic/HEI/nmmaps.html.
   Dominici, F., A. McDermott, S.L. Zeger, and J.M. Samet. 2002. "On the Use of
Generalized Additive Models in Time-Series Studies of Air Pollution and Health."  Am J
Epidemiol  156(3): 193-203.
   Driscoll, C.T., K.M. Postek, D. Mateti, K. Sequeira, J.D. Aber, W.J. Kretser, M.J. Mitchell,
and D.J. Raynal. 1998. "The Response of Lake Water in the Adirondack Region of New York
State to Changes in Acid  Deposition." Environmental Science and Policy 1:185-198.
   Driscoll, C.T., G. Lawrence, A. Bulger, T. Butler, C. Cronan, C.  Eagar, K.F. Lambert, G.E.
Likens, J. Stoddard, and K. Weathers.  2001. "Acid Deposition in the Northeastern U.S.:
Sources and Inputs, Ecosystem Effects, and Management Strategies." Bioscience 51:180-198.
                                        9-186

-------
                                                              Cost-Benefit Analysis
   Eisenstein, E.L., L.K. Shaw, KJ. Anstrom, C.L. Nelson, Z. Hakim, V. Hasselblad and D.B.
Mark. 2001.  "Assessing the Clinical and Economic Burden of cCoronary Artery Disease:
1986-1998." MedCare 39(8):824-35.
   EPA-SAB-COUNCIL-ADV-98-003.  1998.  "Advisory Council on Clean Air Compliance
Analysis Advisory on the Clean Air Act Amendments (CAAA) of 1990 Section 812 Prospective
Study: Overview of Air Quality and Emissions Estimates: Modeling, Health and Ecological
Valuation Issues Initial Studies."
   EPA-SAB-COUNCIL-ADV-00-001.  October 1999. "The Clean Air Act Amendments
(CAAA) Section 812 Prospective Study of Costs and Benefits (1999):  Advisory by the Health
and Ecological Effects Subcommittee on Initial Assessments of Health and Ecological Effects:
Part 2."
   EPA-SAB-COUNCIL-ADV-00-002.  October 1999. "The Clean Air Act Amendments
(CAAA) Section 812 Prospective Study of Costs and Benefits (1999):  Advisory by the Advisory
Council on Clean Air Compliance Analysis: Costs and Benefits of the CAAA. Effects
Subcommittee on Initial Assessments of Health and Ecological Effects: Part 2."
   EPA-SAB-COUNCIL-ADV-99-012.  July 1999. "The Clean Air Act Amendments (CAAA)
Section 812 Prospective Study of Costs and Benefits (1999):  Advisory by the Health and
Ecological Effects Subcommittee on Initial Assessments of Health and Ecological Effects: Part
1."
   EPA-SAB-COUNCIL-ADV-99-05. February 1999. "An SAB Advisory on the Health and
Ecological Effects Initial  Studies of the Section 812 Prospective Study: Report to Congress:
Advisory by the Health and Ecological Effects Subcommittee."
   EPA-SAB-EEAC-00-013. July 2000. "An SAB Report on EPA's White Paper Valuing the
Benefits of Fatal Cancer Risk Reduction."
   EPA-S AB-COUNCIL-ADV-01 -004.  2001.  Review of the Draft Analytical Plan for EPA 's
Second Prospective Analysis—Benefits and Costs of the Clean Air Act 1990-2020: An Advisory
by a Special Panel of the Advisory Council on Clean Air Compliance Analysis.  September.
   Evans, William N. and W. Kip Viscusi.  1993.  "Income Effects and the Value of Health."
Journal of Human Resources 28(3):497-518.
   Ferrier, R.C., R.C. Helliwell, B.J. Cosby, A.  Jenkins, and R.F. Wright.  2001. "Recovery
from Acidification of Lochs in Galloway, South-west Scotland, UK: 1979-1998." Hydrology
and Earth System Sciences 5:421 -431.
   Fox, S. and R.A. Mickler. 1995. "Impact of Air Pollutants on Southern Pine Forests."
Ecological Studies 118. New York:  Springer Verlag.
   Freeman, A.M. III. 1993. The Measurement of Environmental and Resource Values: Theory
and Methods. Washington, DC: Resources for the Future.
   Garcia, P., B. Dixon, and J. Mjelde. 1986. "Measuring the Benefits of Environmental
Change Using a Duality Approach: The Case of Ozone and Illinios Cash Grain Farms." Journal
of Environmental Economics and Management.
   Gilliland, F.D., K. Berhane, E.B. Rappaport, D.C. Thomas, E. Avol, W.J. Gauderman, S.J.
London, H.G. Margolis, R. McConnell, K.T. Islam, and J.M. Peters. 2001. "The Effects of
                                        9-187

-------
Final Regulatory Impact Analysis
Ambient Air Pollution on School Absenteeism due to Respiratory Illnesses." Epidemiology
12(l):43-54.
    Gold D.R., A. Litonjua, J. Schwartz, E. Lovett, A. Larson, B. Nearing, G. Allen, M. Verrier,
R. Cherry., and R. Verrier. 2000. "Ambient Pollution and Heart Rate Variability." Circulation
101(ll):1267-73
    Goldstein, R.A., S.A. Gherini, C.W. Chen, L. Mak, and R.J.M. Hudson.  1984.  "Integrated
acidification study (ILWAS): A Mechanistic Ecosystem Analysis. " Phil. Trans. R. Soc.
London., Ser. B, 305:409-425.
    Grandjean, P., P. Weihe, R.F. White, F. Debes, S. Araki, K. Yokoyama, K. Murata, N.
Sorensen, R Dahl, and P. J., Jorgensen.  1997.  "Cognitive Deficit in 7-year-old Children with
Prenatal Exposure to Methylmercury."  Neurotoxicology and Teratology 19(6):417-28.
    Grandjean, P., P. Weihe, R.F. White, and F. Debes.  1998. "Cognitive Performance of
Children Prenatally Exposed to "Safe" Levels of Methylmercury." Environmental Research
77(2): 165-72.
    Grandjean, P., R.F. White, A. Nielsen, D. Cleary, and E.C. De Oliveira Santos. 1999.
"Methylmercury Neurotoxicity  in Amazonian Children Downstream from Gold Mining."
Environmental Health Perspectives 107 (7):587-91.
    Greenbaum, D. 2002. Letter to colleagues dated May 30, 2002.  [Available at
www.healtheffects.org]. Letter from L.D. Grant, Ph.D. to Dr. P. Hopke re: external review of
EPA's Air Quality Criteria for Particulate Matter, with copy of 05/30/02 letter from Health
Effects Institute re: re-analysis of National Morbidity, Mortality and Air Pollution Study data
attached. Docket No. A-2000-01. Document No. IV-A-145.
    Grosclaude, P. and N.C.  Soguel.  1994. "Valuing Damage to Historic Buildings Using a
Contingent Market: A Case Study of Road Traffic Externalities." Journal of Environmental
Planning and Management 37:  279-287.
    Guo, Y.L., Y.C. Lin, F.C. Sung, S.L. Huang, Y.C. Ko, J.S. Lai, HJ.  Su, C.K. Shaw, R.S. Lin,
and D.W. Dockery.  1999. "Climate, Traffic-Related Air Pollutants, and Asthma Prevalence in
Middle-School Children in Taiwan." Environmental Health Perspectives 107:1001-1006.
    Hall, J.V., V. Brajer, and F.W. Lurmann.  2003.  "Economic Valuation of Ozone-related
School Absences in the South Coast Air Basin of California." Contemporary Economic Policy
21(4):407-417.
    Hammitt, J.K. 2002. "Understanding Differences in Estimates of the Value of Mortality
Risk." Journal of Policy Analysis and Management 21 (2):271 -273.
    Harrington, W. and P.R. Portney. 1987. "Valuing the Benefits of Health and Safety
Regulation." Journal of Urban Economics 22:101-112.
    Health Effects Institute (HEI). 2003.  Revised Analyses of Time-Series Studies of Air
Pollution and Health; Revised Analyses of the National Morbidity, Mortality and Air Pollution
Study, Part II; Revised Analyses of Selected Time-Series Studies. Health Effects Institute,
Boston, MA.
                                         9-188

-------
                                                                Cost-Benefit Analysis
   Herlihy, A.T., P.R. Kaufmann, M.R. Church, P.J. Wigington, Jr., J.R. Webb, and MJ. Sale.
1993. "The Effects of Acid Deposition on Streams in the Appalachian Mountain and Piedmont
Region of the Mid-Atlantic United States."  Water Re sour. Res. 29:2687-2703.
   Herlihy, A.T., P.R. Kaufmann, J.L. Stoddard, K.N. Eshleman, and AJ. Bulger.  1996.
Effects of Acidic Deposition on Aquatic Resources in the Southern Appalachians with a Special
Focus on Class I Wilderness Areas. The Southern Appalachian Mountain Initiative (SAMI).
   Hollman, F.W., T.J. Mulder, and I.E. Kalian.  January 2000.  "Methodology and
Assumptions for the Population Projections  of the United States:  1999 to 2100." Population
Division Working Paper No. 38, Population Projections Branch,  Population Division, U.S.
Census Bureau, Department of Commerce.
   Hornberger, G.M., BJ. Cosby, and R.F.  Wright.  1989. "Historical Reconstructions and
Future Forecasts of Regional Surface Water Acidification in Aouthernmost Norway." Water
Resourc.Res. 25:2009-2018.
   Horsley, S.B., RP. Long, S.W. Bailey, RA. Hallett, and T.J. Hall. 2000.  "Factors
Associated with the Decline Disease of Sugar maple on the Allegheny Plateau." Canadian
Journal of Forest Research 30:1365-1378.
   Howarth.  1998.  "An Assessment of Human Influences on Fluxes of Nnitrogen from the
Terrestrial Landscape to the Estuaries and Continental Shelves of the North Atlantic Ocean."
Nutrient Cycling in Agroecosystems 52(2/3):213-223.
   Huntington, T.G., R.P. Hooper, C.E. Johnson, B.T. Aulenbach, R. Cappellato, and A.E.
Blum. 2000.  "Calcium Depletion in  a Southeastern United States forest Ecosystem." Soil
Science  Society of America Journal 64:1845-1858.
   Ibald-Mulli, A., J. Stieber, H.-E. Wichmann, W. Koenig, and A. Peters. 2001.  "Effects of
Air Pollution on Blood Pressure:  A Population-Based Approach." American Journal of Public
Health 91:571-577.
   Industrial Economics, Incorporated (TEc). March 31., 1994.  Memorandum to Jim
DeMocker, Office of Air and Radiation, Office of Policy Analysis and Review, U.S.
Environmental Protection Agency.
   Ito, K.  2003. "Associations of Parti culate Matter Components with Daily Mortality and
Morbidity in Detroit, Michigan." In Revised Analyses of Time-Series Studies of Air Pollution
and Health. Special Report. Health Effects Institute, Boston, MA.
   Ito, K. and G.D. Thurston. 1996.  "Daily PMlO/Mortality Associations: An Investigations of
At-Risk Subpopulations."  Journal of Exposure Analysis and Environmental Epidemiology
6(l):79-95.
   Jenkins, A., P.O. Whitehead, B.J.  Cosby, and H.J.B. Birks. 1990.  "Modelling Long-Term
Acidification: A Comparison with Diatom Reconstructions and the Implications for
Reversibility." Phil.Trans.RSoc.Lond.B 327:435-440.
   Jones-Lee, M.W.  1989. The Economics of Safety and Physical Risk.  Oxford: Basil
Blackwell.
   Jones-Lee, M.W., M. Hammerton, and P.R. Philips.  1985. "The Value of Safety: Result of
a National Sample Survey." Economic Journal. 95(March):49-72.
                                         9-189

-------
Final Regulatory Impact Analysis
   Jones-Lee, M.W., G. Loonies, D. O'Reilly, and P.R. Phillips. 1993. "The Value of
Preventing Non-fatal Road Injuries: Findings of a Willingness-to-pay National Sample Survey".
TRY Working Paper, WP SRC2.
   Kahl, I, S. Norton, I. Fernandez, L. Rustad, and M. Handley. 1999. "Nitrogen and Sulfur
Input-Output Budgets in the Experimental and Reference Watersheds, Bear Brook Watershed,
Maine (BBWM)." Environmental Monitoring and Assessment 55:113-131.
   Kahl, J.S., S.A. Norton, I.J. Fernandez, K.J. Nadelhoffer, C.T. Driscoll, and J.D. Aber. 1993.
"Experimental Inducement of Nitrogen Saturation at the Watershed Scale." Environmental
Science and Technology 27:565-568.
   Kjellstrom, T., P. Kennedy, S. Wallis, and C. Mantell. 1986. Physical and Mental
Development of Children with Prenatal Exposure to Mercury from Fish. Stage 1: Preliminary
Tests at Age 4. Sweden: Swedish National Environmental Protection Board.
   Kleckner, N. and J. Neumann. June 3, 1999. "Recommended Approach to Adjusting WTP
Estimates to Reflect Changes in Real Income." Memorandum to Jim Democker, U.S.
EPA/OPAR.
   Krewski D., R.T. Burnett, M.S. Goldbert, K. Hoover, J.  Siemiatycki, M. Jerrett, M.
Abrahamowicz, and W.H. White. July 2000. Reanalysis of the Harvard Six Cities Study and the
American Cancer Society Study of Paniculate Air Pollution and Mortality. Special Report to the
Health Effects Institute, Cambridge MA.
   KrupnickA. 2002. "The Value of Reducing Risk of Death: A Policy Perspective." Journal
of Policy Analysis and Management 2:275-282.
   Krupnick, A.J. and M.L. Cropper.  1992. "The Effect of Information on Health Risk
Valuations." Journal of Risk and Uncertainty 5(2):29-48.
   Krupnick, A., M. Cropper, A. Alberini, N. Simon, B. O'Brien, R. Goeree, and M.
Heintzelman.  2002. "Age, Health and the Willingness to Pay for Mortality Risk Reductions: A
Contingent Valuation Study of Ontario Residents." Journal of Risk and Uncertainty 24:161-186.
   Kunzli N., S. Medina, R. Kaiser, P. Quenel, F. Horak Jr, and M. Studnicka. 2001.
"Assessment of Deaths Attributable to Air Pollution: Should We Use Risk Estimates Based on
Time  Series or on  Cohort Studies?" AmJEpidemiol 153(11): 1050-55.
   Kunzli, N., R. Kaiser, S. Medina, M. Studnicka, O. Chanel, P. Filliger, M. Kerry, F. Horak
Jr., V. Puybonnieux-Texier, P. Quenel, J. Schneider,  R. Seethaler, J-C Vergnaud, and H.
Sommer. 2000. "Public-Health Impact of Outdoor and Traffic-Related Air Pollution:  A
European Assessment." The Lancet 356:795-801.
   Lareau, T.J. and D.A. Rae.  1989.  "Valuing WTP for Diesel Odor Reductions: An
Application of Contingent Ranking Techniques." Southern Economic Journal 55: 728-742.
   Lave, L.B. and E.P. Seskin.  1977. Air Pollution and Human Health. Baltimore: Johns
Hopkins University Press for Resources for the Future.
   Lawrence, G.B., M.B. David, and W.C. Shortle.  1995.  "A New Mechanism for Calcium
Loss in Forest-Floor Soils." Nature 378:162-165.
                                        9-190

-------
                                                                Cost-Benefit Analysis
   Lawrence, G.W., M.B. David, S.W. Bailey, and W.C. Shortle.  1997.  "Assessment of
Calcium Status in Soils of Red Spruce Forests in the Northeastern United States."
Biogeochemistry 38:19-39.
   Lawrence, G.B., M.B. David, G.M. Lovett, P.S. Murdoch, D.A. Burns, J.L. Stoddard, B.P.
Baldigo, J.H. Porter, and A.W. Thompson. 1999. "Soil Calcium Status and the Response of
Stream Chemistry to Changing Acidic Deposition Rates in the Catskill Mountains of New
York." Ecological Applications 9:1059-1072.
   Levy, J.L, J.K. Hammitt, Y. Yanagisawa, and J.D. Spengler. 1999. "Development of a New
Damage Function Model for Power Plants: Methodology and Applications."  Environmental
Science and Technology  33:4364-4372.
   Levy, J.L, T.J. Carrothers, J.T. Tuomisto, J.K. Hammitt, and J.S. Evans. 2001. "Assessing
the Public Health Benefits of Reduced Ozone Concentrations."  Environmental Health
Perspectives 109:1215-1226.
   Liao D., J. Cai, W.D. Rosamond, R.W. Barnes, R.G. Hutchinson, E.A. Whitsel, P.
Rautaharju, and G. Heiss. 1997. "Cardiac Autonomic Function and Incident Coronary Heart
Disease: A Population-Based Case-Cohort Study. The ARIC Study. Atherosclerosis Risk in
Communities Study." American Journal of Epidemiology 145(8):696-706.
   Liao D., J. Creason, C. Shy, R. Williams, R. Watts, and R. Zweidinger. 1999. "Daily
Variation of Parti culate Air Pollution and Poor Cardiac Autonomic Control in the Elderly."
Environ Health Per sped  107:521-5
   Likens, G.E., C.T. Driscoll,  and D.C. Buso. 1996. "Long-Term Effects of Acid Rain:
Responses and Recovery of a Forest Ecosystem." Science 272:244-246.
   Lipfert, F.W., S.C. Morris, and R.E. Wyzga.  1989. "Acid Aerosols - the Next Criteria Air
Pollutant." Environmental Science & Technology  23(11):1316-1322.
   Lipfert, F.W., H. Mitchell Perry Jr., J. Philip Miller, Jack D. Baty, Ronald E. Wyzg, and
Sharon E. Carmody. 2000. "The Washington University-EPRI Veterans' Cohort Mortality
Study: Preliminary Results." Inhalation Toxicology 12:41-74.
   Lippmann, M., K. Ito, A. Nadas, and R.T. Burnett. August  2000.  "Association of Particulate
Matter Components with Daily Mortality and Morbidity in Urban Populations." Health Effects
Institute Research Report Number 95.
   MacDonald, N.W., A.J. Burton, H.O. Liechty, J.A. Whitter, K.S. Pregitzer, G.D. Mroz, and
D.D. Richter. 1992. "Ion Leaching in Forest Ecosystems along a Great Lakes Air Pollution
Gradient." Journal of Environmental Quality 21:614-623.
   Magari S.R., R. Hauser, J. Schwartz, P.L. Williams, T.J. Smith, and D.C. Christiani. 2001.
"Association of Heart rate Variability with Occupational and Environmental Exposure to
Particulate Air Pollution." Circulation 104(9):986-91
   McClelland, G., W. Schulze, D. Waldman, J. Irwin, D. Schenk, T. Stewart, L. Deck, and M.
Thayer. September 1993. Valuing Eastern Visibility: A Field Test of the Contingent  Valuation
Method. Prepared for Office of Policy, Planning and Evaluation, U.S. Environmental  Protection
Agency.
                                         9-191

-------
Final Regulatory Impact Analysis
   McConnell, R., K. Berhane, F. Gilliland, S J. London, H. Vora, E. Avol, WJ. Gauderman,
H.G. Margolis, F. Lurmann, D.C. Thomas, and J.M. Peters. 1999. "Air Pollution and Bronchitic
Symptoms in Southern California Children with Asthma." Environmental Health Perspectives
107(9):757-760.
   McConnell R., K. Berhane, F. Gilliland, SJ. London, T. Islam, WJ. Gauderman, E. Avol,
H.G. Margolis, J.M. Peters. 2002. "Asthma in Exercising cChildren Exposed to Ozone: A
Cohort Study." Lancet 359(9309):896.
   McDonnell, W.F., D.E. Abbey, N. Nishino, and M.D. Lebowitz.  1999. "Long-Term
Ambient Ozone Concentration and the Incidence of Asthma in Nonsmoking Adults: The
Ahsmog Study." Environmental Research 80(2 Pt 1): 110-21.
   McLaughlin S.B. andR. Wimmer.  1999.  "Tansley Review No. 104, Calcium Physiology
and Terrestrial eEcosystem Processes."  New Phytologist 142:373-417.
   Miller, T.R. 2000.  "Variations between Countries in Values of Statistical Life. " Journal of
Transport Economics and Policy 34:169-188.
   Mitchell, M.J., M.B. David, IJ. Fernandez, R.D. Fuller, K. Nadelhoffer, L.E. Rustad, and
A.C. Stam.  1994.  "Response of Buried Mineral Soil Bags to Experimental Acidification of
Forest Ecosystem." Soil Science Society of America Journal 58:556-563.
   Mitchell, M.J., C.T. Driscoll, J.S. Kahl, GE. Likens, P.S. Murdoch, and L.H. Pardo.  1996.
"Climate Control on Nitrate Loss from Forested Watersheds in the Northeast United States."
Environmental Science and Technology 30:2609-2612.
   Moldan, F., R.F. Wright, R.C. Ferrier, B.I., Andersson, and H. Hultberg.  1998.  "Simulating
the Gardsjon covered catchment experiment with the MAGIC model," In Experimental Reversal
of Acid Rain Effects.  The Gardsjon Roof Project., Hultberg, H. and Skeffington, R. A., eds., p.
351-362, Chichester, UK: Wiley and Sons, 466 pp.
   Moolgavkar, S.H. 2000. "Air Pollution and Hospital Admissions for Diseases of the
Circulatory System in Three U.S. Metropolitan Areas." JAir WasteManagAssoc 50:1199-206.
   Moolgavkar, S.H..  2003.  "Air Pollution and Daily Deaths and Hospital Admissions in Los
Angeles and Cook Counties." In Revised Analyses of Time-Series Studies of Air Pollution and
Health.  Special Report. Boston, MA:  Health Effects Institute.
   Moolgavkar S.H., E.G. Luebeck, and E.L. Anderson.  1997. "Air Pollution and Hospital
Admissions for Respiratory Causes in Minneapolis-St. Paul and Birmingham." Epidemiology
8:364-70
   MrozekJ.R., and L.O. Taylor.  2002. "What Determines the Value of Life?  AMeta-
Analysis."  Journal of Policy Analysis and Management 21(2):253-270.
   Murdoch, P.S., D.S. Burns, and G.B. Lawrence. 1998. "Relation of Climate Change to the
Acidification of Surface Waters by Nitrogen Deposition." Environmental Science and
Technology 32:1642-1647.
   National Acid Precipitation Assessment Program( NAPAP).  1991. 1990 Integrated
Assessment Report. Washington, DC: National Acid Precipitation Assessment Program Office
of the Director.
                                        9-192

-------
                                                               Cost-Benefit Analysis
   National Academy of Sciences (NAS). 2000. lexicological Effects ofMethylmercury.
Washington, DC:  National Academy Press. Available at
http://books.nap.edu/catalog/9899.html?onpi_newsdoc071100.
   National Center for Education Statistics.  1996. "The Condition of Education 1996, Indicator
42: Student Absenteeism and Tardiness."  U.S. Department of Education National Center for
Education Statistics. Washington DC.
   National Research Council (NRC).  1998. Research Priorities for Airborne Particulate
Matter: I. Immediate Priorities and a Long-Range Research Portfolio. Washington, DC: The
National Academies Press.
   National Research Council (NRC).  2002. Estimating the Public Health Benefits of
Proposed Air Pollution Regulations.  Washington, DC: The National Academies Press.
   NCLAN.  1988. Assessment of Crop Loss from Air Pollutants. Walter W. Heck, O. Clifton
Taylor and David  T. Tingey, (eds.), pp. 1-5. (ERL,GB 639). New York: Elsevier Science
Publishing Co.
   Neumann, I.E., M.T. Dickie, and R.E. Unsworth. March 31, 1994.  "Linkage Between
Health Effects Estimation and Morbidity Valuation in the Section 812 Analysis — Draft
Valuation Document." Industrial Economics Incorporated (lEc) Memorandum to Jim
DeMocker, U.S. Environmental Protection Agency, Office of Air and Radiation, Office of Policy
Analysis and Review.
   Norris, G., S.N. YoungPong, J.Q. Koenig, T.V. Larson, L. Sheppard, and J.W. Stout.  1999.
"An Association between Fine Particles and Asthma Emergency Department Visits for cChildren
in Seattle." Environ Health Perspect. 107(6):489-93.
   Norton, S.A., J.S. Kahl, I.J. Fernandez, L.E. Rustad, J.P. Schofield,  and T.A. Haines.  1994.
"Response of the West Bear Brook Watershed, Maine, USA, to the addition of (NH4)2SO4: 3-
year results." Forest and Ecology Management 68:61 -73.
   Norton, S.A., R.F.  Wright, J.S. Kahl, and J.P. Scofield. 1998. "The MAGIC Simulation of
Surface Water Acidification at, and First Year RSesults from, the  Bear Brook Watershed
Manipulation, Maine, USA."  Environ.?ollut. 77:279-286.
   Norton, S.A., J.S. Kahl, I.J. Fernandez, T.A. Haines, L.E. Rustad, S. Nodvin, J.P. Scofield,  T.
Strickland, H. Erickson, P. Wiggington, and J. Lee.  1999.  "The Bear Brook Watershed, Maine,
(BBWP) USA." Environmental Monitoring and Assessment 55:7-51.
   Ostro, B.D.  1987. "Air Pollution and Morbidity Revisited: A Specification Test." Journal
of Environmental Economics Management 14:87-98.
   Ostro B.D. and S. Rothschild.  1989. "Air Pollution and Acute Respiratory Morbidity: An
Observational Study of Multiple Pollutants." Environmental Research 50:238-247.
   Ostro, B.D., M.J. Lipsett, M.B. Wiener, and J.C. Seiner. 1991. "Asthmatic Responses to
Airborne Acid Aerosols." Am J Public Health 81(6):694-702.
   Ostro, B. and L. Chestnut.  1998.  "Assessing the Health Benefits of Reducing Particulate
Matter Air Pollution in the United States." Environmental Research, Section A, 76: 94-106.
                                         9-193

-------
Final Regulatory Impact Analysis
   Ostro, B., M. Lipsett, J. Mann, H. Braxton-Owens ,and M. White.  2001.  "Air Pollution and
Exacerbation of Asthma in African-American Children in Los Angeles." Epidemiology 12(2):200-
   Ozkaynak, H. and G.D. Thurston.  1987. "Associations between 1980 U.S. Mortality Rates
and Alternative Measures of Airborne Particle Concentration." Risk Analysis 7(4): 449-61.
   Peters A., D.W. Dockery, I.E. Muller, and M.A. Mittleman. 2001. "Increased Particulate
Air Pollution and the Triggering of Myocardial Infarction."  Circulation 103:2810-2815.
   Poloniecki J.D., R.W Atkinson., A.P de Leon., and H.R. Anderson. 1997. "Daily Time
Series for Cardiovascular Hospital Admissions and Previous Day's Air Pollution in London,
UK." Occup Environ Med 54(8):535-40.
   Pope, C.A. 2000. "Invited Commentary: Particulate Matter-Mortality Expsoure-Response
Relations and Thresholds." American Journal of Epidemiology 152:407-412.
   Pope, C.A., III, MJ. Thun, M.M. Namboodiri, D.W.  Dockery, J.S. Evans, F.E. Speizer, and
C.W. Heath, Jr.  1995. "Particulate Air Pollution as a Predictor of Mortality in a Prospective
Study of U.S. Adults." American Journal of Respiratory Critical Care Medicine  151:669-674.
   Pope, C.A., III, R.T. Burnett, MJ. Thun, E.E. Calle, D. Krewski, K. Ito, and G.D. Thurston.
2002. "Lung Cancer, Cardiopulmonary Mortality, and Long-term Exposure to Fine Particulate
Air Pollution." Journal of the American Medical Association 287:1132-1141.
   Pope, C.A., III, D.W. Dockery, J.D. Spengler, and M.E. Raizenne. 1991. "Respiratory
Health and PM10 Pollution: A Daily Time Series Analysis." American Review of Respiratory
Diseases 144:668-674.
   Pratt, J.W. and RJ. Zeckhauser. 1996. "Willingness to Pay and the Distribution  of Risk and
Wealth." Journal of Political Economy 104:747-763.
   Ransom, M.R. and C.A. Pope.  1992. "Elementary School Absences and PM(10) Pollution
in Utah Valley."  Environmental Research 58(2):204-219.
   Rodier, P.M.  1995.  "Developing Brain as a Target of Toxicity." Environmental Health
Perspectives 103 Suppl 6:73-6.
   Rosamond, W., G. Broda, E. Kawalec, S. Rywik,  A. Pajak, L. Cooper, and L.  Chambless.
1999. "Comparison of Medical Care and Survival of Hospitalized Patients with Acute
Myocardial Infarction in Poland and the United States."  American Journal of Cardiology
83:1180-5.
   Rossi G., M.A. Vigotti, A. Zanobetti, F. Repetto,  V. Gianelle, and J. Schwartz. 1999. "Air
Pollution and Cause-Specific Mortality in Milan, Italy, 1980-1989." Arch Environ Health
54(3): 158-64
   Rowe, R.D. and L.G. Chestnut. 1986. "Oxidants and Asthmatics in Los Angeles: A Benefits
Analysis—Executive Summary." Prepared by Energy and Resource Consultants, Inc. Report to
the U.S. Environmental Protection Agency, Office of Policy Analysis. EPA-230-09-86-018.
Washington, DC.
   Rowlattetal. 1998.  "Valuation of Deaths from Air Pollution."  NERA and CASPAR for
DETR.
   Russell, M.W., D.M. Huse, S. Drowns, E.C. Hamel, and S.C. Hartz.  1998.  "Direct Medical
Costs of Coronary Artery Disease in the United States."  AmJCardiol. 81(9): 1110-5.
                                         9-194

-------
                                                                Cost-Benefit Analysis
   Rustad, L.E., IJ. Fernandez, M.B. David, MJ. Mitchell, KJ. Nadelhoffer, and R.D. Fuller.
1996. "Experimental Soil Acidification and Recovery at the Bear Brook Watershed in Maine."
Soil Science of America Journal 60:1933-1943.
   Samet, J.M., S.L. Zeger, I.E. Kelsall, J. Xu, and L.S. Kalkstein. March 1997.  "Air Pollution,
Weather, and Mortality in Philadelphia 1973-1988." Cambridge, MA: Health Effects Institute.
   Samet J.M., S.L. Zeger, F. Dominici, F. Curriero, I. Coursac, D.W. Dockery, J. Schwartz,
and A. Zanobetti.  June 2000.  The National Morbidity, Mortality and Air Pollution Study: Part
II: Morbidity, Mortality and Air Pollution in the United States.  Research Report No. 94, Part II.
Health Effects Institute, Cambridge MA.
   Schnoor, J.L., W.D. Palmer, Jr., and G.E. Glass.  1984.  "Modeling Impacts of Acid
Precipitation for Northeastern Minnesota." In Schnoor, J.L. (ed.), Modeling of'Total Acid
Precipitation Impact, pp. 155-173. Boston: Butterworth.
   Schwartz, J.  1993. "Particulate Air Pollution and Chronic Respiratory Disease."
Environmental Research 62:7-13.
   Schwartz, J.  1994a.  "PM(10) Ozone, and Hospital Admissions for the Elderly in
Minneapolis-St Paul, Minnesota." Archives of Environmental Health 49(5):366-374.
   Schwartz, J.  1994b.  "Air Pollution and Hospital Admissions for the Elderly in Detroit,
Michigan." American Journal of Respiratory and Critical Care Medicine 150(3):648-655.
   Schwartz, J.  1995. "Short Term Fluctuations in Air Pollution and Hospital  Admissions of
the Elderly for Respiratory Disease."  Thorax 50(5):531-8
   Schwartz, J. 2000. "Assessing Confounding, Effect Modification, and Thresholds in the
Association between Ambient Particles and Daily Deaths." Environmental Health Perspectives
108(6):563-8.
   Schwartz J. 2000.  "The Distributed Lag between Air Pollution and Daily Deaths."
Epidemiology 11 (3): 320-6.
   Schwartz, J., D.W. Dockery, L.M. Neas, D. Wypij, J.H. Ware, J.D. Spengler, P. Koutrakis,
F.E.  Speizer, and E.G. Ferris, Jr. 1994. "Acute Effects of Summer Air Pollution on Respiratory
Symptom Reporting in Children." American Journal of Respiratory Critical Care Medicine
150:1234-1242.
   Schwartz J., D.W. Dockery, and L.M. Neas. 1996. "Is Daily Mortality Associated
Specifically with Fine Particles?" J Air Waste Manag Assoc.  46:927-39.
   Schwartz J. and A. Zanobetti.  2000. "Using Meta-Smoothing to Estimate Dose-Response
Trends across Multiple Studies, with Application to Air Pollution and Daily Death."
Epidemiology 11:666-72.
   Schwartz J., and L.M. Neas. 2000. "Fine Particles are More Strongly Associated than
Coarse Particles with Acute Respiratory Health Effects in Schoolchildren." Epidemiology
11:6-10.
   Schwartz J., F. Laden, and A. Zanobetti. 2002. "The Concentration-Response Relation
between PM(2.5) and Daily Deaths." Environmental Health Perspectives 110:1025-9.
                                         9-195

-------
Final Regulatory Impact Analysis
    Seigneur, C., G. Hidy, I. Tombach, J. Vimont, and P. Amar. 1999. Scientific Peer Review of
the Regulatory Modeling System for Aerosols and Deposition (REMSAD). Prepared for the
KEVRIC Company, Inc.
    Sheppard, D.C. and Zeckhauser, RJ. 1984.  "Survival Versus Consumption." Management
Science 30(4).
    Sheppard, L. 2003. "Ambient Air Pollution and Nonelderly Asthma Hospital Admissions in
Seattle, Washington, 1987-1994." In Revised Analyses of Time-Series Studies of Air Pollution
and Health.  Special Report. Boston, MA:  Health Effects Institute.
    Sheppard, L., D. Levy, G. Norris, T.V. Larson, and J.Q. Koenig.  1999.  "Effects of Ambient
Air Pollution on Nonelderly Asthma Hospital Admissions in Seattle, Washington, 1987-1994."
Epidemiology 10: 23-30.
    Shogren, J. and T. Stamland. 2002.  "Skill and the Value of Life." Journal of Political
Economy  110:1168-1197.
    Shortle, W.C. and K.T. Smith.  1988. "Aluminum-Induced Calcium Deficiency Syndrome in
Declining Red Spruce Trees." Science 240:1017-1018.
    Shortle, W.C., K.T. Smith, R. Minocha, G.B. Lawrence, and M.B. David.  1997. "Acidic
Deposition, Cation Mobilization, and Biochemical Indicators of Stress in Healthy Red Spruce."
Journal of Environmental Quality 26:871-876.
    Sisler, J.F.  July 1996. Spatial and Seasonal Patterns and Long Term Variability of the
Composition of the Haze  in the United States: An Analysis of Data from the IMPROVE Network.
Fort Collins, CO:  Cooperative Institute for Research in the Atmosphere, Colorado State
University.
    Smith, A., T. Kim, M. Fuentes,  and D. Spitzner.  2000. "Threshold Dependence of Mortality
Effects for Fine and Coarse Particles in Phoenix,  Arizona." Journal of the Air and Waste
Management Association 5:1367-1379.
    Smith, D.H., D.C. Malone, K.A. Lawson, LJ. Okamoto,  C. Battista, and W.B. Saunders.
1997.  "A National Estimate of the Economic Costs of Asthma."  Am J Respir Crit Care Med.
156(3 Ptl):787-93.
    Smith, K.T.  and W.C. Shortle. 2001. "Conservation of Element Concentration in Xylem
Sap of Red Spruce." Trees 15:148-153.
    Smith, V.K., G. Van Houtven, and S.K. Pattanayak.  2002.  "Benefit Transfer via Preference
Calibration." Land Economics 78:132-152.
    S0rensen, N., K. Murata,, E. Budtz-J0rgensen,, P. Weihe, and P. Grandjean.  1999. "Prenatal
Methylmercury  Exposure as a Cardiovascular Risk Factor at Seven Years of Age." Epidemiology
10 (4):370-5.
    Southern Appalachian Man and the Biosphere (SAMAB). 1996.  The Southern Appalachian
Assessment: Summary Report.  Atlanta, GA: U.S. Department of Agriculture, Forest Service,
Southern Region.
    Stanford, R., T. McLaughlin, and LJ. Okamoto.  1999.  "The Cost of Asthma in the
Emergency Department and Hospital." Am J Respir Crit Care Med.  160(1):211-5.
                                        9-196

-------
                                                               Cost-Benefit Analysis
    Stieb, D.M., R.T. Burnett, R.C. Beveridge, and J.R. Brook.  1996.  "Association between
Ozone and Asthma Emergency Department Visits in Saint John, New Brunswick, Canada."
Environmental Health Perspectives 104(12): 1354-1360.
    Stieb D.M., S. Judek, and R.T. Burnett. 2002.  "Meta-Analysis of Time-Series Studies of Air
Pollution and Mortality:  Effects of Gases and Particles and the Influence of Cause of Death,
Age, and Season."  JAir Waste Manag Assoc 52(4):470-84
    Sweet, L.I. and J.T. Zelikoff. 2001. "Toxicology and Immunotoxicology of Mercury: A
Comparative Review in Fish and Humans." Journal of Toxicology and Environmental Health.
Part B, Critical Reviews 4(2): 161 -205.
    Taylor, C.R., K.H. Reichelderfer, and S.R.  Johnson.  1993.  "Agricultural Sector Models for
the United States: Descriptions and Selected Policy Applications." Ames, IA: Iowa State
University Press.
    Thurston, G.D. and K. Ito.  2001. "Epidemiological Studies of Acute Ozone Exposures and
Mortality." JExpo Anal Environ Epidemiol ll(4):286-94.
    Tolley, G.S. et al. January 1986. Valuation of Reductions in Human Health Symptoms and
Risks. University of Chicago. Final Report for the U.S. Environmental Protection Agency.
    Tsuji H., M.G. Larson, F.J. Venditti, Jr., E.S. Manders, J.C. Evans, C.L. Feldman, D. Levy.
1996. "Impact of Reduced Heart Rate Variability on Risk for Cardiac Events. The Framingham
Heart Study." Circulation 94(11):2850-5.
    U.S. Bureau of Census. 2000.  Population Projections of the United States by Age, Sex,
Race, Hispanic Origin and Nativity: 1999 to 2100.  Population Projections Program, Population
Division, U.S. Census Bureau, Washington, DC, Available at
http://www.census.gov/population/projections/nation/summary/np-t.txt.
    U.S. Department of Commerce, Bureau of Economic Analysis. July  1995. BEA Regional
Projections to 2045: Vol. 1, States. Washington, DC:  U.S. Govt.  Printing Office.
    U.S. Environmental Protection Agency. 1991.  "Ecological Exposure and Effects of
Airborne Toxic Chemicals: An Overview. EPA/6003-91/001." Corvallis, OR: Environmental
Research Laboratory.
    U.S. Environmental Protection Agency. December 1992. Regulatory Impact Analysis for
the National Emissions Standards for Hazardous Air Pollutants for Source Categories: Organic
Hazardous Air Pollutants from the Synthetic Organic Chemical Manufacturing Industry and
Seven Other Processes.  Draft Report. Office of Air Quality Planning and Standards. Research
Triangle Park, NC. EPA-450/3-92-009.
    U.S. Environmental Protection Agency. 1993.  External Draft, Air Quality Criteria for
Ozone and Related Photochemical Oxidants. Volume II.  U.S. EPA, Office of Health and
Environmental Assessment.  Research Triangle Park, NC, EPA/600/AP-93/004b.3v.
    U.S. Environmental Protection Agency. 1996a. Review of the National Ambient Air Quality
Standards for Ozone: Assessment of Scientific and Technical Information.  Office of Air Quality
Planning and Standards, Research Triangle Park, NC, EPA report no. EPA/4521R-96-007.
    U.S. Environmental Protection Agency. 1996b. Review of the National Ambient Air Quality
Standards for Particulate Matter: Assessment of Scientific and Technical Information.  Office of
                                         9-197

-------
Final Regulatory Impact Analysis
Air Quality Planning and Standards, Research Triangle Park, NC EPA report no. EPA/4521R-96-
013.
    U.S. Environmental Protection Agency.  1996. Review of the National Ambient Air Quality
Standards for Paniculate Matter: Assessment of Scientific and Technical Information.  Office of
Air Quality Planning and Standards, Research Triangle Park, NC; EPA report no. EP A/4521R-
96-013.
    U.S. Environmental Protection Agency.  1996. Mercury Study Report to Congress Volumes I
to VII. Washington, DC: U.S. Environmental Protection Agency Office of Air Quality Planning
and Standards. EPA-452-R-96-001b.  Available at http://www.epa.gov/oar/mercury.html.
    U.S. Environmental Protection Agency.  1997. Mercury Report to Congress. Office of Air
Quality Planning and Standards.  EPA report no. XXX.
    U.S. Environmental Protection Agency.  1997. The Benefits and Costs of the Clean Air Act,
1970 to 1990. Prepared for U.S. Congress by U.S. EPA, Office of Air and Radiation/Office of
Policy Analysis and Review, Washington, DC.
    U.S. Environmental Protection Agency.  1998. Utility Air Toxics Report to  Congress.
Office of Air Quality Planning and Standards. EPA report no.  XXX.
    U.S. Environmental Protection Agency (EPA). 1999. "An SAB Advisory:  The Clean Air
Act Section 812 Prospective Study Health and Ecological Initial Studies."  Prepared by the
Health and Ecological Effects Subcommittee (HEES) of the Advisory Council on the Clean Air
Compliance Analysis, Science Advisory Board,  U.S. Environmental Protection Agency.
Washington DC. EPA-SAB-COUNCIL-ADV-99-005, 1999.
    U.S. Environmental Protection Agency.  1999. The Benefits and Costs of the Clean Air Act,
1990-2010. Prepared for U.S. Congress by U.S. EPA, Office of Air and Radiation/Office of
Policy Analysis and Review, Washington, DC, November; EPA report no. EPA-410-R-99-001.
    U.S. Environmental Protection Agency.  2000a. Regulatory Impact Analysis: Heavy-Duty
Engine and Vehicle Standards and Highway Diesel Fuel Sulfur Control Requirements. Prepared
by: Office of Air and Radiation. Available at http://www.epa.gov/otaq/diesel.htm.  Accessed
March 20, 2003.
    U.S. Environmental Protection Agency.  2000b.  "Valuing Fatal Cancer Risk Reductions."
White Paper for Review by the EPA Science Advisory Board.
    U.S. Environmental Protection Agency.  September 2000c. Guidelines for Preparing
Economic Analyses.  EPA 240-R-00-003.
    U.S. Environmental Protection Agency.  2000. Integrated Risk Information  System; website
access available at www.epa.gov/ngispgm3/iris. Data as of December 2000.
    U.S. Environmental Protection Agency.  200la. "Mercury  White Paper."  Available at
http://www.epa.gov/ttn/oarpg/t3/memoranda/whtpaper.pdf
    U.S. Environmental Protection Agency.  2001b. Integrated Risk Information System (IRIS)
Risk Information for Methylmercury (MeHg). Washington, DC: National Center for
Environmental Assessment.  Available at http://www.epa.gov/iris/subst/0073.htm.
                                        9-198

-------
                                                               Cost-Benefit Analysis
   U.S. Environmental Protection Agency.  2002a.  "Technical Addendum: Methodologies for
the Benefit Analysis of the Clear Skies Initiative. September." Available at
http://www.epa.gov/air/clearskies/tech_adden.pdf. Accessed March 20, 2003.
   U.S. Environmental Protection Agency.  2002b.  "Final Regulatory Support Document:
Control of Emissions from Unregulated Nonroad Engines.  EPA Office of Air and Radiation.
EPA420-R-02-022, Docket number A-2000-01, Document V-B-4, September 2002.
   U.S. Environmental Protection Agency.  2003a.  Clear Skies Act—Technical Report:
Section B.
   U.S. Environmental Protection Agency.  2003b.  America's Children and the Environment:
Measures of Contaminants, Body Burdens, and Illnesses. EPA report no. 240R03001. Available
at http://www.epa.gov/envirohealth/children/report/ index.html.
   U.S. Environmental Protection Agency.  2003 c.  Fourth External Review Draft of Air Quality
Criteria for Paniculate Matter, EPA ORD, National Center for Environmental Assessment,
RTF. Volume II. EPA/600/P-99/002aD. June 2003.
   U.S. Environmental Protection Agency (EPA). 2003d.  "Advisory on Plans for Health
Effects Analysis." Presented in the May 12, 2003, Analytical Plan for EPA's Second
Prospective Analysis—Benefits and Costs of the Clean Air Act, 1990-2020: An  Advisory by the
Health Effects Subcommittee (HES) of the Advisory Council for Clean Air Compliance
Analysis, U.S. Environmental Protection Agency, Washington DC.
   U.S. Bureau of the Census. 2002. Statistical Abstract of the United States:  2001.
Washington DC.
   U.S. Office of Management and Budget. October 1992. "Guidelines and Discount Rates for
Benefit-Cost Analysis  of Federal Programs." Circular No.  A-94.
   U.S. Department of Health and Human Services, Centers for Disease Control and Prevention,
National Center for Health Statistics. 1999.  National Vital Statistics Reports. 47(19).
   Valigura, R.A., R.B. Alexander, M.S. Castro, T.P. Meyers, H.W. Paerl, P.E. Stacy, and R.E.
Turner. 2001. Nitrogen Loading in Coastal Water Bodies: An Atmospheric Perspective.
Washington, DC: American Geophysical Union.
   Van Sickle, J. and M.R. Church. 1995. Methods for Estimating the Relative  Effects of Sulfur
and Nitrogen Deposition on Surface Water Chemistry. EPA/600/R-95/172. Washington, DC:
U.S. Environmental Protection Agency.
   Vedal, S., J.  Petkau, R. White, and J. Blair.  1998.  "Acute Effects of Ambient Inhalable
Particles in Asthmatic  and Nonasthmatic Children."  American Journal of Respiratory and
Critical Care Medicine 157(4): 1034-1043.
   Viscusi, W.K.  1992.  Fatal Tradeoffs: Public and Private Responsibilities for Risk. New
York: Oxford University Press.
   Viscusi, W.K. and  More MJ. 1989.  "Rates of Time Preference and Valuations of the
Duration of Life." Journal of Public Economics 3 8:297-317.
   Viscusi, W.K., W.A. Magat, and J. Huber.  1991. "Pricing Environmental Health Risks:
Survey Assessments of Risk-Risk and Risk-Dollar Trade-Offs for Chronic Bronchitis."  Journal
of Environmental Economics and Management 21:32-51.
                                        9-199

-------
Final Regulatory Impact Analysis
   Viscusi W.K. and I.E. Aldy.  2003 forthcoming. "The Value of A Statistical Life: A Critical
Review of Market Estimates Throughout the World." Journal of Risk and Uncertainty.
   Webb, J.R., F.A. Deviney, Jr., BJ. Cosby, AJ. Bulger, and J.N. Galloway.  2000. Change in
Acid- Base Status in Streams in the Shenandoah National Park and the Mountains of Virginia.
American Geophysical Union, Biochemical Studies of the Shenandoah National Park.
   Webb, J.R., F.A. Deviney, J.N. Galloway, C.A. Rinehart, P.A Thompson, and S. Wilson.
1994. The Acid-Base Status of Native Brook  Trout Streams in the Mountains of Virginia. A
Regional Assessment Based on the Virginia Trout Stream Sensitivity Study.  Charlottesville, VA:
Univ. of Virginia.
   Weisel, C.P., R.P. Cody, and P.J. Lioy. 1995. "Relationship between Summertime Ambient
Ozone Levels and Emergency Department Visits for Asthma in Central New Jersey." Environ
Health Perspect. 103 Suppl 2:97-102.
   Whitehead, P.G., B. Reynolds, G.M. Hornberger, C. Neal, B.J.  Cosby, and P. Paricos.  1988.
"Modelling Long Term Stream Acidification  Trends in Upland Wales at Plynlimon."
HydrologicalProcesses 2:357-368.
   Whitehead, P.G., J. Barlow, E.Y. Haworth, and J.K. Adamson.  1997.  "Acidification in
Three Lake District tarns: Historical Long Term Trends and Modelled Future Behaviour under
Changing Sulphate and Nitrate Deposition." Hydrol.Earth System  Sci. 1:197-204.
   Whittemore, A.S. and E.L. Korn.  1980. "Asthma and Air Pollution in the Los Angeles
Area." American Journal of Public Health 70:687-696.
   Wittels, E.H., J.W. Hay, and A.M. Gotto,  Jr. 1990.  "Medical Costs of Coronary Artery
Disease in the United States." AmJCardiol 65(7):432-40.
   Woodruff,  T.J., J. Grille, and K.C. Schoendorf. 1997. "The Relationship Between Selected
Causes of Postneonatal Infant Mortality and Paniculate Air Pollution in the United States."
Environmental Health Perspectives 105(6):608-612.
   Woods & Poole Economics Inc. 2001. "Population by Single Year of Age CD." Woods &
Poole Economics, Inc.
   World Health Organization. 2002. "Global Burden of Disease Study." World Health
Organization.
   Wright, R.F., B.J. Cosby,  M.B. Flaten, and J.O. Reuss.  1990. "Evaluation of an
Acidification Model with Data from Manipulated Catchments in Norway." Nature 343: 53-55.
   Wright, R.F., B.J. Cosby,  R.C. Ferrier, A  Jenkins, A.J. Bulger,  and R. Harriman.  1994.
"Changes in the Acidification of Lochs in Galloway, Southwestern Scotland, 1979-1988: The
MAGIC Model Used to Evaluate the Role of  Afforestation, Calculate Critical Loads, and Predict
Fish Status." J.Hydrol 161:257-285.
   Wright, R.F., B.A. Emmett, and A. Jenkins. 1998. "Acid Deposition, Land-Use Change and
Global Change: MAGIC7 Model Applied to  Risdalsheia, Norway  (RAIN and CLIMEX
projects)  and Aber, UK (NITREX project)." Hydrol.Earth System Sci. 2:385-397.
   Yu, O., L. Sheppard, T. Lumley, J.Q. Koenig, and G.G. Shapiro. 2000.  "Effects of Ambient
Air Pollution on Symptoms of Asthma in Seattle-Area Children Enrolled in the CAMP Study."
Environ Health Perspect 108(12): 1209-1214.
                                         9-200

-------
                                                               Cost-Benefit Analysis
   Zanobetti, A., J. Schwartz, E. Samoli, A. Gryparis, G. Touloumi, R. Atkinson, A. Le Tertre,
J. Bobros, M. Celko, A. Goren, B. Forsberg, P. Michelozzi, D. Rabczenko, E. Aranguez Ruiz,
and K. Katsouyanni. 2002. "The Temporal Pattern of Mortality Responses to Air Pollution: A
Multicity Assessment of Mortality Displacement." Epidemiology 13(l):87-93.
                                        9-201

-------
Final Regulatory Impact Analysis
APPENDIX 9B: Supplemental Analyses Addressing Uncertainties in the
Concentration-Response and Valuation Functions for Particulate Matter Health Effects
   9B. 1 Introduction	9-205
   9B.2  Monte Carlo Based Uncertainty Analysis Using Classical Statistical Sources of
       Uncertainty	9-206
       9B.2.1  Monte Carlo Analysis Using Pope et al. (2002) to Characterize the Distribution
          of Reductions in Premature Mortality	9-210
   9B.3 Expert Elicitation of PM Mortality	9-215
          9B.3.1 Elicitation Method	9-216
       9B.3.3  Experts' Views of Sources of Uncertainty 	9-223
       9B.3.4  Advisory Council Comments on the Preliminary Design of the Elicitation . 9-223
       9B.3.5 Limitations in Pilot Elicitation Design	9-224
       9B.3.6  Combining the Expert Judgments for Application to Economic Benefit Analyses
           	 9-227
       9B.3.5 Limitations of Combining Expert Judgments 	9-235
   9B.4.  Illustrative Application of Pilot Expert Elicitation Results	9-235
   9B.5.3 Limitations of the Application of the Pilot Elicitation Results to the Nonroad Scenario
        	 9-246
                                         9-202

-------
                                                                 Cost-Benefit Analysis
9B.1 Introduction

   In this appendix, we describe our progress toward improving our approach to characterizing
the uncertainties in our economic benefits estimates, with particular emphasis on the
concentration-response (C-R) function.  We present two types of probabilistic approaches
designed to illustrate how some aspects of the uncertainty in the C-R function might be handled
in a PM benefits analysis. The first approach generates a probabilistic estimate of statistical
uncertainty based on standard errors reported in the underlying studies used in the benefit
modeling framework.  The second approach uses the results from a pilot expert elicitation
designed to characterize certain aspects of uncertainty in the ambient PM2.5/mortality
relationship. For the reasons discussed earlier in Chapter 9, neither the primary benefit estimate
nor these approaches have been used to inform any regulatory decisions in this rulemaking.

   In any benefit analyses of air pollution regulations, estimation of pre-mature mortality
accounts for 85 to 95 percent of total benefits.  Therefore, it is an endpoint that will be an
important focus for characterizing the uncertainty related to the estimates of total benefits. As
part of a collaboration with the EPA's Office of Air and Radiation (OAR) and the Office of
Management and Budget (OMB) on the Non-Road Diesel Rule, EPA extended it's collaboration
with OMB in 2003 to conduct a pilot expert elicitation intended to more fully characterize
uncertainty in the effect estimates used to estimate mortality resulting from exposure to PM.

   It should be recognized that in addition to uncertainty, the annual benefit estimates for the
Final Non-Road Diesel Rule also are inherently variable, due to the truly random processes that
govern pollutant emissions and ambient air  quality in a given year. Factors such as hourly use of
engines and daily weather display constant variability regardless of our ability to accurately
measure them. As such, the primary estimates of annual benefits presented in this chapter and
the sensitivity analysis estimates presented in this and other appendices should be viewed as
representative of the types of benefits that will  be realized,  rather than the actual benefits that
would occur every year. As such, the distributions of the estimate of annual benefits should be
viewed as representative of the types of benefits that will be realized, rather than the actual
benefits that would occur every year.
9B.2  Monte Carlo Based Uncertainty Analysis Using Classical Statistical
Sources of Uncertainty

   The recent NAS report on estimating public health benefits of air pollution regulations
recommended that EPA begin to move the assessment of uncertainties from its ancillary analyses
into its primary analyses by conducting probabilistic, multiple-source uncertainty analyses.

                                         9-203

-------
Final Regulatory Impact Analysis
However, for this proposal we did not attempt to assign probabilities to all of the uncertain
parameters in the model due to a lack of resources and reliable methods.  At this time, we simply
generate estimates of the distributions of dollar benefits for PM health effects and for total dollar
benefits including visibility. We provide a likelihood distribution for the total benefits estimate,
based solely on the statistical uncertainty surrounding the estimated C-R functions and the
assumed distributions around the unit values.

    Our estimate of the likelihood distribution for total benefits should be viewed as an
approximate result because of the wide range of sources of uncertainty that we have not
incorporated. The  5th and 95th percentile points of our estimate are based on statistical error and
cross-study variability provides some insight into how uncertain our estimate is with regards to
those sources of uncertainty. However, it does not capture other sources  of uncertainty regarding
other inputs to the model, including emissions, air quality, and aspects of the health science not
captured in the studies, such as the  likelihood that PM is causally related  to premature mortality
and other serious health effects..

    Although there are several sources of uncertainty affecting estimates of endpoint-specific
benefits, the sources of uncertainty that are most readily quantifiable in this analysis are the C-R
relationships and uncertainty about unit dollar values. The total dollar benefit associated with a
given endpoint depends on how much reducing risk of the endpoint will change due to the final
standard (e.g., how many premature deaths will be avoided) and how much each unit of change
is worth (e.g., how much a premature death avoided is worth)."^ However, as we have noted,
this omits important sources of uncertainty, such as the contribution of air quality changes,
baseline population incidences,  projected populations exposed, transferability of the C-R
function to diverse locations, and uncertainty  about the C-R relationship for premature mortality.
Thus, a likelihood description based on the standard error would provide a misleading picture
about the overall uncertainty in the estimates. The empirical evidence about uncertainty is
presented where it is available.

    Both the uncertainty about the incidence changes and uncertainty about unit dollar values can
be characterized by distributions.   Each " likelihood distribution" characterizes our beliefs about
what the true value of an unknown  variable (e.g., the true change in incidence of a given health
effect in relation to PM exposure) is likely to be, based on the available information from
relevant studies.™1 Unlike a sampling distribution (which describes the possible values that an
    MM Because this is a national analysis in which, for each endpoint, a single C-R function is applied everywhere,
there are two sources of uncertainty about incidence: (1) statistical uncertainty (due to sampling error) about the true
value of the pollutant coefficient in the location where the C-R function was estimated, and (2) uncertainty about
how well any given pollutant coefficient approximates p*.

    NN Although such a "likelihood distribution" is not formally a Bayesian posterior distribution, it is very similar
in concept and function (see, for example, the discussion of the Bayesian approach in Kennedy 1990, pp. 168-172).

                                           9-204

-------
                                                                  Cost-Benefit Analysis
estimator of an unknown variable might take on), this likelihood distribution describes our
beliefs about what values the unknown variable itself might be.  Such likelihood distributions
can be constructed for each underlying unknown variable (such as a particular pollutant
coefficient for a particular location) or for a function  of several underlying unknown variables
(such as the total dollar benefit of a regulation). In either case, a likelihood distribution is a
characterization of our beliefs about what the unknown variable  (or the function of unknown
variables) is likely to be, based on all the available relevant information.   A likelihood
description based on such distributions are typically expressed as the interval from the fifth
percentile point of the likelihood distribution to the ninety-fifth percentile point. If all
uncertainty had been included, this range would be the "credible range" within which we believe
the true value is likely to lie with 90 percent probability.

   The uncertainty about the total dollar benefit associated with any single endpoint combines
the uncertainties from these two sources (the C-R relationship and the valuation), and is
estimated with  a Monte Carlo method. In each iteration of the Monte Carlo procedure, a value is
randomly drawn from the incidence distribution and a value is randomly drawn from the unit
dollar value distribution, and the total dollar benefit for that iteration is the product of the two.00
If this is repeated for many (e.g., thousands of) iterations, the distribution of total dollar benefits
associated with the endpoint is generated.

   Using this Monte Carlo procedure, a distribution of dollar benefits may be generated for each
endpoint.  As the number of Monte Carlo draws gets larger and larger, the Monte Carlo-
generated distribution becomes a better and better approximation of a joint likelihood
distribution for the considered likelihood distributions making up the overall model of total
monetary benefits for the endpoint.

   After endpoint-specific distributions are generated, the same Monte Carlo procedure can then
be used to combine the dollar benefits from different  (non-overlapping) endpoints to generate a
distribution of total dollar benefits.

   The estimate of total benefits may be thought of as the end result of a sequential process in
which, at each step, the estimate of benefits from an additional source is added. Each time an
estimate of dollar benefits from a new source (e.g.,  a new health endpoint) is added to the
previous estimate of total dollar benefits, the estimated total dollar benefits increases.  However,
our bounding or likelihood description of where the true total value lies also increases as we add
more sources.
   00 This method assumes that the incidence change and the unit dollar value for an endpoint are stochastically
independent.

                                          9-205

-------
Final Regulatory Impact Analysis
   As an example, consider the benefits from reductions in PM-related hospital admissions for
cardiovascular disease. Because the actual dollar value is unknown, it may be  described using a
variable, with a distribution describing the possible values it might have. If this variable is
denoted as Xl, then the mean of the distribution, E(Xt) and the variance of Xl3 denoted Var(Xj),
and the 5th and 95th percentile points of the distribution (related to Var(Xj)), are ways to
describe the likelihood for the true but unknown value for the benefits reduction.

   Now suppose the benefits from reductions in PM-related hospital admissions for respiratory
diseases are added. Like the benefits from reductions in PM-related hospital admissions for
cardiovascular disease, the likelihood distribution for where we expect the true value to  be  may
be considered a variable, with a distribution. Denoting this variable as X2, the benefits from
reductions in the incidence of both types of hospital admissions is Xl + X2. This variable has a
distribution with mean E(Xl + X2) = E(Xj) + E(X2), and a variance of Var(Xx + X2) = Var(Xj) +
Var(X2) + 2Cov(X1,X2); if Xl and X2 are stochastically independent, then it has a variance of
Var(Xj + X2) = Var(Xj) + Var(X2), and the covariance term is zero.

   The benefits from  reductions in all non-overlapping PM-related health and welfare endpoints
    j,  ..., XJ is X = Xj + ... + X,,.  The mean of the distribution of total benefits, X, is:
                        E(X) =  E(Xl) +  E(X2) +  ...  + E(XJ                        (1)


and the variance of the distribution of total benefits — assuming that the components are
stochastically independent of each other (i.e., no covariance between variables) — is:
                   Var(X)  =  Var(X^ +  Vartf^  +  ...  +
   If all the means are positive, then each additional source of benefits increases the point
estimate (mean) of total benefits. However, with the addition of each new source of benefits, the

      E(XJ < E(X,  + X2) < E(X, + X2  + XJ  < ...< E(X,  +  ... + XJ  = E(X)      (3)


variance of the estimate of total benefits also increases.  That is,
but:
                       X2) <  Var(X1  + X2 + X3) < ...< Var(Xl +  ...  + Xn)  =  Var(X)
That is, the addition of each new source of benefits results in a larger mean estimate of total
benefits (as more and more sources of benefits are included in the total) about which there is less
certainty.  This phenomenon occurs whenever estimates of benefits are added.
                                          9-206

-------
                                                                  Cost-Benefit Analysis
    Calculated with a Monte Carlo procedure, the distribution of X is composed of random draws
from the components of X. In the first draw, a value is drawn from each of the distributions, Xl3
X2, through Xj,, these values are summed, and the procedure is repeated again, with the number
of repetitions set at a high enough value (e.g., 5,000) to reasonably trace out the distribution of
X.  The fifth percentile point of the distribution of X will be composed of points pulled from all
points along the distributions of the individual components, and not simply from the fifth
percentile.  While the sum of the fifth percentiles of the components would be represented in the
distribution of X generated by  the Monte Carlo, it is likely that this value would occur at a
significantly lower percentile.  For a  similar reason, the 95th percentile of X will be less than the
sum of the  95th percentiles of the components, and instead the 95th percentile of X will be
composed of component values that are  significantly lower than the 95th percentiles.

    The physical effects estimated in this analysis are assumed to occur independently. It is
possible that, for any given pollution level, there is some correlation between the occurrence of
physical effects, due to say avoidance behavior or common causal pathways and treatments (e.g.,
stroke, some kidney disease, and heart attack are related to treatable blood pressure).  Estimating
accurately any such  correlation, however, is beyond the scope of this analysis, and instead it is simply
assumed that the physical effects  occur independently.

    We conduct two different Monte Carlo analyses, one based on the distribution of reductions in
premature mortality  characterized by the mean effect estimate and standard error from the epidemiology
study of PM-associated mortality associated with long-term exposure used in the primary estimate in
Chapter 9 (Pope et al, 2002), and one based on the results from a pilot expert elicitation project
(Industrial Economics, 2004).  In both analyses, the distributions of all other health endpoints are
characterized by the reported mean and standard deviations from the epidemiology literature.
Distributions for unit dollar values are based on reported ranges or distributions of values in the
economic literature and are summarized in Table 9B-1. We are unable at this time to
characterize the uncertainty in  the estimate of benefits of improvements in visibility at Class I
areas. As such, we treat the visibility benefits as fixed and add them to all percentiles of the PM
health benefits distribution.  Results of the Monte Carlo analysis based on the Pope et al. (2002)
distribution are presented in the next section.  Results  of the Monte Carlo analysis based on the
pilot expert elicitation are presented in section 9B.3.
9B.2.1  Monte Carlo Analysis Using Pope et al. (2002) to Characterize the Distribution of
Reductions in Premature Mortality

   Based on the Monte Carlo techniques described earlier, we generated likelihood distributions
for the dollar value of reductions in PM-related health endpoints and a similar distribution for
total annual PM-related benefits including PM health and visibility benefits for the nonroad
diesel modeled preliminary control option.  For this analysis, the  likelihood descriptions for the

                                          9-207

-------
Final Regulatory Impact Analysis
true value for each of the PM health endpoint incidence measures, including premature
mortality, were based on classical statistical uncertainty measures, including the mean and
standard deviation for the C-R relationships in the epidemiological literature, and assumption of
particular likelihood distribution shapes for the valuation for each health endpoint values based
on reported values in the economic literature.  Table 9B-1 summarizes the chosen parameters
for likelihood distributions for unit values for each of the PM health effects included in the
Monte Carlo simulation. The distributions for the value used to represent  incidence of a health
effect in the total benefits valuation represent both the simple statistical uncertainty surrounding
individual effect estimates and, for those health endpoints with multiple effects from different
epidemiology studies, interstudy variability. Visibility benefits are also included in the
distribution of total benefits, however, we were unable  to characterize a distribution for visibility
benefits. As such, they are simply added to each percentile of the distribution of PM health
benefits.
                                           9-208

-------
                    Table 9B-1.  Distributions for Unit Values of Health Endpoints
Health
Endpoint
Mean Value,
Adjusted for
Income
Growth to
2030
Derivation of Distribution
Premature Mortality (Value of
a Statistical Life)
                                $5,500,000
                    Normal distribution anchored at 2.5th and 97.5th percentiles of  $1
                   and $10 million, respectively. Confidence interval is based on two
                   meta-analyses of the wage-risk VSL literature. $1 million represents
                   the lower end of the interquartile range from the Mrozek and Taylor
                   (2000) meta-analysis. $10 million represents the upper end of the
                   interquartile range from the Viscusi and Aldy (2003) meta-analysis.
                   The VSL represents the value of a small change in mortality risk
                   aggregated over the affected population. Normal distribution chosen
                   through best professional judgment.	
Chronic Bronchitis (CB)
  $430,000
The distribution of WTP to avoid a case of pollution-related
CB was generated by Monte Carlo methods, drawing from each of
three distributions: (1) WTP to avoid a severe case of CB is assigned
a 1/9 probability of being each of the first nine deciles of the
distribution of WTP responses in Viscusi et al., 1991; (2) the severity
of a pollution related case of CB (relative to the case described in the
Viscusi study) is assumed to have a triangular distribution, centered
at severity level 6.5 with endpoints at 1.0 and 12.0 (see text for
further explanation); and (3) the constant in the elasticity of WTP
with respect to severity is normally distributed with mean = 0.18 and
standard deviation = 0.0669 (from Krupnick and Cropper, 1992). This
process and the rationale for choosing it is described in detail in the
Costs and Benefits of the Clean Air Act, 1990 to 2010 (U.S. EPA,
1999)	
Nonfatal Myocardial Infarction
(heart attack)
     3% discount rate
     Age 0-24
     Age 25-44
     Age 45-54
     Age 55-65
     Age 66 and over

     7% discount rate
     Age 0-24
     Age 25-44
     Age 45-54
     Age 55-65
     Age 66 and over
   $66,902
   $74,676
   $78,834
  $140,649
   $66,902
   $65,293
   $73,149
   $76,871
  $132,214
   $65,293
No distribution available. Age specific cost-of-illness values
reflecting lost earnings and direct medical costs over a 5 year period
following a non-fatal MI. Lost earnings estimates based on Cropper
and Krupnick (1990). Direct medical costs based on simple average
of estimates from Russell et al. (1998) and Wittels et al. (1990).

Lost earnings:
Cropper and Krupnick (1990). Present discounted value of 5 yrs of
lost earnings:
age of onset:    at 3%      at 7%
25-44        $8,774       $7,855
45-54        $12,932     $11,578
55-65        $74,746     $66,920

Direct medical expenses: An average of:
1. Wittels et al., 1990 ($102,658-no discounting)
2. Russell et al., 1998, 5-yr period. ($22,331 at 3% discount rate;
$21,113 at 7% discount rate)	
Hospital Admissions
Chronic Obstructive
Pulmonary Disease (COPD)
(ICD codes 490-492, 494-496)
   $12,378
No distribution available. The COI point estimates (lost earnings plus
direct medical costs) are based on ICD-9 code level information (e.g.,
average hospital care costs, average length of hospital stay, and
weighted share of total COPD category illnesses) reported in Agency
for Healthcare Research and Quality, 2000 (www.ahrq.gov).

-------
Health
Endpoint
Mean Value,
Adjusted for
Income
Growth to
2030
Derivation of Distribution
Pneumonia
(ICD codes 480-487)
   $14,693
No distribution available. The COI point estimates  (lost earnings plus
direct medical costs) are based on ICD-9 code level information (e.g.,
average hospital care costs, average length of hospital stay, and
weighted share of total pneumonia category illnesses) reported in
Agency for Healthcare Research and Quality, 2000 (www.ahrq.gov).
Asthma admissions
   $6,634
The COI estimates (lost earnings plus direct medical costs) are based
on ICD-9 code level information (e.g., average hospital care costs,
average length of hospital stay, and weighted share of total asthma
category illnesses) reported in Agency for Healthcare Research and
Quality, 2000 (www.ahrq.gov).
All Cardiovascular
(ICD codes 390-429)
   $18,387
No distribution available. The COI point estimates  (lost earnings plus
direct medical costs) are based on ICD-9 code level information (e.g.,
average hospital care costs, average length of hospital stay, and
weighted share of total cardiovascular category illnesses) reported in
Agency for Healthcare Research and Quality, 2000 (www.ahrq.gov).
Emergency room visits for
asthma
    $286
No distribution available. The COI point estimate is the simple
average of two unit COI values:
(1) $311.55, from Smith etal, 1997, and
(2) $260.67, from Stanford et al., 1999.
Respiratory Ailments Not Requiring Hospitalization
Upper Respiratory Symptoms
(URS)
     $27
Combinations of the 3 symptoms for which WTP estimates are
available that closely match those listed by Pope, et al. result in 7
different "symptom clusters," each describing a "type" of URS. A
dollar value was derived for each type of URS, using mid-range
estimates of WTP (lEc,  1994) to avoid each symptom in the cluster
and assuming additivity of WTPs. In the absence of information
surrounding the frequency with which each of the seven types of
URS occurs within the URS symptom complex, we assume a uniform
distribution between $10 and $45.
Lower Respiratory Symptoms
(LRS)
     $17
Combinations of the 4 symptoms for which WTP estimates are
available that closely match those listed by Schwartz, et al. result in
11 different "symptom clusters," each describing a "type" of LRS. A
dollar value was derived for each type of LRS, using mid-range
estimates of WTP (lEc, 1994) to avoid each symptom in the cluster
and assuming additivity of WTPs. The dollar value for LRS is the
average of the dollar values for the 11 different types of LRS. In the
absence of information surrounding the frequency with which each of
the eleven types of LRS occurs within the  LRS symptom complex,
we assume a uniform distribution between $8 and $25.
Asthma Exacerbations
    $45
Asthma exacerbations are valued at $45 per incidence, based on the
mean of average WTP estimates for the four severity definitions of a
"bad asthma day," described in Rowe and Chestnut (1986).  This
study surveyed asthmatics to estimate WTP for avoidance of a "bad
asthma day," as defined by the subjects. For purposes of valuation,
an asthma exacerbation is assumed to be equivalent to a day in which
asthma is moderate or worse as reported in the Rowe and Chestnut
(1986) study. The value is assumed have a uniform distribution
between $17 and $73.

-------
 Health
 Endpoint
Mean Value,
Adjusted for
Income
Growth to
2030
Derivation of Distribution
 Acute Bronchitis
    $390
Assumes a 6 day episode, with the distribution of the daily value
specified as uniform with the low and high values based on those
recommended for related respiratory symptoms in Neumann, et al.
1994. The low estimate is the sum of the midrange values
recommended by lEc (1994) for two symptoms believed to be
associated with acute bronchitis: coughing and chest tightness.  The
high estimate was taken to be twice the value of a minor respiratory
restricted activity day.
 Restricted Activity and Work Loss Days
 Work Loss Days (WLDs)
  Variable
No distribution available. Point estimate is based on county-specific
median annual wages divided by 50 (assuming 2 weeks of vacation)
and then by 5 - to get median daily wage. U.S. Year 2000 Census,
compiled by Geolytics, Inc.
 Minor Restricted Activity
 Days (MRADs)
    $55
Median WTP estimate to avoid one MRAD from Tolley, et al.
(1986). Distribution is assumed to be triangular with a minimum of
$22 and a maximum of $83. Range is based on assumption that value
should exceed WTP for a single mild symptom (the highest estimate
for a single symptom-for eye irritation-is $16.00) and be less than
that for a WLD. The triangular distribution acknowledges that the
actual value is likely to be closer to the point estimate than either
extreme.
    Results of the Monte Carlo simulations are presented in Table 9B-2.  The table provides the
estimated means of the distributions and the estimated 5th and 95th percentiles of the  distributions.
The contribution of mortality to the mean benefits and to both the 5th and 95th percentiles of total
benefits is substantial, with mortality accounting for over 90 percent of the mean estimate, and
even the 5th percentile of mortality benefits dominating the 95th percentile of all other benefit
categories. Thus, the choice of value and the shape for likelihood distribution for VSL should be
examined closely and is key information to provide to decision makers for any decision
involving this variable. The 95th percentile of total benefits is approximately twice the mean,
while the 5th percentile is approximately one fourth of the mean.  The overall range from 5th to
95th represents about one order of magnitude.
                                                9-211

-------
Final Regulatory Impact Analysis
                                                    Table 9B-2.
     Distribution of Value of Annual Human Health and Welfare Benefits in 2030 for the
                Modeled Preliminary Control Option of the Non-Road Diesel RuleA
Endpoint
Premature mortality0
Long-term exposure, (adults, >30yrs)
Long-term exposure (child 
-------
                                                                 Cost-Benefit Analysis
9B.3 Expert Elicitation of PM Mortality

   In its 2002 report, the NAS provides a number of recommendations on how EPA might
improve the characterization of uncertainty in its benefits analyses.  One recommendation was
that "EPA should begin to move the assessment of uncertainties from its ancillary analyses into
its primary analyses by conducting probabilistic, multiple-source uncertainty analyses.  This shift
will require specification of probability distributions for major sources of uncertainty.  These
distributions should be based on available data and expert judgment."(NAS, 2002: 14)  The NAS
elaborated on this recommendation by  saying "although the specific methods for selection and
elicitation of experts may need to be modified somewhat, the protocols that have been developed
and tested by OAQPS [in prior EPA projects — see below] provide a solid foundation for future
work in the area. EPA may also consider having its approaches reviewed and critiqued by
decision analysts, biostatisticians, and psychologists from other fields where expert judgment is
applied." (NAS, 2002: 140).  They recommended the use of formally elicited expert judgments,
but noted that a number of issues must be addressed, and that sensitivity analyses would be
needed for distributions that are based on expert judgment. They also recommended that EPA
clearly distinguish between data-derived  components of an uncertainty assessment and those
based on expert opinions. As a first step  in addressing the NAS recommendations regarding
expert elicitation, EPA, in collaboration with OMB, conducted a pilot expert elicitation to
characterize uncertainties in the relationship between ambient PM2.5 and mortality. While it is
premature to include the results of the pilot in the primary analysis for this rulemaking, EPA and
OMB believe this pilot is an important step in moving toward the goal of incorporating
additional uncertainty analyses in its future primary benefits analyses.

   This pilot was designed to provide EPA with an opportunity to improve its understanding of
the design and application of expert elicitation methods to economic benefits analysis and lay the
groundwork for a more comprehensive elicitation.  For instance, the pilot was designed to
provide feedback on the  efficacy of the protocol developed and the analytic challenges, as well
as to provide insight regarding potential implications of the results on the degree of uncertainty
surrounding the C-R function for PM2 5 mortality. The scope of the pilot was limited in that we
focused the elicitation on the C-R function of PM mass rather than on individual issues
surrounding an estimate  of the change in  mortality due to PM exposure. Also, to meet time
constraints placed on the pilot, we selected expert s for participation from two previously
established expert panels of the NAS, and chose not to conduct a workshop with the experts
prior to the elicitation. The limited scope of the pilot meant that a full expert elicitation process
was truncated and many aspects of the uncertainty surrounding the PM2 5-mortality relationship
could not be quantitatively characterized. Recognizing this, the results of the pilot are only used
in this benefits estimation for illustrative  purposes.  A full description of the pilot is contained in
a report titled, "An Expert Judgment Assessment of the Concentration-Response Relationship
between PM2.5 Exposure and Mortality," (TEc, 2004) available in the public docket for this rule.

                                         9-213

-------
Final Regulatory Impact Analysis
    The analytic plan for the pilot was developed based on established elicitation methods as
suggestedby the NAS and published in the peer-reviewed literature. The plan was internally
reviewed by EPA and OMB scientists with experience using expert elicitation methods. The
Health Effect Subcommittee (HES) of the Council on Clean Air Compliance Analysis (the
"Council") then provided additional suggestions, which led to further changes in the elicitation
protocol. However, it should be noted that the Council did not provide a complete peer review
of the elicitation methods or interpretation of results.  Finally,  the protocol was tested on PM
scientists from within EPA and external to the Agency, who would not be part of the final
elicitation process. The project team that implemented the pilot consisted of individuals with
experience in expert elicitation and individuals with expertise  in PM health effects and health
benefits.

    As a final step in this carefully designed pilot, the EPA and OMB will sponsor an external
peer review of the methods used in this pilot expert elicitation as well as the approaches to
presenting the results (particularly with respect to combining results across experts), in
accordance with EPA's peer review guidelines.  Until the peer review is complete andthe
comments of the reviewers addressed, we do not recommend use of these results for other
regulatory analysis.

    9B.3.1 Elicitation Method

Expert elicitation  is a formal, highly structured and well documented process whereby expert
judgments, usually of multiple experts, are obtained (U.S. NRC, 1996).  Formal expert elicitation
usually involves experts with training and expertise in statistics, decision analysis, and
probability encoding who work with subject matter experts to  structure questions about uncertain
relationships or parameters  and who design and implement the process used to obtain
probability and other judgments from subject matter experts. Several academic traditions -
judgment and decision-making, human factors, cognitive sciences, expert systems, management
science, to name a few - have sought to understand how to successfully elicit probabilistic
judgments from both lay people and experts (Morgan and Henri on 1990,_Cooke 1991,; Wright
and Ayton 1994, Ayyub 2002). Over the past two decades, there has been an increasing number
of studies that have used expert judgment techniques to characterize uncertainty in quantities of
interest to environmental risk analysis and decision-making. North and Merkhofer (1976)
considered the use of expert judgment in evaluating emission control strategies. As referred to
by the NAS, the EPA's Office of Air Quality Planning and Standards (OAQPS) successfully
used expert judgment to characterize uncertainty in the health  effects of exposure to lead
(McCurdy and Richmond, 1983; Whitfield and Wallsten,  1989) and to ozone (Whitfield et al.
1991; Winkler et al.,  1995). Amaral (1983) and Morgan et al.  (1984) used expert judgment in
the  evaluation of the transport and impacts of sulfur air pollution.  Several studies have been
done in the area of climate change (Manne and Richels, 1994;  Nordhaus, 1994; Morgan and
                                         9-214

-------
                                                                 Cost-Benefit Analysis
Keith, 1995; Reilly et al, 2001). Hawkins and Evans (1989) used industrial hygienists to predict
toluene exposures to workers involved in a batch chemical process.  In a more recent use of
expert judgment in exposure analysis, Walker et al. (2001, 2003) asked experts to estimate
ambient, indoor and personal air concentrations of benzene.  A few studies have used expert
judgment to characterize uncertainty in chemical dose response:  Hawkins and Graham (1988)
and Evans et al. (1994) for formaldehyde and Evans et al. (1994b) for risk of exposure to
chloroform in drinking water. Expert judgment has also been used in the characterization of
residential radon risks (Krewski et al., 1999).

   The literature (Granger and Morgan, 1990) suggest there are several steps involved in the
design and implementation of an expert elicitation, including:

   developing  a protocol that contains the specific content of the elicitation and the questions
   that will be  asked of the experts,
   selection of experts,
   compiling a briefing book of materials that can be used by the experts as background
   information to respond to the elicitation,
•  pilot testing the protocol,
•  conducting  the elicitation and  summarizing the findings.

       The pilot expert elicitation consisted of a series of structured questions, both quantitative
and qualitative, about the nature of the PM2 5/mortality relationship.  The objective was to obtain
experts' quantitative, probabilistic judgments about the average expected decrease in mortality
rates associated with decreases in  PM25 exposures in the United States. These judgments were
expressed in terms of median estimates and associated percentile values of an uncertainty
distribution. The quantitative questions in the protocol asked experts to provide judgments about
changes in mortality due exposure to PM25. Specifically, they were asked to estimate:!) the
percent change  in annual non-accidental mortality associated with a 1 ng/m3 change in annual
average PM2 5 (long-term exposure); and 2) the percent change in daily non-accidental mortality
associated with a 10 |J,g/m3 change in daily 24-hour average  PM2 5 (short-term exposure). For
each type of exposure, each expert provided minimum, maximum, and median estimates, plus
5th, 25th, 75th, and 95th percentile values for the distribution describing his uncertainty in the
mortality effect of the specified change in PM25.

       The pilot focused on eliciting judgments about the C-R function for PM2.5 mass (without
regard to source) and their solicited opinions about the key factors influencing the uncertainty in
estimating the PM2.5/mortality relationship.  As a warm-up  to answering the quantitative
question, experts were asked their views on several key issues including: cause of death,
mechanisms, thresholds, lag/cessation period, the relative effect of PM components and their
sources, confounding, and effect modification. This discussion allowed the experts to articulate
                                          9-215

-------
Final Regulatory Impact Analysis
the way they interpreted the underlying issues, thus what would form the conceptual framework
of their quantitative judgments.  Their responses also provided EPA with information that would
be useful for designing a more comprehensive and disaggregated elicitation assessment in the
future.

   The pilot elicitation consisted of personal interviews with five experts. The five experts were
selected from an initial pool defined by the membership on two PM-related NAS committees. The
rosters of  both NRC committees  included  recognized experts in pertinent  fields such  as
epidemiology and toxicology who had already undergone extensive review of their qualifications
by the NRC, producing a reasonable initial list of experts likely to meet our expert selection criteria.
The five experts selected for participation in the elicitation include: Dr. Roger McClellan, Dr. Bart
Ostro, Dr. Jonathan Samet, Dr. Mark Utell, and Dr. Scott Zeger. The specific process used to select
experts is detailed in the technical report of the elicitation (ffic, 2004) along with additional
information about the experts' affiliations and fields of expertise. The size of the final expert panel
was dictated by time and resource constraints, and the decision to restrict the initial expert pool to
the NRC committees was made to help expedite the expert selection process. The experts were
provided a briefing book of reference materials and a copy of the elicitation protocol prior to the
interviews.  Each interview lasts 6-8 hours.

       9.B.3.2    Elicitation Results

       Figure 9B-1 displays the responses of the experts to the quantitative elicitation question for
the mortality effects of changes in long-term PM25 exposures. The distributions provided by  each
expert, identified by the letters A through E, are depicted as box plots with the diamond symbol
showing the median (50th percentile), a circle symbol showing the mean estimate, the box defining
the interquartile range (bounded by the 25th and 75th percentiles), and the whiskers defining  each
expert's 90 percent confidence interval (bounded by the 5th and 95th percentiles of the distribution).
                                          9-216

-------
             Figure 9B-1.  Summary of Experts' Judgments About the Percent Increase in Annual Average Non-Accidental
                       Mortality Associated with a 1 ng/m3 Increase in Annual Average Exposures to PM2.5
2 -
m

1 ;
1
ra 1.5-

o.
£
.{S 1
o
c
(fl
£ 0.5
u
55
0 -



























I


O
















I




A































B *















1


C--8 u g / m 3 C--10 u g / m 3 C--15 u g / m 3 C --
Ex p e r t
























O




















20 u g / m 3





























o
I


D






















O





























E

            *Expert B specified this distribution for the PM/mortality coefficient above an uncertain threshold which he characterized as ranging between 4 and
15 with a modal value of 12 ng/m3.As illustrated here, considerable variation exists in both the median values and the spread of uncertainty provided by the
experts. The median value of the percent change in annual non-accidental mortality per unit change in annual PM2 5 concentration (within a range of PM2 5
concentrations from 8 to 20 i-ig/m3) ranged from values at or near zero to a value of 0.7 percent. The variation in the responses largely reflects differences in the
amount of uncertainty each expert considered inherent in the key epidemiological results from long-term cohort studies, the likelihood of a causal relationship,
and the shape of the C-R function.  The technical report (lEc, 2004) provides detailed descriptions of the experts' judgments about these factors, but we present a
few brief observations relative to their responses below.
        **
          Expert C specified a non-linear model and provided distributions for the slope of the curve at four discrete concentrations within the range.

-------
Final Regulatory Impact Analysis
       As illustrated by the figure, the experts exhibited considerable variation in both the
median values they reported and in the spread of uncertainty about the median. In response to the
question concerning the effects of changes in long-term exposures to PM2 5, the median value
ranged from values at or near zero to a 0.7 percent increase in annual non-accidental mortality
per 1 jig/m3 increase in annual mean PM25 concentration (within a range of PM25 concentrations
from 8 to 20 |ig/m3). The variation in the responses for the effects of long-term exposures
largely reflects differences of opinion among the experts on a number of factors such as key
epidemiological results from long-term cohort studies, the likelihood of a causal relationship,
and the shape of the C-R function. Some observations concerning the outcome of the individual
expert judgments are provided below:

       Key Cohort Studies. The experts' non-zero responses for the  percent change in annual
mortality were mostly influenced by the Krewski et al., (2000) reanalysis of the original American
Cancer Society (ACS) cohort study and by the later Pope et al. (2002) update of the ACS study that
included  additional years of follow-up.  None of the experts placed  substantial weight on the
mortality estimates from the Six-Cities study (Dockery et al., 1993) in composing their quantitative
responses, despite citing  numerous strengths of that analysis.  Concern about sample size and
representativeness of the Six Cities study for the entire U.S. appeared to be a major reason for de-
emphasizing those results.

       Causality for Long-Term Effects. Three of the five experts gave distributions more heavily
weighted towards zero.  Those experts were also the ones who gave the lowest probability of a
causal effect of long-term exposure to PM2 5 in the preliminary questions. All of the experts placed
at least a 5 percent probability on the possibility that there is no causal relationship between fine PM
exposure and mortality; as a result, all experts gave a fifth percentile value for the C-R coefficient
of zero. For most of the experts, this was based primarily on residual concerns about the strength
of the mechanistic link between the exposures and mortality.

       Shape of the C-R Function for Long-Term Effects.  The other key determinant of each
expert's responses for long-term effects was his assumption about the nature of the C-R function
across the range of baseline annual average PM2 5 concentrations assumed in the pilot (8 to 20
|ig/m3).  Three experts (A, D,  and E) assumed that the function  relating mortality with  PM
concentrations would be log-linear with constant slope over the specified range. They therefore
gave a single estimate of the distribution of the slope describing that log-linear function. The other
two experts provided more complex responses.

       Expert B assumed a population threshold in his model, below which there would be no effect
of increased PM2 5 exposure and above which the relationship would be log-linear. He characterized
his estimate  of a possible threshold as uncertain, ranging between 4 |ig/m3 and 15 i-ig/m3, with a
modal value of 12 |ig/m3.  He then described a distribution for the slope for the log-linear function
                                         9-218

-------
                                                                  Cost-Benefit Analysis
that might exist above the threshold; this distribution is depicted in Figure 9B-2. The effect of
incorporating the uncertain threshold is essentially to shift his entire distribution downward.

       Expert C believed that the increased relative risks for mortality observed in the cohort studies
were likely to be the result of exposures at the higher end of the exposure range, and he expected
there to be a declining effect on mortality with decreasing levels of PM2 5. He also argued that some
practical concentration threshold was likely to exist below which we would not observe any increase
in mortality. He reflected these beliefs by developing a non-linear model within the range from 8
to 20 jig/m3; he described the model by providing distributions for the slope of the curve at four
discrete concentrations within the range.
                                          9-219

-------
Final Regulatory Impact Analysis
                                        Figure 9B-2.
                  Expert B's Distributions for the Percent Increase in Annual
                 Non-Accidental Mortality Associated with a 1 (J,g/m3 Increase
              in Long-term Exposures to PM2.5: Comparison of His Distribution
           Above a Threshold to His Expected Distribution* for the Range 8-20 (j,g/m3
               IO
               es
               Q.
               fO
               I
               0)
               Q.
               re
               r
               0
ncrease
1.5
  1 -
                     0.5
                                   +
                            Above threshold
                                Incorporating
                                 threshold**
                                            Expert B
                    * Expert B specified the threshold as uncertain between 4 and
             15 i-ig/m3 with a modal value at 12 |ig/m3. He assumed the percent
             increase in mortality to increase linearly with concentration above the
             threshold. His effective distribution was simulated using Monte Carlo
             techniques assuming an underlying distribution of population-weighted
             annual average PM25 concentrations for the U.S. generated from the
             BenMAP model (see the technical report (lEc, 2004) for details).
                                        9-220

-------
                    Cost-Benefit Analysis
9-221

-------
Final Regulatory Impact Analysis
       9B.3.3 Experts' Views of Sources of Uncertainty

       The experts were asked at several points during the interview to discuss the key sources of
potential bias and uncertainty in current evidence on which they relied for their judgments. In the
context of the quantitative discussion they were asked to list the top  five issues.  They were
encouraged to think about how these issues would affect the uncertainty surrounding their best
estimate of the potential impact on total mortality of a small change in long-term exposure to PM2 5.
The tables summarizing the factors identified by each expert may be found in Appendix E of the
technical report (lEc, 2004).

       Many of the same factors appeared in the list of the five experts. However, the experts often
differed on whether a particular factor was a source of potential bias or uncertainty. Some of the
common concerns raised as either sources of bias or uncertainty, include:

       •   Residual confounding by smoking,
       •   Residual confounding by "life-style" or other personal factors or "stressors,"
       •   Exposure errors/misclassification,
           The role of co-pollutants as confounders or effect modifiers,
       •   Impact of the relative toxicity of PM components,
       •   Representativeness of the cohort populations with respect to the general U. S.
           population, and
       •   Investigator/publication biases.

           Despite the many qualitative discussions about sources of uncertainty, because the
pilot study did not elicit quantitative judgments about the size and nature of impacts of each
source of uncertainty and bias, we were unable to systematically evaluate the nature of the
influence of these factors on the quantitative results provided by each expert unless an expert
explicitly adjusted his estimates by a particular factor.

       9B.3.4 Advisory Council Comments on the Preliminary Design of the Elicitation

       As  part of a review of the  analytical blueprint of the EPA's Second Prospective Analysis
of the Costs and Benefits of the Clean Air Act under section 812 of the Act, a panel of outside
experts - the Health Effects Subcommittee (HES) of the Advisory Council on Clean Air
Compliance Analysis (Council)pp - provided a limited™ and preliminary review of the
    pp The Council is an advisory committee with an independent statutory charter that is organized and supported
under the EPA's Science Advisory Board.

    QQ  Council/HES report: "...in view of the fact that the pilot project is well-underway, the experts have already
been selected, and many (if not all) of the interviews have been conducted, the HES sees little potential benefit in
providing detailed suggestions about the design or conduct of the pilot study." (EPA-SAB-COUNCIL-ADV-04-002,

                                           9-222

-------
                                                                 Cost-Benefit Analysis
methodology and design of the expert elicitation. In an Advisory issued by the Council to the
EPA (EPA-SAB-COUNCIL-ADV-04-002, March 2004), the Council-HES provided the
following comments with regard to the elicitation:

          "We applaud the Agency's interest in exploring the use of formal expert judgment as
          a tool for improving uncertainty analysis and believe that the proposed pilot study has
          great potential to yield important insights. The pilot is well designed to inform
          subsequent and more comprehensive expert elicitation projects, but relies on the
          opinions of a relatively small group of experts. It may provide preliminary
          information about the general magnitude of the mortality effects, and may yield a
          sense of both the uncertainty inherent in these estimates and the factors largely
          responsible for such uncertainty. However, until the pilot study methods and results
          have been subjected to peer review, it may be unwise for the Agency to rely directly
          on these preliminary results in key policy decisions."

       •   In presenting results of the pilot elicitation, "the HES advises the EPA to present the
          entire collection of individual judgments; to carefully examine the collection of
          individual judgments noting the extent of agreement or disagreement; to thoughtfully
          assess the reasons for any disagreement; and to consider formal combinations of
          judgments only after such deliberation and with full awareness of the context..."

          "The HES recognizes that in order to make the pilot tractable it was necessary to limit
          participation, and is aware of the many factors which must be balanced in the
          selections of expert panels (Hawkins and Graham, 1988), but is concerned about
          whether the judgments of such a limited group can reasonably be interpreted as
          representing a fair and balanced view of the current state of knowledge."

       9B.3.5 Limitations in Pilot Elicitation Design

          The pilot elicitation has afforded many opportunities for learning about expert
elicitation in the context of economic benefits analysis. However, because this was an initial
assessment that was limited in scope (as is discussed in section 9B.1), this section briefly
discusses some of the limitations in the design of the pilot.  Additional detail on the strengths and
weaknesses of the pilot are provided in the technical report (lEc, 2004).

       •   Short time-period to design and conduct the elicitation - The scope of the pilot was
          limited in order to complete the assessment and present our findings as part of the
          Final Nonroad Diesel Rule. Thus, there was a one-year time period in which were
March 2004, page 34).

                                         9-223

-------
Final Regulatory Impact Analysis
          designed the elicitation, conducted the interviews, and provided an interpretation of
          the results in this RIA and the technical report (lEc, 2004). In addition to designing
          the elicitation with specific limitations as are discussed below, the experts we given
          short notice of the elicitation (some experts were interested but not available in our
          time frame), and we were required to process the results rapidly to meet the
          rulemaking schedule.

          The design and implementation of the elicitation has not undergone a complete
          external peer review. While EPA is planning to conduct a peer review of the
          elicitation process, we were not able to complete the review prior to the promulgation
          of the final rule. The results of the pilot should be viewed tentatively until the full
          peer review is complete.

          Small panel of experts - Due  to resource constraints we limited the pilot to a panel of
          five experts. As noted above, the SAB-HES expressed their concern "about whether
          the judgements of such a limited group can reasonably be interpreted as representing
          a fair and balanced view of the current state of knowledge." They point to the many
          factors which must be balanced in the selection of expert panels (Hawkins and
          Graham, 1988)  and there are numerous opinions among a large set of experts.

          Little analytical research has  been conducted on the more difficult question of how to
          determine the ideal number of experts for a particular application. We have not found
          any analyses of the effect of expert panel size based on comparisons of empirical
          results of expert judgment studies. A theoretical analysis  by Clemen and Winkler
          (1985) suggests that where data sources are moderately positively dependent there are
          diminishing marginal returns to the value of information associated with each
          additional data source. In the context of expert judgment studies, such a result
          implies that when dealing with experts of similar backgrounds who rely on the same
          models and studies,  a larger expert panel may not provide  significantly higher quality
          results than a smaller one.  However, the addition of an expert expected to provide a
          more independent assessment, such as an expert from a different, but pertinent field,
          would be expected to exhibit a much greater value of information. Clemen and
          Winkler (1999) note that "heterogeneity among experts is highly desirable."  These
          findings would appear to support addressing complex issues using a panel comprised
          of relatively small subgroups (perhaps three to five experts each) from multiple
          disciplines. Although the decision analysis field tends to use relatively small sample
          sizes (i.e., typically 5-10 experts), some are not comfortable with obtaining a
          combined distribution from such small numbers in the absence of an apriori
          assessment of the degree to which the expert panel is likely to be representative of the
                                         9-224

-------
                                                                    Cost-Benefit Analysis
           overall population of relevant experts on the question of interest. The panel we used
           may not have captured the full range of reasonable opinions.

           Use of an aggregate elicitation question - The expert judgment literature discusses
           two broad approaches to elicitation of judgments; an aggregated and a disaggregated
           approach. As the term implies, an aggregated approach asks the expert to estimate
           the quantity of interest directly; for example, the numbers of newspapers sold in the
           U.S. in a particular year. In a disaggregated approach, the expert (or group of
           experts) would be asked to construct a model for estimating the quantity of interest
           and would be asked directly about the inputs to that model (e.g. population in each
           state, percentage of the population that reads newspapers, etc.) The intuition is that it
           is easier for experts to answer questions about the intermediate quantities than about
           the total quantity.

           The project team carefully considered the relative advantages and disadvantages of
           the two approaches. A major advantage of the disaggregated approach is a more
           structured and transparent characterization of the key inputs and sources of
           uncertainty in the final quantity of interest. However, the method does  require
           additional time and resources to develop a model structure (or in some cases, multiple
           models) and set of inputs on which the experts can agree prior to the individual
           elicitations.

           The limited time frame available to complete this assessment drove the  decision to
           undertake an aggregate approach to elicit the C-R coefficient for the PM2 5/mortality
           relationship." Nonetheless, a major goal of the preliminary and follow-up questions
           in the protocol was to identify critical issues that could be addressed through the
           development of a more disaggregated approach in a future assessment.

           Thus, the design of the pilot limits our ability to determine the influence of any one
           key factor over others in a large list of issues that the experts were to consider prior to
           answering the quantitative question. It also limited the ability of the experts to
           express their views about the difference in the C-R function based on the location in
           the U.S. (i.e., the demographics of the exposed population, the air concentration of
           PM and/or PM mixture).
       RR While the Project Team initially considered using a highly aggregated approach that would have asked
experts to characterize a single overall PM / mortality effect due to both short- and long-term exposures to PM2.5.
However, based on advice from the SAB-HES, we opted to disaggregate effects due to long- and short-term exposures.
The Project Team felt that separate questions to address effects of long- and short-term exposures, though still at a high
level of aggregation, would prove to be easier for experts to address than a question that "rolled up" all the effects into
a single estimate. This level of disaggregation also enabled the elicitation team to explore with experts possible overlap
in reported mortality effects detected using long-term and short-term epidemiological studies.

                                            9-225

-------
Final Regulatory Impact Analysis
          No workshop was conducted - It is customary to conduct a workshop prior to the
          elicitation interview with the experts. This allows the experts to become familiar
          with the protocol, the background materials contained in the briefing book, and to
          discuss methods to limit bias during the interview. Due to time constraints for the
          pilot, we did not conduct a pre-elicitation workshop.

          No calibration of experts - We do not have calibration measures that could be used to
          assess the results of this pilot. At this point, we can only assess the process - did the
          pilot assessment employ a structure, supporting materials, and a process that enabled
          experts to make judgments that would be likely to be well calibrated? The peer
          review for this aspect is still underway.  Nevertheless, without calibration measures,
          we cannot weight experts based on their performance on calibration tasks.

          Full-day elicitation - The elicitation interview with each expert took  a full-day to
          complete. Again, experts were given short notice of the elicitation and found time  in
          their schedules to participate, yet not all of the experts were available for the full-day
          interview. The length of the interview could lead to response fatigue that could affect
          the outcome of the experts'  response.
       9B.3.6 Combining the Expert Judgments for Application to Economic Benefit
Analyses

       Analysts must give careful thought to whether and how to combine the results individual
expert judgments into a single distribution.  When dealing with a small sample number of
experts, the analyst must be particularly careful to identify the influence of each expert's
response on the combined distribution. Therefore, we considered four alternative methods for
combining the pilot results. However, the Project Team identified significant issues associated
with each  of the methods.  In this section, we discuss the issues we considered in combining the
results of the pilot and how we came to the conclusion that for the illustrative benefits analysis
presented  in Section 9B.5 below, we would present both the individual quantitative distributions
of the C-R coefficient elicited from the five experts interviewed as well as results based on a
probabilistic estimate that represents the combined results of the pilot based on an equal
weighting of the calculated change in mortality incidence based on the individual judgments.

       9B.3.6.1 Background

       Combination  of expert judgments is not strictly necessary;  some investigators (e.g.,
Hawkins and Graham, 1990; Winkler and Wallsten, 1995; and Morgan et al.,  1984) have
preferred to keep expert opinions separate in order to preserve the  diversity of opinion on the
                                          9-226

-------
                                                                  Cost-Benefit Analysis
issues of interest. In such situations, the range of values expressed by the experts can help
decision-makers to understand the sensitivity of their analyses to the analytical model chosen,
thereby bounding possible outcomes. Individual judgments can also illustrate varying opinions
arising from different disciplinary perspectives or from the rational selection of alternative
theoretical models or data sets (Morgan and Henri on, 1990). Nonetheless, analysts are often
interested in developing a single distribution of values that reflects a synthesis of the judgments
elicited from a group of experts.

       There are also some advantages to combining the results across experts. An extensive
literature exists concerning methods for combining expert judgments.  These methods  can be
broadly classified as either mathematical or behavioral (Clemen and Winkler, 1999).
Mathematical approaches range from simple averaging of responses to much more complex
models incorporating information about the quality of expert responses, potential dependence
among expert judgments, or (in the case of Bayesian methods) prior probability distributions
about the variable of interest.  Behavioral approaches require the interaction of experts in an
effort to encourage them to achieve consensus, either through face-to-face meetings or through
the exchange of information about judgments among experts.  As noted in the technical report
(ffic, 2004), there are both methodological and practical issues arguing against a behavioral
approach. Therefore, we used a mathematical combination process to derive a single
distribution.

       One advantage of mathematical combination over behavioral approaches is the ability to be
completely transparent about how weights have been assigned to the judgments of specific experts
and about what assumptions have been made concerning the degree of correlation between experts.
Several approaches  can be used to assign weights to individual experts. Weights can be assigned
based on the analyst's opinion of the relative expertise of each expert; on a quantitative  assessment
of the calibration and informativeness (i.e., precision) of each expert based on their responses to a
set of calibration questions (as described in Cooke, 1991); or on weights assigned by each expert,
either to him or herself or to the other experts  on the panel (see Evans et al., 1994 for an example
of this approach). Ideally, such a weighting system would address problems of uneven calibration
and informativeness across experts, as well as potential motivational biases (Cooke,  1991).a In
practice, appropriate weights can be difficult to determine, though Cooke and others have conducted
considerable research on this issue.

       At the design stages of the pilot, we decided that the resulting  expert judgments would be
combined using equal weights, essentially calculating the arithmetic mean of the expert responses,
for simplicity and transparency.  The reasons for  choosing equal weights were both practical and
       A "Motivational bias" refers to the willful distortion of an expert's true judgments. The origins of this bias can
vary, but could include, for example, a reluctance to contradict views expressedby one's employer or a deliberate attempt
to skew the outcome of the study for political gain.

                                          9-227

-------
Final Regulatory Impact Analysis
methodological. Development of defensible differential  weights was not possible given the
expedited schedule for this project.  Although we did conduct a sample calibration exercise with
each expert, the purpose of the exercise was to train the experts in providing quantitative responses,
not to develop calibration scores that would be used to weight experts.  Some empirical evidence
suggests that the simple  combination rules, like equal weighting, perform equally well when
compared to more complex methods in terms of calibration scores for the combined results (Clemen
and Winkler, 1999).  The methods to combine the  expert judgments will be explicitly addressed
during the peer review of the pilot assessment.

       9B.3.6.2 Alternative Combination Methods

       While   a combination method using  equal weights for the results of each expert  is
straightforward in principle, applying it in this context of the results of the pilot was complicated
by the fact that the elicitation protocol  gave the experts freedom to specify  different forms for the
C-R function. If all the experts had chosen the same  form of the C-R function,(e.g., if each expert
had specified a log-linear  C-R function with a constant, but uncertain, C-R coefficient (i.e., slope)
over the PM range specified in the  protocol) the combination of their distributions for the C-R
coefficient would require a simple averaging across experts at each elicited percentile. However,  in
this assessment, three experts specified log-linear functions with constant C-R coefficients over the
specified range of PM2 5 concentrations,  and two of the experts  specified  the C-R coefficient  as
likely to vary  over the range of specified PM25 concentrations (as discussed in Section 9B.4.2
above).  These more complex C-R functions necessitated some additional steps in the calculation
of the combined results.

       As discussed in the technical report for the pilot (ffic, 2004), individual response either can
be  combined before application of the benefits model or during the application  of the model,
allowing  each expert's  C-R function to be  estimated in the  benefits  model independently.
Specifically, we derive the total mortality incidence for each expert, and  combine (or pool) the
estimates into an aggregate value before taking an average of the mortality incidence.  This  is
referred  to  as a "pooled" approach and  is used in our modeling framework for other benefit
endpoints that have multiple C-R function (due to multiple studies). We prefer the pooled approach
because it seems to reduce the amount  of alteration of the actual step-function responses provided
by Experts B and C (although some adjustments must still be made)b. Details of the illustration are
provided in  Section 9B.6.
    B Expert B specified a distribution for the C-R coefficient for PM2 5 concentrations above a threshold and
assigned the coefficient a value of zero for all PM concentrations below the threshold. He then specified a
probability distribution to describe the uncertainty about the threshold value. Expert C specified separate
distributions for the C-R coefficient at four discrete points within the concentration ranges defined in the protocol, to
represent a continuous C-R function whose slope varied with the PM2 5 concentration. Expert C indicated that the
coefficient value between these points was best modeled as a continuous function, rather than a step function. Both
experts assumed the same functional forms in responding to elicitation question.

                                           9-228

-------
                                                                   Cost-Benefit Analysis
       The alternative would be to combine the individual expert judgments into a single C-R
function before applying the results to the benefits model.  Below, we present three approaches
we considered for combining the expert judgments before applying the benefits model.  Among
the three approaches to combining expert judgments before the benefits analysis, the primary
difference is how they account for the underlying particulate air pollution levels. The first option
assumes a uniform distribution and equal weighting, which involves taking a simple average of
responses across experts for each percentile.  In a second combination method, we combined the
results using a normal distribution describing population-weighted annual average PM2 5
concentration data generated from EPA's Environmental Benefits and Mapping Analysis
Program (BenMAP), the model EPA currently uses for economic benefit analyses of air quality
regulations affecting PM and other criteria pollutants.0

       As discussed above, for the two of the experts that specified a C-R function that varied
over the range of PM concentrations., their estimated C-R function necessitated some additional
steps in the calculation of the combined results. To derive a single distribution across all experts
for a particular range  of exposures (e.g. 8-20 |ig/m3 annual average PM25), we first needed to
estimate an "effective" distribution of uncertainty about the C-R coefficient for both Experts B
and C across that range by using Monte Carlo simulation (Crystal Ball® software) to estimate the
expected value of each percentile elicited across the full PM2 5 range specified. Specifically, the
additional steps we took for this combination method are as follows:

       •  Expert B specified a distribution for the C-R coefficient fo rPM2.5 concentrations
           above a stated threshold and assigned the coefficient a value of zero for all PM
           concentrations below the threshold.  He then specified a probability distribution to
           describe the uncertainty about the threshold value.  Thus, we conducted Monte Carlo
           sampling using two distributions: his uncertainty distribution for the threshold, and an
           assumed distribution of baseline PM2 5 concentrations for the PM2 5 range specified in
          the elicitation protocol. On each iteration, we selected a value from each of these two
           distributions and compared them. If the selected baseline concentration was less than
           or equal to the selected threshold value, each of the percentiles of Expert B's
          uncertainty distribution was assigned a zero value (no mortality effect); if the
           concentration was greater than the threshold, we assigned each percentile the "above-
          the-threshold" value specified by Expert B in his interview.d We repeated this
       c To facilitate Monte Carlo sampling, we evaluated the fit of the BENMAP data to several distributional forms,
ultimately selecting a normal distribution, truncated at zero, with a mean of 11.04 |J.g/m3 and a standard deviation of 2.32
|ag/m3.

       D An example for mortality effects from long-term exposures helps illustrate this approach. Expert B estimated
that he was 75 percent sure (i.e., his 75th percentile) that the percent increase in mortality would be less than or equal to
0.5 percent per 1 |J.g/m3 change in PM2 5 concentration if the baseline concentration were above the threshold, but zero

                                           9-229

-------
Final Regulatory Impact Analysis
           process for thousands of iterations and then took the average value for each of the
           percentiles to obtain Expert B's "effective" distribution of uncertainty about the C-R
           coefficient across each range of exposures.

       •   Expert C specified separate distributions for the C-R coefficient at four discrete
           points within the concentration ranges defined in the protocol, to represent an
           continuous function whose slope varied with the PM concentration.  Thus, we first
           randomly sampled from the assumed distribution of baseline PM concentrations.  We
           then linearly interpolated between Expert C's responses at the two points nearest to
           the sampled PM concentration, to estimate his uncertainty distribution for the C-R
           coefficient at the sampled concentration.  For example, Expert C provided slope
           values at PM25 concentrations  of 8, 10, 15 and 20 for mortality effects of long-term
           exposure. If, on a given iteration we selected a PM2 5 concentration of 12 ng/m3, we
           would generate a  slope at each percentile of his uncertainty distribution by
           interpolating between Expert C's responses  at 10 and 15  |J,g/m3.  We repeated this
           process for thousands of iterations and then took the average value for each of the
           percentiles to obtain the "effective" distribution of the average slope of Expert C's C-
           R function.

       While the uniform distribution is the simplest method of combining the expert judgments,
it required us to alter the true  responses of Experts  B and C.  It is also based on a uniform
distribution, which does not match the observed PM2 5 concentrations that tend to be skewed toward
the lower concentration values.  The estimates of Expert B and C's "effective" distributions, and
thus the combined expert distribution, are all sensitive to the probability density function chosen to
describe the U.S. baseline PM25 concentrations in the simulations. This sensitivity arises because
both Experts B and C assume that the effect of an increase in PM2 5 concentration on mortality
depends on the initial PM25 concentration.  Table 9B-3 presents the resulting values of the
distribution for these two methods of combining the results of the pilot.
percent if it were below the threshold. If on a given iteration, the program selects a baseline concentration of 12 |ag/m3
and a threshold level of 10 |ag/m3, we assign his 75th percentile the value of 0.5. If the threshold level selected were 15
M.g/m3, the 75th percentile would be assigned a value of zero.

                                           9-230

-------
                                                                Cost-Benefit Analysis
        Table 9B-3. Methods for Combining Expert Judgments:  Combined C-R function
           with Uniform Distribution and a Population-Weighted Distribution



Percentiles


95th %ile
75th %ile
50th %ile
25th %ile
5th %ile
Minimum
Maximum
Combined
Expert Judgments using
a Uniform Distribution
of Baseline Annual
Mean PM2 5
Concentrations
1.05
0.65
0.33
0.17
0.00
0.00
1.71
Combined Expert
Judgments Based on
Population-Weighted
Distribution of Baseline
Annual Mean PM2 5
Concentrations in U.S.
0.93
0.59
0.3
0.16
0
0
1.5
       Given the differences in the responses given by Experts B and C at various levels of PM
concentrations (i.e., a conditional C-R function), we considered a third combination method in
which we calculate combined expert distributions at four different PM25 baseline concentrations.
Using the methods described above, we first calculated Expert B's and C's distributions at the
four concentration points and then averaged them with the distributions of the other three experts
(which remain constant over the concentration range) using equal weights. This method reduces
the level of adjustments that are made to Expert B's and C's response function in that we estimate
four C-R function for each individual, rather than one  smoothed function.  The functions for the
three other experts remain log-linear. Results of this combination method are provided in Table
9B-4.
                                         9-231

-------
Final Regulatory Impact Analysis
          Table 9B-4.  Combined Concentration-Response Function Conditional to PM
                                     Concentrations
Percentiles
95th percentile
75th percentile
50th percentile
25th percentile
5th percentile
8 ug/m3
0.82
0.56
0.30
0.16
0
12 ug/m3
0.99
0.61
0.30
0.16
0
15 ug/m3
1.08
0.64
0.30
0.16
0
20 ug/m3
1.20
0.76
0.42
0.24
0
          Overall, the combination methods considered result in fairly similar results at the
median and mean relative risk estimate. However, slight differences occur in the tails of the
distribution in their characterization of uncertainty. In figure 9B-2, the C-R function for the
population-weighted combination method was compared to the existing cohort epidemiological
studies of the long-term PM2 5/mortality relationship.  We observe that the results of the pilot
elicitation are generally within the range of findings from these epidemiological studies.
However, as expected, the elicitation results in a larger spread of uncertainty than is given by the
standard errors of the individual studies.
                                         9-232

-------
                                                        Cost-Benefit Analysis
   Figure 9B-2. Comparison of Combined Expert Judgment Distribution to Selected Published Studies
IO
CNI

D.
CO
E
t
o
s
_c
0)
co
0)
          2 --
        1 .5 --
        0.5 --
                 Com b in e d
                    Expert
                D istrib u tio n
Pope  e t a I., 1995
  Pope  e t a I.,
 Reanalysis, 63
citie s, K re w ski,
      2000
Pope  et a I., 2002
 Dockery etal.,
1993  Reana lysis,
 K re w ski, 2000
                                                  9-233

-------
Final Regulatory Impact Analysis
       9B.3.5 Limitations of Combining Expert Judgments

       Although we present several methods for combining the results of the pilot, there are
several limitations in interpreting the pilot results that should be considered.

          The conditional functions of Experts B and C required us to estimate some values on
          the C-R function between the points that were elicited, which requires an
          extrapolation from the response provided in the pilot to create continuous
          distributions.

       •   There are many methods available to combine the responses from the experts.  Each
          method has advantages and disadvantages from a statistical viewpoint.  The project
          team is not aware of any rule-of-thumb in statistics that would provide guidance for
          combining linear and non-linear functions. Therefore, we present four alternative
          methods for combining the results as an illustration of potential combinations of the
          results, and have asked for a peer review of these methods.

       •   In designing the pilot, there was a decision to combine the  results of the individual
          experts using an equal weighting.  In some elicitation studies, the authors use a
          calibration measure to weight the experts appropriately.  Because we did not conduct
          a calibration exercise, we present only an equal weighting of the responses.

       •   We have used a normal distribution to characterize the pilot results, but the
          distribution could potentially be skewed due to the bounding at zero. The C-R
          functions are bounded by zero, and anchored to one data source. There is a concern
          that the upper-end of the distribution resulting from the pilot may not fully reflect the
          available data and knowledge on the PM/mortality relationship.  There may have
          been some anchoring to the study results from the ACS cohort, and  less use of the
          Six-Cities study in the characterization of uncertainty upper-bounds. However, the
          experts were provided the Six Cities results in their briefing books as background
          material.

       9B.4. Illustrative Application of Pilot Expert Elicitation Results

          In this section, we apply the pilot expert elicitation results, using the pooled approach
discussed above for combining results across participants to the VSL  distribution discussed in
Chapter 9 (section 9.3.4), thereby providing an illustrative example of how one might translate
the results from the pilot elicitation into quantified estimates of economic benefits. The analysis
is based on the modeled air quality changes conducted for the preliminary nonroad  diesel control
option in 2030. As such, the results are comparable to the point estimates provided in Appendix

                                         9-234

-------
                                                                 Cost-Benefit Analysis
9A, but not to those in Chapter 9.  The values generated below do not reflect the Agency's
estimates of the benefits of the emissions reductions expected from the Final Non-Road Diesel
rule and are included solely as an illustration of the impacts of using expert elicitation based
distributions for premature mortality associated with long-term exposure to PM2 5 rather than a
data-derived distribution.

       9B.5.1 Method

       9B.5.1.1 Concentration-Response Distribution Based on Combined Results Across
Experts

       As discussed in Section 9B.4.5, we converted each expert's percentile responses about
mortality associated with long-term exposure into a custom distribution such that each percentile
is correctly represented and percentiles in between are represented as continuous functions
(custom distributions were generated using Crystal Ball and are represented as 15,000 equally
probable points).
       For experts A, D, and E, we used a standard log-linear functional form:

                                                        (4)
       where we set p equal to ln(l+B/100), where B is the percent change in all cause mortality
associated with a one |ig reduction in PM25. BenMAP then represents the distribution of Ay
based on the custom distribution of p.

          Expert C provided a set of conditional C-R functions for different baseline levels of
PM25.  Expert C provided four conditional responses, one for 8 |ig/m3, one for 10 |ig/m3, one for
15 |J.g/m3, and one for 20 |ig/m3.  In order to "fill-in" the C-R function for intermediate baseline
PM25 values, we linearly interpolated between the responses for each pair of points, e.g. 10 to 15
or 15 to 20.  We calculated interpolated values for 13 points, ranging from 8 |ig to 20 |ig. For
baseline values less than 8 |ig, we assigned a value of zero (essentially assuming a threshold at 8
|ig).  For baseline values greater than 20, we assigned the values provided by Expert C for 20 |ig.
This may result in an underestimate of the incidence of mortality for Expert C. For each of the
conditional functions, we used a log-linear specification, similar to A, D,  and E. Total incidence
of mortality for Expert C is the sum of the conditional estimates over the range of baseline air
concentrations.

          Expert B provided a log-linear C-R function, conditional on an unknown threshold
characterized by a triangular distribution bounded by 4 |ig and 15 |ig, with a mode at 12 |ig.  We
discretized the triangular distribution into 12 ranges of unit length (e.g. 4 to 5, 5 to 6, etc.) and
                                          9-235

-------
Final Regulatory Impact Analysis
calculated the expected value of the response at each population gridcell based on the observed
baseline PM2 5 and the probability of that baseline value exceeding the potential threshold. We
assume that if a grid cell has a baseline value above the threshold, then the full value of the
reduction in PM2 5 at that grid cell is associated with a reduction in mortality. This may result in
an overestimate of the mortality impact for Expert B because for grid cells where the baseline
level is only marginally above the threshold, a benefit might only accrue to the change in PM2 5
down to the threshold. The rest of the change would not result in any mortality reduction.
Because most of the changes in air quality are relatively small (population weighted change in
annual mean PM25 is -0.59 |ig), this should not be a large issue.

           To put these estimates in perspective, it is useful to summarize the projected baseline
(pre-nonroad diesel regulations) air quality in 2030. Table 9B-5 lists the population distribution
of baseline concentrations of PM25 in 2030:

                 Table 9B-5. Population Distribution of Baseline Ambient PM2 5
Baseline PM2 5 (jig/m3)
PM25<5
5
-------
                                                                  Cost-Benefit Analysis
values.  The Monte Carlo process was conducted using the estimated distribution for each expert
individually and for the combined (pooled) distribution, as well as for the distribution derived
from the Pope et al. (2002) study.

       9B.5.2 Results

       Figure 9B-4 presents box plots that display the distribution of the reduction in PM25
related premature mortality based on the concentration response distributions provided by each
expert, as well as that based on the pooled response.a For comparison, the figure also displays
the distribution derived from the statistical error associated with Pope et al (2002). The figure
shows that the average annual number of premature deaths avoided for the "modeled
preliminarily control option" ranges from approximately 4000 to 19,000, depending on the
concentration response function used.  The medians span zero to 16,000,  with the zero value due
to the low threshold associated with one of the expert's distributions. Specifically, because less
than a quarter of the population is expected to live in areas with PM2.5 levels above the
threshold specified by expert C, and much of the decrease in PM2.5 predicted by the preliminary
control option occurs  below that threshold, a much smaller decrease in premature morality is
predicted for expert C than those experts who provided continuos C-R functions down to zero
(PM2.5) as well as for expert B who provided an uncertain threshold. Furthermore, note that at
the 50th and 75th percentiles, the  C-R functions provided by all of the experts predict positive
benefits from the modeled control option.

       The boxplots displayed in Figure 9B-4 are derived by applying the C-R distributions
specified by each expert (as presented in Figure 9B-1) to the change in air quality predicted by
the preliminary non-road diesel control option.  Although the figures 9B-3 and 9B-1 show
similar patterns, there are important differences. Specifically, the ratio of 75th percentiles of the
C-R functions specified by experts A and B (as denoted in Figure 9B-1) is 0.4, whereas the ratio
of the predicted change in incidence of premature mortality associated with the modeled
preliminary control option is 0.5.  This 25% increase in the ratio suggests  a larger effective
difference in the distributions between the experts than was evident before applying the expert's
C-R functions to a predicted change in  air quality and  highlights the impact of the air quality
change predicted on the choice of C-R function used in the benefits analysis.

       The combined expert distribution depicted in Figure 9B-4 provides additional  insights.
The combined (average) distribution has a 90 percent credible interval between zero and 24,000.
When compared with  results derived from the Pope et al. (2002) study, it is clear that the
combined expert distribution reflects greater uncertainty about the estimated reduction in
A As discussed above, the elicitation results were combined assuming equal weight for each expert's distribution.
We assumed complete dependence of the expert's distributions for this illustrative analysis, so that each percentile of
the pooled distribution is simply the average of the corresponding percentiles of the 5 experts.

                                          9-237

-------
Final Regulatory Impact Analysis
premature mortality, as well as placing more weight on the lower end of the distribution. The
mean estimate from the combined expert distribution is almost 30 percent lower than the mean
derived from the Pope et al. (2002) distribution. However, the 90 percent confidence interval
based on the standard error from Pope et al. (2002) is completely contained within the 90 percent
credible interval of the combined expert distribution.

       Figure 9B-5 shows the same data using cumulative distribution functions (CDFs). This
figure is valuable for demonstrating differences in degree of certainty in achieving specific
reductions in premature mortality. For instance, the Pope et al. 2002 concentration response
distribution predicts a  20% chance that there will be at least 10,000 fewer premature deaths,
whereas the pooled distribution predicts a 60% chance of the same reduction in premature
deaths. The probabilities associated with the individual experts for avoiding 10,000 premature
deaths range from about 28% to 98%, demonstrating once again the sensitivity of the estimate to
assumptions regarding the concentration response function.       The CDFs of the estimated
reductions in premature mortality shows that for several experts, there is a small probability of a
substantially higher estimate.  For example, the 75th percentile of the distribution based on
Expert B's responses is at 8,800, while the 99th percentile for that distribution is almost 4 times
higher, at 34,000. The CDF also shows that while most of the experts provided fairly wide
distributions, reflecting a lack of confidence in the precision of the empirical data, the  CDF
based on Expert C's responses is much narrower, reflecting the high degree of confidence he
placed on the existence of a threshold below 15 |ig.

       Figures 9B-6 and 9B-7 use box plots and CDFs to display the estimated dollar value of these
annual reductions  in premature mortality. Whereas the  average based on the Pope  et al  2002
distribution is  $94 billion, the average based on the pooled estimate is $67 billion, a difference of
approximately one-third. Once the concentration response distributions are combined with the VSL
distributions, not only are the mean values closer to one another, but the distributions show
considerably more overlap.

          Because these distributions are the result of a Monte Carlo simulation combining the
non-normal distributions for reductions in mortality with a normal distribution for VSL, the
resulting distributions  will also be non-normal, but the shape depends on the skewness of the
input distribution of mortality reductions. For example, the ratio of the 95th to 75th percentile of
mortality reductions for Expert B is 3.1, while the same ratio for the value of mortality
reductions is 4.2, indicating the value distribution is more skewed than the reductions
distribution. In general, combining normal or left skewed distributions in a mulitplicative
fashion will result in left skewed distributions with greater skewness than the input distributions.
So even for the normally distributed estimates based on Pope et al. (2002), the value distribution
is somewhat skewed, because it is the result of multiplying two normally distributed random
variables.
                                          9-238

-------
                                                                   Cost-Benefit Analysis
           The shapes of the two distributions are more similar in this case because both reflect
the same additional information in the VSL distribution.  This demonstrates that as additional
sources of uncertainty are added to the analysis, the influence of any one source of uncertainty
will fall. Because VSL is a large source of uncertainty, the influence on overall uncertainty
relative to the distribution of the mortality reduction is also large. All of the distributions of the
value of mortality reductions have a small negative tail, this time due to propagation of the
normally distributed VSL, which has a small amount of the distribution below zero. Again, we
interpret this as a statistical artifact rather than a true probability that the value of a statistical life
is negative (implying that individuals would pay to increase the risk of death).

           We used additional Monte Carlo simulations to combine the expert-based
distributions for the dollar benefits of mortality with the distributions of dollar benefits for the
remaining health  and welfare endpoints to derive estimates of the overall distribution of total
dollar benefits'3.  The box plots for these distributions of overall dollar benefits associated with
the modeled nonroad diesel preliminary control options are presented in Figure B-8. Because
mortality accounts for over 90 percent of the benefits, the addition of other endpoints has little
impact on the overall distributions. The overall mean annual total dollar benefits in 2030 for the
distribution incorporating the combined expert distribution for reductions in premature mortality
is $70 billion, compared to $96 billion for the results derived from the Pope et al. (2002) study
for the nonroad diesel modeled preliminary control option.

           For clarity of presentation, in Figure 9B-9, we present CDFs for total dollar benefits
only for the combined expert distribution and results derived from the Pope et al. (2002) study.
These again suggest that the use of the expert elicitation based representation of uncertainty in
the relationship between PM2 5 and premature mortality has a large impact on the shape and
range of the distribution of total benefits.  The Pope et al. (2002) derived results have an
approximately Weibull shaped distribution with a range from 5th to 95th percentiles of $23 billion
to $190 billion, or about one order of magnitude.  The distribution of total  dollar benefits
incorporating the combined expert distribution for reductions in premature mortality has a much
more skewed shape with an elongated positive tail above the 75th percentile with a range from 5th
to 95th percentiles of $3 billion to $240 billion, or about two orders of magnitude.
    BNote that visibility benefits are treated as fixed for this illustrative analysis.  We are working on methods to
characterize the uncertainty in visibility and other non-health benefits.

                                           9-239

-------
          Figure 9B-4  Results of Illustrative Application of Pilot Expert Elicitation: Annual Reductions in
Premature Mortality in 2030 Associated with the Modeled Preliminary Control Option for the Nonroad Diesel Rule
  40,000
  35,000
Note: Distributions labeled Expert A - Expert E are based on individual
expert responses. The distribution labeled Combined Experts is based
on the averaged distributions of reduced incidence of premature mortality
across the set of experts. The distribution labeled Pope et al. (2002)
Statistical Error is based on the mean and standard error of the C-R
function from the study.
              Expert A       Expert B        Expert C        Expert D
                                                              Expert E       Combined      Pope et al
                                                                              Experts          (2002)
                                                                                           Statistical Error

-------
         Figure 9B-5.  Cumulative Distribution Functions for Annual Reductions in Premature Mortality in 2030 Associated
with the Nonroad Diesel Modeled Preliminary  Control Option
    0.1
                                                                                         Note: Distributions labeled Expert A - Expert E are based on individual
                                                                                         expert responses. The distribution labeled Combined Experts is based
                                                                                         on the averaged distributions of reduced incidence of premature mortality
                                                                                         across the set of experts. The distribution labeled Pope et al. (2002)
                                                                                         Statistical Error is based on the mean and standard error of the C-R
                                                                                         function from the study.
                         10,000
20,000               30,000                40,000               50,000

       Annual Reductions in Incidence of Premature Mortality in 2030
                                                             60,000
                                                      70,000
                — - — - Expert A — — Expert B
    - Expert C
Expert D  — -  - Expert E
-Combined Experts '
•Pope et al (2002) Statistical Error

-------
Final Regulatory Impact Analysis
      Figure 9B-6.  Results of Illustrative Application of Pilot Expert Elicitation: Dollar Value of Annual Reductions in
     Premature Mortality in 2030 Associated with the Modeled Preliminary Control Option for the Nonroad Diesel Rule
        $400,000
        $350,000
        $300,000
                 _
        $250,000
    o
    o
    o
    CM
    If*
     o  $200,000
     tfi
     c
     o
        $150,000
        $100,000
         $50,000
Note: Mortality distributions labeled Expert A - Expert E are based on
individual expert responses. The mortality distribution labeled Pooled
Expert Estimate is based on the averaged distributions of reduced
incidence of premature mortality across the set of experts. The
mortality distribution labeled Pope et al. (2002) statistical error is based
on the mean and standard error of the C-R function from the study.
Mortality valuation is based on a normally distributed VSL with a mean
of $5.5 million and a 95% Cl between $1 and $10 million. The VSL
distribution has then been adjusted for  income growth out to 2030 using
an adjustment factor of 1.23.
                                                                                                                                   $94,000
                      Expert A
                Expert B
Expert C
Expert D
Expert E      Combined Experts Pope et al (2002)
                               Statistical Error

-------
            Figure 9B-7. Cumulative Distribution Functions for Dollar Value of Annual Reductions in
     Premature Mortality in 2030 Associated with the Nonroad Diesel Modeled Preliminary Control Option
                                                                     Note: Mortality distributions labeled Expert A - Expert E are based on
                                                                     individual expert responses. The mortality distribution labeled Pooled
                                                                     Expert Estimate is based on the averaged distributions of reduced
                                                                     incidence of premature mortality across the set of experts. The
                                                                     mortality distribution labeled Pope et al. (2002) statistical error is based
                                                                     on the mean and standard error of the C-R function from the study.
                                                                     Mortality valuation is based on a normally distributed VSL with a mean
                                                                     of $5.5 million and a 95% Cl between $1 and $10 million. The VSL
                                                                     distribution has then been adjusted for income growth out to 2030 using
                                                                     an adjustment factor of 1.23.
0.1
  0
-$10,000               $90,000               $190,000              $290,000
                                                           Million ($2000)
  $390,000
$490,000
           - Expert A                           	Expert B
            Expert D                           	Expert E
            Pope et al (2002) Statistical Error
• Expert C
•Combined Experts

-------
 Figure 9B-8.  Results of Illustrative Application of Pilot Expert Elicitation: Dollar Value of Total Annual PM-related
 Health and Visibility Benefits in 2030 Associated with the Modeled Preliminary Control Option for the Tier 4 Rule
o
s
CM
   $400,000
   $350,000
   $300,000
   $250,000
   $200,000
   $150,000
   $100,000
    $50,000
          $0
 Note: All non-mortality distributions are based on classical statistical
- error derived from the standard errors reported in epidemiology
 studies and distributions of unit values based on empirical data.
 Visibility benefits are included as a constant. Mortality distributions
 labeled Expert A - Expert E are based on individual expert responses.
 The mortality distribution labeled Pooled Expert Estimate is based on
I the averaged distributions of reduced incidence of premature mortality
 across the set of experts. The mortality distribution labeled Pope et
 al. (2002) statistical error is based on the mean and standard error of
 the C-R function from the study.
                          $80,000
                                          $47,000
                                                            $32,000
                                                                             $69,000
                                                                                              $130
                                                                                    000
                                                                                                                $70,000
                                                                                                                                 $96,000
                  Expert A         Expert B         Expert C          Expert D          Expert E      Pooled Expert    Pope et al 2002
                                                                                                        Estimate      Statistical Error

-------
  Figure 9B-9.  Cumulative Distribution Functions of Dollar Value of Total Annual PM-related Health and Visibility
                Benefits in 2030 Associated with the Nonroad Diesel Modeled Preliminary Control Option
   0.9
   0.8
   0.7
1 0.6
   0.5
I 0.4
O
   0.3
   0.2
   0.1
     0
                                                                   Note: All non-mortality distributions are based on classical statistical error
                                                                   derived from the standard errors reported in epidemiology studies and
                                                                   distributions of unit values based on empirical data. Visibility benefits are
                                                                   included as a constant. Mortality distributions labeled Expert A - Expert E
                                                                   are based on  individual expert responses.  The mortality distribution labeled
                                                                   Pooled Expert Estimate is based on the averaged distributions of reduced
                                                                   incidence of premature mortality across the set of experts. The mortality
                                                                   distribution labeled Pope et al. (2002) statistical error is based on the mean
                                                                   and standard  error of the C-R function from the study.
-$6,000       $44,000       $94,000      $144,000     $194,000     $244,000     $294,000
                                            Annual Total Benefits in 2030 (Million $2000)
                                                                                                      $344,000      $394,000      $444,000
                                             •Combined Experts
                                                                    Pope et al (2002) Statistical Error

-------
Final Regulatory Impact Analysis
       9B.5.3 Limitations of the Application of the Pilot Elicitation Results to
the Nonroad Scenario

          The results presented in this section should be viewed cautiously given the limited
scope of the pilot, and the limitations of the elicitation design and methods used to combine the
expert judgments discussed above. Therefore, the results presented above should be considered
"illustrative" until both the peer review of the pilot is complete and the methods used to interpret
and apply the results of the pilot have been peer-reviewed and accepted.  Until this occurs, we do
not recommend applying this method in other regulatory analyses.

          Specific limitations of the illustrative application include:

       •   Extrapolation of percentile responses provided by individual experts. Each expert
          provided minimum and maximum values, as well as the 5th, 25th, 50th, 75th, and 95th
          percentiles. In order to generate the continuous distributions of mortality impacts, we
          had to make assumptions about the continuity of the distributions between the
          reported percentiles.  This adds uncertainty to the results.

       •   Interpolation of C-R relationship across PM2 5 levels.  Expert C provided a set of
          conditional distributions of the C-R relationship conditioned on the baseline level of
          PM25.  Because he only provided functions for a limited number of baseline levels,
          we had to interpolate the values between  levels, introducing additional uncertainty.
          In addition, Expert C provided no information on the  C-R function for baseline PM2 5
          levels below 8 |ig/m3 or above 20 |ig/m3. We assumed no mortality impacts for
          baseline levels lower than 8 and no increase in the C-R  function above 20. This
          likely biased our results downward.

       •   Interpretation of Expert B results.  Expert B provided a  conditional distribution for
          the C-R function, conditioned on an uncertain threshold. Expert B provided
          additional information about the shape of the distribution for the threshold.  To
          develop an applied function, we assumed that the uncertain threshold could be
          incorporated into the C-R function through the construction of an expected value
          function.  The specific functions may lead to a slight overestimate of mortality
          impacts.
          Use of simple averaging of expert results.  To develop the combined expert
          distribution, we used equal weights for each expert.  Given the lack of calibration
          questions in the pilot elicitation, this is the most defensible approach. However,
          many expert elicitation applications have use more complex weighting schemes based
          on how well experts are calibrated.

                                         9-246

-------
                                                        Cost-Benefit Analysis
Ranges based on individual experts should be viewed with caution as they represent
only a single individual's interpretation of the state of knowledge about PM and
mortality. Results for individual experts should not be extracted and presented
without reference to the full range of results across the five experts.

Any range of results presented based on this application should be presented along
with their relative likelihood (i.e., the percentile represented in the distribution).
                                9-247

-------
Final Regulatory Impact Analysis
       References for Appendix 9B
       Amaral, D. 1983. Estimating Uncertainty in Policy Analysis: Health Effects from
inhaled Sulfur Oxides. Ph.D. Thesis, Department of Engineering and Public Policy, Carnegie
Mellon University, Pittsburgh, PA.

       Ayyub, B.M. Elicitation of Expert Opinions for Uncertainty and Risks. CRC Press,
Florida, 2002..

       Cooke, R. Experts in Uncertainty. Oxford University Press, New York, 1991.

       Evans, J.S., J.D. Graham, G.M. Gray, R.L. Sielken 1994a. "A distributional approach to
characterizing low-dose cancer risk." Risk Analysis 14(1): 25-34.

       Evans, J.S., G.M Gray, R.L. Sielken, Jr., A.E. Smith, C. Valdez-Flores, and J.D. Graham.
1994b. "Use of Probabilistic Expert Judgment in Uncertainty Analysis of Carcinogenic Potency,"
Regulatory Toxicology and Pharmacology. 20:25-36.

       Hawkins, N.C. and J.S. Evans 1989. "Subjective Estimation of Toluene Exposures: A
Calibration Study of Industrial Hygienists" Applied. Ind. Hygiene 4:61-68.
       Hawkins, N.C. and J.D. Graham 1988 "Expert scientific judgment and cancer risk
assessment: a pilot study of pharmacokinetic data.": Risk Analysis 8(4): 615-625.
       Krewski et al., D., S.N. Ras, J.M. Zielmski, and P.K.  Hopke 1999. "Characterization of
uncertainty and variability in residential radon cancer risks." Ann. N.Y. Acad. Sci. 895:245-272.

       Manne, A.S.  and R.G. Richels, 1994. "The Costs of stabilizing global CO2 emissions: a
probabilistic analysis based on expert judgments." The Energy Journal 15(l):31-56.

       McCurdy, T  and H. Richmond, 1983. Description of the OAQPS Risk Program and the
ongoing Lead NAAQS Risk Assessment Project. Paper 83-74.1. Presented at the  76th Annual
Meeting of the Air Pollution Control Association, June 19-24, Atlanta, Georgia. As cited in NAS.

       Morgan, G.  and M.  Henrion, Uncertainty; A Guide to  Dealing with  Uncertainty in
Quantitative Risk and Policy Analysis, Cambridge University Press, Cambridge.

       Morgan, M.G.  and D.W Keith 1995. "Subjective  judgments  by climate  experts."
Environmental Science and Technology 29: 468A-476A.

       Nordhaus, W.D.  1994." Expert Opinion on Climatic Change. "American Scientist. 82:45-51.

       North, W.  and M.W. Merkhofer 1976.  "A methodology for analyzing emission control
strategies." Comput.  Oper. Res. 3:187-207.
                                         9-248

-------
                                                               Cost-Benefit Analysis
       Reilly, J., P.H. Stone, C.E. Forest, M.D. Webster,  H.E. Jacoby, and R.G. Prinn. 2001.
Climate Change. Uncertainty and climate change assessments. Science 293(5529):430-433.

       U.S. Nuclear Regulatory Commission.  1996. "Branch Technical Position on the Use of
Expert Elicitation in the High-Level Radioactive Waste Program." November, 1996.

       Walker, K.D., P. Catalano, J.K. Hammitt, and J.S. Evans. 2003. "Use of expert judgment in
exposure assessment: Part 2. Calibration of expert judgments about personal exposures to benzene."
J Expo Anal Environ Epidemiol. 13(1):1-16.

       Walker, K.D.,  J.S.  Evans, D. Macintosh. 2001. "Use of expert judgment in exposure
assessment. Part 1.  Characterization of personal exposure to benzene."J Expo Anal Environ
Epidemiol. ll(4):308-22.

       Whitfield and Wallsten 1989.  "A risk assessment for selected lead-induced health effects:
an example of a general methodology." Risk Analysis, 9(2): 197-207.

       Whitfield, R.G., T.S.  Wallsten, R L. Winkler, H.M. Richmond, and S.R Hayes.  1991.
Assessing the Risks of Chronic Lung Injury Attributable to Long-Term Ozone Exposure. Argonne
National Laboratory Report ANL/EAIS-2.  NTIS/DE91016814. Argonne, IL. July.

       Winkler, R.L., T.S. Wallsten, R.G Whitfield,  H.M. Richmond, S.R. Hayes, and A.S.
Rosenbaum. 1995. An assessment of the risk of chronic lung injury attributable to long-term ozone
exposure. Operations Research, Vol.  43 (1), pp. 19-28.

       Wright, G. and P. Ayton (eds.) 1994. Subjective Probability. John Wiley, Chi Chester.
                                         9-249

-------
Final Regulatory Impact Analysis
       APPENDIX 9C: Sensitivity Analyses of Key Parameters in the Benefits Analysis

          The primary analysis is based on our current interpretation of the scientific and
economic literature. That interpretation requires judgments regarding the best available data,
models, and modeling methodologies; and assumptions we consider most appropriate to adopt in
the face of important uncertainties.  The majority of the analytical assumptions used to develop
the Base Estimate have been reviewed and approved by EPA's Science Advisory Board (SAB).
However, we recognize that data and modeling limitations as well as simplifying assumptions
can introduce significant uncertainty into the benefit results and that alternative choices exist for
some inputs to the analysis, such as the mortality C-R functions.

          We supplement our primary estimates of benefits with a series of sensitivity
calculations that make use of other sources of health effect estimates and valuation data for key
benefits categories. These estimates examine sensitivity to both valuation issues (e.g. the
appropriate income elasticity)  and for physical effects issues (e.g., possible recovery from chronic
illnesses). These estimates are not meant to be comprehensive. Rather, they reflect some of the
key issues identified by EPA or commentors as likely to have a significant impact on total
benefits. Individual adjustments in the tables should not be added together without addressing
potential issues of overlap and low joint probability among the endpoints.

       9C.1 Premature Mortality—Long term exposure

          Given current evidence regarding their value,  reduction in the risk of premature
mortality is the most important PM-related health outcome in terms of contribution to dollar
benefits. There are at least three important analytical assumptions that may significantly impact
the estimates of the number and valuation of avoided premature mortalities. These include
selection of the C-R function, structure of the lag between reduced exposure and reduced
mortality risk, and effect thresholds.  Results of this set of sensitivity analyses are presented in
Table 9C.1.

       9C.1.1 Alternative C-R Functions

          Following the advice of the EPA Science Advisory Board Health Effects
Subcommittee (SAB-HES), we used the Pope, et al. (2002) all-cause mortality model exclusively
to derive our primary estimate of avoided premature mortality. While the SAB-HES
"recommends that the base case rely on the Pope et al. (2002) study and that EPA use total
mortality concentration-response functions (C-R), rather than separate cause-specific C-R
functions, to calculate total PM mortality cases," they also suggested that "the cause-specific
estimates can be used to communicate the relative contribution of the main air pollution related
causes of death." As such, we provide the estimates of cardiopulmonary and lung cancer deaths
based on the Pope et al. (2002).

          In addition, the SAB-HES has noted that the American Cancer Society cohort used in
Pope et al. (2002) "has some inherent deficiencies, in particular the imprecise exposure data, and
the non-representative (albeit very  large) population. Thus, ACS is not necessarily "the better
                                         9-250

-------
                                                                 Cost-Benefit Analysis
study," but, at this point in time, is a prudent choice for the base case estimates in the Second
Prospective Analysis. The Harvard Six-Cities C-R functions are valid estimates on a more
representative, although geographically selected, population, and its updated analysis has not yet
been published. The Six Cities estimates may be used in a sensitivity analysis to demonstrate that
with different but also plausible selection criteria for C-R functions, benefits may be considerably
larger than suggested by the ACS study." (EPA-SAB-COUNCIL-ADV-04-002). In previous
advice, the SAB has noted that "the [Harvard Six Cities] study had better monitoring with less
measurement error than did most other studies"  (EPA-SAB-COUNCIL-ADV-99-012, 1999).
The demographics of the ACS study population, i.e., largely white and middle-class, may also
produce a downward bias in the estimated PM mortality coefficient, because short-term studies
indicate that the effects of PM tend to be significantly greater among groups of lower
socioeconomic  status.  The Harvard Six Cities study also covered a broader age category (25 and
older compared to 30 and older in the ACS study) and followed the cohort for a longer period (15
years compared to 8 years in the ACS study). We emphasize, that based on our understanding of
the relative merits of the two datasets, the Pope, et al. (2002) ACS model based on mean PM2 5
levels in 63 cities is the most appropriate model for analyzing the premature mortality impacts of
the nonroad  standards.  It is thus used for our base estimate of this important health effect.

       9C.1.2 Alternative Lag Structures

          As noted by the SAB (EPA-SAB-COUNCIL-ADV-00-001, 1999), "some of the
mortality effects of cumulative exposures will occur over short periods of time in individuals
with compromised health status, but other effects are likely to occur among individuals who, at
baseline, have reasonably good health that will deteriorate because of continued exposure. No
animal models have yet been developed to quantify these cumulative effects, nor are there
epidemiologic studies bearing on this question." However, they also note that "Although there is
substantial evidence that a portion of the mortality effect of PM is manifest within a short period
of time, i.e.,  less than one year, it can be argued that, if no lag assumption is made, the entire
mortality excess observed in the cohort studies will be analyzed as immediate effects, and this
will result in an overestimate of the health benefits of improved air quality. Thus some time lag is
appropriate for  distributing the cumulative mortality effect of PM in the population." In the
primary analysis, based on previous SAB advice, we assume that mortality occurs over a five
year period,  with 25 percent of the deaths occurring in the first year, 25 percent in the second
year, and 16.7 percent in each of the third, fourth, and fifth years.  Readers should note that the
selection of a 5  year lag is not supported by any scientific literature on PM-related mortality
(NRC 2002). Rather it is intended to be a reasonable guess at the appropriate distribution of
avoided incidences of PM-related mortality. The SAB-HES has recently noted that "empirical
evidence is lacking to inform the choice of the lag distribution directly and agrees with the NAS
report that there is little empirical justification for the 5-year cessation lag structure used in  the
previous analyses."  The SAB-HES suggests that appropriate lag structures may be developed
based on the distribution of cause specific deaths within the overall all-cause estimate.  Diseases
with longer progressions should be characterized by longer term lag structures, while air
pollution impacts occurring in populations with existing disease may be characterized by shorter
term lags.
                                          9-251

-------
Final Regulatory Impact Analysis
          A key question is the distribution of causes of death within the relatively broad
categories analyzed in the long-term cohort studies. While we may be more certain about the
appropriate length of cessation lag for lung cancer deaths, it is not at all clear what the
appropriate lag structure should be for cardiopulmonary deaths, which include both respiratory
and cardiovascular causes.  Some respiratory diseases may have a long period of progression,
while others, such as pneumonia, have a very short duration.  In the case of cardiovascular
disease, there is an important question of whether air pollution is causing the disease,  which
would imply a relatively long cessation lag, or whether air pollution is causing premature death in
individuals with preexisting heart disease, which would imply very short cessation  lags.  The
SAB-HES provides several recommendations for future research that could support the
development of defensible lag structures, including the use of disease specific lag models, and the
construction of a segmented lag distribution to combine differential lags across causes of death.
The SAB-HES indicated support for using "a Weibull distribution or a simpler distributional
form made up of several segments to cover the response mechanisms outlined above,  given our
lack of knowledge on the specific form  of the distributions."  However, they noted  that "an
important question to be resolved is what the relative magnitudes of these segments should be,
and how many of the acute effects are assumed to be included in the cohort effect estimate."
They conclude their discussion of cessation lags by stating that "given the current lack of direct
data upon which to  specify the lag function, the HES recommends that this question be
considered for inclusion in future expert elicitation efforts and/or sensitivity analyses." (EPA-
SAB-COUNCIL-ADV-04-002)  EPA  will continue to investigate this important issue for future
benefits analyses and in the upcoming 2nd Prospective Analysis of the Costs and Benefits of the
Clean Air Act. For this RIA, we investigate alternative cessation lag structures as senstivity
analyses, noting that these might be as likely as the previous 5-year distributed lag in  the base
analysis.

          Although the prior SAB recommended the five-year distributed lag be used for the
primary analysis, the SAB has also recommended that alternative lag structures be explored  as a
sensitivity analysis (EPA-SAB-COUNCIL-ADV-00-001,  1999). Specifically, they  recommended
an analysis of 0, 8, and 15 year lags.  The 0 year lag is representative of EPA's assumption in
previous RIAs. The 8 and 15 year lags  are based on the study periods from the Pope, et al.
(1995) and Dockery, et al. (1993) studies, respectively0. However, neither the Pope, et al. or
Dockery, et al. studies assumed any lag structure when estimating the relative risks  from PM
exposure.  In fact, the Pope, et al. and Dockery, et al. studies do not contain any data either
supporting or refuting the existence of a lag. Therefore, any  lag structure applied to the avoided
incidences estimated from either of these studies will be an assumed structure.  The 8  and 15 year
lags implicitly assume that all premature mortalities occur at the end of the study periods, i.e. at 8
and 15 years.

          In addition to the simple 8 and 15 year lags, we have added an additional senstivity
analysis examining  the impact of assuming a segmented lag of the type suggested by the SAB-
    GAlthough these studies were conducted for 8 and 15 years, respectively, the choice of the duration of the study
by the authors was not likely due to observations of a lag in effects, but is more likely due to the expense of
conducting long-term exposure studies or the amount of satisfactory data that could be collected during this time
period.

                                          9-252

-------
                                                                   Cost-Benefit Analysis
HES. This illustrative lag structure is characterized by 20 percent of mortality reductions
occuring in the first year, 50 percent occuring evenly over years 2 to 5 after the reduction in
PM2 5, and 30 percent occurring evenly over the years 6 to 20 after the reduction in PM2 5.  The
distribution of deaths over the latency period is intended to reflect the contribution of short term
exposures in  the first year, cardiopulmonary deaths in the 2 to 5 year period, and longer term lung
disease and lung cancer in the 6 to 20 year period. For future analyses, the specific distribution
of deaths over time will need to be determined through research on causes of death and
progression of diseases associated with air pollution. It is important to keep in mind that changes
in the lag assumptions do not change the total number of estimated deaths, but rather the timing
of those deaths.

          The estimated impacts of alternative lag structures on the monetary benefits associated
with reductions in PM-related premature mortality (estimated with the Pope et al. ACS impact
function) are presented in Table 9C.2. These estimates are based on the value of statistical lives
saved approach, i.e. $5.5 million per incidence, and are presented for both a 3 and 7 percent
discount rate over the lag period.

       9C.1.3 Thresholds

          Although the consistent advice from EPA's Science Advisory Board has been to
model premature mortality associated with PM exposure as a non-threshold effect, that is,  with
harmful effects to exposed populations regardless of the absolute level of ambient PM
concentrations, some analysts have hypothesized the presence of a threshold relationshipd.  The
nature of the  hypothesized relationship is that there might exist a PM concentration level below
which further reductions no longer yield premature mortality reduction benefits.6 EPA does not
necessarily endorse any particular threshold and, as discussed in Appendix 9A, virtually every
study to consider the issue indicates  absence of a threshold.

          We construct a senstivity analysis by assigning different cutpoints below which
changes in PM2 5  are assumed to have no impact on premature mortality.  The sensitivity analysis
illustrates how our estimates of the number of premature mortalities in the Base Estimate might
change under a range of alternative assumptions for a PM mortality threshold. If, for example,
there were no benefits of reducing PM concentrations below the PM2 5 standard of 15 |ig/m3, our
estimate of the total number of avoided PM-related premature mortalities in 2030 from the
preliminary modeling would be reduced by approximately 70 percent, from approximately
    DThe most recent advice from the SAB-HES is characterized by the following: "For the studies of long-term
exposure, the HES notes that Krewski et al. (2000) have conducted the most careful work on this issue. They report
that the associations between PM2.5 and both all-cause and cardiopulmonary mortality were near linear within the
relevant ranges, with no apparent threshold. Graphical analyses of these studies (Dockery et al., 1993, Figure 3 and
Krewski et al., 2000, page 162) also suggest a continuum of effects down to lower levels. Therefore, it is reasonable
for EPA to assume a no threshold model down to, at least, the low end of the concentrations reported in the studies."
    EThe illustrative example in Appendix 9B presents the potential implications of assuming some probability of a
threshold on the benefits estimate.

                                           9-253

-------
Final Regulatory Impact Analysis
14,000 annually to approximately 4,000 annually. However, this type of cutoff is unlikely, as
supported by the recent NRC report, which stated that "for pollutants such as PM10 and PM2 5,
there is no evidence for any departure of linearity in the observed range of exposure, nor any
indiciation of a threshold. (NRC, 2002)"  Another possible senstivity analysis which we have not
conducted at this time might examine the potential for a nonlinear relationship at lower exposure
levels/

          One important assumption that we adopted for the threshold sensitivity analysis is that
no adjustments are made to the shape of the C-R function above the assumed threshold. Instead,
thresholds were applied by simply assuming that any changes in ambient concentrations below
the assumed threshold have no impacts on the incidence of premature mortality.  If there were
actually a threshold, then the shape of the C-R function would likely change and there would be
no health benefits to reductions in PM below the threshold.  However, as noted by the NRC, "the
assumption of a zero slope over a portion of the curve will force the slope in the remaining
segment of the positively sloped concentration-response function to be greater than was indicated
in the original study" and that "the generation of the steeper slope in the remaining portion of the
concentration-response function may fully offset the effect of assuming a threshold." The NRC
suggested that the treatment of thresholds should be evaluated in a formal uncertainty analysis.

          The results of these sensitivity analyses demonstrate that choice of effect estimate can
have a large impact on benefits, potentially doubling benefits if the effect estimate is derived
from the HEI reanalysis of the Harvard Six-cities data (Krewski et al., 2000). Due to discounting
of delayed benefits, the lag structure may also have a large impact on monetized benefits,
reducing benefits by 30 percent if an extreme assumption that no effects occur until after 15 years
is applied. The overall impact of moving from the 5-year distributed lag to a segmented lag is
relatively modest, reducing benefits by approximately 8 percent when a three percent discount
rate is used and 22 percent when a seven percent discount rate is used. If no lag is assumed,
benefits are increased by around five percent. The threshold analysis indicates that
approximately 85 percent of the premature mortality related benefits are due to changes in PM25
concentrations occurring above  10 i-ig/m3, and around 30 percent are due to changes above 15
l-ig/m3, the current PM2 5 standard.
   FThe pilot expert elicitation discussed in Appendix 9B provides some information on the impact of applying
nonlinear and threshold based C-R functions.

                                          9-254

-------
                                     Table 9C-1.
          Sensitivity of Benefits of Premature Mortality Reductions to Alternative
Assumptions (Relative to Base Case Benefits of Modeled Preliminary Control Option)


Description of Sensitivity
Analysis

Avoided
IncidencesA

20
20

20
30
Value (million
2000$)B

20
20

20
30
Alternative Concentration-Response Functions for PM-related Premature Mortality
Pope/ACS Study (2002)c

Lung Cancer
Cardiopulmonary
Krewski/Harvard Six-city Study

I
200
6,
000
17
,000


2,100
11
,000
30
,000

$7
,700
$3
7,000
$1
10,000

$1
3,000
$6
7,000
$1
90,000
Alternative Lag Structures for PM-related Premature Mortality
N
one
8-
year




1
5 -year




S
egmented





Incidences all occur in
the first year
Incidences all occur in
the 8th year
3% Discount Rate

7% Discount Rate

Incidences all occur in
the 15th year
3% Discount Rate

7% Discount Rate

20 percent of
incidences occur in 1st year,
50 percent in years 2 to 5, and
30 percent in years 6 to 20
3% Discount Rate

7% Discount Rate
7,
800


7,
800
7,
800


7,
800
7,
800




7,
800
7,
14
,000


14
,000
14
,000


14
,000
14
,000




14
,000
14
$5
2,000


$4
2,000
$3
2,000


$3
4,000
$2
0,000




$4
5,000
$3
$9
4,000


$7
6,000
$6
2,000


$6
2,000
$3
6,000




$8
2,000
$6

-------
Final Regulatory Impact Analysis
Alternative Thresholds
No Threshold (base estimate)
5
10
15
20
25
7,
800
7,
800
6,
300
1,
700
63
0
19
0
14
,000
14
,000
12
,000
4,
000
1,
300
52
0
$4
9,000
$4
9,000
$4
0,000
$1
1,000
$4
,000
$1
,200
$8
9,000
$8
9,000
$7
7,000
$2
6,000
$8,
400
$3,
400
       A Incidences rounded to two significant digits.
       B Dollar values rounded to two significant digits.
       c Note that the sum of lung cancer and cardiopulmonary deaths will not be equal to the total all cause death
estimate.  There is some residual mortality associated with long term exposures to PM2 5 that is not captured by the
caridopulmonary and lung cancer categories.
       9C.2 Other Health Endpoint Sensitivity Analyses

       9C.2.1 Overlapping Endpoints

       In Appendix 9 A, we estimated the benefits of the modeled preliminary control options
using the most comprehensive set of endpoints available.  For some health endpoints, this meant
using a health impact function that linked a larger set of effects to a change in pollution, rather
than using health impact functions for individual effects. For example, for premature mortality,
we selected an impact function that captured reductions in incidences due to long-term exposures
to ambient concentrations of particulate matter, assuming that most incidences of mortality
associated with short-term exposures would be captured.  In addition, the long-term exposure
premature mortality impact function for PM2 5 is expected to capture at least some of the
mortality effects associated with exposure to ozone.

           In order to provide the reader with a fuller understanding of the health effects
associated with reductions in air pollution associated with the preliminary control options, this
set of sensitivity estimates examines those health effects which, if included in the primary
estimate, could result in double-counting of benefits. For some endpoints, such as ozone
mortality, additional research is needed to provide separate estimates of the effects for different
pollutants, i.e. PM and ozone. These supplemental estimates should not be considered as additive
to the total estimate of benefits, but illustrative of these issues and uncertainties. Sensitivity
estimates  included in this  appendix include premature mortality associated with short-term
exposures to ozone,  and acute respiratory symptoms in adults. Results of this set of sensitivity
                                           9-256

-------
                                                                  Cost-Benefit Analysis
analyses are presented in Table 9.C-3.

          There has been a great deal of research recently on the potential effect of ozone on
premature mortality. While the air pollutant most clearly associated with premature mortality is
particulate matter, with dozens of studies reporting such an association, repeated ozone exposure
is a likely contributing factor for premature mortality, causing an inflammatory response in the
lungs which may predispose elderly and other sensitive individuals to become more susceptible.
The findings of three recent analyses provide consistent data suggesting that ozone exposure is
associated with increased mortality. Although the National Morbidity, Mortality, and Air
Pollution Study (NMMAPS) did not find an effect of ozone on total mortality across the full
year, Samet et al. (2000), who conducted the NMMAPS study, did observe an effect after
limiting the  analysis to summer when ozone levels are highest.  Similarly, Thurston and Ito
(1999) have shown associations between ozone and mortality. Toulomi et al. (1997) found that
1-hour maximum ozone levels  were associated with daily numbers of deaths in 4 cities (London,
Athens, Barcelona, and Paris),  and  a quantitatively similar effect was found in a group of 4
additional cities (Amsterdam, Basel, Geneva, and Zurich). Fairly  et al. (2003) also found a
relatively strong association between maximum 8-hour average ozone concentrations and
mortality in Santa Clara County, CA, even after controlling for PM2 5 exposure.

          While not as  extensive as the data base for particulate matter, these recent studies
provide supporting evidence for inclusion of mortality in the ozone health benefits analysis.  A
recent analysis by Thurston and Ito (2001) reviewed previously published time series studies of
the effect of daily ozone levels on daily mortality and found that previous EPA estimates of the
short-term mortality benefits of the ozone NAAQS (U.S. EPA, 1997) may have been
underestimated by up to a factor of two.  Thurston and Ito hypothesized that much of the
variability in published estimates of the ozone/mortality effect could be explained by how well
each model controlled for the influence of weather, an important confounder of the
ozone/mortality effect, and that earlier studies using less sophisticated approaches to controlling
for weather consistently under-predicted the ozone/mortality effect.

          Thurston and Ito (2001) found that models incorporating a non-linear temperature
specification appropriate for the "U-shaped" nature of the temperature/mortality relationship
(i.e., increased deaths at both very low and very high temperatures) produced ozone/mortality
effect estimates that were both  more strongly positive (a two percent increase in relative risk
over the pooled estimate for all studies evaluated) and consistently statistically significant.
Further accounting for the interaction effects between temperature and relative humidity
produced even more strongly positive results.  Inclusion of a PM index to control for
PM/mortality effects had little effect on these results, suggesting an ozone/mortality relationship
independent of that for PM. However, most of the studies examined by Thurston and Ito only
controlled for PM10  or broader  measures of particles  and did not directly control for PM25. As
such, there may still be potential for confounding of PM25 and ozone mortality effects, as ozone
and PM2 5 are highly correlated during summer months in some  areas.

          A recent World Health Organization (WHO) report found that "recent
epidemiological  studies  have strengthened the evidence that there are short-term O3 effects on
                                          9-257

-------
Final Regulatory Impact Analysis
mortality and respiratory morbidity and provided further information on exposure-response
relationships and effect modification." (WHO, 2003). Based on a preliminary meta-analysis, the
WHO report suggests an effect estimate of between 0.2 and 0.4 percent increase in premature
death per 10 |ig/m3 increase in 1 hour maximum ozone and between 0.4 and 0.6 percent increase
in premature death per 10 |ig/m3 increase in daily average.  This is equivalent to a relative risk of
between 1.04 and 1.08 per 100 ppb increase in 1 hour maximum and between 1.08 and 1.12 per
100 ppb increase in daily average. The WHO report provides effect estimates for both all
seasons and summer seasons.  Because our analysis is limited to the summer ozone season, the
most appropriate effect estimate is for the summer season.  The WHO summer season relative
risk estimate is 1.08 per 100 ppb increase in 1 hour maximum ozone and 1.12 per 100 ppb
increase in daily average ozone.

          Levy et al. (2001) assessed the epidemiological evidence examining the link between
short term exposures to ozone and premature mortality. Based on four U.S. studies (Kellsall et
al., 1997; Moolgavkar et al., 1995; Ito and Thurston, 1996; and Moolgavkar, 2000), they
conclude that an appropriate pooled effect estimate is a 0.5 percent increase in premature deaths
per 10 |ag/m3 increase in 24-hour average ozone concentrations, with a 95 percent confidence
interval between 0.3 percent and 0.7 percent.  This is equivalent to a relative risk of 1.10 per 100
ppb increase in daily average, which falls in the middle of the range of relative risks from the
WHO report.  Levy et al. also note that there are a number of studies which did not report a
quantitative effect estimate but did indicate that ozone was insignificant. They suggest that the
uncertainty surrounding the ozone-mortality effect estimate is greater than that reflected in the
confidence interval around their pooled estimate.

          In its September 2001 advisory on the draft analytical blueprint for the second
Section 812 prospective analysis, the SAB Health Effects Subcommittee (HES) cited the
Thurston and Ito study as a significant advance in understanding the effects of ozone on daily
mortality and recommended re-evaluation of the ozone mortality endpoint for inclusion in the
next prospective study (EPA-SAB-COUNCIL-ADV-01-004, 2001). Based on these new
analyses and recommendations,  EPA is sponsoring three independent meta-analyses of the
ozone-mortality epidemiology literature to inform a determination on inclusion of this important
health endpoint. Publication of these meta-analyses will significantly enhance the scientific
defensibility of benefits estimates for ozone which include the benefits of premature mortality
reductions. In its 2003  review of the analysis plans for the second Prospective Analysis, the
HES indicated support for EPAs new meta-analyses  of the ozone mortality literature and EPA's
plans to consider adding ozone mortality to the base  case analysis, subsequent to review of the
results of the meta-analyses.  Thus, recent evidence suggests that by not including an estimate of
reductions in short-term mortality due to changes in ambient ozone, the Base Estimate may
underestimate the benefits of implementation of the Nonroad Diesel Engine rule.

          The ozone mortality sensitivity estimate is calculated using results from four U.S.
studies (Ito and Thurston, 1996; Kinney et al., 1995; Moolgavkar  et al., 1995; and Samet et al.,
1997), based on the assumption that demographic and environmental conditions on average
would be more similar between these studies and the conditions prevailing when the nonroad
standards are implemented. We include the Kinney et al., 1995 estimate for completeness, even
                                         9-258

-------
                                                                 Cost-Benefit Analysis
though Levy et al. (2001) reject the results because the study only included a linear term for
temperature. Because the Kinney et al. (1995) study found no significant effect of ozone, this
has the effect of reducing the estimated mortality impacts and increasing the uncertainty
surrounding the estimated mortality reductions. We combined these studies using probabilistic
sampling methods to estimate the impact of ozone on mortality incidence.  The technical support
document for this analysis provides additional details of this approach (Abt Associates, 2003).
The estimated  incidences of short-term premature mortality are valued using the value of
statistical lives saved method, as described in Appendix 9A.

                                          Table 9C-2.
                  Sensitivity Estimates for Potentially Overlapping EndpointsA
Description of Sensitivity
Analysis




Avoided
Incidences


20
20
20
30
Monetized
Value
(Million
2000$)
20
20
20
30
Mortality from Short-term Ozone Exposure8
Ito and Thurston( 1996)

Kinney etal. (1995)
Moolgavkar et al. (1995)

Sametetal. (1997)

Pooled estimate (random effects
weights)
44
0
0
77

12
0
94

1,
000
0
24
0
36
0
28
0
$2
,900
$0
$5
10
$7
90
$6
20
$6,
800
$0
$1,
600
$2,
400
$1,
900
Any of 19 Acute Respiratory Symptoms, Adults 18-64 (Krupnick et al. 1990)
Ozone

PM

1,
500,000
14
,000,000
2,
800,000
19
,000,000
$3
8
$3
40
$7
1
$4
90
       A All estimates rounded to two significant digits.
       B Mortality valued using Base estimate of $5.5 million per premature statistical death,
adjusted for income growth.
                                         9-259

-------
Final Regulatory Impact Analysis
       9C.2.2 Alternative and Supplementary Estimates

          We also examine how the value for individual endpoints or total benefits would
change if we were to make a different assumption about specific elements of the benefits
analysis. Specifically, in Table 9C.3, we show the impact of alternative assumptions about other
parameters, including treatment of reversals in chronic bronchitis as lowest severity cases,
alternative impact functions for PM hospital and ER admissions, valuation of residential
visibility, valuation of recreational visibility at Class I areas outside of the study regions
examined in the Chestnut and Rowe (1990a, 1990b) study, and valuation of household soiling
damages.
                                         9-260

-------
                                   Cost-Benefit Analysis
             Table 9C-3.
Additional Parameter Sensitivity Analyses

Alternative
Calculation





1






2





3





4
~




5
_/

Reversal
s in chronic
bronchitis
treated as
lowest severity
cases



Value of
visibility
changes in all
Class I areas


Value of
visibility
changes in
Eastern U.S.
residential areas

Value of
visibility
changes in
Western U.S.
residential areas

Househ
old soiling
damage


Description of Estimate


Instead of omitting cases of
chronic bronchitis that reverse after
a period of time, they are treated as
being cases with the lowest severity
rating. The number of avoided
chronic bronchitis incidences in
2020 increases from 4,300 to 8,000
(87%). The increase in 2030 is
from 6,500 to 12,000 (87%).
Values of visibility changes
at Class I areas in California, the
Southwest, and the Southeast are
transferred to visibility changes in
Class I areas in other regions of the
country.
Value of visibility changes
outside of Class I areas are
estimated for the Eastern U.S.
based on the reported values for
Chicago and Atlanta from
McClelland etal. (1990).
Value of visibility changes
outside of Class I areas are
estimated for the Western U.S.
based on the reported values for
Chicago and Atlanta from
McClelland etal. (1990).
Value of decreases in
expenditures on cleaning are
estimated using values derived
from Manuel, et al. (1983).
Impact on Base Benefit
Estimate (million 2000$)

2020
+$730
(+1.4%)







+$640
(+1.2%)




+$700
(+1.3%)




+$530
(+1.0%)




+$170
(+0.3%)



2030
+$1,10
0(+1.2%)







+$970
(+1.1%)




+$1,10
0(+1.1%)




+$830
(+0.9%)




+$260
(+0.3%)


             9-261

-------
Final Regulatory Impact Analysis
          An important issue related to chronic conditions is the possible reversal in chronic
bronchitis incidences (row 1 of Table 9C-3). Reversals are defined as those cases where an
individual reported having chronic bronchitis at the beginning of the study period but reported
not having chronic bronchitis in follow-up interviews at a later point in the study period.  Since,
by definition, chronic diseases are long-lasting or permanent, if the disease goes away it is not
chronic. However, we have not captured the benefits of reducing incidences of bronchitis that
are somewhere in-between acute and chronic.  One way to address this is to treat reversals as
cases of chronic bronchitis that are at the lowest severity level. These cases thus get the lowest
value for chronic bronchitis.

          The alternative calculation for recreational visibility (row 2 of Table 9C-3) is an
estimate of the full value of visibility in the entire region affected by the nonroad emission
reductions. The Chestnut and Rowe study from which the primary valuation estimates are
derived only examined WTP for visibility changes  in the southeastern portion of the affected
region. In order to obtain estimates of WTP for visibility changes in the northeastern and central
portion of the affected region, we have to transfer the southeastern WTP values.  This introduces
additional uncertainty into the estimates.  However, we have taken steps to adjust the WTP
values  to account for the possibility that a visibility improvement in parks in one region, is not
necessarily the same environmental quality good as the same visibility improvement at parks in a
different region.  This may be due to differences in the scenic vistas at different parks,
uniqueness of the parks, or other factors,  such as public familiarity with the park resource. To
take this potential difference into account, we adjusted the WTP being transferred by the ratio of
visitor  days in the two regions.

          The alternative calculations for residential visibility (rows 3 and 4 of Table 9C-3) are
based on the McClelland, et al. study of WTP for visibility changes in Chicago and Atlanta. As
discussed in Appendix 9A, SAB advised EPA that  the residential visibility  estimates from the
available literature are inadequate for use in a primary estimate in a benefit-cost analysis.
However, EPA recognizes that residential visibility is likely to have some value and the
McClelland, et al. estimates are  the most useful in providing an estimate of the likely magnitude
of the benefits of residential visibility improvements.

          The alternative calculation for household soiling (row 5  of Table 9C-3) is based on
the Manuel,  et al. study of consumer expenditures on cleaning and household maintenance. This
study has been cited as being "the only study that measures welfare benefits in a manner
consistent with economic principals (Desvouges et al., 1998)." However, the data used to
estimate household soiling damages in the Manuel, et al. study are from a 1972 consumer
expenditure survey and as such may not accurately represent consumer preferences in 2030.
EPA recognizes this limitation, but believes the Manuel, et al. estimates are still useful in
providing an estimate of the likely magnitude of the benefits of reduced PM household soiling.

       9C.3 Income Elasticity of Willingness to Pay

          As discussed in Appendix 9A, our estimate of monetized benefits accounts for growth
in real  GDP per capita by adjusting the WTP for individual endpoints based on the central
                                          9-262

-------
                                                                  Cost-Benefit Analysis
estimate of the adjustment factor for each of the categories (minor health effects, severe and
chronic health effects, premature mortality, and visibility). We examine how sensitive the
estimate of total benefits is to alternative estimates of the income elasticities. Table 9C-4 lists
the ranges elasticity values used to calculate the income adjustement factors, while Table 9C-5
lists the ranges of corresponding adjustement factors.  The results of this sensitivity analysis,
giving the monetized benefit subtotals for the four benefit categories, are presented in Table 9C-
6.

          Consistent with the impact of mortality on total benefits, the adjustment factor for
mortality has the largest impact on total benefits.  The value  of mortality ranges from 81  percent
to  150 percent of the primary estimate based on the lower and upper sensitivity bounds on the
income adjustment factor.  The effect on the value of minor and chronic health effects is  much
less pronounced, ranging from 93 percent to 111 percent of the primary estimate for minor
effects and from 88 percent to 110 percent for chronic effects.
                                          9-263

-------
Final Regulatory Impact Analysis
                                              Table 9C-4.
          Ranges of Elasticity Values Used to Account for Projected Real Income GrowthA
Benefit Category
Minor Health Effect
Severe and Chronic
Health Effects
Premature Mortality
Visibility8
Lower Sensitivity
Bound
0.04
0.25
0.08
—
Upper Sensitivity
Bound
0.30
0.60
1.00
—
        A Derivation of these ranges can be found in Kleckner and Neumann (1999) and Chestnut (1997). Cost of Illness
(COI) estimates are assigned an adjustment factor of 1.0.
        B No range was applied for visibility because no ranges were available in the current published literature.

                                              Table 9C-5.
        Ranges of Adjustment Factors Used to Account for Projected Real Income GrowthA
Benefit
Category
Minor Health
Effect
Severe and
Chronic Health
Effects
Premature
Mortality
Visibility8
Lower Sensitivity Bound
2020
1.018
1.121
1.037
—
2030
1.021
1.139
1.043
—
Upper Sensitivity Bound
2020
1.147
1.317
1.591
—
2030
1.170
1.371
1.705
—
        A Based on elasticity values reported in Table 9A-11, US Census population projections, and projections of real gross
domestic product per capita.
        B No range was applied for visibility because no ranges were available in the current published literature.
                                              9-264

-------
                                                                   Cost-Benefit Analysis
                                        Table 9C-6.
                 Sensitivity Analysis of Alternative Income ElasticitiesA
Benefit Category
Minor Health Effect
Severe and Chronic
Health Effects
Premature Mortality
Visibility and Other
Welfare EffectsA
Total Benefits
Lower Sensitivity
Bound
2020
$510
$2,50
0
$42,0
00
$1,40
0
$47,0
00
2030
$760
$3,90
0
$75,0
00
$2,20
0
$82,0
00
Upper Sensitivity
Bound
2020
$540
$2,80
0
$65,0
00
$1,40
0
$70,0
00
2030
$810
$4,40
0
$123,
000
$2,20
0
$131,
000
1 All estimates rounded to two significant digits.
! No range was applied for visibility because no ranges were available in the current published literature.
                                        9-265

-------
Final Regulatory Impact Analysis
       Appendix 9C References

       Abt Associates, Inc. 2003. ProposedNonroad Landbased Diesel Engine Rule: Air
          Quality Estimation, Selected Healthand Welfare Benefits Methods, and Benefit
          Analysis Results. Prepared for Office of Air Quality Planning and Standards, U.S.
          EPA. April, 2003.
       Alberini, A., M.  Cropper, A. Krupnick, and N.B. Simon. 2002. Does the Value of a
          Statistical Life Vary with Age and Health Status? Evidence from the United States
          and Canada.  Resources for the Future Discussion Paper 02-19. April.
       Blumenschein, K. and M. Johannesson. 1998. "Relationship Between Quality of Life
          Instruments,  Health State Utilities, and Willingness to Pay in Patients with Asthma."
          Annals of Allergy, Asthma, and Immunology 80:189-194.
       Chestnut, L.G.  1997. Draft Memorandum: Methodology for Estimating Values for
          Changes in Visibility at National Parks. April 15.
       Chestnut, L.G. and R.D. Rowe. 1990a. Preservation Values for Visibility Protection at
          the National Parks: Draft Final Report. Prepared for Office of Air Quality Planning
          and Standards, US Environmental Protection Agency, Research Triangle Park, NC
          and Air Quality Management Division, National Park Service, Denver, CO.
       Chestnut, L.G., and R.D. Rowe. 1990b. A New National Park Visibility Value Estimates.
          In Visibility and Fine Particles, Transactions of an AWMA/EPA International
          Specialty Conference, C. V. Mathai, ed. Air and Waste Management Association,
          Pittsburgh.
       Desvousges, W.H., F. R. Johnson, H.S. Banzhaf. 1998.  Environmental Policy Analysis
          With Limited Information: Principles and Applications of the Transfer Method (New
          Horizons in Environmental Economics.) Edward Elgar Pub: London.
       Dockery, D.W., C.A. Pope, X.P. Xu, J.D. Spengler, J.H. Ware, M.E. Fay, E.G. Ferris and
          F.E. Speizer. 1993. "An association between air pollution and mortality in six U.S.
          cities." New England Journal of Medicine. 329(24): 1753-1759.
       EPA-SAB-COUNCIL-ADV-00-001, 1999. The Clean Air Act Amendments (CAAA)
          Section 812 Prospective Study of Costs and Benefits (1999): Advisory by the Health
          and Ecological Effects Subcommittee on Initial Assessments of Health and
          Ecological Effects; Part 2. October.
       EPA-SAB-COUNCIL-ADV-99-012, 1999. The Clean Air Act Amendments (CAAA)
          Section 812 Prospective Study of Costs and Benefits (1999): Advisory by the Health
          and Ecological Effects Subcommittee on Initial Assessments of Health and
          Ecological Effects; Part 1. July.
       EPA-SAB-COUNCIL-ADV-01-004. 2001. Review of the  Draft Analytical Plan for
          EPA's Second Prospective Analysis - Benefits and Costs of the Clean Air Act
          1990-2020: An Advisory by a Special Panel of the Advisory Council on Clean Air
          Compliance Analysis. September.
       Ito, K. and G.D.  Thurston. 1996.  "Daily PMlO/mortality associations: an investigations
          of at-risk subpopulations." Journal of Exposure Analysis and Environmental
          Epidemiology 6(1): 79-95.
       Jones-Lee, M.W. 1989. The Economics of Safety and Physical Risk. Oxford: Basil
Blackwell.
                                        9-266

-------
                                                         Cost-Benefit Analysis
Jones-Lee, M.W., G. Loonies, D. O'Reilly, and P.R. Phillips. 1993. The Value of
   Preventing Non-fatal Road Injuries: Findings of a Willingness-to-pay National
   Sample Survey. TRY Working Paper, WP SRC2.
Kinney, P.L., K. Ito and G.D. Thurston. 1995. A Sensitivity Analysis of Mortality Pm-
   10 Associations in Los Angeles. Inhalation Toxicology  7(1): 59-69.
Kleckner, N. and J. Neumann.  1999. Recommended Approach to Adjusting WTP
   Estimates to Reflect Changes in Real Income. Memorandum to Jim Democker, US
   EPA/OPAR, June3.
Krewski D, Burnett RT, Goldbert MS, Hoover K, Siemiatycki J, Jerrett M,
   Abrahamowicz M, White WH.  2000.  Reanalysis of the Harvard Six Cities Study and
   the American Cancer Society Study of Particulate Air Pollution and Mortality.
   Special Report to the Health Effects Institute, Cambrid.ge MA, July 2000.
Krupnick, A., M. Cropper., A. Alberini, N. Simon, B. O'Brien, R. Goeree, and M.
   Heintzelman. 2002. Age, Health and the Willingness to Pay for Mortality Risk
   Reductions: A Contingent Valuation Study of Ontario Residents, Journal of Risk and
   Uncertainty, 24, 161-186.
Manuel, E.H., R.L. Horst, K.M. Brennan, W.N. Lanen, M.C. Duff and J.K. Tapiero.
   1982. Benefits Analysis of Alternative Secondary National Ambient Air Quality
   Standards for Sulfur Dioxide and Total Suspended Particulates, Volumes I-IV.
   Prepared for U.S. Environmental Protection Agency, Office of Air Quality Planning
   and Standards. Research Triangle Park, NC.
McClelland, G., W. Schulze, D. Waldman, J. Irwin, D. Schenk, T. Stewart, L. Deck and
   M.  Thayer. 1991. Valuing Eastern Visibility: A Field Test of the Contingent
   Valuation Method. Prepared for U.S. Environmental Protection Agency, Office of
   Policy, Planning and Evaluation. June.
McDonnell, W.F., D.E. Abbey, N. Nishino and M.D. Lebowitz. 1999. Long-term
   ambient ozone concentration and the incidence of asthma in nonsmoking adults: the
   ahsmog study. Environmental Research. 80(2 Pt 1): 110-21.
Moolgavkar, S.H., E.G.  Luebeck, T.A. Hall and E.L. Anderson.  1995. Air Pollution and
   Daily Mortality in Philadelphia.  Epidemiology 6(5): 476-484.
National Research Council  (NRC). 2002. Estimating the Public Health Benefits of
   Proposed Air Pollution Regulations. The National Academies Press: Washington,
   D.C.
O'Connor, R.M. and G.C. Blomquist.  1997. Measurement of Consumer-Patient
   Preferences Using a Hybrid Contingent Valuation Method.  Journal of Health
   Economics. Vol. 16: 667-683.
Ostro, B.D., M.J. Lipsett, M.B. Wiener and J.C. Seiner. 1991. Asthmatic Responses to
   Airborne Acid Aerosols. American Journal of Public Health 81(6): 694-702.
Pope, C.A., M.J. Thun, M.M. Namboodiri, D.W.  Dockery, J.S. Evans, F.E. Speizer and
   C.W. Heath.  1995. Parti culate air pollution as a predictor of mortality in a
   prospective study of U.S. adults. American Journal of Respiratory Critical Care
   Medicine 151(3): 669-674.
Pope, C.A., III, R.T. Burnett, M.J. Thun, E.E. Calle, D. Krewski, K. Ito, G.D. Thurston.
   2002. Lung Cancer,  Cardiopulmonary Mortality, and Long-term Exposure to Fine
   Particulate Air Pollution. Journal of the American Medical Association. 287:  1132-
                                  9-267

-------
Final Regulatory Impact Analysis
          1141.
       Samet, J.M., S.L. Zeger, I.E. Kelsall, J. Xu and L.S. Kalkstein.  1997.  Air Pollution,
          Weather, and Mortality in Philadelphia 1973-1988. Health Effects Institute.
          Cambridge, MA. March.
       Scultze. W. 2003. Personal Communication. January.
       Smith, V.K., M.F. Evans, H. Kim, and D.H. Taylor, Jr. 2003. Do the "Near" Elderly
          Value Mortality Risks Differently? Review of Economics and Statistics
          (forthcoming).
       Thurston, G.D. and K. Ito. 2001. Epidemiological studies of acute ozone exposures and
          mortality.  J Expo Anal Environ Epidemiol. Vol. 11(4): 286-94.
       U.S. EPA. 1997. Regulatory Impact Analyses for the Particulate Matter and Ozone
          National Ambient Air Quality Standards and Proposed Regional Haze Rule.  U.S.
          EPA, Office of Air Quality Planning and Standards. Research Triangle Park, NC.
          July.
       US Environmental Protection Agency, 2000. Valuing Fatal Cancer Risk Reductions.
          White Paper for Review by the EPA Science Advisory Board.
       Woodruff, T.J., J. Grillo and K.C. Schoendorf. 1997. The relationship between  selected
          causes of postneonatal infant mortality and parti culate air pollution in the United
          States. Environmental Health Perspectives. 105(6): 608-612.
                                         9-268

-------
APPENDIX 9D: Visibility Benefits Estimates for Individual Class I
   Areas
                              Table
         Apportionment Factors for 2020
9D-1
Park Specific Visibility Benefits
PARK
Shenandoah
Anaconda-Pintlar W
Boundary Waters
Breton W
Isle Royale
Jarbidge W
Medicine Lake W
Red Rock Lakes W
Roosevelt Campobello
Selway-Bitterroot W
Seney W
Wolf Island W
Agua Tibia W
Black Canyon of the
Caribou W
Chiricahua
Cucamonga W
Dome Land W
Flat Tops W
Grand Canyon
Hoover W
John Muir W
Kaiser W
La Garita W
Mazatzal W
Mesa Verde
Petrified Forest
Pine Mountain W
Pinnacles
Point Reyes
Rawah W
Rocky Mountain
Saguaro
San Gabriel W
San Gorgino W
San Jacinto W
San Rafael W
Sequoia-Kings
Sycamore Canyon W
Ventana W
Yolla-Bolly-Middle-
COUNTY
Lawrence
Cochise Co
Gila Co
Gila Co
Coconino
Apache Co
Apache Co
Graham Co
Pima Co
Maricopa
Coconino
Yavapai Co
Tuolumne
San
Calaveras
Trinity Co
Fresno Co
Mono Co
Inyo Co
Marin Co
Los
Monterey
San Benito
Riverside
Siskiyou
San
Del Norte
Shasta Co
Fresno Co
Lassen Co
Riverside
San Diego
Shasta Co
El Dorado
Mariposa
Fresno Co
Tuolumne
Tulare Co
Siskiyou
Santa
Tulare Co
STAT
E
AL
AZ
AZ
AZ
AZ
AZ
AZ
AZ
AZ
AZ
AZ
AZ
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
Percent ot 2020 Visibility Benelit Due to
Changes in:
SO2 NOx direct PM
0.428
0.337
0.396
0.396
0.336
0.469
0.469
0.302
0.224
0.061
0.336
0.216
0.090
0.074
0.049
0.367
0.051
0.195
0.145
0.060
0.099
0.071
0.057
0.040
0.469
0.074
0.518
0.146
0.051
0.285
0.040
0.068
0.146
0.050
0.085
0.051
0.090
0.052
0.469
0.111
0.052
0.234
0.061
0.054
0.054
0.053
0.049
0.049
0.038
0.061
0.014
0.053
0.140
0.580
0.158
0.520
0.239
0.101
0.302
0.098
0.577
0.143
0.563
0.633
0.314
0.220
0.158
0.097
0.469
0.101
0.347
0.314
0.497
0.469
0.487
0.374
0.101
0.580
0.478
0.220
0.156
0.478
0.338
0.602
0.550
0.550
0.612
0.481
0.481
0.660
0.715
0.924
0.612
0.644
0.330
0.768
0.432
0.394
0.848
0.504
0.757
0.363
0.758
0.366
0.310
0.646
0.311
0.768
0.385
0.385
0.848
0.368
0.646
0.435
0.385
0.463
0.541
0.848
0.330
0.470
0.311
0.733
0.470

-------
PARK
Yo Semite
Carlsbad Caverns
GilaW
Joyce Kilmer- Slickrock
Kalmiopsis W
Linville Gorge W
Lostwood W
Pecos W
Presidential Range-Dry
Salt Creek W
Shining Rock W
Wheeler Peak W
Wichita Mountains W
Fitzpatrick W
Glacier Peak W
Mount Adams W
Dolly Sods W
North Absaroka W
Olympic
Lye Brook W
Bridger W
Goat Rocks W
Otter Creek W
Pasayten W
Bandelier
Bosque del Apache W
Brigantine W
Crater Lake
Mount Hood W
Mount Washington W
San Pedro Parks W
Swanguarter W
Theodore Roosevelt
Maroon Bells-
Mount Rainier
North Cascades
Bob Marshall W
Gates of the Mountain
Glacier
St. Marks W
Voyageurs
Teton W
Yellowstone
Grand Teton NP
Washakie W
STAT
COUNTY -?
E
Modoc Co CA
San Juan CO
GarfieldCo CO
Routt Co CO
Larimer Co CO
Pitkm Co CO
Alamosa CO
Gunnison CO
Montezuma CO
Montrose CO
Summit Co CO
Mineral Co CO
Larimer Co CO
Monroe Co FL
WakullaCo FL
Citrus Co FL
Charlton GA
Mclntosh GA
Edmonson KY
Stone Co MS
Hyde Co NC
Haywood NC
Avery Co NC
Graham Co NC
Sandoval NM
Rio Arriba NM
Grant Co NM
Chaves Co NM
Mora Co NM
Eddy Co NM
Socorro Co NM
Taos Co NM
Lincoln Co NM
Elko Co NV
Polk Co TN
Blount Co TN
San Juan UT
Grand Co UT
San Juan UT
Washington UT
GarfieldCo UT
Botetourt VA
Madison VA
Grant Co WV
Tucker Co WV
Percent ot 2020 Visibility Benefit Due to
Changes in:
SO2 NOx direct PM
0.277
0.522
0.335
0.420
0.449
0.425
0.458
0.437
0.353
0.355
0.525
0.589
0.449
0.546
0.535
0.416
0.543
0.500
0.415
0.539
0.344
0.476
0.516
0.564
0.426
0.512
0.414
0.471
0.568
0.417
0.409
0.538
0.603
0.311
0.405
0.384
0.373
0.354
0.373
0.219
0.295
0.485
0.385
0.533
0.568
0.407
0.114
0.246
0.140
0.120
0.098
0.097
0.152
0.077
0.175
0.042
0.048
0.120
0.020
0.048
0.148
0.058
0.052
0.246
0.112
0.327
0.191
0.184
0.138
0.034
0.047
0.017
0.094
0.081
0.052
0.025
0.057
0.056
0.301
0.237
0.184
0.048
0.038
0.048
0.096
0.052
0.151
0.316
0.190
0.118
0.316
0.364
0.420
0.440
0.431
0.477
0.445
0.411
0.570
0.470
0.433
0.364
0.431
0.434
0.417
0.436
0.399
0.448
0.338
0.349
0.329
0.333
0.300
0.298
0.540
0.441
0.569
0.434
0.352
0.531
0.565
0.405
0.341
0.388
0.358
0.432
0.579
0.608
0.579
0.685
0.652
0.364
0.300
0.278
0.314

-------
                        Table 9D-2
Apportionment Factors for 2030 Park Specific Visibility Benefits
PARK
Shenandoah
Anaconda-Pintlar W
Boundary Waters
Breton W
Isle Royale
Jarbidge W
Medicine Lake W
Red Rock Lakes W
Roosevelt Campobello
Selway-Bitterroot W
Seney W
Wolf Island W
Agua Tibia W
Black Canyon of the
Caribou W
Chiricahua
Cucamonga W
Dome Land W
Flat Tops W
Grand Canyon
Hoover W
John Muir W
Kaiser W
La Garita W
Mazatzal W
Mesa Verde
Petrified Forest
Pine Mountain W
Pinnacles
Point Reyes
Rawah W
Rocky Mountain
Saguaro
San Gabriel W
San Gorgino W
San Jacinto W
San Rafael W
Sequoia-Kings
Sycamore Canyon W
Ventana W
Yolla-Bolly-Middle-
Yo Semite
Carlsbad Caverns
GilaW
COUNTY
Lawrence
Cochise
Gila Co
Gila Co
Coconino
Apache
Apache
Graham
Pima Co
Maricopa
Coconino
Yavapai
Tuolumne
San
Calaveras
Trinity Co
Fresno Co
Mono Co
Inyo Co
Marin Co
Los
Monterey
San Benito
Riverside
Siskiyou
San
Del Norte
Shasta Co
Fresno Co
Lassen Co
Riverside
San Diego
Shasta Co
El Dorado
Mariposa
Fresno Co
Tuolumne
Tulare Co
Siskiyou
Santa
Tulare Co
Modoc Co
San Juan
Garfield
STATE
AL
AZ
AZ
AZ
AZ
AZ
AZ
AZ
AZ
AZ
AZ
AZ
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CO
CO
Percent ot 2030 Visibility Benefit Due to
Changes in:
SO. NOx rHrnp.t PM
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
376
313
277
293
342
429
429
188
207
342
057
293
055
226
065
129
039
046
070
070
049
0.033
0
0
0
0
0
0
049
049
116
411
411
158
0.043
0
0
0
0
0
0
047
053
468
090
065
033
0.099
0
0
0
0
0
0
0
0
046
225
116
059
321
073
312
464
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
297
075
048
089
107
069
069
173
072
107
019
089
571
407
191
111
520
493
616
616
109
376
109
109
518
270
270
344
535
663
588
133
175
191
376
179
493
452
518
593
292
400
203
087
0
0
0
0
0
0
0
0
0
0
0
0
0
327
612
675
619
551
503
503
639
721
551
924
619
375
0.368
0
0
0
0
0
0
0
0
0
0
0
0
0
745
759
441
461
314
314
842
591
842
842
366
320
320
0.498
0
0
422
289
0.360
0
0
0
0
0
0
0
0
0
399
735
745
591
722
461
323
366
348
0.386
0
0
0
527
485
449

-------
PARK
Joyce Kilmer- Slickrock
Kalmiopsis W
Linville Gorge W
Lostwood W
Pecos W
Presidential Range-Dry
Salt Creek W
Shining Rock W
Wheeler Peak W
Wichita Mountains W
Fitzpatrick W
Glacier Peak W
Mount Adams W
Dolly Sods W
North Absaroka W
Olympic
Lye Brook W
Bridger W
Goat Rocks W
Otter Creek W
Pasayten W
Bandelier
Bosque del Apache W
Brigantine W
Crater Lake
Mount Hood W
Mount Washington W
San Pedro Parks W
Swanguarter W
Theodore Roosevelt
Maroon Bells-
Mount Rainier
North Cascades
Bob Marshall W
Gates of the Mountain
Glacier
St. Marks W
Voyageurs
Teton W
Yellowstone
Grand Teton NP
Washakie W
COUNTY STATE
Routt Co CO
Larimer CO
Pitkin Co CO
Alamosa CO
Gunnison CO
Montezum CO
Montrose CO
Summit CO
Mineral CO
Larimer CO
Monroe FL
Wakulla FL
Citrus Co FL
Charlton GA
Mclntosh GA
Edmonson KY
Stone Co MS
Hyde Co NC
Haywood NC
Avery Co NC
Graham NC
Sandoval NM
Rio Arriba NM
Grant Co NM
ChavesCo NM
Mora Co NM
Eddy Co NM
Socorro NM
Taos Co NM
Lincoln NM
Elko Co NV
Polk Co TN
BlountCo TN
San Juan UT
Grand Co UT
San Juan UT
Washingto UT
Garfield UT
Botetourt VA
Madison VA
Grant Co WV
Tucker Co WV
Percent ot 2030 Visibility Benefit Due to
Changes in:
SO. NOx Hirer.t PM
0.289
0.407
0.537
0.391
0.320
0.367
0.397
0.397
0.471
0.385
0.365
0.503
0.497
0.503
0.463
0.365
0.486
0.515
0.455
0.436
0.309
0.389
0.374
0.378
0.387
0.525
0.421
0.472
0.481
0.553
0.261
0.359
0.345
0.322
0.265
0.337
0.337
0.190
0.445
0.331
0.455
0.487
0.286
0.123
0.074
0.103
0.091
0.180
0.156
0.156
0.140
0.188
0.204
0.033
0.070
0.085
0.082
0.304
0.166
0.183
0.252
0.232
0.371
0.051
0.037
0.069
0.021
0.100
0.124
0.059
0.092
0.078
0.345
0.295
0.232
0.046
0.065
0.064
0.064
0.129
0.193
0.387
0.275
0.200
0.425
0.470
0.389
0.505
0.589
0.452
0.447
0.447
0.389
0.428
0.431
0.464
0.433
0.412
0.456
0.332
0.348
0.302
0.293
0.332
0.320
0.560
0.589
0.553
0.592
0.375
0.455
0.469
0.427
0.369
0.394
0.346
0.423
0.632
0.671
0.600
0.600
0.680
0.361
0.282
0.270
0.313

-------
CHAPTER 10:  Economic Impact Analysis
    10.1 Overview and Results  	  10-1
        10.1.1 What is an Economic Impact Analysis?  	  10-1
        10.1.2 What Methodology Did EPA Use in this Economic Impact Assessment? 	  10-2
        10.1.3 What are the key features of the NDEIM?  	  10-5
            10.1.3.1 Brief Description of the NDEIM	  10-5
            10.1.3.2 Product Markets Included in the Model 	  10-6
            10.1.3.3 Supply and Demand Elasticities 	  10-9
            10.1.3.4 Fixed and Variable Costs	  10-11
            10.1.3.5 Compliance Costs  	  10-11
            10.1.3.6 Other NDEIM Features	  10-12
        10.1.4 Summary of Economic Analysis	  10-14
            10.1.4.1  What are the Rule's Expected Market Impacts?  	  10-15
            10.1.4.2  What are the Rule's Expected Social Costs? 	  10-19
    10.2 Economic Methodology	  10-28
        10.2.1 Behavioral Economic Models 	  10-28
        10.2.2 Conceptual Economic Approach  	  10-29
            10.2.2.1 Types of Models: Partial vs.  General Equilibrium Modeling Approaches  	  10-29
            10.2.2.2 Market Equilibrium in a Single Commodity Market	  10-31
            10.2.2.3 Incorporating Multimarket Interactions	  10-32
        10.2.3 Key Modeling Elements  	  10-37
            10.2.3.1 Perfect vs. Imperfect Competition	  10-37
            10.2.3.2 Short- vs. Long-Run Models	  10-38
            10.2.3.3 Variable vs. Fixed Regulatory Costs   	  10-42
            10.2.3.4 Substitution	  10-45
        10.2.4  Estimation of Social Costs  	  10-47
    10.3  NDEIM Model Inputs and Solution Algorithm	  10-50
        10.3.1 Description of Product Markets	  10-51
            10.3.1.1 Engine Markets	  10-51
            10.3.1.2 Equipment Markets  	  10-51
            10.3.1.3 Application Markets   	  10-54
            10.3.1.4 Diesel Fuel Markets	  10-55
            10.3.1.5  Locomotive and Marine Transportation Markets	  10-57
        10.3.2 Market Linkages  	  10-58
        10.3.3 Baseline Economic Data  	  10-58
            10.3.3.1 Baseline Quantities: Engines, Equipment and Fuel	  10-58
            10.3.3.2  Baseline  Prices: Engines, Equipment and Fuel  	  10-64
            10.3.3.3  Baseline  Quantities and Prices for Transportation and Application Markets  	  10-65
        10.3.4 Calibrating the Fuel Spillover Baseline	  10-67
        10.3.5 Compliance Costs   	  10-67
            10.3.5.1  Engine and Equipment Compliance Costs	  10-68
            10.3.5.2  Nonroad Diesel Fuel Compliance Costs	  10-76
            10.3.5.3  Changes  in Operating Costs	  10-77
        10.3.6 Growth Rates	  10-80
        10.3.7  Market Supply and Demand Elasticities	  10-80
        10.3.8  Model Solution	  10-84
            10.3.8.1 Computing Baseline and With-Regulation Equilibrium Conditions	  10-84
            10.3.8.2 Solution Algorithm 	  10-86
    10.4  Estimating Impacts	  10-87
    APPENDIX 10A:  Impacts on the Engine Markets  	  10-93
    APPENDIX 10B:  Impacts on Equipment Markets	  10-102
    APPENDIX IOC:  Impacts on Application Markets	  10-153
    APPENDIX 10D:  Impacts on the Nonroad Fuel Market	  10-159
    APPENDIX 10E: Time Series of Social Cost 	  10-164
    APPENDIX 10F: Model Equations	  10-168
    APPENDIX 10G:  Elasticity Parameters for Economic Impact Modeling	  10-173
    APPENDIX 10H:  Derivation of Supply Elasticity  	  10-189
    APPENDIX 101: Sensitivity Analysis 	  10-190

-------
                                                     Economic Impact Analysis
             CHAPTER 10:  Economic  Impact Analysis
   This chapter contains the Economic Impact Analysis (EIA) prepared to estimate the
economic impacts of this rule on producers and consumers of nonroad engines, equipment, fuel
and related industries.  This EIA relies on the Nonroad Diesel Economic Impact Model
(NDEIM), developed for this  analysis, to estimate market-level changes in prices and outputs for
affected engine, equipment, fuel, and application markets as well as the social costs and their
distribution across economic sectors affected by the program. The basis for this analysis is
provided in the Economic Impact Analysis technical support document prepared for the proposal
for this rule, as updated by a technical memoranda (RTI, 2003a, RTI2004).

   This analysis is based on an earlier version of the engineering costs developed for this rule.
The final cost estimates for the engine program are slightly higher ($142 million) and the final
fuel costs are slightly lower ($246 million), resulting in a 30-year net present value of $27.1
billion (30 year net present values in the year 2004, using a 3% Discount Rate, $2002) or $104
million less than the engineering costs used in this analysis.  We do not expect that the revised
engineering costs would change the overall results of this economic impact analysis given the
small portion of engine, equipment, and fuel costs to total production costs for goods and
services using these inputs and given the inelastic value of the estimated demand elasticities for
the application markets.

   The first section of this chapter briefly describes the methodology we used to estimate the
economic impacts of this rule and  presents an overview of the results. According to this analysis,
this rule would be highly beneficial to society, with a net present value of social costs of about
$27.2 billion, compared to net present value benefits through 2036 of $780 billion (30 year net
present values in the year 2004 using 3% discount rate, $2002).  The impact of these costs on
society should be minimal, with the average price of goods and services produced using
equipment and fuel affected by the final rule expected to increase by about 0.1 percent.  The
second section of this chapter presents a more detailed description of the economic methodology
behind the NDEIM and certain modeling assumptions. The third section describes the markets
included in the model and data inputs as well as the solution algorithm. Finally, the appendices
to this chapter contain detailed results, additional information on the model, and a sensitivity
analysis.

10.1 Overview and Results

10.1.1 What is an Economic Impact Analysis?

   An Economic Impact Analysis is prepared to inform decision makers within the Agency
about the potential economic consequences of a regulatory action. The analysis contains
estimates of the social costs of a regulatory program and explores the distribution of these costs
across stakeholders.  These estimated social costs can then be compared with estimated social
benefits (as presented in Chapter 9).  As defined in EPA's Guidelines for Preparing Economic

                                         10-1

-------
Final Regulatory Impact Analysis
Analyses (EPA 2000, p. 113), social costs are the value of the goods and services lost by society
resulting from a) the use of resources to comply with and implement a regulation and b)
reductions in output.  In this analysis, social costs are explored in two steps. In the first step,
called the market analysis, we estimate how prices and quantities of good directly and indirectly
affected by the emission control program can be expected to change once the emission control
program goes into effect.  The estimated price and quantity changes for engines, equipment, fuel,
and goods produced using these inputs are examined separately. In the second step, called the
economic welfare analysis, we look at the total social costs associated with the program and their
distribution across stakeholders.

10.1.2 What Methodology Did EPA Use in this Economic Impact Assessment?

   The Nonroad Diesel Economic Impact Model (NDEEVI) developed for this EIA estimates
how producers and consumers can be expected to respond to the regulatory compliance costs
associated with this rule.  The NDEEVI uses a multi-market analysis framework that considers
interactions between regulated markets and other markets to estimate how compliance costs can
be expected to ripple through these markets. The analysis provides an estimate of the average
increase in price and decrease in quantity of output produced for the markets examined. It also
allows us to estimate the social costs of the rule and  identify how the various groups of affected
stakeholders will bear them.  The economic theory on which the NDEEVI is based is described in
Section 10.2. Market characteristics, compliance costs, and other data used in the NDEEVI are
described in Section 10.3.

   The NDEIM tracks average price and quantity changes for 62 integrated product markets.
Figure 10.1-1 illustrates the connections between the industry segments included in the model
and the flow of regulatory compliance costs through the economic  system. The rule will
increase the cost of producing nonroad diesel engines. Engine manufacturers are expected to
attempt to pass some or all of their direct compliance costs on to equipment manufacturers in the
form of higher diesel engine prices. Similarly, equipment manufacturers are expected to attempt
to pass some or all of their direct compliance costs (in the form of equipment redesign costs) and
indirect compliance costs (in the form of increased engine costs) to application market producers
through higher diesel equipment prices.  Petroleum refiners are also expected to attempt to pass
some or all of their direct compliance costs  on to application market producers and to the
locomotive and marine transportation service sectors through higher prices for diesel fuel.
Finally, application market producers are expected to pass on some or all of their increased
equipment and diesel fuel costs to consumers of final application market products and services.
It is the interaction of suppliers and purchasers in each of the markets that determines the extent
to which prices and quantities of goods produced change in response to the compliance costs
associated with the rule.  The amount of the compliance costs that can be passed on is affected
by the price sensitivity of purchasers in the relevant  market. The NDEIM explicitly models
market linkages and behavioral responses and estimates new equilibrium prices and output and
the resulting distribution of social costs across affected stakeholders.

   Diesel engines, equipment, and fuel represent only a small  portion of the total production
costs for each of the three application market sectors (the final users of the engines, equipment

                                          10-2

-------
                                                               Economic Impact Analysis
and fuel affected by this rule). Other more significant production costs include land, labor, other
capital, raw materials, insurance, profits, etc.  These other production costs are not affected by
     Application
      Imports
     Application
      Exports
                             Application Consumers
                       Application Markets included in NDEIM
                    • Agriculture • Construction 'Manufacturing
Application Suppliers not included in NDEIM
i
t

                       Application Suppliers included in NDEIM
Locomotive and
Marine Service
   Markets
                       Diesel Equipment Markets (by hp size)
                 Agriculture     • Refrigeration       • Pumps and
                 Construction    . Generator Sets      Compressors
                              • Lawn and Garden
                               Diesel Equipment
                                Manufacturers
                             Diesel Engine Markets
                             . <25hp  . 101-175 hp
                             . 26-50 hp .176-600 hp
                            . 51-75 hp .>601 hp
                             • 76-100 hp
1
t
Diesel Engine
Manufacturers
Locomotive and
Marine Service
  Providers
                                                                                 Locomotive and Marine
                                                                                 Fuel Consumption
                                                                           Diesel Fuel Markets~\
                                                                           (by PADD Region)   \

                                                                           500 ppm sulfur content I
                                                                           15 ppm sulfur content /
               Petroleum Refineries
                Figure 10.1-1. Market Linkages Included in the Economic Model


this emission control program.  This is important because it means that this rule directly affects
only a small part of total inputs for the relevant markets.  Therefore, rule is not expected to have
a large adverse impact on output and prices of goods produced in the three application sectors.
                                                 10-2

-------
Final Regulatory Impact Analysis
10.1.3 What are the key features of the NDEIM?

   10.1.3.1 Brief Description of the NDEIM

   The NDEIM is a computer model comprised of a series of spreadsheet modules that define
the baseline characteristics of supply and demand for the relevant markets and the relationships
between them. The basis for this analysis is provided in the EIA technical support document, as
updated by a technical memo (RTI, 2003a, RTI2004).  The model methodology, as explained in
Section 10.2.2, is firmly rooted in applied microeconomic theory and was developed following
the OAQPS Economic Analysis Resource Document (EPA, 1999). The NDEIM uses a multi-
market partial equilibrium approach to track changes in price and quantity for the modeled
product markets. As explained in the EPA Guidelines for Preparing Economic Analyses (EPA
2000, 125-6; see also  Section 10.2.2, below), 'partial' equilibrium refers to the fact that the
supply and demand functions are modeled for just one or a few isolated markets and that
conditions in other markets are assumed either to be unaffected by a policy or unimportant for
social cost estimation. Multi-market models go beyond partial equilibrium analysis by extending
the inquiry to more than just a single market. Multi-market analysis attempts to capture at least
some of the interactions between markets.

   The NDEIM uses  an intermediate run time frame and assumes perfect competition in the
market sectors.  These model features are explained in Sections 10.2.3.1 and 10.2.3.2.  The use
of the intermediate run means that some factors of production are fixed and some are variable.
This modeling period allows analysis of the economic effects of the rule's compliance costs on
current producers. The short run, in contrast, imposes all compliance costs on the manufacturers
(no pass-through to consumers), while the long run imposes all costs on consumers (full cost
pass-through to  consumers). The perfect competition assumption is appropriate given the
number of firms and other key characteristics for each market (no indications of barriers to entry;
the firms are not price setters; there is no evidence of high levels of strategic behavior in the
price and quantity decisions of the firms; the products produced within each market are
somewhat homogeneous in that engines from one firm can be purchased instead of engines from
another firm; see Section 10.2.3.1, below).  The use of the intermediate run time frame and the
assumption of perfect competition are based on widely accepted economic practice for this type
of analysis (see  for example EPA, 2000, p. 126).

   The NDEIM is constructed based on the market characteristics and inter-connections
described in this chapter. The model is shocked by applying the engineering compliance cost
estimates to the  appropriate market suppliers, and then numerically solved using an iterative
auctioneer approach by "calling out" new prices until a new equilibrium is reached in all markets
simultaneously.  The output of the model is new equilibrium prices and quantities for all affected
markets.  This information is used to estimate the social costs of the model and how those costs
are shared among affected markets.
                                         10-4

-------
                                                      Economic Impact Analysis
   10.1.3.2 Product Markets Included in the Model

   There are 62 integrated product markets included in the model, as follows:
   •   7 diesel engine markets: less than 25  hp, 26 to 50 hp, 51 to 75 hp, 76 to 100 hp, 101 to
       175 hp, 176 to 600 hp, and greater than 600 hp. The EIA includes more horsepower
       categories than the standards to allow more efficient use of the engine compliance costs
       estimates.  The additional categories also allow estimating economic impacts for a more
       diverse set of markets.
   •   42 diesel equipment markets: 7 horsepower categories within 7 application categories:
       agricultural, construction, general industrial, pumps and compressors,  generator and
       welder sets, refrigeration and air conditioning, and lawn and garden. There are 7
       horsepower/application categories that did not have sales in 2000 and  are not included in
       the model, so the total number of diesel equipment markets is 42 rather than 49.
   •   3 application markets: agricultural, construction, and manufacturing.
   •   8 nonroad diesel fuel markets:  2 sulfur content levels (15 ppm and 500 ppm) for each of
       4 PADDs. PADDs 1 and 3 are combined for the purpose of this analysis. It should be
       noted that PADD 5 includes Alaska and Hawaii. Also, California fuel volumes that are
       not affected by the program (because they are covered by separate California nonroad
       diesel fuel standards) are not included in the analysis.
   •   2 transportation service markets: locomotive and marine.

   Table 10.1-1 summarizes the characteristics of each of these five groups of markets. More
detailed information on NDEIM model inputs in provided in Section 10.3.
                                          10-5

-------
                                                         Table 10.1-1
                          Summary of Markets in Nonroad Diesel Economic Impact Model (NDEIM)
Model
Dimension
Geographic
scope
Product
groupings
Market
structure
Baseline
population
Growth
projections
Supply
elasticity
Demand
elasticity
Regulatory
shock
Markets (number)
Diesel Engines (7)
National
7 horsepower categories
consistent with emission
standards8
Perfectly competitive
Power Systems Research
(PSR) database for 2000
as modified by EPA
Growth rates used in
cost analysis; see
Section 8.1
Econometric estimate
(elastic)
Derived demand
Direct compliance costs
cause shift in supply
function
Diesel Equipment (42)
National
7 horsepower categories
within seven application
categories'5'0
Perfectly competitive
Assume one-to-one
relationship with engine
populationf
Growth rates used in cost
analysis; see Section 8.1
Econometric estimate
(elastic)
Derived demand
Direct compliance costs
and higher diesel engine
prices cause shift in
supply function
Diesel Fuel (8)
Regional by PADDs
2 diesel fuels by sulfur
content (500, 15 ppm)
for 4 regional markets'1
Perfectly competitive
Based on Energy
Information
Administration (EIA)
2000 fuel consumption
data
Based on nonroad
model and EIA
Published econometric
estimate (inelastic)
Derived demand
Direct compliance
costs cause shift in
supply function
Application (3)
National
Three broad
commodity categories6
Perfectly competitive
Value of shipments
for 2000 from U.S.
Census Bureau
Average of equipment
growth rates
consumed by these
markets
Published econometric
estimate (inelastic)
Econometric estimate
(inelastic)
No direct compliance
costs but higher prices
for diesel equipment
and fuel cause shift in
supply function
Locomotive and Marine
Transportation Sectors (2)
National
2: rail and marine
transportation services
Perfectly competitive
Service expenditures, BEA.
1997 Benchmark I-O
Supplementary Make, Use
and Direct Requirements
Tables at the Detailed Level,
Table 4
EPA's SO2 inventory
projections for marine
engines that use diesel
distillate fuel (50-state
annual inventory, 1999-
2003)
Published econometric
estimate (elastic)
Derived demand
No direct compliance costs
but higher prices for diesel
fuel cause shift in supply
function
Horsepower categories are 0-25, 26-50, 51-75, 76-100, 101-175, 176-600, and 601 hp and greater; the EIA includes more horsepower categories than the
standards, allowing more efficient use of the engine compliance cost estimates.

-------
Engine categories are agricultural, construction, pumps and compressors, generator and welder sets, refrigeration and air conditioning, general industrial, and
lawn and garden.
There are seven horsepower/application categories that do not have sales in 2000 and are not included in the model.  These are: agricultural equipment >600
hp; gensets & welders > 600 hp; refrigeration & A/C > 71 hp (4 hp categories); and lawn & garden >600 hp. Therefore, the total number of diesel equipment
markets is 42 rather than 49.
PADDs 1 and 3 are combined for the purpose of this analysis). Note that PADD 5 includes Alaska and Hawaii. Also, California fuel volumes that are not
affected by the program (because they are covered by separate California nonroad diesel fuel standards) are not included in the analysis.
Application market categories are construction, agriculture, and manufacturing.
See Section 10.3.1 for an explanation of how the engines were allocated to the seven categories.

-------
Final Regulatory Impact Analysis
   Analysis of the three application markets is limited to market output.  The economic impacts
on particular groups of application market suppliers (e.g., the profitability of farm production
units or manufacturing or construction firms) or particular groups of consumers (e.g., households
and companies that consume agricultural goods, buildings, or durable or consumer goods) are
not estimated.  In other words, while we estimate that the application markets will bear most of
the burden of the regulatory program and we apportion the decrease in application market
surplus between  application market producers and application market consumers, we do not
estimate how those social costs will be shared among specific application market producers and
consumers (e.g.,  farmers and households). In some cases, application market producers may be
able to pass most if not all of their increased costs to the ultimate consumers of their products; in
other cases, they may be obliged to absorb a portion of these costs. The focus on market-level
impacts in this analysis is appropriate because the standards in this emission control program are
technical standards that apply to nonroad engines, equipment, and fuel regardless of how they
are used and the  structure of the program does not suggest that different sectors will be affected
differently by the requirements.

   10.1.3.3 Supply and Demand Elasticities

   The estimated social costs of this emission control program are a function of the ways in
which producers and consumers of the engines, equipment, and fuels affected by the standards
change their behavior in response to the costs incurred in complying with the standards. As the
compliance costs ripple through the markets, producers and consumers change their production
and purchasing decisions in response to changes in prices.  In the NDEIM, these behavioral
changes are modeled by the demand and supply elasticities (behavioral-response parameters),
which measure the price sensitivity of consumers and producers.

   The supply elasticities for the equipment, engine,  diesel fuel, and transportation service
markets and the demand and supply elasticities for the application markets used in the NDEIM
were obtained from peer-reviewed literature sources or were estimated using econometric
methods.  These  econometric methods are well-documented and are consistent with generally
accepted econometric practice. Details on sources and estimation method are provided in
Section 10.3 and Appendix 10H.

   The equipment and engine supply elasticities are elastic, meaning  that quantities supplied are
expected to be fairly sensitive to price changes.  This  means that manufacturers are more likely
(better able) to change production levels in response to price changes.

   The supply elasticities for the fuel, transportation  service, and the  supply and demand
elasticities for the three application  markets are inelastic or unit elastic, meaning that the quantity
supplied/demanded is expected to be fairly insensitive to price changes or will vary one-to-one
with price changes. For the agricultural application market, the inelastic  supply and demand
elasticities reflect the relatively constant demand for food products and the high fixed cost nature
of food production. For the construction and manufacturing application markets, the estimated
demand and supply elasticities are less inelastic, because consumers have more flexibility to
substitute away from construction and manufactured products and producers have more

                                          10-8

-------
                                                      Economic Impact Analysis
flexibility to adjust production levels. The estimated supply elasticity for the diesel fuel market
is inelastic, reflecting the fact that most refineries operate near capacity and are therefore less
responsive to fluctuations in market prices. Note that these elasticities reflect intermediate run
behavioral changes. In the long run, supply and demand are expected to be more elastic since
more substitutes may become available.

   The inelastic values for the demand elasticities for the application markets are consistent
with the Hicks-Allen derived demand relationship, according to which a low cost-share in
production combined with limited substitution yields inelastic demand.A  As noted above, diesel
engines, equipment, and fuel represent only a small portion of the total production costs for each
of the three application sectors.  The limited ability to substitute for these inputs is discussed in
Section 10.2.3.4.

   Because the elasticity estimates are a key input to the model, a sensitivity analysis for supply
and demand elasticity parameters was performed as part of this EIA. The results are presented in
Appendix 101. In general, varying the elasticity values across the range of values reported in the
literature or using the upper and lower bounds of a 90 percent confidence interval around
estimated elasticities has no impact on the magnitude of the total social costs and only a minimal
impact on the distribution of costs across stakeholders. This is not surprising because equipment
and diesel fuel costs are a relatively small share of total production costs in the construction,
agriculture, and manufacturing industries.

   In contrast to the above, the demand elasticities for the engine, equipment, fuel, and
transportation markets are internally  derived as part of the process of running the model. This is
an important feature of the NDEIM, which allows it to link the separate market components of
the model and simulate how compliance costs can be expected to ripple through the affected
economic sectors. In the real world,  for example, the quantity of nonroad equipment units
produced in a particular period depends on the price of engines (the engine market) and the
demand for equipment (the application markets). Similarly, the number of engines produced
depends on the demand for engines (the equipment market) which depends on the demand for
equipment (the application markets).  Changes in conditions in one of these markets will affect
the others. By designing the model to derive the engine, equipment, transportation market, and
fuel demand elasticities, the NDEIM simulates these connections between supply and demand
among all the product markets and replicates the economic interactions between producers and
consumers.

   10.1.3.4 Fixed and Variable Costs

   Engines and Equipment.  The EIA treats the fixed costs expected to be incurred by engine
and equipment manufacturers differently in the market and social costs analyses.  This feature of
AIf the elasticity of demand for a final product is less than the elasticity of substitution between
   an input and other inputs to the final product, then the demand for the input is less elastic the
   smaller its cost share. Hicks, J.R., 1961, 1963.

                                          10-9

-------
Final Regulatory Impact Analysis
the model is described in greater detail in Section 10.2.3.3. In the market analysis, estimated
engine and equipment market impacts (changes in prices and quantities) are based solely on the
expected increase in variable costs associated with the standards.  Fixed costs are not included in
the market analysis reported in Table 10.1-2 because in an analysis of competitive markets the
industry supply curve is based on its marginal cost curve and fixed costs are not reflected in
changes in the marginal cost curve. In addition, the fixed costs associated with the rule are
primarily R&D costs for design and engineering changes. Firms in the affected industries
currently allocate funds for R&D programs and this rule is not expected to lead firms to change
the size of their R&D budget.  Therefore, changes in fixed costs for engine and equipment
redesign associated with this rule are not likely to affect the prices of engines or equipment.
These fixed costs are reported in the social cost analysis reported in Table 10.1-4, however, as an
additional cost to producers.  This is appropriate because even though firms currently allocated
funds to R&D those resources are intended for other purposes such as increasing engine power,
ease of use, or comfort. These improvements will therefore be postponed for the length of the
rule-related R&D program. This is a cost to society.

   It may be the case, however, that some firms will maintain their current R&D budget and
allocate additional funds to comply with the this rule.  Therefore, a sensitivity analysis was
performed as part of this EIA in which fixed costs are included in intermediate-run decision-
making.  The results of this sensitivity analysis are presented in Appendix 10.1.  In this scenario,
including fixed costs in the model results in a transfer of economic welfare losses from engine
and equipment markets to the application markets (engine and equipment producer surplus losses
decrease; consumer surplus losses increase), but does not change the overall  social costs
associated with the rule.

   Fuels. Unlike for engines and equipment, most of the petroleum refinery fixed costs  are for
production hardware. Refiners are expected to have to make physical changes to their refineries
and purchase additional equipment to produce 500 ppm and then 15 ppm fuel.  Therefore, fixed
costs are included in the market analysis for fuel price and quantity impacts.

   10.1.3.5 Compliance Costs

   Engine and Equipment Compliance Costs. The NDEIM uses the engine and equipment
compliance costs described in Chapter 6. Engine and equipment costs vary over time because
fixed costs are recovered over five to ten year periods while total variable costs, despite learning
effects that serve to reduce costs on a per unit basis, continue to increase at a rate consistent with
new sales increases.  Similarly, engine operating costs also vary over time because oil change
maintenance savings, PM filter maintenance, and fuel economy effects, all of which  are
calculated on the basis of gallons of fuel consumed, change over time consistent with the growth
in nationwide fuel consumption.

   The relative magnitude of engine and equipment compliance costs is expected to have a
predictable relationship on market prices and quantities. Generally, the estimated price increases
and quantity reductions for engines and equipment are expected to vary depending on the
magnitude of compliance costs relative to total engine or equipment costs. In general, higher

                                          10-10

-------
                                                      Economic Impact Analysis
(lower) price increases are expected as a result of a high (low) relative level of compliance costs
to market price. The change in price is also expected to be highest when compliance costs are
highest.

   Fuel Compliance Costs. The NDEIM uses the fuel  compliance costs described in Chapter 7.
Fuel-related compliance costs (costs for refining and distributing regulated fuels) also change
over time. These changes are more subtle than the engine costs, however, as the fuel provisions
are largely implemented in discrete steps instead of phasing in over time. Compliance costs
were developed on a eVgallon basis; total compliance costs are determined by multiplying the
eVgallon costs by the relevant fuel volumes. Therefore, total fuel costs increase as the demand
for fuel increases. The variable operating costs are based on the natural gas cost of producing
hydrogen and for heating diesel fuel for the new desulfurization equipment, and thus would
fluctuate along with the price of natural gas.

   10.1.3.6 Other NDEIM Features

   Substitution. In modeling the market impacts and social costs of this rule, the NDEIM
considers only diesel equipment and fuel inputs to the production of goods in the applications
markets.  It does not explicitly model alternate production inputs that would serve as substitutes
for new nonroad equipment or nonroad diesel fuel.  In the model, market changes in the final
demand for application market goods and services directly correspond to changes in the demand
for nonroad equipment and fuel (i.e., in normalized terms there is a one-to-one correspondence
between the quantity of the final goods produced and the quantity of nonroad diesel  equipment
and fuel used as inputs to that production).  We believe modeling the market in this manner is
economically sound and reflects the general experience for the nonroad market. Section 10.2.3.4
describes several alternative means of production that could serve as substitutes for new nonroad
equipment and fuel and explains why they are not included in the NDEIM.

   Operating Savings. Operating savings refers to changes in operating costs that are expected
to be realized by users of both existing and new nonroad diesel equipment as a result of the
reduced sulfur content of nonroad diesel fuel. These include operating savings (cost reductions)
due to fewer oil changes, which accrue to nonroad, marine and locomotive engines that are
already in use as well as new nonroad engines that will  comply with the standards (see Section
6.2.3). These also include any extra operating costs associated with the new PM emission
control technology which may accrue to certain new engines that use this technology.  Operating
savings are not included in the market analysis because some of the savings accrue to existing
engines and because, as explained in Chapter 6, these savings are not expected to affect
consumer decisions with respect to new engines. Operating savings are included in the social
cost  analysis, however, because they accrue to society.  They are added into the estimated social
costs as an additional savings to the application and transportation service markets, since it is the
users of these engines and fuels who will see these savings.  A sensitivity analysis was
performed as part of this EIA that includes the operating savings in the market analysis. The
results of this sensitivity analysis are presented in Appendix 10.1.
                                          10-11

-------
Final Regulatory Impact Analysis
   Fuel Marker Costs.  Fuel marker costs refers to costs associated with marking high sulfur
heating oil to distinguish it from high sulfur diesel fuel produced after 2007 through the use of
early sulfur credits or small refiner provisions. Only heating oil sold outside of the Northeast is
affected. The higher sulfur NRLM fuel is not allowed to be sold in most of the Northeast, so the
marker need not be added in this large heating oil market. These costs are expected to be about
$810,000 in 2007, increasing to $1.38 million in 2008, but steadily decreasing thereafter to about
$940,000 in 2040 (see Chapter 10 of the RIA). Because these costs are relatively small, they are
incorporated into the estimated compliance costs for the fuel program (see discussion of fuel
costs, above).  They are therefore not counted separately in this economic impact analysis. This
means that the costs of marking heating fuel are allocated to all users of the fuel affected by this
rule (nonroad, locomotive, and marine) instead of uniquely to heating oil users. This is a
reasonable approach since it is likely that refiners will pass the marker costs along their complete
nonroad diesel product line and not just to heating oil.


                                      Figure  10.1-2
                       Heating Oil Marker Costs ($Million, $2002)
     $2,400

     $2,200

     $2,000

     $1,800

     $1,600

     $1,400

     $1,200

     $1,000

       $800

       $600

       $400

       $200

        $0
  Z
Z
           OMrMrMrMrMOMrMrMrMrMrMrMrMrMrMrMrM
   Fuel Spillover. Spillover fuel is highway grade diesel fuel consumed by nonroad equipment,
stationary diesel engines, boilers, and furnaces.  As described in 7.1, refiners are expected to
produce more 15 ppm fuel than is required for the highway diesel market. This excess 15 ppm
fuel will be sold into markets that allow fuel with a higher sulfur level (i.e., nonroad for a limited
period of time, locomotive, marine diesel and heating oil). This spillover fuel is affected by the
                                          10-12

-------
                                                       Economic Impact Analysis
diesel highway rule and is not affected by this regulation. Therefore, it is important to
differentiate between spillover and nonspillover fuel to ensure that the compliance costs for that
fuel pool are not counted twice.  In the NDEIM, this is done by incorporating the impact of
increased fuel costs associated with the highway rule prior to analysis of the final nonroad rule
(see Section 10.3.8).

   Compliance Flexibility Provisions. Consistent with the engine and equipment cost
discussion in Chapter 6, the EIA does not include any cost savings associated with the equipment
transition flexibility program or the nonroad engine ABT program.  As a result, the results of this
EIA can be viewed as somewhat conservative.

   Locomotive and Marine Fuel Costs.  The locomotive and marine transportation sectors are
affected by this rule through the sulfur limits on the diesel fuel used by these engines. These
sectors provide transportation to the three application markets as well as to other markets not
considered in the NDEIM (e.g.,  public utilities, nonmanufacturing service industries,
government).  As explained in Section 10.3.1.5, the NDEIM applies only a portion of the
locomotive and marine fuel costs to the three application markets. The rest of the locomotive
and marine fuel costs are added  as a separate item to the total social cost  estimates (as
Application Markets Not Included in NDEIM).

10.1.4 Summary of Economic Analysis

   Economic impact results for 2013, 2020, 2030, and 2036 are presented in this section.  The
first of these years, 2013, corresponds to the first year in which the  standards affect all engines,
equipment, and fuels.  It should  be noted that, as illustrated in Table 10.1-3, aggregate program
costs peak in 2014; increases in  costs  after that year are due to increases in the population of
engines over time.  The other years, 2020, 2030 and 2036, correspond to  years analyzed in our
benefits analysis. Detailed results for all years are included in the appendices for this chapter.

   In the following discussion,  social costs are computed as the sum of market surplus offset by
operating savings.  Market surplus is equal to the aggregate change in consumer and producer
surplus based on the estimated market impacts associated with the rule. As explained above,
operating savings are not included in the market analysis but instead are listed as a separate
category in the social cost results tables.

   In considering the results of this analysis, it should be noted that the estimated output
quantities for diesel engines, equipment,  and fuel are not identical to those estimated in the
engineering cost discussions in Chapters 6 and 7.  The difference is due to the different
methodologies used to estimate these  costs.  As noted above, social costs are the value of goods
and services lost by society resulting from a) the use of resources to comply with and implement
a regulation (i.e., compliance costs) and b) reductions in output. Thus, the social cost analysis
considers both price and output (quantity) effects associated with consumer and producer
reaction to increased prices associated with the regulatory compliance costs. The engineering
cost analysis, on the other hand, is based on applying additional technology to comply with the
new regulations. The engine population in the engineering  cost analysis does not adjust to

                                          10-13

-------
Final Regulatory Impact Analysis
reflect consumer and producer reactions to the compliance costs. Consequently, the estimated
output quantities from the cost analysis are slightly larger than the estimated output quantities
from the social cost analysis.

    10.1.4.1  What are the Rule's Expected Market Impacts?

    The estimated market impacts for 2015, 2020, 2030, and 2036 are presented in Table 10.1-2.
The market-level impacts presented in this table represent production-weighted averages of the
individual market-level impact estimates generated by the model: the average expected price
increase and quantity decrease across all of the units in each of the engine, equipment, fuel, and
final application markets. For example, the model includes seven individual engine markets that
reflect the seven different horsepower size categories. The 21.4 percent price change for engines
shown in Table 10.1-2 for 2013 is an average price change across all engine markets weighted
by the number of production units.  Similarly, the equipment impacts presented in Table 10.1-2
are the weighted averages of 42 equipment-application markets, such as small (< 25hp)
agricultural equipment and large (>600hp) industrial equipment. Note that price increases and
quantity decreases for specific types of engines,  equipment, application sectors, or diesel fuel
markets are likely to be different. The aggregated data presented in this table provide a broad
overview of the expected market impacts that is  useful when considering the impacts of the rule
on the economy as a whole. Individual market-level impacts are presented in Appendix 10A
through Appendix 10D.B

    The market impacts of this rule suggest that the overall economic impact on society is
expected be small, on average. With regard to the market analysis, the average price of goods
and services produced using affected equipment and fuel is expected to increase by less than 0.1
percent  despite the high level of cost pass-through to those markets.

    Engine Market Results: This analysis suggests that most of the variable costs associated with
the rule  will be passed along in the form of higher prices. The average price increase in 2013 for
engines is estimated to be about 21.4 percent.  This percentage is expected to decrease to about
18.3 percent by 2020. In 2036, the last year considered, the average price increase is expected to
be about 18.2 percent. This expected price increase varies by engine size because compliance
costs are a larger share of total production costs  for smaller engines. In 2013, the largest
expected percent price increase is for engines between 25 and 50 hp: 29 percent or $850; the
average price for an engine in this category is about $2,900.  However, this price increase is
expected to drop to 22 percent, or about $645, for 2015 and later.  The smallest expected percent
BThe NDEIM distinguishes between "merchant" engines and "captive" engines.  "Merchant"
   engines are produced for sale to another company and are sold on the open market to anyone
   who wants to buy them. "Captive" engines are produced by a manufacturer for use in its own
   nonroad equipment line (this equipment is said to be produced by "integrated"
   manufacturers). The market analysis for engines includes compliance costs for merchant
   engines only. The market analysis for equipment includes equipment compliance costs plus a
   portion of the engine compliance costs attributable to captive engines.

                                         10-14

-------
                                                      Economic Impact Analysis
price increase in 2013 is for engines in the greater than 600 hp category.  These engines are
expected to see price increases of about 3 percent increase in 2013, increasing to about 7.6
percent in 2015 and then decreasing to about 6.6 percent in 2017 beyond.  The expected price
increase for these engines is about $2,240 in 2013, increasing to about $6,150 in 2015 and then
decreasing to $5,340 in 2017 and later, for engines that cost on average about $80,500.

   The market impact analysis predicts that even with these increased in engine prices, total
demand is not expected to change very much. The expected average change in quantity is less
than 150 engines per year, out of total sales  of more than 500,000 engines. The estimated
change in market quantity is small because as compliance costs are passed along the supply
chain they become a smaller share of total production costs. In other words, firms that use these
engines and equipment will continue to purchase them even at the higher cost because the
increase in costs will not have a large impact on their total production costs (diesel equipment is
only one factor of production for their output of construction,  agricultural, or manufactured
goods).

   Equipment Market Results:  Estimated price changes for the equipment markets reflect both
the direct costs of the new standards on equipment production and the indirect cost through
increased engine prices.  In general, the estimated percentage  price changes for the equipment
are less than that for engines because the engine is only one input in the production of
equipment.  In 2013, the average price increase for nonroad diesel equipment is estimated to be
about 2.9 percent.   This percentage is expected to decrease to  about 2.5 percent for 2020 and
beyond.  The range of estimated price increases across equipment types parallels the share of
engine costs relative to total equipment price, so the estimated percentage price increase among
equipment types also varies.  For example, the market price in 2013 for agricultural equipment
between  175 and 600 hp is estimated to increase about 1.2 percent, or $1,740 for equipment with
an average cost of $143,700. This compares with an estimated engine price increase of about
$1,700 for engines of that size.  The largest  expected price increase in 2013 for equipment is
$2,290, or 2.6 percent, for pumps and compressors over 600 hp. This compares with an
estimated engine price increase of about $2,240 for engines of that size. The smallest expected
price increase in 2013 for equipment is $120, or 0.7 percent, for construction equipment less than
25 hp.  This compares with an estimated engine price increase of about $120 for engines of that
size.

   Again, the market analysis predicts that  even with these increased equipment prices total
demand is not expected to change very much. The expected average change in quantity is less
than 250 pieces of equipment per year, out of a total sales of more than 500,000 units. The
average decrease in the quantity of nonroad diesel equipment  produced as a result of the
regulation is estimated to be about 0.02 percent for all years.  The largest expected decrease in
quantity in 2013 is 18 units of construction equipment per year for construction equipment
between  100 and 175 hp, out of about 63,000 units. The smallest expected decrease in quantity
in 2013 is less than one unit per year in all hp categories of pumps and compressors.

   It should be noted that the absolute change in the number of engines and equipment does not
match. This is because the absolute change in the quantity of engines represents only engines

                                         10-15

-------
Final Regulatory Impact Analysis
sold on the market.  Reductions in engines consumed internally by integrated engine/equipment
manufacturers are not reflected in this number but are captured in the cost analysis.
                                        10-16

-------
                       Economic Impact Analysis
         Table 10.1-2
Summary of Market Impacts ($2002)
Market
Engineering Cost
Per Unit
Change in Price
Absolute Percent
($million)
Change in Quantity
Absolute Percent
2013
Engines
Equipment
Loco/Marine Transpb
Application Markets'1
Mo. 2 Distillate Nonroad
$1,052
$1,198


$0.06
$821 21.4
$975 2.9
0.009
0.097
$0.07 6.0
-79a -0.014
-139 -0.017
-0.007
-0.015
-2.75° -0.019
2020
Engines
Equipment
Loco/Marine Transpb
Application Markets'1
Mo. 2 Distillate Nonroad
$950
$1,107


$0.07
$761 18.3
$976 2.5
0.01
0.105
$0.07 7.0
-98a -0.016
-172 -0.018
-0.008
-0.017
-3.00° -0.021
2030
Engines
Equipment
Loco/Marine Transpb
Application Markets'1
Mo. 2 Distillate Nonroad
$937
$968


$0.07
$751 18.2
$963 2.5
0.010
0.102
$0.07 7.0
-114" -0.016
-200 -0.018
-0.008
-0.016
-3.53° -0.022
2036
Engines
Equipment
Loco/Marine Transpb
Application Markets'1
Mo. 2 Distillate Nonroad
$931
$962


$0.07
$746 18.2
$956 2.5
0.010
0.101
$0.07 7.0
-124a -0.016
-216 -0.018
-0.008
-0.016
-3.85° -0.022
            10-17

-------
Final Regulatory Impact Analysis
a The absolute change in the quantity of engines represents only engines sold on the market. Reductions in engines
consumed internally by integrated engine/equipment manufacturers are not reflected in this number but are captured in
the cost analysis.  For this reason, the absolute change in the number of engines and equipment does not match.
b The model uses normalized commodities in the application markets because of the great heterogeneity of products.
Thus, only percentage changes are presented.
c Units are in million of gallons.

    Transportation Market Results: The estimated price increase associated with the proposed
standards in the locomotive and marine transportation markets is negligible, at 0.01 percent for
all years. This means that these transportation service providers are expected to pass along
nearly all of their increased costs to the agriculture, construction, and manufacturing  application
markets, as well as other application markets not explicitly modeled in the NDEIM.  This price
increases represent a small share of total application market production costs, and therefore are
not expected to affect demand for these services.

    Application Market Results: The estimated price increase associated with the new standards
in all three application markets  is very small and averages about 0.1 percent for all years. In
other words, on average, the prices of goods and services produced using the affected engines,
equipment, and fuel are expected to increase negligibly.  This results from the observation that
compliance costs passed on through price increases represent a very small share of total
production costs in all the application markets.  For example, the construction industry realizes
an increase in  production costs  of approximately $580 million in 2013 because of the price
increases for diesel equipment and fuel. However, this represents less than 0.001 percent of the
$820 billion value of shipments in the construction industry in 2000. The estimated average
commodity price increase in 2013 ranges from 0.08 percent in the manufacturing application
market to about 0.5 percent in the construction application market. The percentage change in
output is also estimated to be very small and averages less than 0.02 percent for all years. Note
that these estimated price increases and quantity decreases are average for these sectors and may
vary for specific subsectors. Also, note that absolute changes in price and quantity are not
provided for the application markets in Table 10.1-2 because normalized commodity values are
used in  the market model. Because of the great heterogeneity of manufactured or agriculture
products, a normalized commodity ($1 unit) is used in the application markets.  This  has no
impact on the  estimated percentage change impacts but makes interpretation of the absolute
changes less informative.

    Fuel Markets Results:  The  estimated average price increase across all nonroad diesel fuel is
about 7 percent for all years.  For 15 ppm fuel, the estimated price increase for 2013 ranges from
5.6 percent in  the East Coast region (PADD 1&3) to 9.1 percent in the mountain region (PADD
4).  The average national output decrease for all fuel is estimated to be about 0.02 percent for all
years, and is relatively constant across all four regional fuel markets.

    10.1.4.2 What are the Rule's Expected Social Costs?

    Social costs include the changes in market surplus estimated by the NDEIM and changes in
operating costs associated with the regulation. Table 10.1-3 shows the time series of engineering

                                           10-18

-------
                                                      Economic Impact Analysis
compliance costs and social cost estimates for 2004 through 2036. As shown, these estimates for
engineering and social costs are of similar magnitude for each year of the analysis.  However, the
compliance costs are distributed differently than the social costs.  As illustrated in Figure 10.1-
3a, engineering compliance costs are distributed evenly across engine, equipment, and fuel
producers. However, as illustrated in Figure 10.1-3b, the social costs that result from those
compliance costs are borne mostly by producers and consumers in the application markets (about
84 percent when the operating savings are not considered) due to the increased prices for diesel
engines, equipment, and fuel.  This means that engine, equipment, and fuel producers are
expected to be able to pass on most of their compliance costs.  Specifically, engine  producers are
expected to be able to pass on about 91.3 percent their compliance costs through higher prices.
The remaining 8.7 percent are primarily fixed R&D costs that are internalized by engine
manufacturers and not passed into the market.  Equipment manufacturers are expected to retain a
slightly higher share of compliance costs (28.5 percent) because they have greater fixed costs.
Diesel fuel refiners are expected to pass about 99 percent of their compliance costs  on to the
application producers and consumers because,  as discussed in Chapter 6, refiners pass both fixed
and variable costs into the market.
                                          10-19

-------
Final Regulatory Impact Analysis
                             $1,796 million
           Application Producers and
                 Consumers
                     0%
      Fuel Refiners
          45%
           Locomotive
           and Marine
          Transportation
            Services
               0%
Application Markets Not
  Included in NDEIM
         0%
                                                  Engine Producers
                                                        27%
    Equipment Producers
           28%
                    a) Engineering Cost Distribution3
                             $1,795 million
 Application Producers
   and Consumers
        84%
       Application Markets Not
         Included in NDEIM
                6%

                 Engine Producers
                       2%

              Equipment Producers
                      8%
               Locomotive and Marine
              Transportation Services
                    b) Social Cost Distribution3

 Figure 10.1-3.  Comparing the Distribution of Engineering Compliance Costs with Social Cost
                                Estimates by Industry (2013)

 Costs do not include operating cost savings, which represent negative $285 million in costs (i.e., benefits).
                                           10-20

-------
                                     Economic Impact Analysis
                       Table 10.1-3
         National Engineering Compliance Costs and
Social Costs Estimates for the Rule (2004 - 2036)($2002; SMillion)
Year Engineering Compliance Costs
2004
2005
2006
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2031
2032
2033
2034
2035
2036
NPV at 3%
NPV at 7%
$0
$0
$0
($17)
$54
$54
$328
$923
$1,305
$1,511
$1,691
$1,742
$1,743
$1,763
$1,778
$1,795
$1,829
$1,816
$1,819
$1,844
$1,858
$1,888
$1,921
$1,954
$1,985
$2,017
$2,047
$2,078
$2,108
$2,139
$2,169
$2,198
$2,228
$27,247
$13,876
Total Social Costs
$0
$0
$0
($18)
$54
$54
$327
$922
$1,304
$1,510
$1,690
$1,741
$1,743
$1,762
$1,778
$1,795
$1,828
$1,815
$1,818
$1,843
$1,857
$1,887
$1,920
$1,952
$1,984
$2,016
$2,046
$2,077
$2,107
$2,137
$2,167
$2,197
$2,227
$27,232
$13,868
                          10-21

-------
Final Regulatory Impact Analysis
   Figure 10.1-4 shows the time series of total social costs from 2004 through 2036. Social
costs increase rapidly between 2007 and 2014 as engine, equipment and fuel costs are phased
into the regulation. Estimated net annual social costs (including operating savings) in 2014 are
about $1,690 million.  After 2014, per unit compliance costs decrease as fixed costs are
depreciated. However, due to growth in engine and equipment sales and related fuel
consumption, net social costs are expected continue to increase, but at a slower rate, from 2015
to 2036.  The estimated net present  value of social costs over the time period 2004 through 2036
based on a social discount rate of 3  percent is reported in Table 10.1-3 and is about $27.2 billion.
The present value over this same period based on a social discount rate of 7 percent is about
$13.9 billion. As shown in Table 10.1-3, these results suggest that total engineering costs exceed
compliance costs by a small amount.  This is due primarily to the fact that the estimated output
quantities for diesel engines, equipment, and fuel are not identical to those estimated in the
engineering cost analysis, which is due to the different methodologies used to estimate these
costs (see previous discussion in this Section 10.1.4).

   Estimated social costs are  disaggregated by market in Table 10.1-4, for 2015, 2020, 2030,
and 2036. A more detailed time series from 2007 to 2030 provided is in Appendix 10E.  The
data in Table 10.1-4 shows that in 2013, social costs are expected to be about $1,510 million
($2002).  About 83 percent of the total social costs is expected to be borne by producers and
consumers in the application markets in 2013, indicating that the majority of the compliance
costs associated with the rule are expected to be passed on in the form of higher prices.  When
these estimated impacts are broken  down, about 58.5 percent of the social costs are expected to
be borne by consumers in the application markets and about 41.5 percent are expected to be
borne by producers in the application markets.  Equipment manufacturers are expected to bear
about 9.5 percent of the total social  costs. Engine manufacturers and diesel fuel refineries are
expected to bear 2.8 percent and 0.5 percent, respectively.  The remaining 4.2 percent of the
social costs is expected to be borne  by the locomotive and marine transportation service sector.
In this last sector, about 97 percent  of the gross decrease in market surplus is expected to be
borne by the application markets that are not included in the NDEIM but that use these services
(e.g., public utilities, nonmanufacturing service industries, government) while about 3 percent is
expected to be borne by locomotive and marine service providers. Because of the way the
NDEIM is structured,  with the fuel  savings added separately, the results imply that locomotive
and marine service provides would  see net benefits from the rule due to the operating savings
associated with low sulfur fuel.  In fact, they are likely to pass along  some or all of those
operating savings to the users  of their services, reducing the size of the welfare losses for those
users.

   Total social costs continue to increase over time and are projected to be about $2,046  million
by 2030 and $2,227 million in 2036 ($2002).  The increase is due to the projected annual  growth
in the engine and equipment populations. Producers and consumers in the application markets
are expected to bear an even larger portion of the costs, approximately 96 percent.  This is
consistent with economic theory, which states that, in the long run, all costs are passed on to the
consumers of goods and services.

                                      Table  10.1-4

                                         10-22

-------
Summary of Social Costs Estimates Associated with Primary Program
         2015, 2020, 2030, and 2036 ($2002, $Million)a'b
2013

Engine Producers Total
Equipment Producers Total
Construction Equipment
Agricultural Equipment
Industrial Equipment
Application Producers & Consumers Total
Total Producer
Total Consumer
Construction
Agriculture
Manufacturing
Fuel Producers Total
PADD I&III
PADDII
PADD IV
PADDV
Transportation Services, Total
Locomotive
Marine
Application markets not included in NDEIM
Total
Market Surplus
($106)
$42.0
$143.1
$64.0
$51.8
$27.2
$1,496.7
$620.9
$875.7
$584.3
$430.0
$482.4
$8.0
$4.1
$3.3
$0.0
$0.6
$104.9
$1.6
$0.9
$102.4
$1,794.7
Operating
Savings
($106)





($243.2)


($115.2)
($78.2)
($49.8)





($41.5)
($12.4)
($9.9)
($19.2)
($284.7)
Total
$42.0
$143.1
$64.0
$51.8
$27.2
$1,253.5


$469.2
$351.8
$432.5
$8.0
$4.1
$3.3
$0.0
$6.0
$63.4
($10.8)
($9.0)
$83.2
$1,510.0
Percent
2.8%
9.5%



83.0%
41.5%
58.5%



0.5%




4.2%



100.0%
2020

Engine Producers Total
Equipment Producers Total
Construction Equipment
Agricultural Equipment
Industrial Equipment
Application Producers & Consumers Total
Total Producer
Total Consumer
Construction
Agriculture
Manufacturing
Fuel Producers Total
PADD I&III
PADDII
Market Surplus
($106)
$0.1
$122.7
$57.8
$39.7
$25.2
$1,826.1
$762.2
$1,063.8
$744.0
$524.3
$557.8
$11.2
$5.6
$4.6
Operating
Savings
($106)





($192.3)


($91.1)
($61.8)
($39.4)



Total
$0.1
$122.7
$57.8
$39.7
$25.2
$1,633.8


$653.0
$462.5
$518.3
$11.2
$5.6
$4.6
Percent
0.0%
6.7%



89.4%
41.7%
58.3%



0.6%



-------
Final Regulatory Impact Analysis
PADDIV
PADDV
Transportation Services, Total
Locomotive
Marine
Application markets not included in NDEIM
Total
$0.2
$0.8
$95.7
$2.0
$1.1
$92.6
$2,055.7


($35.1)
($7.2)
($11.6)
($16.3)
($227.4)
$0.2
$0.8
$60.6
($5.2)
($10.5)
$76.3
$1,828.3


3.3%



100.0%
2030

Engine Producers Total
Equipment Producers Total
Construction Equipment
Agricultural Equipment
Industrial Equipment
Application Producers & Consumers Total
Total Producer
Total Consumer
Construction
Agriculture
Manufacturing
Fuel Producers Total
PADD I&III
PADDII
PADDIV
PADDV
Transportation Services, Total
Locomotive
Marine
Application markets not included in NDEIM
Total
$0.1
$5.9
$4.0
$1.9
$0.1
$2,112.3
$882.2
$1,230.1
$863.8
$606.8
$641.6
$13.2
$6.7
$5.2
$0.3
$1.0
$109.1
$2.5
$1.4
$105.2
$2,240.6





($154.2)


($73.0)
($49.6)
($31.6)





($39.9)
($7.8)
($13.6)
($18.5)
($194.1)
$0.1
$5.9
$4.0
$1.9
$0.1
$1,958.1


$790.8
$557.2
$610.0
$13.2
$6.7
$5.2
$0.3
$1.0
$69.2
($5.3)
($12.2)
$86.7
$2,046.4
0.0%
0.3%



95.7%
41.7%
58.3%



0.6%




3.4%



100.0%
2036

Engine Producers Total
Equipment Producers Total
Construction Equipment
Agricultural Equipment
Industrial Equipment
Application Producers & Consumers Total
Total Producer
Market Surplus
($106)
$0.2
$6.4
$4.3
$2.0
$0.1
$2,287.4
$955.5
Operating
Savings
($106)





($155.7)

Total
$0.2
$6.4
$4.3
$2.0
$0.1
$2,131.7

Percent
0.0%
0.3%



95.7%
41.7%
                                   10-24

-------
Total Consumer
Construction
Agriculture
Manufacturing
Fuel Producers Total
PADD I&III
PADDII
PADD IV
PADDV
Transportation Services, Total
Locomotive
Marine
Application markets not included in NDEIM
Total
$1,331.9
$936.4
$657.8
$693.2
$14.5
$7.3
$5.8
$0.3
$1.0
$116.9
$2.8
$1.6
$112.5
$2,425.3

($50.0)
($73.7)
($31.9)





($42.6)
($8.2)
($14.6)
($19.8)
($198.4)

$862.7
$607.8
$661.3
$14.5
$7.3
$5.8
$0.3
$1.0
$74.3
($5.4)
($13.0)
$92.7
$2,227.0
58.3%



0.7%




3.3%



100.0%
Figures are in 2002 dollars.
Operating savings are shown as negative costs.

-------
Final Regulatory Impact Analysis
                                       Table 10.1-5
            Summary of Social Costs Estimates Associated with Primary Program:
                             NPV, 3%, 2004-2036 ($million)a'b

Engine Producers Total
Equipment Producers Total
Construction Equipment
Agricultural Equipment
Industrial Equipment
Application Producers & Consumers
Total
Total Producer
Total Consumer
Construction
Agriculture
Manufacturing
Fuel Producers Total
PADD I&III
PADDII
PADD IV
PADDV
Transportation Services Total
Locomotive
Marine
Application markets not included in
NDEIM
Total
Market Surplus
($106)
$256.0
$1,162.0
$545.0
$397.0
$220.0
$28,429.0
$11,838.0
$16,591.0
$11,526.0
$8,181.0
$8,723.0
$169.0
$85.0
$69.0
$3.0
$12.0
$1,653.0
$31.0
$18.0
$1,604.0
$31,669.0
Operating
Savings
($106) Total
$256.0
$1,162.0
$545.0
$397.0
$220.0
($3,757.0) $24,672.0


($1,779.0) $9,746.0
($1,208.0) $6,973.0
($770.0) $7,953.0
$169.0
$85.0
$69.0
$3.0
$12.0
($679.0) $973.0
($160.0) ($129.0)
($204.0) ($187.0)
($315.0) $1,228.0
($4,437.0) $27,232.0
Percent
0.9%
4.3%



90.6%
41.6%
58.4%



0.6%




3.6%



100.0%
a   Figures are in 2002 dollars.
b   Operating savings are shown as negative costs.
                                          10-26

-------
                                                      Economic Impact Analysis
                                     Figure 10.1-4
                     Total Social Costs (2004-2036; $2002; $Million)
    $2,400

    $2,200

    $2,000

    $1,800

    $1,600

    $1,400

    $1,200

    $1,000

     $800

     $500

     $400

     $200

       $0
              Z
            I
          O
          CM
O
CM
OO
O
O
CM
O
CM
O
CM
O
CM
                                     CD
O
CM
O
CM
O
CM
O
CM
a
o
CM
o
CM
o
CM
OO
CM
O
CM
                                                                   O   CM   •*    CD
                                                                   CO   CO   CO    CO
O
CM
O
CM
O
CM
O
CM
10.2 Economic Methodology


   Economic impact analysis uses a combination of theory and econometric modeling to
evaluate potential behavior changes associated with a new regulatory program.  As noted above,
the goal is to estimate the impact of the regulatory program on producers and consumers. This is
done by creating a mathematical model based on economic theory and populating the model
using publically available price and quantity data. A key factor in this type of analysis is
estimating the responsiveness of the quantity of engines, equipment, and fuels demanded by
consumers or supplied by producers to a change in the price of that product. This relationship is
called the elasticity of demand or supply. This section discusses the economic theory underlying
the modeling for this EIA and several key issues that affect the way the model was developed.
                                         10-27

-------
Final Regulatory Impact Analysis
10.2.1 Behavioral Economic Models

   Models incorporating different levels of economic decision making can generally be
categorized as w/Y/z-behavior responses or without-bohavior responses (engineering cost
analysis).  Engineering cost analysis is an example of the latter and provides detailed estimates
of the cost of a regulation based on the projected number of affected units and engineering
estimates of the annualized costs.

   The behavioral approach builds on the engineering cost analysis and incorporates economic
theory related to producer and consumer behavior to estimate changes in market conditions.
Owners of affected plants are economic agents that can make adjustments, such as changing
production rates or altering input mixes, that will generally affect the market environment in
which they operate.  As producers change their production levels in response to a regulation,
consumers are typically faced with changes in prices that cause them to alter the quantity that
they are willing to purchase. These changes  in price and output from the market-level impacts
are used to estimate the distribution of social costs between consumers and producers.

   Generally, the behavioral approach and engineering cost approach yield approximately the
same total cost impact. However, the advantage of the behavioral approach is that it illustrates
how the costs flow through the economic system and identifies which stakeholders, producers,
and consumers are most affected.

10.2.2 Conceptual Economic Approach

   This EIA models basic economic relationships between supply and demand to estimate
behavioral changes expected to occur as a result of the rule.  An overview of the basic economic
theory used to develop the model to estimate the potential effect of the rule on market outcomes
is presented in this section.  Following the OAQPSEconomic Analysis Resource Document
(EPA, 1999),  standard concepts in microeconomics are used to model the supply of affected
products and the impacts of the regulations on production costs and the operating decisions.

   10.2.2.1 Types of Models:  Partial vs. General Equilibrium Modeling Approaches

   In the broadest sense, all markets are directly or indirectly linked in the economy; thus, the
rule will affect all commodities and markets to some extent. The appropriate level of market
interactions to be included in an EIA is determined by the number of industries directly affected
by the requirements and the ability of affected firms to pass along the regulatory costs in the
form of higher prices.  Alternative approaches for modeling interactions between economic
sectors can generally be divided into three groups:

   •   Partial equilibrium model—Individual  markets are modeled in isolation.  The only factor
       affecting the market is the cost of the regulation on facilities in the industry being
       modeled; there are no interaction effects with other markets.
   •   General equilibrium model—All sectors of the economy are modeled together,
       incorporating interaction effects between all sectors included in the model. General

                                         10-28

-------
                                                      Economic Impact Analysis
       equilibrium models operationalize neoclassical microeconomic theory by modeling not
       only the direct effects of control costs but also potential input substitution effects,
       changes in production levels associated with changes in market prices across all sectors,
       and the associated changes in welfare economy-wide. A disadvantage of general
       equilibrium modeling is that substantial time and resources are required to develop a new
       model or tailor an existing model  for analyzing regulatory alternatives.
   •   Multimarket model—A subset of related markets is modeled together, with sector
       linkages, and hence selected interaction effects, explicitly specified.  This approach
       represents an intermediate step between a simple, single-market partial equilibrium
       approach and a full general equilibrium approach. This technique has most recently been
       referred to in the literature as "partial equilibrium analysis of multiple markets" (Berck
       and Hoffmann, 2002).

   This analysis uses a behavioral multimarket framework because the benefits of increasing the
dimensions of the model outweigh the cost associated with additional  model detail. As Bingham
and Fox (1999) note, this increased scope provides "a richer story" of the expected distribution
of economic welfare changes across producers and consumers.  Therefore, the NDEIM
developed for this analysis consists of a spreadsheet model that links a series of standard partial
equilibrium models by specifying the interactions between the supply  and demand for products.
Changes in prices and quantities are then solved across all markets simultaneously. The
following markets were included in the model; their linkages are illustrated in Figure 10.2-1 and
they are described in detail in Section 10.3.3 below:
   •   seven diesel engine markets categorized by engine size;
   •   42 equipment markets, including construction, agriculture, refrigeration, lawn and
       garden, pumps and compressors, generators and welder sets, and general industrial
       equipment types—with five to seven horsepower size categories for each equipment type;
   •   eight fuel markets, four regions (PADDs) each with two nonroad diesel fuel markets
       (500 ppm and 15 ppm); and
   •   three application markets  (construction, agriculture, and manufacturing).
                                          10-29

-------
Final Regulatory Impact Analysis
                                         Figure 10.2-1
                         Market Equilibrium without and with Regulation
                               +  P
                 =  P
                                                                             Q
             Domestic Supply
Foreign Supply
                                  a) Baseline Equilibrium
Market
        P'

        P
               S'.
                                       SM/
                                                                          I   I   D^
                                                                          I
                                                                            J_
                                                                         Q'  Q
             Domestic Supply
Foreign Supply
                               b) With-Regulation Equilibrium
Market
   10.2.2.2 Market Equilibrium in a Single Commodity Market

   A graphical representation of a general economic competitive model of price formation, as
shown in Figure 10.2-l(a), posits that market prices and quantities are determined by the
intersection of the market supply and market demand curves.  Under the baseline scenario, a
market price and quantity (p,Q) are determined by the intersection of the downward-sloping
market demand curve (DM) and the upward-sloping market supply curve (SM). The market
supply curve reflects the sum of the domestic (Sd) and import (Sf) supply curves.

   With the regulation, the costs of production increase for suppliers.  The imposition of these
regulatory control costs is represented as an upward shift in the supply curve for domestic and
                                         10-30

-------
                                                      Economic Impact Analysis
import supply, by the estimated compliance costs. As a result of the upward shift in the supply
curve, the market supply curve will also shift upward as shown in Figure 10.2-l(b) to reflect the
increased costs of production.

   At baseline without regulation, the industry produces total output, Q, at price, p, with
domestic producers supplying the amount qd and imports accounting for Q minus qd,  or qf. With
the regulation, the market price increases from p to p', and market output (as determined from
the market demand curve) declines from Q to Q'.  This reduction in market output is the net
result of reductions in domestic and import supply.

   10.2.2.3 Incorporating Multimarket Interactions

   The above description is typical of the expected market effects for a single product market
(e.g., diesel engine manufacturers) considered in isolation. However, the modeling problem  for
this EIA is more complicated because of the need to investigate affected equipment
manufacturers and fuel  producers as well as engine manufacturers.

   For example, the Tier 4 standards will affect equipment producers in two ways.  First, these
producers are affected by higher input costs (increases in the price of diesel  engines) associated
with the rule.  Second, the standards will also impose additional production  costs on equipment
producers associated with equipment changes necessary to accommodate changes in  engine
design.

   The demand for diesel engines is directly linked to the production of diesel equipment. A
single engine is typically used in each piece of equipment, and there are no substitutes (i.e., to
make diesel equipment one needs a diesel engine). For this reason, it is reasonable to assume
that the input-output relationship between the diesel  engines and the equipment is strictly fixed
and that the demand for engines varies directly with the demand for equipment.0

   The demand for diesel equipment is directly linked to the production of final goods and
services that use diesel  equipment. For example, the demand for agricultural equipment depends
on the final demand for agricultural products and the total price of supplying these products.
Thus, any change in the price of agricultural  equipment will shift the agriculture supply curve,
leading to a decrease in agricultural production and hence decreased consumption of agricultural
equipment. Assuming a fixed input-output relationship, the percentage change in agricultural
production will equal the percentage change in agricultural equipment production.

   These relationships link the demand for engines and equipment directly to the level of
production of goods and services in the application markets. A demand curve specified in terms
of its downstream consumption is referred to as  a derived demand curve. Figure 10.2-2
graphically illustrates how a derived demand curve is identified. Consider an event in the
cThis one-to-one relationship holds for engines sold on the market and for engines consumed
   internally by integrated engine/equipment manufacturers.

                                         10-31

-------
Final Regulatory Impact Analysis
construction equipment market that causes the price of equipment to increase by AP (such as an
increase in the price of engines). This increase in the price of equipment will cause the supply
curve in the construction market to shift up, leading to a decreased quantity of construction
activity (AQC). The change in construction activity leads to a decrease in the demand for
construction equipment (AQE).  The new point (QE - AQE, P - AP) traces out the derived demand
curve. Note that the supply and demand curves in the construction applications market are
needed to identify the derived demand in the construction equipment market. The construction
application market supply and demand curves are functional form and elasticity parameters
described in Appendix 1 OF.
                                         10-32

-------
                                             Economic Impact Analysis
                             Figure 10.2-2
              Derived Demand for Construction Equipment
  Unit Price of
  Construction
                              AQ,
                                     Construction
                                       Output
Price Equipment
         AP
t
1 1
1 1
1 1
1 1
•<—
AQE
~- -- Derived
Demand

Equipment
Output
           APrice
         Equipment
                Upward Shift
                Construction
                Supply Curve
AQ,
AQr
                                10-33

-------
Final Regulatory Impact Analysis
   Each point on the derived demand curve equals the construction industry's willingness to pay
for the corresponding marginal input. This is typically referred to as the input's net value of
marginal product (VMP), which is equal to the price of the output (Px) times the input's marginal
physical product (MPP). MPP is the incremental construction output attributable to a change in
equipment inputs:

                       Value Marginal Product (VMP) = Px * MPP.

   An increase in regulatory costs ©) associated with equipment will lower the VMP of all
inputs, leading to a decrease in the net marginal product:

                       Net Value Marginal Product = (Px - c) * MPP.

This decrease in the VMP of equipment, as price increases,  is what leads the downward-sloping
derived demand curve in the equipment market.

   Similarly, derived demand curves are developed for the engine markets that supply the
equipment markets. As shown in Figure 10.2-3, the increased price of engines resulting from
regulatory costs shifts the supply curve for engines and leads to a shift in the supply curve for
equipment. The resulting increased price of equipment leads to a shift in the supply curve for the
construction industry, decreasing construction output. The decrease in  construction output flows
back through the equipment market, resulting in decreased demand for  engines (AQeng).
                                         10-34

-------
                                          Economic Impact Analysis
Unit Price of
Construction
    Price
  Equipment
       APr
    Price
  Engines
     tAPeng
                             Figure 10.2-3
                      Derived Demand for Engines
                          AQ,
                                               Q - Construction
                                            • -.  Derived
                                                Demand
                            AQr
                                                Q - Equipment
                                            eng
                                             —  Derived
                                                Demand
                           AQ
                              eng
                                              Q- Engines
                             10-35

-------
Final Regulatory Impact Analysis
10.2.3 Key Modeling Elements

   In addition to specifying the type of model used and the relationships between the markets, it
is also necessary to specify several other key model characteristics. These characteristics
include the degree of competition in each market, the time horizon of the analysis, and how fixed
costs affect firms' production decisions. The specification of the industry/market characteristics
and how regulatory costs are introduced into the model has an impact on the size and
interpretation of the estimated economic impacts. These modeling issues are discussed below.

   10.2.3.1 Perfect vs. Imperfect Competition

   For all markets that are modeled, the analyst must characterize the degree of competition
within each market.  The discussion generally focuses on perfect competition (price-taking
behavior) versus imperfect competition (the lack of price-taking behavior).  The central issue is
whether individual firms have sufficient market power to influence the market price.

   Under imperfect (such as monopolistic) competition, firms produce products that have
unique attributes that differentiate them from competitors' products.  This allows them to limit
supply, which in turn increases the market price, given the traditional downward-sloping demand
curve.  Decreasing the quantity produced increases the monopolist's  profits but  decreases total
social surplus because a less than optimal amount of the product is being consumed. In the
monopolistic equilibrium, the value society (consumers) places on the marginal  product, the
market price, exceeds the marginal cost to society (producers) of producing the last unit. Thus,
social welfare is increased by inducing the monopolist to increase production.

   Social cost estimates associated with a regulation are larger with  monopolistic market
structures because the regulation exacerbates an already existing social inefficiency of too little
output from a social perspective. The Office of Management and Budget (OMB) explicitly
mentions the need to  consider these market power-related welfare costs in evaluating regulations
under Executive Order 12866 (OMB, 1996).

   However, as discussed in the industry profiles in Chapter 1, most of the diesel engine and
equipment markets have significant levels of domestic and international competition. Even in
markets where a few  firms dominate the market, there is significant excess capacity enabling
competitors to quickly respond to changes in price. In addition, there are no indications of
barriers to entry, the firms in these markets are not  price  setters, and there is no evidence of high
levels of strategic behavior in the price and quantity decisions of the  firms.  Also, the products
produced within each market are somewhat homogeneous in that engines from one firm can be
purchased instead of engines from another firm.  Finally, according to contestable market theory,
oligopolies and even  monopolies will behave very much like firms in a competitive market if it
is possible to enter particular markets costlessly (i.e.,  there are no sunk costs associated with
                                          10-36

-------
                                                      Economic Impact Analysis
market entry or exit).0 With regard to the nonroad engine market, production capacity is not
fully utilized. This means that manufacturers could potentially switch their product line to
compete in another segment of the market without a significant investment.  For these reasons,
for the nonroad diesel rule analysis, it is assumed that within each modeled engine and
equipment market the commodities of interest are similar enough to be considered homogeneous
(e.g., perfectly substitutable) and that the number of buyers and sellers is large enough so that no
individual buyer or seller has market power or influence on market prices (i.e., perfect
competition). As a result of these conditions, producers and consumers take the market price as
given when making their production  and consumption choices. The assumption of perfect
competition in this case is consistent with widely accepted economic practice for this type of
analysis (see for example EPA 2000, p. 126).

   With regard to the fuel market, the Federal Trade Commission (FTC) has developed an
approach to ensure competitiveness in this sector.  The FTC reviews oil company mergers and
frequently requires divestiture of refineries, terminals, and gas stations to maintain a minimum
level of competition. Therefore, it is reasonable to assume a competitive structure for this
market.  At the same time, however, there are several  ways in which refiners may pass along
their fuel compliance costs.  This analysis explores three approaches.  The primary modeling
scenario is the average cost scenario, according to which the change in market price is driven by
the average total (variable + fixed) regional cost of the regulation. The two other approaches are
modeled in a sensitivity analysis and reflect the case in which the highest-cost producer sets the
market price in a region. The first of these is the maximum variable cost scenario, according to
which the market price is drive by the maximum variable regional cost of the regulation. The
second is the maximum total (fixed + variable) regional cost of the regulation.  The results of the
sensitivity analyses for these two fuel scenarios are contained in Appendix 101.

   10.2.3.2 Short- vs. Long-Run Models

   In developing the multimarket partial equilibrium  model, the choices available to producers
must be considered. For example,  are producers able  to increase their factors of production (e.g.,
increase production capacity) or alter their production mix (e.g., substitution between materials,
labor, and capital)? These modeling  issues are largely dependent on the time horizon for which
the analysis is performed. Three benchmark time horizons are discussed below: the very short
run, the long run, and the intermediate run.  This discussion relies in large part on the material
contained in the OAQPSEconomic Analysis Resource Guide (U.S. EPA, 1999).
DA monopoly or firms in oligopoly may not behave as neo-classical economic theories of the
   firm predict because they may be fearful of new entrants to the market. If super-normal profits
   are earned potential competitors may enter the market, so it is argued that the existing firm(s)
   will keep prices and output at a level where only normal profits are made, setting price and
   output at or close to the competitive price and output. Baumol W J, Panzer J and Willig R D,
   (1982); Baumol, 1982.

                                          10-37

-------
Final Regulatory Impact Analysis
   In the very short run, all factors of production are assumed to be fixed, leaving the directly
affected entity with no means to respond to increased costs associated with the regulation.
Within a very short time horizon, regulated producers are constrained in their ability to adjust
inputs or outputs due to  contractual, institutional, or other factors and can be represented by a
vertical supply curve as  shown in Figure 10.2-4.  In essence, this is equivalent to the
nonbehavioral model described earlier. Neither the price nor quantity change and the
manufacturer's compliance costs become fixed or sunk costs. Under this time horizon, the
impacts of the regulation fall entirely on the regulated entity. Producers incur the entire
regulatory burden as a one-to-one reduction in their profit. This is referred to as the "full-cost
absorption" scenario and is equivalent to the engineering cost estimates.  While there is no hard
and fast rule for determining what length of time constitutes the very short run, it is inappropriate
to use this time horizon for this analysis because it  assumes economic entities have no flexibility
to adjust factors of production.

   In the long run,  all factors of production are variable, and producers can be expected to adjust
production plans in response to cost changes imposed by a regulation.  Figure 10.2-5 illustrates a
typical, if somewhat simplified, long-run industry supply function.  The function is horizontal,
indicating that the marginal and average costs of production are constant with respect to output.E
This horizontal slope reflects the fact that, under long-run constant returns to scale, technology
and input prices ultimately determine the market price, not the level of output in the market.
EThe constancy of marginal costs reflects an underlying assumption of constant returns to scale
   of production, which may or may not apply in all cases.

                                           10-38

-------
                                                        Economic Impact Analysis
                              Figure.10.2-6

               Partial Cost Pass-ThroTOtfelulator  Costs
                      ~
                                                         With Regulation



                                                Unit Cost Increase

                                                       SQ: Without Regulation
                                                           Output
      Price  f
     Inaease \
                                    Figure 10.2-5
                     Full-Cost Pass-Through of Regulatory Costs
                                                                With Regulation
Unit Cost Increase
                                                                Without Regulation
                                                                Output
   Market demand is represented by the standard downward-sloping curve. The market is
assumed here to be perfectly competitive; equilibrium is determined by the intersection of the
                                           10-39

-------
Final Regulatory Impact Analysis
supply and demand curves. In this case, the upward parallel shift in the market supply curve
represents the regulation's effect on production costs. The shift causes the market price to
increase by the full amount of the per-unit control cost (i.e., from P0 to Pj). With the quantity
demanded sensitive to price, the increase in market price leads to a reduction in output in the
new with-regulation equilibrium (i.e., Q0 to Qj). As a result, consumers incur the entire
regulatory burden as represented by the loss in consumer surplus (i.e., the area P0 ac Pj).  In the
nomenclature of EIAs, this long-run scenario is typically referred to as "full-cost pass-through,"
and is illustrated in Figure 10.2-5.

   Taken together, impacts modeled under the long-run/full-cost-pass-through scenario reveal
an important point: under fairly general economic conditions, a regulation's impact on producers
is transitory.  Ultimately, the costs are passed on to consumers in the form of higher prices.
However, this does not mean that the impacts of a regulation will have no impact on producers
of goods and services affected by a regulation.  For example, the long run may cover the time
taken to retire all of today's capital vintage, which could take decades. Therefore, transitory
impacts could be protracted and could dominate long-run impacts in terms of present value. In
addition, to evaluate impacts on current producers, the long-run is approach is not appropriate.
Consequently an time horizon that falls between the very short-run/full-cost-absorption case and
the long-run/full-cost-pass-through case is most appropriate for this EIA.

   The intermediate run can best be defined by what it is not. It is not the very short run and it
is not the long run. In the intermediate run, some factors are fixed; some are variable.F The
existence of fixed production  factors generally leads to diminishing returns to those fixed factors.
This typically manifests  itself in the form of a marginal cost (supply)  function that rises with the
output rate, as shown in Figure 10.2-6.

   Again, the regulation causes an upward shift in the supply function. The lack of resource
mobility may cause producers to suffer profit (producer surplus) losses in the face of regulation;
however, producers are able to pass through some of the associated costs to consumers, to the
extent the market will allow.  As shown, in this case, the market-clearing process generates an
increase in price (from P0 to Px) that is less than the per-unit increase in costs (fb), so that the
regulatory burden is shared by producers (net reduction  in profits) and consumers (rise in price).
In other words there is a loss of both producer and consumer surplus.

   10.2.3.3 Variable vs. Fixed Regulatory Costs

   Related to short-run versus long-run modeling issues is the question of how fixed and
variable cost increases affect market prices and quantities.  The engineering estimates of fixed
R&D and capital costs and variable material and operating and maintenance (O&M) costs
provide an initial measure  of total annual compliance costs without accounting for behavioral
FAs a semantical matter, the situation where some factors are variable and some are fixed is often
   referred to as the "short run" in economics, but the term "intermediate run" is used here to
   avoid any confusion with the term "very short run."

                                          10-40

-------
                                                       Economic Impact Analysis
responses.  The starting point for assessing the market impacts of a regulatory action is to
incorporate the regulatory compliance costs into the production decision of the firm.

   In general, shifting the supply curve by the total cost per unit implies that both capital and
operating costs vary with output levels. At least in the case of capital, this raises some questions.
In the long run, all inputs (and their costs) can be expected to vary with output.  But a
short(er)-run analysis typically holds some capital factors fixed.  For instance, to the extent that a
market supply function is tied to existing facilities, there is an element of fixed capital (or one-
time R&D). As indicated above, the current market supply function might reflect these fixed
factors with an upward slope. As shown in Figure 10.2-7, the MC curve will only be affected, or
shift upwards, by the per-unit variable compliance costs, while the AT AC curve will shift up by
the per-unit total compliance costs (c2). Thus, the variable costs will directly affect the
production decision (optimal output rate), and the fixed costs will affect the closure decision by
establishing a new higher reservation price for the firm (i.e., Pm). In other words, the fixed costs
are important in determining whether the firm will stay in this line of business (i.e., produce
anything at all), and the variable costs determine the level (quantity) of production.
                                       Figure 10.2-7
                                   Modeling Fixed Costs
                    $/q
                                                           MC'
                   prrf
                   pm
            qm   qm       qlvl
(a) Upward-sloping supplyfunction
                                                                           q/t
   In the EIA for this rule, it is assumed that only the variable cost influences the firm's
production decision level and that the fixed costs are absorbed by the firm.  Fixed costs
associated with the engine emission standards are not included in the market analysis, because in

                                          10-41

-------
Final Regulatory Impact Analysis
an analysis of competitive markets the industry supply curve is based on its marginal cost curve,
and fixed costs are not reflected in changes in the marginal cost curve. In addition, fixed costs
are primarily R&D costs associated with design and engineering changes, and firms in the
affected industries currently allocate funds for these costs (see below). These costs are still a
cost to society because they displace other R&D activities that may improve the quality or
performance of engines and equipment.  However, in this example, the fixed costs do not
influence the market price or quantity in the intermediate run.  Therefore, fixed costs are not
likely to affect the prices of engines or equipment.

   R&D costs are a long-run concern, and decisions to invest or not invest in R&D are made in
the long run.  If funds have to be diverted from some other activity into R&D needed to meet the
environmental regulations, then these costs represent a component of the social costs of the rule.
Therefore, fixed R&D costs are included in the welfare impact estimates reported in Table 10.1-
4 as unavoidable costs that reduce producer surplus. In other words, engine manufacturers
budget for research and development programs and include these charges in their long-run
strategies. In the absence of new standards, these resources would be focused on design changes
to increase customer satisfaction. Engine manufacturers are expected to redirect these resources
toward compliance with the standards, instead of adding additional resources to research and
development programs.

   Operationally, the model used in this EIA shifts the diesel engines' and equipment markets'
supply curves by the variable cost per unit only. The rule's estimated fixed costs are calculated
to reflect their opportunity costs and then added to the producer surplus decrease after the new
market (with-regulation) equilibrium has been established.0 The primary fixed costs in these
markets are associated with one-time expenditures to redesign products and retool production
lines to comply with the regulation. These fixed costs can be recovered as part of the industry's
routine R&D budget and hence are not likely to lead to additional price increases.  This
assumption is supported by information received from a number of nonroad engine and
equipment manufacturers, with whom EPA met to discuss redesign and equipment costs. The
manufacturers indicated that their redesign budgets (for emissions or other product changes) are
constrained by R&D budgets that are set annually as a percentage of annual revenues. While the
decision to redesign may be driven by anticipated future revenues for an individual piece of
equipment, the resources from with the redesign budget is allocated are determined from the
current year's R&D budget.  Thus, redesigns to meet emission standards represent a reallocation
of resources that would have been spent for other kinds of R&D (i.e., a lost opportunity cost).
To account for the value to the  company of this loss, the engineering cost analysis includes a 7
percent rate of return for all fixed costs"recovered" over a defined period for the emission
compliant products.

   An alternative approach for R&D expenditures can be used, in  which these costs are included
in intermediate-run decision-making. This alternative assumes that manufacturers will change
GThe fixed R&D costs capture the lost opportunity of forgone investments to the firm.

                                          10-42

-------
                                                      Economic Impact Analysis
their behavior based on the R&D required for compliance with the standards. A sensitivity
analysis in Appendix 101 reflects this approach.

   Fixed costs on the refiner side are treated differently in the NDEIM. Unlike for engines and
equipment where the fixed costs are primarily for up-front R&D, most of the petroleum refinery
fixed costs are for production hardware. The decision to invest to increase, maintain, or decrease
production capacity may be made in response to anticipated or actual changes in price.  To
reflect the different ways in which refiners can pass costs through to refiners, three scenarios
were run for the following supply shifts in the diesel fuel markets:
   •   shift by average total (variable + fixed cost)
   •   shift by max total (variable + fixed cost)
   •   shift by max variable cost.

The first, shift by average total cost (variable + fixed), is the primary scenario and is included in
the NDEIM.  The other two are investigated using sensitivity analyses. These supply shifts  are
discussed further in sensitivity  analysis presented in AppendixlOI.

   10.2.3.4 Substitution

   In modeling the market impacts and social costs of this rule, the NDEIM considers only
diesel equipment and fuel inputs to the production of goods in the applications markets. It does
not explicitly model alternate production inputs that could serve as substitutes for new nonroad
equipment or nonroad diesel fuel. In the model, market changes in the final demand for
application goods and services  directly correspond to changes in the demand for nonroad
equipment and fuel (i.e., in normalized terms there is a one-to-one correspondence between  the
quantity of the final goods produced and the quantity of nonroad diesel equipment and fuel used
as inputs to that production). We believe modeling the market in this manner is economically
sound and reflects the general experience for the nonroad market.

   Alternate means of production include pre-buying, delayed buying, extending the life of a
current machine, and substituting with different (e.g., gasoline-powered) equipment.  For the
reasons described below, we conclude that revising the NDEIM to include these effects would be
inappropriate.

   The term "pre-buying" refers to the possibility that the suppliers in the application market
could choose to buy additional  unneeded quantities of nonroad equipment prior to the beginning
of the Tier 4 program and then  use that equipment as an alternate means of production during the
time period of the Tier 4 program, thus avoiding the higher cost for the Tier 4 equipment.
Although such pre-buying may be economically rational in some very limited situations, its  use
as a substitute is severely limited. First, it should be clear that this form of pre-buying only
applies to equipment and not to nonroad diesel fuel. The high cost to storing any significant
quantity of nonroad diesel fuel  prior to Tier 4 makes such pre-buying unlikely. For nonroad
equipment, the logic behind pre-buying is  relatively straightforward and analogous to the
average consumer deciding to buy a new car at the end of the model year in the  anticipation that
next year's model will be more expensive. The critical difference is that the nonroad equipment

                                         10-43

-------
Final Regulatory Impact Analysis
is bought early and then held idle until it is needed as an input to production. The economic
viability of such strategic purchases are limited by the cost of idle capital and the cost for
maintaining unused equipment. In simple terms, if one assumed that the value of capital tied up
in an idle piece of equipment would have returned 7 percent in some other investment and the
cost of equipment were to go up by 7 percent, it would be economically rational to pre-buy
equipment up to one-year earlier than needed.  If the equipment will not be needed as an input to
production in the next year, it would be more rational to invest the money elsewhere and then
purchase the equipment when it is actually needed. In real terms, the window for which pre-
buying can be a rational choice is even more limited due to the cost of maintaining, storing and
insuring equipment that is not being used.  In practice then, such strategic purchases are limited
to a time period of a few months around the start of a new regulation.  The NDEIM is intended to
model market reactions in the intermediate run time frame  and thus models a period of time well
beyond the scope of the short time period during which any potential pre-buy might be rational.
We therefore have not tried to include pre-buying as a means of substitution in NDEIM.

   "Delayed-buying" refers to the possibility that producers in the application market would
defer purchasing new equipment initially but would eventually (after a delay period?) buy new
equipment.  The economic rationality of such a delay is not clear (i.e., what cheaper substitute
might be used). However, since in the end it is assumed that the new more expensive equipment
is purchased, such a substitution method would appear to be inappropriate  for an economic
model designed to model the intermediate run time frame.

   In addition, there are many other factors besides a new regulatory program that may affect a
consumer's decision to pre-buy or delay a purchase.  Specifically, manufacturer short-term
pricing promotions or marketing strategies such as rebates, dealer incentives, and advertising can
change consumer behavior.  These effects are not well captured in a general equilibrium model
such as the one used in the NDEIM, the goal of which is to estimate the rule's impact on
equilibrium prices and quantities.  Distinguishing these effects would require the use of a sales
function, which is beyond the scope of the NDEIM.

   Extending the life of a current machine is suggested as another alternative to purchasing new
equipment.  We believe this would also be  a short term phenomenon that is not relevant for the
intermediate time frame of the NDEIM.  Based on our meetings with equipment users and
suppliers, we do not believe that extending the life of nonroad equipment will prove to be an
economically rational substitute to the purchase of new equipment. Based on our understanding
of the nonroad equipment market, we believe that most users of nonroad equipment already do
this to the maximum extent possible.  That is, we believe it is already economically rational to
extend the life of nonroad equipment as long as possible. It is our understanding that new
nonroad equipment is only bought when: 1) the existing equipment can no longer perform its
function; or 2) when new demand for production requires additional means for production; or 3)
when new equipment offers a cheaper means of production than  existing equipment. The
changes in equipment due to the Tier 4 program will not substantially change these three primary
reasons for purchasing new equipment.  Further, were we to discover that extending equipment
life is economically rational (i.e., if it were cheaper to extend equipment life rather than to buy
new equipment), this would lower the cost of nonroad equipment as an input to production and

                                         10-44

-------
                                                      Economic Impact Analysis
thus would reduce the economic impact of the Tier 4 program compared to our estimate. For all
of the reasons stated here, we have decided not to attempt to model an extended equipment life
in the NDEIM.

    Finally, stakeholders suggested that equipment users may choose to substitute with different
equipment or perhaps more generally different inputs to production.  These could include the use
of gasoline powered equipment, or the use of additional labor (i.e., the use of a laborer and
shovel instead of a backhoe), or some other unknown substitute. We have specifically
considered the possibility of substitution to gasoline technology. Gasoline engines are an
alternative power source for equipment in the lowest power categories (i.e., below 75
horsepower). Above this size range there are very limited viable gasoline engine substitutes
today, and we do not believe that such substitutes will arise in the future.  We should also note
that there are a number of benefits to diesel engines and some stakeholders have argued that
there are no acceptable substitutes for diesel powered equipment.11 The fuel economy advantage
of diesel engines compared to gasoline engines dominates the overall operating costs for larger
equipment and makes the application of large gasoline engines to large nonroad equipment
economically infeasible.1  For smaller nonroad equipment, where the fuel portion of operating
costs are not as important, gasoline and diesel engines are both provided today.  The dominant
reasons  for choosing  diesel engines over the substantially  cheaper gasoline engines include
better performance from diesel  engines, lower fuel consumption from diesel engines, and the
ability to use diesel fuel.  This latter reason is a significant advantage for two reasons: diesel fuel
is safer to  store and dispense due to its lower volatility and, hence, greater resistance to
accidental ignition, and it is compatible with the fuel needed for larger equipment at  the same
worksite.  Thus, the costs for addressing safety issues with gasoline fuel storage and  the costs for
storing two fuels onsite (gasoline for small engines and diesel for large) acts  as a barrier to entry
to the market for gasoline powered equipment. Where such a barrier doesn't exist, gasoline
engines  already enjoy a substantial economic advantage over diesel.  In cases where  the more
expensive diesel powered equipment is currently used, an incremental increase in new
equipment cost is unlikely to provide the  necessary economic incentives for switching to
gasoline based power systems.  In short, we believe that users who can economically dispense
gasoline fuel already  choose the substantially cheaper gasoline technology, while diesel users are
already  choosing a more expensive technology due to reasons that will persist independent of
this rulemaking. The incremental equipment costs are not expected to be large enough to change
these market characteristics.  Therefore, we have not attempted to model the  possibility of
substitution to gasoline equipment in NDEIM.
H "To date, there is no substitute for diesel power." Associated General Contractors of America,
   OAR-2003-0012-0791.

1  Preamble Table VI.C-1  documents the lifetime operating costs (for fuel and oil only) for a 500
   hp bulldozer as $77,850. If simplistically, we assumed that a gasoline engine would have a
   30 percent higher operating cost (in practice it would likely be higher), the extra operating
   cost for a gasoline engine would be in excess of $23,000 dwarfing any additional control cost
   from the Tier 4 program.

                                         10-45

-------
Final Regulatory Impact Analysis
10.2.4 Estimation of Social Costs

   The economic welfare implications of the market price and output changes with the
regulation can be examined by calculating consumer and producer net "surplus" changes
associated with these adjustments. This is a measure of the negative impact of an environmental
policy change and is commonly referred to as the "social cost" of a regulation. It is important to
emphasize that this measure does not include the benefits that occur outside of the market, that
is, the value of the reduced levels of air pollution with the regulations.  Including this benefit will
reduce the net cost of the regulation and even make it positive.

   The demand and supply curves that are used to project  market price and quantity impacts can
be used to estimate the change in consumer, producer, and  total surplus or social cost of the
regulation (see Figure 10.2-8).

   The difference between the maximum price consumers are willing to pay for a good and the
price they actually pay is referred to  as "consumer surplus." Consumer surplus is measured as
the area under the demand curve and above the price of the product.  Similarly, the difference
between the minimum price producers are willing to accept for a good  and the price they actually
receive is referred to as "producer surplus." Producer surplus is measured as the area above the
supply curve below the price of the product. These areas can be thought of as consumers' net
benefits of consumption and producers' net benefits of production, respectively.

   In Figure 10.2-8, baseline equilibrium occurs at the intersection of the demand curve, D, and
supply curve, S.  Price is P[ with quantity Qj.  The increased cost of production with the
regulation will cause the market supply curve to shift upward to S'. The new equilibrium price
of the product is P2.  With a higher price for the product there  is less consumer welfare, all else
being unchanged. In Figure 10.2-8(a), area A represents the dollar value of the annual  net loss
in consumers' welfare associated with the increased price.  The rectangular portion represents
the loss in consumer surplus on the quantity still consumed due to the price increase, Q2, while
the triangular area represents the foregone surplus resulting from the reduced quantity consumed,
Q! - Q2.

   In addition to the  changes in consumers' welfare, there are also changes in producers'
welfare with the regulatory action. With the increase in market price, producers receive higher
revenues on the quantity still purchased, Q2. In Figure 10.2-8(b), area B represents the increase
in revenues due to this increase in price.  The difference in the area under the supply curve up to
the original market price, area C, measures the loss in producer surplus, which includes the loss
associated with the quantity no longer produced.  The net change in producers' welfare is
represented by area B - C.
                                          10-46

-------
                                                     Economic Impact Analysis
   The change in economic welfare attributable to the compliance costs of the regulations is the
sum of consumer and producer surplus changes, that is, - (A) + (B-C).  Figure 10.2-8©) shows
the net (negative) change in economic welfare associated with the regulation as area D.J
JHowever, it is important to emphasize that this measure does not include the benefits that occur
   outside the market, that is, the value of the reduced levels of air pollution with the regulations.
   Including this benefit may reduce the net cost of the regulation or even make it positive.

                                         10-47

-------
Final Regulatory Impact Analysis
                                         Figure 10.2-8
             Market Surplus Changes with Regulation: Consumer and Producer Surplus
                       $/Q
                                                 Q2  Q1          Q/t
                                (a) Change in Consumer Surplus with
                                          Regulation
                       $/Q
                                                 Q2  Q1
                                (b) Change in Producer Surplus with
                                           Regulation
Q/t
                       $/Q
                                                 Q2  Q1          Q/t
                               (c) Net Change in Economic Welfare with
                                           Regulation
                                          10-48

-------
                                                      Economic Impact Analysis
   If not all the costs of the regulation are reflected in the supply shift, then the producer and
consumer surplus changes reflected in Figure 10.2-5 will not capture the total social costs of the
regulation.  As discussed earlier, fixed R&D and capital costs are not included in the supply
curve shift for the engine and equipment markets. The fixed costs in these instances are assumed
to be borne totally by the producers in that none of these costs are passed on to consumers in the
form of higher prices. The costs are added to the producer surplus estimates generated from the
market analysis so that the accounting accurately reflects the total social cost of the regulation.

   Operating savings are included in the total social cost estimates but not integrated into the
market analysis.  Operating savings are changes in operating costs are expected to be realized by
diesel equipment users, for both existing and new equipment, as a result of the reduced sulfur
content of nonroad diesel fuel.  These include operating savings (cost reductions) due to fewer
oil changes, which accrue to nonroad engines that are already in use as well as those that will
comply with new emission standards. These savings (costs) also include any extra operating
costs associated with the new PM emission control technology that may accrue to new engines
that use this new technology. Operating savings are not included in the market analysis because
some of the savings accrue to existing engines and because these savings are not expected to
affect consumer decisions with respect to new engines (see Chapter 6). Operating savings are
included in the social cost analysis,  however, because they accrue to society. They are added
into the estimated social costs as an additional savings to the application and transportation
service markets, since it is the users of these engines and fuels that will see these savings. A
sensitivity analysis was performed in which operating  savings are included as inputs to the
NDEIM market.  The results of this analysis are presented in Appendix 101.

10.3  NDEIM Model Inputs  and Solution Algorithm

   The NDEIM is a computer model comprising a series of spreadsheet modules. The model
equations, presented in Appendix F to this chapter, are based on the economic relationships
described in Section 10.2. The NDEIM analysis consists of four steps:

   •   Define the baseline characteristics of the supply and demand of affected commodities and
       specify the intermarket relationships.
   •   Introduce a policy "shock" into the model based on estimated compliance costs that shift
       the supply functions.
   •   Use a solution algorithm to estimate a new, with-regulation equilibrium price and
       quantity for all markets.
   •   Estimate the change in producer and consumer surplus in all markets included in the
       model.

   This section describes the data inputs used to construct the model, the compliance costs used
to shock it,  and the algorithm used to solve it. The model results are presented in Appendices A
through E.
                                          10-49

-------
Final Regulatory Impact Analysis
10.3.1 Description of Product Markets

   There are 60 integrated engine, equipment, fuel, transportation service, and application
product markets included in the NDEIM.

   10.3.1.1 Engine Markets

   The engine markets are the markets associated with the production and consumption of
engines. The producers in these markets are the engine manufacturers; the consumers are
companies that make the nonroad equipment that use these engines.  Seven engine markets are
modeled, segmented by engine size (in horsepower).

   •   less than 25 hp
   •   26 to 50 hp
   •   51 to 75 hp
   •   76 to 100 hp
   •   101 to 175 hp
   •   176 to 600 hp
   •   greater than 601 hp

   The number of horsepower categories included in the NDEIM is larger than the number of
nonroad engine standard horsepower categories.  This allows more efficient use of the engine
compliance cost estimates developed for this proposal. It also allows a more refined examination
of economic impacts on equipment types.

   The NDEIM distinguishes between "merchant" engines and "captive" engines.  "Merchant"
engines are produced for sale to another company and are sold on the open market to anyone
who wants to buy them. "Captive" engines are produced by a manufacturer for use in its own
nonroad equipment line (this equipment is said to be produced by "integrated" manufacturers).
It is important to differentiate between merchant and captive engines because compliance costs
affect them differently. All compliance costs for captive engines are absorbed into the
equipment costs of integrated suppliers. In contrast, nonintegrated equipment suppliers who buy
merchant engines pay only a portion of the engine compliance costs.  As long as engine demand
is not perfectly inelastic, the increased market price for merchant engines will reflect only a
partial pass through of engine compliance costs.  The rest of the compliance costs will be borne
by the engine manufacturer.

   10.3.1.2 Equipment Markets

   The equipment markets are the markets associated with the production and consumption of
equipment that use nonroad diesel engines. The producers  in these markets are the equipment
manufacturers; the consumers are companies that use this equipment to make goods  sold in the
application markets. Seven equipment markets are modeled:

   •   Construction

                                        10-50

-------
                                                     Economic Impact Analysis
   •   Agricultural
   •   Pumps and compressors
   •   Generators and welder sets
   •   Refrigeration and air conditioning
   •   General industrial, and
   •   Lawn and garden.

   Each of the 60 applications listed in the Power Systems Research OELink Sales Version
2002 (PSR) database were allocated to one of these categories to obtain a manageable number of
equipment markets to be included in the NDEIM (Gallaher, 2003).  The mapping is contained in
Table 10.3-1.  For each of these equipment types, up to seven horsepower size category markets
are included in the model, for a total of 42 individual equipment markets.K
KThere are seven horsepower/application categories that do not have sales in 2000 and are not
   included in the model. These are: agricultural equipment >600 hp; gensets & welders > 600
   hp; refrigeration & A/C > 71 hp (4 hp categories); and lawn & garden >600 hp. Therefore,
   the total number of diesel equipment markets is 42 rather than 49.

                                        10-51

-------
Final Regulatory Impact Analysis
                                Table 10.3-1
            Mapping from PSR Equipment Categories to Equipment Markets
     Application Markets
        Equip Markets
     Equipment Types
       AGRICULTURE
AGRICULTURAL EQUIP
2-WHEEL TRACTORS
                                                AG TRACTORS
                                                BALERS
                                                COMBINES
                                                IRRIGATION SETS
                                                OTHER AG EQUIPMENT
                                                SPRAYERS
                                                WINDROWERS
      CONSTRUCTION
CONSTRUCTION
AERIAL LIFTS
                                                BORE/DRILL RIGS
                                                CRANES
                                                CRAWLERS
                                                EXCAVATORS
                                                FINISHING EQUIPMENT
                                                FOREST EQUIPMENT
                                                GRADERS
                                                LT PLANTS/SIGNAL BDS
                                                MIXERS
                                                OFF-HWY TRACTORS
                                                OFF-HWY TRUCKS
                                                OTHER CONSTRUCTION
                                                PAVERS
                                                PLATE COMPACTORS
                                                ROLLERS
                                                S/S LOADERS
                                                SCRAPERS
                                                TAMPERS/RAMMERS
                                                TRACTR/LOADR/BCKHOES
                                                TRENCHERS
                                                WHEEL LOADERS/DOZERS
     MANUFACTURING
GENERAL INDUSTRIAL
AIRCRFT SUPPRT EQUIP
                                                CHIPPERS/GRINDERS
                                                CONCRETE/IND SAWS
                                                CRUSH/PROC EQUIP
                                                DUMPERS/TENDERS
                                                FORKLIFTS
                                                OIL FIELD EQUIPMENT
                                                OTH MATERIAL HANDLNG
                                                OTHER GEN INDUSTRIAL
                                                RAILWAY MAINTENANCE
                                                ROUGH TRN FORKLFTS
                                                SCRUBBERS/SWEEPERS
                                                SPEC VEHICLES/CARTS
                                                SURFACING EQUIP
                                                TERMINAL TRACTORS
                                                UTILITY VEHICLES
                      GENERATOR SETS & WELDERS
                          GENERATOR SETS
                                  10-52

-------
                                                      Economic Impact Analysis
Application Markets

Equip Markets

LAWN & GARDEN
PUMPS & COMPRESSORS
REFRIGERATION/AC
Equipment Types
WELDERS
COMMERCIAL MOWERS
COMMERCIAL TURF
LEAF BLOWERS/VACS
LN/GDN TRACTORS
OTHER LAWN&GARDEN
TRIMMER/EDGER/CTTERS
AIR COMPRESSORS
GAS COMPRESSORS
HYD POWER UNITS
PRESSURE WASHERS
PUMPS
REFRIGERATION/AC
       Source: Gallaher (2003).
   For the purpose of this analysis, nonroad diesel equipment is assumed to be a fixed factor of
production in the application markets. Applying this assumption, a 1 percent decrease in
agricultural output will lead to a 1 percent decrease in the demand for agricultural equipment
(and fuel). The relationship between the percentage increase in equipment price and the
percentage change in equipment demand (the elasticity of demand) is determined by the input
share of diesel equipment relative to other inputs in the application markets and the  supply and
demand elasticities in the application markets.

   10.3.1.3 Application Markets

   The application markets are the markets associated with the production and consumption of
goods that use the affected diesel engines, equipment, and fuel.  The producers in these markets
include farmers, ranchers, construction firms, industrial firms, and mines; consumers include
other companies and households. Three application markets are modeled:

   •   Construction
   •   Agricultural
   •   Manufacturing

   These three application markets created after considering various economic activity
classification schemes, including theNAICS and SIC (Revelt, 2004; Gallaher, 2003).  These
three markets are included as separate groupings in each of those economic activity  classification
schemes. They are also the most significant categories of activities for which diesel engines,
equipment, and fuel are most likely to be used, as suggested in the PSR data on which the
equipment markets were chosen. Finally, they are a manageable number of markets to use in the
NDEIM. Each of the 7 equipment markets listed above were  allocated to one of these
categories.  The mapping is contained in Table 10.3-2.
                                         10-53

-------
Final Regulatory Impact Analysis
                                      Table 10.3-2
                 Mapping from Equipment Markets to Application Markets
Application Market
Agricultural
Construction
Manufacturing
Equipment Market
Agricultural equipment
Construction equipment
Pumps and compressors
Gen sets and welding equipment
Refrigeration
Lawn and garden
General industrial
   One of the consequences of reducing economic activities that use diesel engines, equipment,
and fuel into such a small number of application market categories is that seemingly unrelated
activities are linked to aggregate trends and market responses.  So, for example, if
manufacturing application market production decreases by one percent, the demand for lawn and
garden equipment, gen sets and welders, and forklifts will all decrease by the same one percent
because they are all linked to the same application market.  Similarly, forest equipment and
signal boards are grouped with cranes and bulldozers in the construction application market. In
addition, gen sets used in agricultural activities are considered to be used in the manufacturing
application market. Unfortunately, this is a problem whenever a large number of different kinds
of products or activities are reduced to a small number of categories.  At the same time, most of
the activity covered by each of the three categories, and thus most of the engines and equipment
that are included in them, is directly related to the application category.

   Analysis of the impacts on the three application markets is limited to market level changes.
The results are reported in terms of average percent change for prices and quantities of goods
sold in each of the three application markets. Changes in producer and consumer surplus at the
market level are also reported. The  economic impacts on suppliers or consumers  in particular
markets (e.g., farm production units or manufacturing or construction firms, or households and
companies that consume agricultural goods, buildings, or durable or consumer goods) are not
estimated.

   10.3.1.4 Diesel Fuel Markets

   The diesel fuel markets are the markets associated with the production and consumption of
nonroad diesel fuel. Eight nonroad diesel fuel markets were modeled: two distinct nonroad
diesel fuel commodities in four regional markets.  The two fuels are:

   •   500 ppm nonroad diesel fuel
   •   15 ppm nonroad diesel fuel


                                         10-54

-------
                                                      Economic Impact Analysis
   The Department of Energy defines five Petroleum Administrative Districts for Defense
(PADDs). For the purpose of this EIA, two of these PADDs are combined, giving four regional
district fuel markets.  These are:

   •   PADD 1 and 3
   •   PADD2
   •   PADD4
   •   PADD 5 (includes Alaska and Hawaii; California fuel volumes that are not affected by
       the program because they are covered by separate California nonroad diesel fuel
       standards are not included in the analysis)

   PADD 1 and PADD 3 are combined because of the high level of interregional trade.
Regional imports and exports across the remaining four regions included in the model are not
included in the analysis.

   Separate compliance costs are estimated for each 500 ppm and 15 ppm regional fuel market.
As a result, the price and quantify impacts, as well as the changes in producer surplus, vary
across the eight fuel markets.

   As discussed in Section 10.2, the NDEEVI is based on the assumption of perfect competition.
Using this assumption, estimated social costs are obtained by using average per-unit variable
compliance costs to shift the market supply curve (see Section 10.2.3.3). In the fuel market case,
however, each regional supply curve is shifted by the average total (variable + fixed) regional
cost of the regulation.  This approach is used for the fuel market because, unlike for engines and
equipment where the fixed costs are primarily for up-front R&D, most of the petroleum refinery
fixed costs are for production hardware. This fuel market scenario (referred to as average total
cost) is used when presenting disaggregated market results in Appendices 10. A through 10.D
and sensitivity analysis results in Appendix 101.

   However, in some fuel regions, it may be more appropriate to let the "high cost" refinery's
compliance cost drive the new market price.  If refiners' investment in desulfurization capacity is
very close to that needed to satisfy demand for 15 ppm NRLM fuel,  then refiners may have to
often operate their equipment at a capacity beyond that which minimizes cost. For example, the
temperature in the reactor can be increased, allowing greater fuel throughput. However, this
speeds up catalyst deactiviation and shortens catalyst life. This effectively increases the
operating cost per gallon of producing 15 ppm fuel. The long-term solution is for refiners  not
producing 15 ppm fuel to invest in desulfurization capacity.  However, according to EPA's cost
methodology, this incremental fuel would have a higher desulfurization cost than that
experienced by those who have already invested.  In order to justify  this new 15  ppm fuel
capacity, refiners have to anticipate  not only covering their operating costs, but their capital costs
as well. For this to occur, they would have to anticipates prices being at or above those of the
"high cost" refineries as estimated here. Under this assumption it is  the high cost producer's
dollars per gallon compliance cost increase that determines the new price. This is referred to as
the max cost scenario and no longer reflects perfect competition because now individual firms
have direct influence on market price. Two max cost scenarios  are explored in the sensitivity

                                         10-55

-------
Final Regulatory Impact Analysis
analysis presented in Appendix 101:  one in which the high-cost refinery's total (variable + fixed)
compliance costs determine price, and a second in which only the high-cost refinery's variable
compliance costs determine price.

    10.3.1.5 Locomotive and Marine Transportation Markets

    The locomotive and marine sectors are affected by this rule through the limits on the sulfur
content of fuel. These sectors provide inputs to a variety of end-use sectors in the form of
transportation services. In this sense, their role is similar to other markets for intermediate goods
already included in the NDEIM.  For example, the equipment markets in the NDEIM are markets
for intermediate goods that provide diesel-powered equipment to agriculture, construction, and
manufacturing application markets.  Using this analogy, locomotive and marine sectors are
included in the NDEIM as two intermediate markets (see Figure 10.1-1). The  intermediate
goods/services in this context are the rail and water transportation services provided to end-use
markets.

    The U.S. Bureau  of Economic Analysis (BEA) Industry Economic Program produces the
input-output tables, which show how industries interact to provide input to, and take output
from, each other. The data set can provide an appropriate measure transportation services
purchased by the application markets included in NDEIM. The BEA data show that
approximately 54 percent of rail and water transportation expenditures are made by the three
application markets in the NDEIM (see Table 10.3-3).  The remaining expenditures for these
services are associated with explicitly modeled sectors not included in the model (e.g. electric
utilities (transporting coal to electric power plants), nonmanufacturing service  industries (public
transportation), and governments). Costs flowing into these "other" sectors are included as a line
item in the economic impact estimates but do not lead to changes in market prices or quantities.

                                       Table  10.3-3
       Distribution  of Rail and Water Costs to Deliver Commodities by Industry: 1997
Application Market
Agriculture
Construction
Manufacturing
Other
Share of Rail Transportation
Expenditures
3.5%
4.3%
45.9%
46.2%
Share of Water Transportation
Expenditures
2.5%
8.3%
42.7%
45.5%
Source: U.S. Bureau of Economic Analysis (BEA). 1997 Benchmark I-O Supplementary Make, Use, and Direct
Requirements Tables at the Detailed Level. Table 4. http://www.bea.goWbea/dn2/i-o_benchmark.htm. Last
updated November 24, 2003.

   Locomotive and fuel costs were added only to the three application markets, even though
equipment and engine manufacturers also use these services. This is a simplifying assumption
                                          10-56

-------
                                                     Economic Impact Analysis
and, is not expected to have an impact on the results of the market or social cost analysis because
the share of these costs in total engine and equipment production is very small.

10.3.2 Market Linkages

    In the national economy, the markets described above are connected in that changes in
demand in one market will  affect the supply of goods in a related market. For example, nonroad
equipment manufacturers consume engines in their production processes in the sense that each
piece  of nonroad equipment has a nonroad engine. This equipment is then supplied to
application market producers through the application markets. A decrease in the demand for
equipment in the application market will lead to a decrease in the quantity of equipment
produced, which will in turn lead to a decrease in the quantity of engines produced.  Similarly,
the fuel markets are also linked to the application markets, with the demand for No. 2 distillate
being specified as a function of the production and consumption decisions made in the
construction, agricultural, and manufacturer application markets. In the NDEIM, increased
equipment costs decrease the demand for fuel, and increased fuel costs decrease the demand for
equipment because both increase the costs of production in the application markets.  This in turn
leads to  a decrease in production in the application markets and hence a decrease in the demand
for inputs (fuel and equipment).

    The linkages between the markets are illustrated in Figure 10.1-1. These interaction effects
are accounted for by designing the model to derive the engine, equipment, transportation, and
fuel market demand elasticities. The derived demand aspect of the model simulates connections
between supply and demand among all the product markets and replicates the economic
interactions between producers and consumers. Detailed specifications of the market model
equations (supply  and demand functions, equilibrium conditions) are provided in Appendix 10F.

10.3.3 Baseline Economic Data

    This section describes the data used to define the baseline conditions in the model.  These
include baseline quantities  and prices for the engines, equipment and fuel affected by the rule
and for the transportation service sectors and application markets that use these engines,
equipment, and fuel.

    10.3.3.1 Baseline Quantities: Engines, Equipment and Fuel

    Engines and Equipment: The NDEIM uses the same engine  sales that are used in the engine
and equipment cost analysis presented in Chapter 6.  The engine sales are based on the Power
Systems Research OELink Sales Version 2002 database, adjusted to eliminate stationary
equipment and to maintain  consistency with the 1998 Nonroad inventory model (see Chapter 8,
Table 8.1-1 and related text).  Sales data are used as a proxy for production data in the NDEIM
because detailed production data by horsepower and equipment application are not available
(modeling inventory decisions of engine and equipment manufacturers is beyond the scope of the
NDEIM). The sales distribution by size and application is the same for equipment as for engines
due to the assumption of a one-to-one relationship between engines and equipment. Engines and

                                         10-57

-------
Final Regulatory Impact Analysis
equipment are allocated to equipment type categories according to the PSR database
categorization scheme (see Section 10.3.1.2 and Table 10.3-1, above).  Table 10.3-4 lists sales
data for affected diesel nonroad engines and equipment sold in the United States in 2000 by
engine horsepower and equipment category.
                                        10-58

-------
                                                        Table 10.3-4
                                              Engine/Equipment Sales in 2000
Engine Market
0 600 hp
Grand Total
Agricultural
Equipment
13,195
38,303
19,156
11,788
35,226
41,678
—
159,347
Construction
17,043
30,233
30,919
30,146
49,503
42,126
4,945
204,915
General
Industrial
3,173
6,933
7,074
14,204
17,757
8,327
576
58,044
Generator
Sets and
Welders
54,971
32,540
13,234
5,567
7,313
1,813
—
115,440
Lawn and
Garden
17,118
10,323
1,456
2,722
1,556
509
—
33,684
Pumps and
Compressors
4,980
4,254
3,930
4,238
985
1,494
16
19,898
Refrigeration/
Air Condition
8,677
10,394
18,145


—
—
37,215
Grand Total
119,159
132,981
93,914
68,665
112,340
95,947
5,537
628,542
Source: Power Systems Research, OELink Sales Version, 2002.; see also Chapter 8, Table 8.1-1 and related text.
                                                            10-59

-------
Final Regulatory Impact Analysis
   Fuel. Baseline nonroad, locomotive, and marine diesel fuel consumption is provided in
Table 10.3-5. Fuel consumption is broken out by region (PADD) and application market
(construction, agriculture, and manufacturing).

   The fuel volumes used in NDEIM were developed from the information contained in Section
7.1 of Chapter 7 of the RIA. Only a brief summary of the methodology used to develop these
volumes is contained here so the reader is directed to Chapter 7 of the RIA for a complete
discussion.  Demand volumes are first estimated for nonroad, locomotive and marine diesel fuel
for 2001 for each PADDL and then grown to 2014.  The analysis of varying regulatory scenarios
always occurs using the 2014 estimated volumes. The three regulatory scenarios associated with
the final rule are:

   •   NRLM meeting a 500 ppm sulfur standard in 2007 to 2010 exempting small refiners

   •   NR meeting a 15 ppm sulfur standard and LM meeting a 50 ppm sulfur standard in 2010
       to 2012 exempting small refiners

   •   NRLM meeting a 15 ppm sulfur standard in 2010 to 2014 exempting some small refiners
       and allowing downgrade to meet demand except in PADD 1

   •   NRLM meeting a 15 ppm sulfur standard in 2014 which is fully phased in.  The
       downgrade can be used in locomotive and marine diesel fuel except in PADD 1

   The volume of pipeline downgrade and highway diesel fuel spillover are estimated and
apportioned to nonroad, locomotive and marine diesel fuel depending on the distribution system
constraints identified for each PADD and consistent with each regulatory scenario.  After the
downgrade and spillover are accounted for, the residual demands in each PADD are met by on-
purpose production of low sulfur fuel.

   The summary tables of 2014 volumes for each regulatory scenario are contained in Chapter
7.  The volumes are summarized in Table 7.1.4-10 for the period from 2007 to 2010, Table 7.1.4-
11 for the period from 2010 to 2012, Table 7.1.4-12 for the period from 2012 to  2014, and Table
7.1.4-13 for the period 2014 and thereafter.

   The 2014 volumes are adjusted to  estimated the volumes in each year from 2007 to 2040
using growth ratios compared to 2014 based on the growth rate factors in Tables 7.1.5-1 and
7.1.5-2. Each substream (i.e., spillover, downgrade, low sulfur fuel) within each fuel category is
adjusted using the same growth factor.

   The results of the volumes analysis are shown in Table 10.3-5. In the first column, the
nonroad, locomotive and  marine diesel fuel volume which must be desulfurized  are summarized.
L Petroleum Administrative Districts for Defense.

                                        10-60

-------
                                                   Economic Impact Analysis
The downgrade and spillover are aggregated together and shown in another column.  Then a total
is presented which represents the total of the two columns.  The volumes are shown for PADDs
1 and 3 together, PADD 2, PADD 4 and PADD 5 without California, as well as a national total
without California.
                                       10-61

-------
                                Table 10.3-5
Nonroad, Locomotive and Marine Diesel Fuel Consumption, 2007-2036 (million gallons)
Year
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2031
2032
2033
2034
2035
2036
PADD Em
Highway
Sulfur,
Nonroad, Downgrad
Locomotiv e and
e, Marine Spillover Total
3,771 4,169 7,940
6,592 1 ,503 8,095
6,720 1 ,532 8,252
7,008 1,405 8,412
7,282 1 ,300 8,582
7,414 1,323 8,737
7,540 1 ,343 8,883
7,669 1 ,365 9,034
7,801 1,384 9,185
7,932 1 ,403 9,336
8,064 1 ,423 9,487
8,200 1 ,442 9,643
8,342 1 ,464 9,806
8,545 1 ,41 1 9,956
8,729 1,375 10,104
8,872 1,395 10,266
9,007 1,413 10,420
9,145 1,432 10,577
9,282 1,451 10,733
9,420 1,469 10,889
9,558 1 ,488 1 1 ,046
9,696 1 ,506 1 1 ,203
9,835 1 ,525 1 1 ,360
9,974 1,544 11,518
10,113 1,563 11,676
10,253 1,582 11,835
10,393 1,601 11,994
10,534 1,620 12,154
10,675 1,639 12,314
10,816 1,659 12,475
PADD H
Highway
Sulfur,
Nonroad, Downgrad
Locomotiv e and
e, Marine Spillover Total
2,573 3,617 6,189
4,503 1,817 6,319
4,597 1 ,855 6,452
4,392 2,195 6,587
4,277 2,450 6,727
4,359 2,498 6,857
4,440 2,544 6,984
4,521 2,591 7,111
4,609 2,631 7,240
4,696 2,673 7,369
4,783 2,714 7,497
4,871 2,753 7,625
4,960 2,796 7,756
4,934 2,948 7,882
4,937 3,069 8,006
5,022 3,114 8,137
5,107 3,159 8,265
5,191 3,204 8,395
5,276 3,249 8,525
5,360 3,294 8,653
5,444 3,338 8,782
5,528 3,382 8,910
5,612 3,427 9,039
5,697 3,472 9,168
5,781 3,516 9,297
5,865 3,561 9,427
5,950 3,606 9,556
6,034 3,651 9,686
6,119 3,696 9,815
6,204 3,742 9,945
PADD I?
Highway
Sulfur,
Nonroad, Downgrad
Locomotiv e and
e, Marine Spillover Total
217 695 912
380 551 931
387 563 950
337 633 970
303 687 991
309 700 1,010
315 713 1,028
321 725 1 ,046
327 737 1 ,065
334 749 1 ,083
340 762 1,102
347 773 1,120
353 785 1,139
353 804 1,157
354 821 1,174
360 833 1,193
366 845 1 ,21 1
372 857 1 ,230
379 870 1 ,249
385 882 1 ,267
391 894 1 ,285
397 907 1 ,304
403 919 1,322
410 931 1,341
416 943 1,359
422 956 1 ,377
428 968 1 ,396
434 980 1,414
441 992 1 ,433
447 1 ,005 1 ,452
PADD V
Highway
Sulfur,
Nonroad, Downgrad
Locomotiv e and
e, Marine Spillover Total
223 785 1 ,007
390 639 1 ,029
398 652 1 ,050
420 652 1 ,072
439 655 1 ,095
448 669 1,116
455 682 1,137
553 605 1,158
629 550 1,179
641 560 1 ,200
652 569 1 ,222
664 579 1 ,243
677 588 1 ,265
688 598 1 ,286
700 607 1 ,307
712 616 1,329
724 626 1 ,350
736 636 1 ,371
748 645 1 ,393
759 655 1,414
771 664 1 ,436
783 674 1 ,457
795 684 1 ,478
807 693 1 ,500
819 703 1,521
831 712 1,543
843 722 1 ,565
855 732 1 ,586
867 741 1 ,608
879 751 1 ,630
Ttotal
Highway
Sulfur,
Nonroad, Downgrad
Locomotiv e and
e, Marine Spillover Total
6,783 9,265 16,048
11,864 4,510 16,374
12,102 4,601 16,704
12,158 4,883 17,041
12,301 5,093 17,394
12,530 5,189 17,719
12,750 5,282 18,032
13,064 5,286 18,350
13,367 5,302 18,669
13,603 5,385 18,988
13,840 5,467 19,307
14,083 5,548 19,630
14,332 5,634 19,965
14,520 5,760 20,280
14,720 5,872 20,592
14,966 5,958 20,925
15,203 6,043 21,246
15,445 6,128 21,573
15,684 6,215 21,899
15,924 6,300 22,224
16,164 6,384 22,548
16,405 6,469 22,874
16,646 6,554 23,200
16,887 6,640 23,527
17,129 6,725 23,854
17,371 6,811 24,182
17,614 6,897 24,511
17,857 6,983 24,840
18,101 7,069 25,171
18,345 7,156 25,501

-------
                                                     Economic Impact Analysis
   10.3.3.2  Baseline Prices: Engines, Equipment and Fuel

   Engines and Equipment:  The baseline engine prices used in the NDEIM are the same as
those contained in Table 6.2-5 in Chapter 6, above, sales weighting those values where
appropriate.  Table 10.3-6 provides the prices for the seven engine categories used in the model.
The baseline equipment prices used in the NDEIM are contained in Table 10.3-7.M These were
estimated by EPA using price data for the seven categories of equipment were complied from a
variety of sources, including the U.S. General Services Administration and various websites. A
relationship between price and horsepower was obtained using a linear interpolation method.
The price estimates for the equipment were obtained using the sales weighted horsepower value
for each power category and the corresponding linear equation (Guerra, 2004).

                                     Table 10.3-6
                                 Baseline Engine Prices








Power Range
0 600 hp
Estimated Price
$1,500
$2,900
$3,000
$4,000
$5,500
$20,000
$80,500
Source: See also Chapter 6, Table 6.2-5.
MIt should be noted that the equipment prices used in this analysis reflect current conditions and
   do not reflect any future price increases associated with EPA's nonroad Tier 3 standards.
                                         10-63

-------
Final Regulatory Impact Analysis
                                       Table 10.3-7
                       Baseline Prices of Nonroad Diesel Equipment"
Application
Agricultural Equip
Construction Equip
Pumps & Compressors
GenSets & Welders
Refrigeration & A/C
General Industrial
Lawn & Garden
<25hp
$6,900
$18,000
$6,000
$6,800
$12,500
$17,300
$9,300
26-50 hp
$14,400
$29,700
$12,200
$8,700
$27,000
$42,300
$21,500
51-75hp
$22
$31
$10
$8
$42
$56
$33
,600
,600
,600
,300
,100
,400
,100
76-100 hp
$33
$57
$12
$18

$74
$38
,400
,900
,500
,000
N/A
,300
,500
101-175 hp
$69,100
$122
$23
$21

$116
$29
,700
,800
,400
N/A
,900
,900
176-600 hp
$143
$312
$53
$35

$154
$64
,700
,900
,000
,700
N/A
,200
,300
>600 hp
N/A
$847,400
$88,000
N/A
N/A
$345,700
N/A
Source: Guerra, 2004.
a These equipment prices reflect current conditions and do not reflect any future price increases associated with
EPA's nonroad Tier 3 standards.
   Fuel Prices: The baseline fuel prices used in the NDEIM are the 2002 market prices for each
PADD obtained from the U.S. Energy Information Administration's Petroleum Market Monthly.
These prices are reported in Table 10.3-8 and are based on the average sales to end-users for
high-sulfur diesel fuel.

                                       Table 10.3-8
                        Average Market Prices for Diesel Fuel": 2002
Market
PADD I&III
PADD II
PADD IV
PADDV
Price ($/gallon)
$0.91
$0.94
$0.91
$0.87
aHigh-Sulfur Diesel Fuel observation for December 2002.
Source: U.S. Energy Information Administration. 2004. Petroleum Marketing Monthly March 2004. Table 41.
    10.3.3.3 Baseline Quantities and Prices for Transportation and Application Markets

    For the three application markets, the NDEIM uses the values of production data reported by
the U.S. Bureau of the Census and the U.S. Department of Agriculture.  The Economic Census
provides official measures of output for industries and geographic areas. It is the best publicly
available data that measures market supply for the broadly defined application markets in the
NDEIM, because its industrial classification system provides aggregate statistics for agriculture,
constructing, and manufacturing.  Trade data for agriculture and manufacturing is reported by
                                           10-64

-------
                                                        Economic Impact Analysis
the USDA and U.S. International Trade Commission (USITC)N. The NDEIM uses normalized
commodities (e.g. price is one and value equals quantity) because of the great heterogeneity of
products within each application market. To estimate production for future years, we applied
average equipment growth rates to the value of output reported in Table 10.3-9 (see discussion of
growth rates in Section 10.3.6).

                                       Table  10.3-9
                   Baseline Data for NDEIM's Application Markets: 2000
Application Market
Agriculture
Construction
Manufacturing
Value ($109)
Domestic Production: $ 219
Imports: $ 39
Domestic Production: $ 820
Domestic Production: $ 4,209
Imports: $ 1,074
Sources: U.S. Department of Agriculture, National Agricultural Statistics Service (USDA-NASS). 2002.
Agricultural Statistics 2002. Washington, DC: U.S. Department of Agriculture.  Table 9-39 and Table 15-1.  U.S.
Census Bureau. 2003b.  Value of Construction Put In Place: December 2002. C30/02-12. Washington, DC: U.S.
Census Bureau. Table 1. U.S. Census Bureau. 2003a.  Annual Survey of Manufactures.  2001 Statistics for
Industry Groups and Industries. M01(AS)-1. Washington, DC: U.S. Census Bureau. Table 1. U.S. International
Trade Commission.  2004. ITC Trade DataWeb. http://dataweb.usitc.gov/ As obtained March, 2004.
    For the transportation service sectors, the NDEIM uses the latest service expenditure data
reported by the U.S. Bureau of Economic Analysis. These values come from the 1997
Benchmark I-O Supplementary Make, Use, and Direct Requirements Tables at the Detailed
Level." BEA's Industry Economic Program produces the input-output tables, which show how
industries interact to provide input to, and take output from, each other.  The data set can provide
an appropriate measure transportation services purchased by the application markets included in
NDEIM.  Similar to the application markets, the model uses normalized commodities (e.g. price
is one and value equals quantity). To estimate production for future years, we applied SO2
growth rates for these sectors to the service expenditures reported in Table 10.3-10 (see
discussion of growth rates in Section 10.3.6).
International trade in construction is not significant.

                                           10-65

-------
Final Regulatory Impact Analysis
                                      Table 10.3-10
              Baseline Data for NDEIM's Transportation Service Markets: 1997
Transportation Service Market
Locomotive
Marine
Value of Services Used by Application
Markets Included in NDEIM ($109)
$19
$4
Source: U.S. Bureau of Economic Analysis (BEA). 1997 Benchmark I-O Supplementary Make, Use, and Direct
Requirements Tables at the Detailed Level. Table 4. http://www.bea.goWbea/dn2/i-o_benchmark.htm. Last
updated November 24, 2003.
10.3.4 Calibrating the Fuel Spillover Baseline

   The economic impact of the nonroad diesel rule is measured relative to the highway diesel
rule. The highway rule is scheduled to be phased in prior to the nonroad rule. Thus, the effect of
the highway rule must be incorporated into the baseline prior to modeling the impact of the
nonroad rule. The main factor to be addressed is "spillover" fuel from the highway market.  The
Agency estimates that approximately one-third of nonroad equipment currently uses highway
grade fuel because of access and distribution factors. Nonroad  equipment currently using
highway  diesel will experience increased fuel costs as a result of the highway rule, but not as a
result of the nonroad rule.  These costs have already been captured in the highway rule analysis;
thus, it is important to discount "spillover" fuel in the nonroad market to avoid double counting
of cost impacts.

   In this analysis, the baseline model is shocked by applying the compliance costs for the
highway  fuel requirements to the spillover fuel volumes included in Table 10.3-5. This provides
an adjusted baseline for the nonroad economic impact analysis from which the incremental
impact of the nonroad rule is estimated.  When this adjustment is performed, increasing the cost
of producing spillover fuel leads to a slight increase in the cost of producing goods and services
in the application markets, and  a decrease in application quantity ripples through the derived-
demand curves of the equipment and engine markets, slightly reducing the baseline equipment
and engine population. We assume that there are no substitutions between spillover diesel fuel
consumption and nonroad diesel fuel consumption as prices change because demand is primarily
driven by availability constraints.

10.3.5 Compliance Costs

   The NDEIM uses the compliance cost estimates described in Chapters 6 and 7.  These cost
are summarized in Tables 10.3-13 through 10.3-15. The compliance cost per unit vary over time
and by industry sector (engine,  equipment, or fuel producer). All costs are presented in 2002
dollars.
                                         10-66

-------
                                                      Economic Impact Analysis
   For the reasons described in Section 10.1 and 10.2, the NDEIM does not handle all
compliance costs in the same way. While all compliance costs are included in the economic
welfare analysis to estimate the total social costs associated with the program, only some
compliance costs are included in the market analysis to estimate changes in price and quantities
of goods produced using the engines, equipment, and fuel affected by the rule. Table 10.3-11
identifies which compliance costs are used as shocks in the market analysis and which are added
to the social cost estimates after changes in market prices and quantifies have been determined.

                                      Table 10.3-11
             How Compliance Costs are Accounted for in the Economic Analysis
Compliance Costs used to
Shock the Market Model
• Variable costs for diesel engines
• Variable costs for diesel equipment
• Fixed and variable costs for nonroad
diesel fuel
Compliance Costs added after
Market Analysis
• Fixed costs for diesel engines
• Fixed costs for diesel equipment
• Changes in operating costs of diesel equipment
   As noted above, marker costs for home heating fuel are included in the estimate of fixed and
variable costs for nonroad diesel fuel (see Section 10.3.3.2, above).

   10.3.5.1  Engine and Equipment Compliance Costs

   For diesel engines, the projected compliance costs are largely due to using new technologies,
such as advanced emissions control technologies and low-sulfur diesel fuel, to meet the proposed
Tier 4 emissions standards. Compliance costs for engines are broken out by horsepower
category and impact year. The method used to estimate these compliance costs is described in
Section 6.4.3; the per unit compliance costs for the 175 to 600 hp range were estimated by sales
weighting the 175 to 300 hp and the 300 to 600 hp per unit costs. The costs per unit change from
year to year because engine standards are implemented differently in each power category.  As
shown in Table 10.3-13, the fixed cost per engine typically decreases after 5 years as these
annualized costs are depreciated. The regulation's market impacts are driven primarily by the
per-engine variable costs that remain relatively constant over time.

   Because the estimated compliance costs for the rule are not directly proportional to engine
price, the relative supply shift in each of the engine size markets is expected to vary.0 As
illustrated in Table 10.3-12, the ratio of variable engine compliance costs to market price ranges
from 29 percent for engines 25 to 50 hp to 3 percent  for engines above 600 hp.  These different
ratios lead to different relative shifts in the supply curves, and different impacts on the changes
in market price and quantity for each market.

                                     Table 10.3-12
°Fixed engine costs are not included in the supply shift; see Section 10.2.3.3.

                                         10-67

-------
Final Regulatory Impact Analysis
                Ratio of Variable Engine Compliance Costs to Engine Price
Power Range
0 600 hp
Variable Engine Compliance
Cost / Engine Price
8.2%
29.3%
27.9%
28.3%
25.0%
8.5%
2.8%
   For nonroad equipment, the majority of the projected compliance costs are due to the need to
redesign the equipment. The method used to estimate these compliance costs is described in
Section 6.4.3.  The fixed cost consists of the redesign cost to accommodate new emissions
control devices. The variable cost consists of the cost of new or modified equipment hardware
and of labor to install the new emissions control devices. The per unit compliance costs are
weighted average costs within the appropriate horsepower range. The equipment sector
compliance costs are broken out by horsepower category and impact year in Table 10.3-14. The
majority of costs per piece of equipment are the fixed costs.  The overall compliance costs per
piece of equipment are less than half the overall costs associated with the same horsepower
category engine.
                                        10-68

-------
                                                    Table 10.3-13
                                            Compliance Costs per Engine"
HP Category Cost Types
0600hp Variable
Fixed
Total
2008
$129
$33
$162
$147
$49
$196
$167
$50
$217
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
2009
$129
$32
$161
$147
$48
$195
$167
$49
$216
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
2010
$123
$31
$154
$139
$47
$187
$158
$49
$206
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
2011
$123
$30
$153
$139
$46
$186
$158
$48
$205
$0
$0
$0
$0
$0
$0
$2,191
$326
$2,517
$2,911
$861
$3,771
2012
$123
$30
$152
$139
$45
$185
$158
$47
$205
$1,133
$80
$1,213
$1,375
$78
$1,453
$2,190
$321
$2,511
$2,910
$848
$3,758
2013
$123
$0
$123
$849
$74
$924
$837
$76
$913
$1,133
$78
$1,212
$1,375
$77
$1,452
$1,697
$316
$2,012
$2,246
$835
$3,081
2014
$123
$0
$123
$849
$73
$922
$837
$75
$912
$1,122
$108
$1,229
$1,351
$106
$1,457
$2,137
$437
$2,574
$2,733
$1,083
$3,817
2015
$123
$0
$123
$645
$71
$716
$636
$73
$710
$1,122
$106
$1,227
$1,351
$105
$1,455
$2,136
$430
$2,567
$6,153
$1,526
$7,679
2016
$123
$0
$123
$645
$70
$715
$636
$72
$709
$1,122
$104
$1,226
$1,351
$103
$1,454
$2,136
$122
$2,258
$6,153
$705
$6,857
2017
$123
$0
$123
$645
$69
$714
$636
$71
$708
$1,122
$29
$1,151
$1,351
$29
$1,380
$2,135
$120
$2,255
$5,347
$695
$6,042
12002 dollars
(continued)
                                                        10-69

-------
                                              Table 10.3-13 (continued)
                                            Compliance Costs per Enginea
HP Category Cost Types
0600hp Variable
Fixed
Total
2018
$123
$0
$123
$645
$0
$645
$636
$0
$636
$1,122
$28
$1,150
$1,351
$29
$1,380
$2,134
$119
$2,253
$5,347
$685
$6,032
2019
$123
$0
$123
$645
$0
$645
$636
$0
$636
$1,122
$0
$1,122
$1,351
$0
$1,351
$2,133
$0
$2,133
$5,347
$433
$5,780
2020
$123
$0
$123
$645
$0
$645
$636
$0
$636
$1,122
$0
$1,122
$1,351
$0
$1,351
$2,132
$0
$2,132
$5,347
$0
$5,347
2021
$123
$0
$123
$645
$0
$645
$636
$0
$636
$1,122
$0
$1,122
$1,351
$0
$1,351
$2,132
$0
$2,132
$5,347
$0
$5,347
2022
$123
$0
$123
$645
$0
$645
$636
$0
$636
$1,122
$0
$1,122
$1,351
$0
$1,351
$2,131
$0
$2,131
$5,347
$0
$5,347
2023
$123
$0
$123
$645
$0
$645
$636
$0
$636
$1,122
$0
$1,122
$1,351
$0
$1,351
$2,130
$0
$2,130
$5,347
$0
$5,347
2024
$123
$0
$123
$645
$0
$645
$636
$0
$636
$1,122
$0
$1,122
$1,351
$0
$1,351
$2,130
$0
$2,130
$5,347
$0
$5,347
2025
$123
$0
$123
$645
$0
$645
$636
$0
$636
$1,122
$0
$1,122
$1,351
$0
$1,351
$2,129
$0
$2,129
$5,347
$0
$5,347
2026
$123
$0
$123
$645
$0
$645
$636
$0
$636
$1,122
$0
$1,122
$1,351
$0
$1,351
$2,128
$0
$2,128
$5,347
$0
$5,347
2027
$123
$0
$123
$645
$0
$645
$636
$0
$636
$1,122
$0
$1,122
$1,351
$0
$1,351
$2,128
$0
$2,128
$5,347
$0
$5,347
12002 dollars
(continued)
                                                        10-70

-------
                                             Table 10.3-13 (continued)
                                           Compliance Costs per Enginea
HP Category Cost Types
0600hp Variable
Fixed
Total
2028
$123
$0
$123
$645
$0
$645
$636
$0
$636
$1,122
$0
$1,122
$1,351
$0
$1,351
$2,127
$0
$2,127
$5,347
$0
$5,347
2029
$123
$0
$123
$645
$0
$645
$636
$0
$636
$1,122
$0
$1,122
$1,351
$0
$1,351
$2,127
$0
$2,127
$5,347
$0
$5,347
2030
$123
$0
$123
$645
$0
$645
$636
$0
$636
$1,122
$0
$1,122
$1,351
$0
$1,351
$2,126
$0
$2,126
$5,347
$0
$5,347
2031
$123
$0
$123
$645
$0
$645
$636
$0
$636
$1,122
$0
$1,122
$1,351
$0
$1,351
$2,126
$0
$2,126
$5,347
$0
$5,347
2032
$123
$0
$123
$645
$0
$645
$636
$0
$636
$1,122
$0
$1,122
$1,351
$0
$1,351
$2,125
$0
$2,125
$5,347
$0
$5,347
2033
$123
$0
$123
$645
$0
$645
$636
$0
$636
$1,122
$0
$1,122
$1,351
$0
$1,351
$2,124
$0
$2,124
$5,347
$0
$5,347
2034
$123
$0
$123
$645
$0
$645
$636
$0
$636
$1,122
$0
$1,122
$1,351
$0
$1,351
$2,124
$0
$2,124
$5,347
$0
$5,347
2035
$123
$0
$123
$645
$0
$645
$636
$0
$636
$1,122
$0
$1,122
$1,351
$0
$1,351
$2,123
$0
$2,123
$5,347
$0
$5,347
2036
$123
$0
$123
$645
$0
$645
$636
$0
$636
$1,122
$0
$1,122
$1,351
$0
$1,351
$2,123
$0
$2,123
$5,347
$0
$5,347
12002 dollars

-------
                                                    Table 10.3-14
                                             Costs per Piece of Equipment"
HP Category Cost Types
0600hp Variable
Fixed
Total
2008
$0
$15
$15
$0
$8
$8
$0
cc
q>o
$8
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
2009
$0
$15
$15
$0
$8
$8
$0
cc
q>o
$8
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
2010
$0
$14
$14
$0
$8
$8
$0
cc
q>o
$8
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
2011
$0
$14
$14
$0
$7
$7
$0
$8
$8
$0
$0
$0
$0
$0
$0
$75
$378
$453
$57
$690
$748
2012
$0
$14
$14
$0
$7
$7
$0
$8
$8
$45
$109
$154
$46
$170
$216
$75
$372
$447
$57
$680
$737
2013
$0
$13
$13
$20
$42
$62
$21
$44
$65
$45
$107
$152
$46
$168
$213
$60
$366
$426
$46
$670
$716
2014
$0
$13
$13
$20
$41
$62
$21
$43
$64
$48
$132
$180
$49
$207
$256
$80
$453
$533
$61
$806
$867
2015
$0
$13
$13
$16
$40
$57
$17
$42
$59
$48
$130
$178
$49
$204
$253
$80
$446
$526
$123
$1,404
$1,527
2016
$0
$12
$12
$16
$40
$56
$17
$42
$59
$48
$128
$176
$49
$201
$250
$80
$439
$519
$123
$1,384
$1,507
2017
$0
$12
$12
$16
$39
$55
$17
$41
$58
$48
$126
$174
$49
$197
$246
$80
$433
$513
$111
$1,365
$1,475
12002 dollars
(continued)
                                                        10-72

-------
                                               Table 10.3-14 (continued)
                                             Costs per Piece of Equipment"
HP Category Cost Types
0600hp Variable
Fixed
Total
2018
$0
$0
$0
$16
$32
$48
$17
$33
$50
$48
$124
$172
$49
$194
$243
$80
$427
$506
$111
$1,346
$1,457
2019
$0
$0
$0
$16
$31
$47
$17
$33
$50
$48
$122
$170
$49
$192
$241
$80
$421
$500
$111
$1,328
$1,438
2020
$0
$0
$0
$16
$31
$47
$17
$32
$49
$48
$120
$168
$49
$189
$238
$79
$415
$494
$111
$1,310
$1,421
2021
$0
$0
$0
$16
$30
$46
$17
$32
$49
$48
$118
$167
$49
$186
$235
$79
$83
$162
$111
$693
$804
2022
$0
$0
$0
$16
$30
$46
$17
$31
$48
$48
$24
$72
$49
$37
$86
$79
$82
$161
$111
$684
$795
2023
$0
$0
$0
$16
$0
$16
$17
$0
$17
$48
$24
$72
$49
$37
$86
$79
$81
$160
$111
$675
$786
2024
$0
$0
$0
$16
$0
$16
$17
$0
$17
$48
$0
$48
$49
$0
$49
$79
$0
$79
$111
$540
$650
2025
$0
$0
$0
$16
$0
$16
$17
$0
$17
$48
$0
$48
$49
$0
$49
$79
$0
$79
$111
$0
$111
2026
$0
$0
$0
$16
$0
$16
$17
$0
$17
$48
$0
$48
$49
$0
$49
$79
$0
$79
$111
$0
$111
2027
$0
$0
$0
$16
$0
$16
$17
$0
$17
$48
$0
$48
$49
$0
$49
$79
$0
$79
$111
$0
$111
12002 dollars
(continued)
                                                        10-73

-------
                                              Table 10.3-14 (continued)
                                            Costs per Piece of Equipment"
HP Category Cost Types
0600hp Variable
Fixed
Total
2028
$0
$0
$0
$16
$0
$16
$17
$0
$17
$48
$0
$48
$49
$0
$49
$79
$0
$79
$111
$0
$111
2029
$0
$0
$0
$16
$0
$16
$17
$0
$17
$48
$0
$48
$49
$0
$49
$79
$0
$79
$111
$0
$111
2030
$0
$0
$0
$16
$0
$16
$17
$0
$17
$48
$0
$48
$49
$0
$49
$79
$0
$79
$111
$0
$111
2031
$0
$0
$0
$16
$0
$16
$17
$0
$17
$48
$0
$48
$49
$0
$49
$79
$0
$79
$111
$0
$111
2032
$0
$0
$0
$16
$0
$16
$17
$0
$17
$48
$0
$48
$49
$0
$49
$79
$0
$79
$111
$0
$111
2033
$0
$0
$0
$16
$0
$16
$17
$0
$17
$48
$0
$48
$49
$0
$49
$79
$0
$79
$111
$0
$111
2034
$0
$0
$0
$16
$0
$16
$17
$0
$17
$48
$0
$48
$49
$0
$49
$79
$0
$79
$111
$0
$111
2035
$0
$0
$0
$16
$0
$16
$17
$0
$17
$48
$0
$48
$49
$0
$49
$79
$0
$79
$111
$0
$111
2036
$0
$0
$0
$16
$0
$16
$17
$0
$17
$48
$0
$48
$49
$0
$49
$79
$0
$79
$111
$0
$111
12002 dollars

-------
                                                     Economic Impact Analysis
   10.3.5.2  Nonroad Diesel Fuel Compliance Costs

   The fuel  compliance costs used in the NDEIM are the same as those described in Chapter 7.
The NDEIM uses different compliance costs for each PADD, and for different fuel sulfur levels
(15 and 500 ppm fuel). Thus, the compliance costs change when the fuel standards change,
reflecting the phase-in of the fuel requirements. From 2007 to 2010, nonroad, locomotive, and
marine diesel fuels are required to meet a 500 ppm sulfur cap.  During this period small refiners
can continue producing high sulfur distillate fuel (-3000 ppm) and sell it into the nonroad,
locomotive and marine diesel fuel pool. In 2010, the sulfur standard for nonroad, locomotive
and marine diesel fuel changes to a 15 ppm sulfur cap.  From 2010 to 2014, small refiners can
provide fuel  complying with a 500 ppm sulfur cap to the nonroad, locomotive and marine diesel
fuel pool, except in most of PADD 1 where 500 ppm small refiner fuel cannot be sold. After
2014, the program is fully  phased-in when the small refinery provisions cease.  Table 10.3-15
presents a summary of the compliance costs used in the model. It should be noted that these
compliance costs are weighted averages of the separate compliance costs for each grade of fuel
sold in that period.

   In contrast to the engine and equipment compliance costs, the fuel compliance costs include
fixed costs. They also include the marker costs described in Section 10.1.3.6. See Chapter 7 for
a more detailed description of the components of the fuel compliance costs and how they are
estimated. See Section 10.2. .2.3 for  a discussion of how fixed and variable costs are handled in
the model.

                                     Table 10.3-15
           Fuel Compliance Costs, Locomotive, and Marine Diesel Fuel by PADD
                                    Selected  Years
Year3
Average Cost
500 ppm
15 ppm
Maximum Total Cost
500 ppm
15 ppm
PADD I and III
2007-9
2010
2011
2014-13
2015
1.8
1.86
2.7
2.7
2.7
—
5.7
5.7
6.0
6.3
4.5
4.57
6.1
6.1
6.1
—
9.4
9.4
9.6
9.8
PADD II
2007-9
2010
2011-13
2.5
2.55
3.5
—
7.4
7.4
3.8
3.94
5.9
—
10.8
10.8
                                        10-75

-------
Final Regulatory Impact Analysis
Year3
2014
2015
Average Cost
500 ppm
3.5
3.5
15 ppm
7.7
7.9
Maximum Total Cost
500 ppm
5.9
5.9
15 ppm
11.1
11.2
PADDIV
2007-9
2010
2011-13
2014
2015
3.5
3.83
9.2
9.2
9.2
—
12.6
12.6
12.8
13
6.1
6.26
9.2
9.2
9.2
—
13.6
13.6
13.8
13.9
PADDVb
2007-9
2010
2011
2014-13
2015
1.5
1.58
3.7
3.7
3.7
—
5.1
5.1
6.1
6.9
1.5
1.62
4.4
4.4
4.4
—
5.2
5.2
6.4
7.3
aNote that the 500 ppm standard begins in 6/06 and the 15 ppm standard begins in 6/10
b Excludes diesel fuel sold for use in California which is regulated by California's regulations.
    10.3.5.3 Changes in Operating Costs

    As described in Section 6.2.3 of Chapter 6, changes in operating costs are expected to be
realized by all diesel equipment users as a result of the reduced sulfur content of nonroad diesel
fuel. These changes in operating costs include the change in maintenance costs associated with
applying new emission controls to the engines; the change in maintenance costs associated with
low-sulfur fuel such as extended oil-change intervals (extended oil change intervals results in
maintenance savings); the change in fuel costs associated with the incrementally higher costs for
low-sulfur fuel (see Chapter 7), and the change in fuel costs due to any fuel consumption impacts
associated with applying new emission  controls to the engines (e.g., cost is attributed to the
CDPF  and its need for periodic regeneration). Some of these changes in operating costs will
accrue to users of existing as well as new equipment.

    The expected changes in operating costs are not included in the market analysis. This is
because, as explained in Chapter 6, these savings are not expected to affect consumer decisions
with respect to new engines. Changes in operating costs are included in the social cost analysis,
however, because they accrue to society. They are added into the estimated social costs as an
                                          10-76

-------
                                                      Economic Impact Analysis
additional savings to the application markets, since it is the users of these engines and fuels who
will see these savings. Appendix 101 contains a sensitivity analysis in which operating cost
savings are introduced into the market analysis as a downward shift in the application supply
functions.

   The operating savings in the social cost analysis were estimated by EPA using the estimated
0/gallon operating savings estimates and the fuel volumes described in Chapter 6 and 7. Total
annual operating savings were estimated for nonroad, locomotive, and marine fuel.  The annual
operating savings associated with nonroad fuel were allocated to the three application markets
(i.e., the users of nonroad equipment) based on the number of gallons of nonroad diesel
consumed in each of the agriculture (32.1 percent), construction (47.4 percent), and
manufacturing sectors (20.5 percent).  A different approach was followed for locomotive and
marine fuel. This is necessary because not all locomotive and marine transportation services are
provided to the three application markets included in the NDEIM (see Section 10.1.5). In this
case, 54 percent of the locomotive and marine operating savings were allocated to the marine
and locomotive transportation services included in the NDEIM and 46 percent were allocated to
marine and locomotive transportation services provided for application markets not included in
the NDEIM.
                                         10-77

-------
Final Regulatory Impact Analysis
                                        Table 10.3-16
                              Operating Cost Savings (SMillions)
Year
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2031
2032
2033
2034
2035
2036
Nonroad
140
246
251
266
271
261
243
257
256
241
228
216
205
192
182
176
171
167
163
160
157
156
155
154
154
154
154
154
155
156
Locomotive
12
21
21
22
23
23
23
17
13
13
13
13
13
13
13
14
14
14
14
14
14
14
14
15
15
15
15
15
15
15
Marine
9
15
16
17
18
18
18
19
20
20
20
20
21
22
23
23
23
23
24
24
24
25
25
25
26
26
26
27
27
27
Total
161
282
288
305
311
302
285
293
288
274
261
249
239
227
218
213
208
204
201
198
196
195
194
194
194
195
195
196
197
198
Source: See Chapter 6 for an explanation of operating savings; the above values are based on the values reported in
Table 6.4-3, applied to the relevant fuel volumes.
                                            10-78

-------
                                                      Economic Impact Analysis
10.3.6 Growth Rates

   The growth rates used in this analysis for engines and equipment are the same as those
provided in Section 8.1.  The growth rate for nonroad diesel fuel is from the Nonroad Model.
The growth rates for locomotive, marine, heating oil, and highway diesel fuel are all from EIA's
Annual Energy Outlook 2003.

   Growth rates for the application markets are the average of the growth rates for equipment
used in the relevant markets. They range from 1.8 percent (>600 HP) to 3.5 percent (<25 HP).
This method was used over a method applying sales weighted averages because it does not
overestimate the application growth rate by giving more weight to higher growth rates of small
HP equipment.  If a weighted average were used, the small engine growth rate would dominate
because there are so many more small engines. Using such a weighted average would then
overstate the growth rate for the larger engines. The difference between the two approach is
about 0.2 percent (about 2.3 percent for unweighted and about 2.5 percent for weighted).

   Finally, for the locomotive and marine sectors, growth is based on EPA's  SO2 inventory
growth projections for marine diesel engines that use distillate fuel (typically engines with
displacement less than 30 liters per cylinder), 50-state annual inventories, 1999-2003.

10.3.7 Market Supply and Demand Elasticities

   To operationalize the market model, supply and demand elasticities are  needed to represent
the behavior adjustments that are likely to be made by market participants.  The following
parameters are needed:
   •   supply and  demand price elasticities for application markets (construction, agriculture,
       and manufacturing),
   •   supply elasticities for equipment markets,
   •   supply elasticities for engine markets, and
   •   supply elasticities for diesel fuel markets.

   Note that, for the equipment, engine, and diesel fuel markets, demand-specific elasticity
estimates are not needed because they are derived internally as a function of changes in output
levels in the applications markets.

   Tables 10.3-17 and 10.3-18 provides a summary of the demand and supply elasticities used
to estimate the economic impact of the proposed rule.  Most elasticities were derived
econometrically using publicly available data, with the exception of the supply elasticities for the
construction and agricultural application markets and the diesel fuel supply elasticity, which
were obtained from previous studies.p The general methodologies for estimating the supply and
PA supply function was estimated as part of the simultaneous equations approach used for the
   construction and manufacturing application markets. However, the supply elasticity estimates
   were not statistically significant and were negative, which is inconsistent with generally

                                          10-79

-------
Final Regulatory Impact Analysis
demand elasticities are discussed below.  The specific regression results are presented in
Appendix 10G. It should be noted that these elasticities reflect intermediate run behavioral
changes. In the long run, supply and demand are expected to be more elastic since more
substitutes may become available.
   accepted economic theory. For this reason, literature estimates were used for the supply
   elasticities in the construction and manufacturing application markets.

                                          10-80

-------
                                                         Economic Impact Analysis
                                      Table 10.3-17
               Summary of Market Demand Elasticities Used in the NDEIM
Market
Applications
Agriculture
Estimate
-0.20
Source
EPA econometric
Method
Productivity shift
Input Data Summary
Annual time series from
Construction
Manufacturing
                           estimate
-0.96      EPA econometric
          estimate
-0.58      EPA econometric
          estimate
                           approach (Morgenstern,
                           Pizer, and Shih, 2002)
Simultaneous equation
(log-log) approach
Simultaneous equation
(log-log) approach.
1958-  1995 developed by
       Jorgenson et al.
(Jorgenson, 1990; Jorgenson,
Gollop, andFraumeni, 1987)

Annual time series from
1958-  1995 developed by
Jorgenson et al. (Jorgenson,
1990; Jorgenson, Gollop, and
Fraumeni, 1987)

Annual time series from
1958-  1995 developed by
Jorgenson et al. (Jorgenson,
1990; Jorgenson, Gollop, and
Fraumeni, 1987)
Transportation
Services
Locomotive
Marine
Equipment
Agriculture
Construction
Pumps/
compressors
Generators and
Welders
Refrigeration
Industrial
Lawn and
Garden
Engines
Diesel fuel

Derived demand
Derived demand
Derived demand
Derived demand
Derived demand
Derived demand

Derived demand
Derived demand
Derived demand

Derived demand
Derived demand

In the derived demand approach,


• compliance costs increase prices and decrease demand
for products and services in the application markets;
• this in turn leads to reduced demand for diesel

equipment, engines and fuel, which are inputs into the
production of products and services in the application
markets
















                                           10-81

-------
Final Regulatory Impact Analysis
                                            Table 10.3-18
                    Summary of Market Supply Elasticities Used in the NDEIM
 Markets
Estimate
Source
Method
Input Data Summary
 Applications
   Agriculture
   Construction
   Manufacturing
  0.32     Literature-based
           estimate
  1.0      Literature-based
           estimate
  1.0      Literature-based
           estimate
 Transportation Services
  Locomotive          0.6
  Marine
           Literature-based
           estimate
  0.6      Literature-based
           estimate
              Production-weighted
              average of individual
              crop estimates ranging
              from 0.27 to 0.55.
              (Lin etal., 2000)
              Based on Topel and
              Rosen, (1988).a
              Literature estimates are
              not available so assumed
              same value as for
              Construction market


              Method based on Ivaldi
              and McCollough (2001)
              Literature estimates not
              available so assumed
              same value as for
              locomotive market
                 Agricultural Census data
                 1991 - 1995
                 Census data, 1963 - 1983
                 Not applicable
                 Association of American
                 Railroads 1978-1997
                 Not applicable
Equipment
Agriculture
Construction
Pumps/
compressors
Generators/
Welder Sets
Refrigeration
Industrial
Lawn and
Garden
Engines
Diesel fuel

2.14
3.31
2.83
2.91
2.83
5.37
3.37
3.81
0.24

EPA econometric
estimate
EPA econometric
estimate
EPA econometric
estimate
EPA econometric
estimate
EPA econometric
estimate
EPA econometric
estimate
EPA econometric
estimate
EPA econometric
estimate
Literature based
estimate

Cobb-Douglas
production function
Cobb-Douglas
production function
Cobb-Douglas
production function
Cobb-Douglas
production function

Cobb-Douglas
production function
Cobb-Douglas
production function
Cobb-Douglas
production function
Based on Considine
(20021b

Census data 1958-1996; SIC
3523
Census data 1958-1996; SIC
3531
Census data 1958-1996; SIC
3561 and 3563
Census data 1958-1996; SIC
3548
Assumed same as
pumps/compressors
Census data 1958-1996; SIC
3537
Census data 1958-1996; SIC
3524
Census data 1958-1996; SIC
3519
From Energy Intelligence
Grourj (EIGY 1987-2000°
a Most other studies estimate ranges that encompass 1.0, including DiPasquale (1997) and DiPasquale and Wheaton
  (1994).
b Other estimates range from 0.02 to 1.0 (Greene and Tishchishyna, 2000). However, Considine (2002) is one of the few
  studies that estimates a supply elasticity for refinery operations. Most petroleum supply elasticities also include
  extraction.
c This source refers to the data used by Considine in his 2002 study.
                                                 10-82

-------
                                                      Economic Impact Analysis
10.3.8 Model Solution

    10.3.8.1 Computing Baseline and With-Regulation Equilibrium Conditions

    To perform the economic impact analysis, the model compares the baseline equilibrium
conditions and the counterfactual with-regulation equilibrium conditions produced under a
changed policy regime. The assumption of an "observable" baseline equilibrium leads directly to
the need for and construction of a data set that fulfills the equilibrium conditions for markets
included in NDEIM.  For this analysis, we examine the impacts of the rule for 29 years (2007 to
2036). As a result, we need to develop an observable baseline for each of these future years.
This section describes the data and approach used to establish these baselines.

    Developing a Baseline Equilibrium:  In order to construct a baseline for each year,
equilibrium market conditions without the rule were computed using the following three steps:

    •   Collect baseline prices and production data for the most recently available year (2000).
   •   Apply appropriate growth rates to estimate future production for markets included in
       NDEIM, and

   •   Incorporate the impact of increased fuel costs associated with the highway rule prior to
       analysis of the final nonroad rule. We incorporate the impact of the highway rule costs in
       the baseline because they have already been captured in the highway rule analysis; thus,
       we avoid double counting of cost impacts of the highway rule. In effect, our baseline
       market projections are "shocked" by the highway rule and a new set of baseline prices
       and quantities is estimated for all linked markets. This new baseline is the appropriate
       point of departure for analysis of the final nonroad rule.

   It is important to note that the baseline analysis of each year does not incorporate the
cumulative regulatory effects from the highway and nonroad rule in previous years. For
example, the regulatory effects impacts from year 2007 do not affect the baseline conditions for
the years 2008 through 2036.  These dynamic interactions would reduce the estimated impact of
the regulation but are beyond the scope of the modeling effort. As a result, the impact estimates
may be viewed as conservative in that they likely over estimate impacts.

   Shifting the Supply Function: The starting point for assessing the market impacts of a
regulatory action is to incorporate the regulatory compliance costs into the production decision
of the firm. In order to quantify this upward shift, the model the per-unit compliance cost
estimates as the measure of additional cost per unit of producing output*2.  Treatment of
compliance costs in this manner is the conceptual equivalent of a unit tax on output.
QWe discuss the calculation of the appropriate per-unit compliance cost measure used in each
   market in Section 10.2.3.3 oftheRIA.

                                         10-83

-------
Final Regulatory Impact Analysis
   Computing With-Regulation Equilibrium Conditions:  The French economist Leon Walras
proposed one early model of market price adjustment by using the following thought experiment.
Suppose there is a hypothetical agent that facilitates market adjustment by playing the role of an
"auctioneer." He announces prices, collects information about supply and demand responses
(without transactions actually taking place), and continues this process until market equilibrium
is achieved.

   For example, suppose the auctioneer calls out a price (P) that is lower than the equilibrium
price (P*) (see Figure 10.3-1).  He then determines that the quantity demanded (A) exceeds the
quantity supplied (B) and calls  out a new (higher) price (P'). This process continues until P=P*.
A similar analysis takes place when excess supply exists.  The auctioneer calls out lower prices
when the price is higher than the equilibrium price.
                                     Figure 10.3-1.
               For Prices Higher (Lower) than P*, Price Will Fall (Rise)
             S/Q
              P*

              P'

              P
                                 Excess Demand
                                                        D
                                                                 Q/t
    10.3.8.2 Solution Algorithm

    Supply responses and market adjustments can be conceptualized as an interactive process.
Producers facing increased production costs due to compliance are willing to supply smaller
quantities at the baseline price. This reduction in market supply leads to an increase in the
market price that all producers and consumers face, which leads to further responses by
producers and consumers and thus new market prices, and so on.  The new with-regulation
equilibrium is the result of a series of iterations in which price is adjusted and producers and
consumers respond, until a set of stable market prices arises where total market supply equals
market demand.  Market price adjustment takes place based on a price revision rule, described
                                         10-84

-------
                                                      Economic Impact Analysis
below, that adjusts price upward (downward) by a given percentage in response to excess
demand (excess supply).

   The NDEIM model uses a similar type of algorithm for determining with-regulation
equilibria and the process can be summarized by six recursive steps:

   1.  Impose the control costs on affected supply segments, thereby affecting their supply
decisions.

   2.  Recalculate the market supply in each market. Excess demand currently exists.

   3.  Determine the new prices via a price revision rule. We use a rule similar to the factor
       price revision rule described by Kimbell and Harrison (1986). P; is the market price at
       iteration I, qd is the quantity demanded, and qs is the quantity supplied. The parameter z
       influences the magnitude of the price revision and speed of convergence. The revision
       rule increases the price when excess demand exists, lowers the price when excess supply
       exists, and leaves the price unchanged when market demand equals market supply.  The
       price adjustment is expressed as follows:
                                                                                  (10.1)
                                               qs;
   4.  Recalculate market supply with new prices,

   5.  Compute market demand in each market.

   6.  Compare supply and demand in each market. If equilibrium conditions are not satisfied,
       go to Step 3, resulting in a new set of market prices. Repeat until equilibrium conditions
       are satisfied (i.e., the ratio of supply and demand is arbitrarily close to one).  When the
       ratio is appropriately close to one, the market- clearing condition of supply equals
       demand is satisfied.

10.4  Estimating Impacts

   Using the static partial equilibrium analysis, the NDEIM model loops through each year
calculating new market equilibriums based on the projected baseline economic conditions and
compliance cost estimates that shift the supply curves in the model. The model calculates price
and quantity changes and uses these measures to estimate the  social costs of the rule and
partition the impact between producers and consumers.  This approach follows the classical
treatment of tax burden distribution in the public finance literature (e.g., Harberger,  1974).
                                         10-85

-------
Final Regulatory Impact Analysis
References for Chapter 10

Allen, R.G.D.  1938. Mathematical Analysis for Economists.  New York:  St. Martin's Press.
See Docket A-2001-28, Document No. IV-B-25 for relevant excerpts.

Baumol, William.  "Contestable Markets: An Uprising in the theory of Industry Structure,"
American Economic Review, 72, March 1982:1-15.

Baumol, William, John Panzer, and Robert Willig.  1982. Contestable Markets and the Theory of
Industry Structure, San Diego, CA: Harcourt, Brace, Jovanovich.

Braeutigam, R. R. 1999.  "Learning about Transport Costs." In Essays inTransportation
Economics and Policy, edited by J. Gomes-Ibanez, William B. Tye, and Clifford Winston.
Washington: Brookings.

Berck, P., and S. Hoffmann. 2002.  "Assessing the Employment Impacts." Environmental and
Resource Economics 22:133-156.

Bingham, T.H., and TJ. Fox. 1999. "Model Complexity and Scope for Policy Analysis."
Public Administration Quarterly 23(3).

Charles River Associates, Inc. and Baker and O'Brien, Inc. 2000. An Assessment of the
Potential Impacts of Proposed Environmental Regulations on U.S. Refinery Supply of Diesel
Fuel. CRA No. 002316-00 (August 2000).  A copy of this document is available in Docket A-
2001-28, Document No. II-A-17.

Considine, Timothy J.  2002. "Inventories and Market Power in the World Crude Oil Market."
Working paper, Department of Energy, Environmental, and Mineral Economics, The
Pennsylvania State University, University Park, PA. A copy of this document is available at
http://www.personal.psu.edU/faculty/c/p/cpw/resume/InventoriesMarketPowerinCrudeOilMarket
s.pdf. A copy is also available in Docket A-2001-28, Document No. II-A-25.

DiPasquale, Denise.  1997.  "Why Don't We Know More about Housing Supply?" Working
paper, University of Chicago. A copy of this document is available at
http://www.cityresearch.com/pubs/supply.pdf. A copy of this document is also available in
Docket A-2001-28, Document No. II-A-24.

DiPasquale, Denise and William C. Wheaton.  1994. "Housing Market Dynamics and the Future
of Housing Prices." Journal of Urban Economics 35(1): 1-27.

Federal Trade Commission. 2001. Final Report of the Federal Trade Commission: Midwest
Gasoline Price Investigation (March 29, 2001). A copy of this document is available at
http://www.ftc.gov/os/2001/03/mwgasrpt.htm.  This document is also available in Docket A-
2001-28, Document No. II-A-23.
                                        10-86

-------
                                                    Economic Impact Analysis
Finizza, Anthony. 2002.  Economic Benefits of Mitigating Refinery Disruptions: A Suggested
Framework and Analysis of a Strategic Fuels Reserve.  Study conducted for the California
Energy Commission pursuant to California State Assembly Bill AB 2076. (P600-02-018D, July
4, 2002). A copy of this document is available at http://www.energy.ca.gov/reports/
2002-07-08_600-02-018D.PDF. A copy is also available in Docket A-2001-28, Document No.
II-A-18.

Gallaher, Michael. 2003.  Memorandum to Todd Sherwood regarding Clarifications on Several
Modeling Issues (March 24, 2003).  A copy of this memorandum can be found in Docket A-
2001-28, Document No. II-A-37.

Greene, D.L. and N.I. Tishchishyna. 2000. Costs of Oil Dependence: A 2000 Update. Study
prepared by Oak Ridge National Laboratory for the U.S. Department of Energy under contract
DE-AC05-OOOR22725 (O RNL/TM-2000/152, May 2000). This document can be accessed at
http ://www. ornl. gov/~webworks/cpr/v823/rpt/l 07319.pdf  A copy of this document is also
available in Docket A-2001-28, Document No. II-A-21.

Guerra, Zuimdi. Memorandum to EPA Air Docket A-2001-28, regarding Price Database for
New Non-Road Equipment, April 21, 2004. This document is available in electronic docket
OAR-2003-0012, Document No. OAR-2003-0012-0960

Harberger, Arnold C.  1974.  Taxation and Welfare. Chicago: University of Chicago Press.

Hicks, J.R., 1961. Marshall's Third Rule: A Further Comment.  Oxford Economic Papers
13:262-65.   See  Docket A-2001-28, Document No. IV-B-25 for relevant excerpts.

Hicks, J.R., 1963. The Theory of Wages. St. Martins Press, NY, pp. 233-247.  See Docket A-
2001-28, Document No. IV-B-25 for relevant excerpts.

Ivaldi, M. and McCullough, G. 2001.  "Density and Integration Effects  on Class I U.S. Freight
Railroads."  Journal of Regulatory Economics 19:161-162.

Jorgenson, Dale W. 1990. "Productivity and Economic Growth."  In Fifty Years of Economic
Measurement:  The Jubilee Conference on Research in Income and Wealth.  Ernst R. Berndt and
JackE. Triplett (eds.).  Chicago, IL: University of Chicago Press.

Jorgenson, Dale W., Frank M. Gollop, and Barbara M. Fraumeni. 1987. Productivity and U.S.
Economic Growth.  Cambridge, MA: Harvard University Press.

Kimbell, L.J., and G.W. Harrison.  1986.  "On the Solution of General Equilibrium Models."
Economic Modeling 3:197-212.

Klein, C. and Kyle, R.  1997. "Technical Change and the Production of Ocean Shipping
Services." Review of Industrial Organization 12:733-750.
                                        10-87

-------
Final Regulatory Impact Analysis
Lin, William, Paul C. Westcott, Robert Skinner, Scott Sanford, and Daniel G. De La Torre
Ugarte.  2000.  Supply Response under the 1996 Farm Act and Implications for the U.S. Field
Crops Sector. U.S. Department of Agriculture, Economics Research Service, Technical Bulletin
No. 1888 (July 2000). A copy of this document is available at
http://www.ers.usda.gov/publications/tb 1888/tb 1888.pdf. A copy is also available in Docket A-
2001-28, Document No. II-A-20.

MathPro, Inc. 2002. Prospects for Adequate Supply of Ultra Low Sulfur Diesel Fuel in the
Transition Period (2006-2007): An Analysis of Technical and Economic Driving Forces for
Investment in ULSD Capacity in the U.S. Refining Sector. Study prepared for The Alliance of
Automobile Manufacturers and The Engine Manufacturers Association (February 26, 2002).  A
copy of this study is available at http://www.autoalliance.org/ul sd_study.pdf. A copy is also
available in Docket A-2001-28, Document No. II-A-19.

Morgenstern, Richard D., William A. Pizer, and Jhih-Shyang Shih. 2002.  "Jobs Versus the
Environment: An Industry-Level Perspective."  Journal of Environmental Economics and
Management 43:412-436.

NBER-CES. National Bureau of Economic Research and U.S. Census Bureau, Center for
Economic Research. 2002. NBER-CES Manufacturing Industry Database, 1958 - 1996.
http://www.nber.org/nberces/nbprod96.htm A copy of this document is available in Docket A-
2001-28, Document No. II-A-70.

Office Management and Budget (OMB).  1996.  Executive Analysis of Federal Regulations
Under Executive Order 12866. Executive Office of the President, Office Management and
Budget. January 11, 1996. A copy of this document is available at
http://www.whitehouse.gov/omb/inforeg/print/riaguide.html. A copy is also available in Docket
A-2001-28, Document No. II-A-22.

Pizer, Bill. Communications between Mike Gallaher and Bill Pizer on November 5, 2002.
Docket A-2001-28, Document No. II-B-18.

Poterba, James M. 1984.  "Tax Subsidies to Owner Occupied Housing: An Asset Market
Approach," Quarterly Journal of Economics 99:4, pp. 729-52.

Revelt, J.M., 2004.  Memorandum to Docket A-2001-28, regarding Identification of Application
Markets for the Nonroad Diesel Economic Impact Model.  This document is available in Docket
A-2001-28, Document No. IV-B-23.

RTI. 2002. Economic Analysis of Air Pollution Regulations: Boilers and Process Heaters.
Final Report. Prepared for the U.S. Environmental Protection Agency by RTI (November 2002).
EPA Contract No. 68-D-99-024; RTI Project No. 7647-004-385. A copy of this document is
available at http://www.epa.gov/ttn/ecas/regdata/economicimpactsanalysis.pdf. A copy is also
available in Docket A-2001-28, Document No. II-A-16.
                                        10-88

-------
                                                    Economic Impact Analysis
RTI. 2003a. Economic Impact Analysis for Nonroad Diesel Tier 4 Rule. Draft Final Report.
Prepared for the U.S. Environmental Protection Agency by RTI (April 2003). EPA Contract No.
68-D-99-024.  A copy of this document is available in Docket A-2001-28, Document No. II-A-
115.

RTI. 2004. Economic Impact Analysis for Nonroad Diesel Tier 4 Rule: NDEIM Enhancements
and Results, Final Report.  Prepared for the U.S. Environmental Protection Agency by RTI
(April 2004). EPA Contract No. 68-D-99-024.  A copy of this document is available in Docket
A-2001-28.

Sherwood, T.  Memorandum to Air Docket A-2001-28 Re: Engine Sales used in Proposed
Nonroad Tier 4 Cost Analysis.  A copy of this document is available in Docket A-2001-28,
Document No. II-B-37.

Topel, Robert and Sherwin Rosen. 1988. "Housing Investment in the United States." Journal of
Political Economy 96(4):718-40.

U.S. Bureau of Economic Analysis (BEA). 1997 Benchmark I-O Supplementary Make, Use,
and Direct Requirements Tables at the Detailed Level.  Table 4.  Last updated November 24, 2003.

U.S. Census Bureau, 2002.  "Annual Value of Construction Put in Place," C30 Table 101. As
accessed on November 12, 2002.    This
document is available in Docket A-2001-28, Document No. II-B-17.

U.S. Census Bureau. 2003a. Annual Survey of Manufactures. 2001 Statistics for Industry
Groups and Industries.  M01(AS)-1. Washington, DC:  U.S. Census Bureau.  Table 1.
. This document is available in Docket A-
2001-28, Document No. IV-B-29.

U.S. Census Bureau. 2003b. Value of Construction Put In Place:  December 2002. C30/02-12.
Washington, DC: U.S. Census Bureau. Table 1.
 This document is available in Docket
A-2001-28, Document No. IV-B-28.

U.S. Department of Agriculture, National Agricultural  Statistics Service (USDA-NASS).  2002.
Agricultural Statistics 2002.  Washington, DC:  U.S. Department of Agriculture. Table 9-39 and
Table 15-1.  These tables are available in
Docket A-2001-28, Document No. IV-B-27.

U.S. Environmental Protection Agency.  1999.  OAQPSEconomic Analysis Resource Document.
Research Triangle Park, NC: EPA. A copy of this document can be found at
http://www.epa.gov/ttn/ecas/econdata/6807-305.pdf. A copy can also be found in Docket A-
2001-28, Document No. II-A-14.
                                        10-89

-------
Final Regulatory Impact Analysis
U.S. Environmental Protection Agency. 2000. Guidelines for Preparing Economic Analyses.
EPA-240-R-00-003, September 2000.

U.S. Environmental Protection Agency. 2000. Regulatory Impact Analysis, Heavy-Duty Engine
and Vehicle Standards and Highway Diesel Fuel Sulfur Control Requirements (EPA420-R-00-
026).  A copy of this document is available at http://www.epa.gov/otaq/diesel.htmtfdocuments.
A copy is also available in Docket A-2001-28, Document No. II-A-01.

U.S. International Trade Commission (USITC). 2004. U.S. Imports for Consumption, 2000:
NAICS 311 to 339. USITC Interactive Tariff and Trade DataWeb Version
2.6.0.. As obtained on March, 2004.
                                        10-90

-------
                                                     Economic Impact Analysis
       APPENDIX 10A: Impacts on the Engine Markets
   This appendix provides the time series of impacts from 2007 through 2036 for the engine
markets.  Seven separate engine markets were modeled segmented by engine size in horsepower
(the EIA includes more horsepower categories than the standards, allowing more efficient use of
the engine compliance cost estimates developed for this rule):

   •   less than 25 hp
   •   26 to 50 hp
   •   51 to 75 hp
   •   76 to 100 hp
   •   101 to 175 hp
   •   176 to 600 hp
   •   greater than 601 hp

   Tables 10A-1 through 10A-7 provide the time series of impacts for the seven horsepower
markets included in the analysis. Each table includes the following:

   •   average engine price
   •   average engineering costs (variable and fixed) per engine
       -  Note that in the engineering cost analysis, fixed costs for engine manufacturers are
          recovered in the first five years (see Chapter 6)
   •   absolute change in the market price ($)
       -  Note that the estimated absolute change in market price is based on variable costs
          only; see Appendix 101 for a sensitivity analysis including fixed costs as well
   •   relative change in market price (%)
   •   relative change in market quantity (%)
   •   total engineering  (regulatory) costs for merchant engines ($)
   •   change in producer surplus from merchant engine manufacturers

       As described in Section 10.3.3.1, approximately 65 percent of engines are sold on the
market and these are referred to as "merchant" engines. The remaining 35 percent are consumed
internally by integrated equipment manufacturers and are referred to as "captive" engines. The
total engineering costs and changes in producer surplus presented in this appendix include only
merchant engines because captive engines never pass through the engines markets. Fixed and
variable engineering costs and changes in producer surplus associated with captive engines are
included in equipment manufacture impact estimates presented in Appendix 10B.

       All prices and costs are presented in $2002, and real engine prices are assumed to be
constant. The engineering cost per engine typically decreases after 5 years as the annualized
fixed costs are recovered. The price increase after that time is driven by the per-engine variable
costs and remains relatively constant over time.
                                         10-91

-------
Final Regulatory Impact Analysis
       For all the engine size categories, the majority of the cost of the regulation is passed
along through increased engine prices.  Price increases in 2036 are estimated to be $123 (8.2
percent) for engines <25 hp, $645 (22.2 percent) for engines 26 to 50 hp, $636 (21.2 percent) for
engines 51 to 75 hp, $1,121 (28 percent) for engines 76 to 100 hp, $1,350 (24.6 percent) for
engines 101 to 175 hp, $2,122 (10.6 percent) for engines 176 to 600 hp, and $5,343 (6.6 percent)
for engines above 601 hp.

       While the cost per engine and market impacts (in terms of percentage change in price  and
quantity) stabilize in the later years of the regulation, the engineering costs and producer surplus
changes continue to gradually increase  because the projected baseline population of engines
increases over time.
                                         10-92

-------
                                                          Economic Impact Analysis
Table 10A-1. Impacts on the Engine Market and Engine Manufacturers: <25hp
(Average Price per Engine = $l,500)a
Engine (^25Hp)
Year
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2031
2032
2033
2034
2035
2036
NPVb
Engineering
Cost/Unit
—
$162
$161
$154
$153
$152
$123
$123
$123
$123
$123
$123
$123
$123
$123
$123
$123
$123
$123
$123
$123
$123
$123
$123
$123
$123
$123
$123
$123
$123

Absolute
Change in
Price
—
$129
$129
$123
$123
$123
$123
$123
$123
$123
$123
$123
$123
$123
$123
$123
$123
$123
$123
$123
$123
$123
$123
$123
$123
$123
$123
$123
$123
$123

Change in
Price (%)
0.0%
8.6%
8.6%
8.2%
8.2%
8.2%
8.2%
8.2%
8.2%
8.2%
8.2%
8.2%
8.2%
8.2%
8.2%
8.2%
8.2%
8.2%
8.2%
8.2%
8.2%
8.2%
8.2%
8.2%
8.2%
8.2%
8.2%
8.2%
8.2%
8.2%

Change in
Quantity
(%)
-0.001%
-0.002%
-0.002%
-0.004%
-0.007%
-0.009%
-0.010%
-0.011%
-0.011%
-0.011%
-0.011%
-0.011%
-0.011%
-0.011%
-0.011%
-0.011%
-0.011%
-0.011%
-0.011%
-0.011%
-0.011%
-0.011%
-0.011%
-0.011%
-0.011%
-0.011%
-0.011%
-0.011%
-0.011%
-0.011%

Total Change in Producer
Engineering Surplus for Engine
Costs (103) Manufacturers (103)
—
$20,017
$20,449
$20,007
$20,417
$20,827
$17,195
$17,605
$18,015
$18,425
$18,835
$19,245
$19,654
$20,064
$20,474
$20,884
$21,294
$21,704
$22,114
$22,524
$22,934
$23,344
$23,753
$24,163
$24,573
$24,983
$25,393
$25,803
$26,213
$26,623
$370.428
—
-$4,043
-$4,043
-$4,044
-$4,045
-$4,047
-$5
-$6
-$6
-$6
-$6
-$6
-$6
-$7
-$7
-$7
-$7
-$7
-$7
-$7
-$7
-$8
-$8
-$8
-$8
-$8
-$8
-$8
-$9
-$9
-$17.043
a  Figures are in 2002 dollars.
b  Net present values are calculated using a social discount rate of 3 percent over the 2004 to 2036 time period.
                                            10-93

-------
Final Regulatory Impact Analysis
Table 10A-2. Impacts on the Engine Market and Engine Manufacturers: 26-50hp
(Average Price per Engine = $2,900)a
Engine (26hp to 50hp)
Year
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2031
2032
2033
2034
2035
2036
NPVb
Engineering
Cost/Unit
—
$196
$195
$187
$186
$185
$924
$922
$716
$715
$714
$645
$645
$645
$645
$645
$645
$645
$645
$645
$645
$645
$645
$645
$645
$645
$645
$645
$645
$645

Absolute
Change in
Price
—
$147
$147
$139
$139
$139
$849
$849
$645
$645
$645
$645
$645
$645
$645
$645
$645
$645
$645
$645
$645
$645
$645
$645
$645
$645
$645
$645
$645
$645

Change in
Price (%)
0.0%
5.1%
5.1%
4.8%
4.8%
4.8%
29.3%
29.3%
22.2%
22.2%
22.2%
22.2%
22.2%
22.2%
22.2%
22.2%
22.2%
22.2%
22.2%
22.2%
22.2%
22.2%
22.2%
22.2%
22.2%
22.2%
22.2%
22.2%
22.2%
22.2%

Change in
Quantity
(%)
-0.002%
-0.003%
-0.003%
-0.006%
-0.011%
-0.014%
-0.015%
-0.016%
-0.016%
-0.016%
-0.016%
-0.016%
-0.016%
-0.016%
-0.016%
-0.016%
-0.016%
-0.016%
-0.016%
-0.016%
-0.016%
-0.016%
-0.016%
-0.016%
-0.016%
-0.016%
-0.016%
-0.016%
-0.016%
-0.016%

Total Change in Producer
Engineering Surplus for Engine
Costs (103) Manufacturers (103)
—
$26,163
$26,589
$25,943
$26,347
$26,750
$136,464
$138,927
$110,004
$111,875
$113,746
$104,651
$106,522
$108,392
$110,263
$112,134
$114,005
$115,875
$117,746
$119,617
$121,488
$123,359
$125,229
$127,100
$128,971
$130,842
$132,712
$134,583
$136,454
$138,325
$1.722.675
-$1
-$6,592
-$6,592
-$6,595
-$6,600
-$6,604
-$10,981
-$10,983
-$10,983
-$10,984
-$10,984
-$19
-$19
-$19
-$20
-$20
-$20
-$21
-$21
-$21
-$22
-$22
-$22
-$23
-$23
-$23
-$24
-$24
-$24
-$25
-$67.561
a Figures are in 2002 dollars.
b Net present values are calculated using a social discount rate of 3 percent over the 2004 to 2036 time period.
                                             10-94

-------
                                                          Economic Impact Analysis
TablelO.A-3. Impacts on the Engine Market and Engine Manufacturers: 51-75hp
(Average Price per Engine = $3,000)a
Engine (51hp to 75hp)
Year
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2031
2032
2033
2034
2035
2036
NPVb
Engineering
Cost/Unit
—
$217
$216
$206
$205
$205
$913
$912
$710
$709
$708
$636
$636
$636
$636
$636
$636
$636
$636
$636
$636
$636
$636
$636
$636
$636
$636
$636
$636
$636

Absolute
Change in
Price
—
$167
$167
$158
$158
$158
$837
$837
$636
$636
$636
$636
$636
$636
$636
$636
$636
$636
$636
$636
$636
$636
$636
$636
$636
$636
$636
$636
$636
$636

Change in
Price (%)
0.0%
5.6%
5.6%
5.3%
5.3%
5.3%
27.9%
27.9%
21.2%
21.2%
21.2%
21.2%
21.2%
21.2%
21.2%
21.2%
21.2%
21.2%
21.2%
21.2%
21.2%
21.2%
21.2%
21.2%
21.2%
21.2%
21.2%
21.2%
21.2%
21.2%

Change in
Quantity
(%)
-0.002%
-0.004%
-0.004%
-0.006%
-0.011%
-0.014%
-0.015%
-0.017%
-0.017%
-0.017%
-0.017%
-0.017%
-0.017%
-0.017%
-0.017%
-0.017%
-0.017%
-0.017%
-0.017%
-0.017%
-0.017%
-0.017%
-0.017%
-0.017%
-0.017%
-0.017%
-0.017%
-0.017%
-0.017%
-0.017%

Total Change in Producer
Engineering Surplus for Engine
Costs (103) Manufacturers (103)
—
$18,388
$18,650
$18,102
$18,350
$18,597
$84,465
$85,780
$67,870
$68,869
$69,868
$63,844
$64,843
$65,842
$66,841
$67,840
$68,840
$69,839
$70,838
$71,837
$72,836
$73,835
$74,834
$75,833
$76,832
$77,832
$78,831
$79,830
$80,829
$81,828
$1.052.492
-$1
-$4,259
-$4,259
-$4,261
-$4,264
-$4,267
-$7,033
-$7,035
-$7,035
-$7,035
-$7,035
-$13
-$13
-$13
-$13
-$13
-$13
-$14
-$14
-$14
-$14
-$14
-$15
-$15
-$15
-$15
-$15
-$16
-$16
-$16
-$43.432
a  Figures are in 2002 dollars.
b  Net present values are calculated using a social discount rate of 3 percent over the 2004 to 2036 time period.
                                            10-95

-------
Final  Regulatory Impact Analysis
TablelOA-4. Impacts on the Engine Market and Engine Manufacturers:
(Average Price per Engine = $4,000)a
76-1 OOhp
Engine (76hp to lOOhp)
Year
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2031
2032
2033
2034
2035
2036
NPVb
Engineering
Cost/Unit
—
—
$1,213
$1,212
$1,229
$1,227
$1,226
$1,151
$1,150
$1,122
$1,122
$1,122
$1,122
$1,122
$1,122
$1,122
$1,122
$1,122
$1,122
$1,122
$1,122
$1,122
$1,122
$1,122
$1,122
$1,122
$1,122

Absolute
Change in
Price
—
—
$1,133
$1,133
$1,121
$1,121
$1,121
$1,121
$1,121
$1,121
$1,121
$1,121
$1,121
$1,121
$1,121
$1,121
$1,121
$1,121
$1,121
$1,121
$1,121
$1,121
$1,121
$1,121
$1,121
$1,121
$1,121

Change in
Price (%)
0.0%
0.0%
0.0%
0.0%
0.0%
28.3%
28.3%
28.0%
28.0%
28.0%
28.0%
28.0%
28.0%
28.0%
28.0%
28.0%
28.0%
28.0%
28.0%
28.0%
28.0%
28.0%
28.0%
28.0%
28.0%
28.0%
28.0%
28.0%
28.0%
28.0%

Change in
Quantity
(%)
-0.002%
-0.004%
-0.004%
-0.006%
-0.011%
-0.015%
-0.016%
-0.017%
-0.017%
-0.017%
-0.017%
-0.017%
-0.017%
-0.017%
-0.017%
-0.017%
-0.017%
-0.017%
-0.017%
-0.017%
-0.017%
-0.017%
-0.017%
-0.017%
-0.017%
-0.017%
-0.017%
-0.017%
-0.017%
-0.017%

Total Change in Producer
Engineering Surplus for Engine
Costs (103) Manufacturers (103)
—
—
$69,454
$70,577
$72,815
$73,926
$75,037
$71,580
$72,691
$72,001
$73,112
$74,223
$75,334
$76,445
$77,556
$78,667
$79,778
$80,889
$82,000
$83,111
$84,222
$85,333
$86,444
$87,555
$88,666
$89,777
$90,889
$1.098.490
-$1
-$1
-$2
-$3
-$6
-$4,576
-$4,577
-$6,379
-$6,379
-$6,379
-$1,812
-$1,812
-$11
-$11
-$11
-$11
-$12
-$12
-$12
-$12
-$12
-$12
-$13
-$13
-$13
-$13
-$13
-$13
-$14
-$14
-$23.502
a  Figures are in 2002 dollars.
b  Net present values are calculated using a social discount rate of 3 percent over the 2004 to 2036 time period.
                                             10-96

-------
                                                          Economic Impact Analysis
Table 10A-5. Impacts on the Engine Market and Engine Manufacturers:
(Average Price per Engine = $5,500)a
101-175hp
Engine (lOlhp to 175hp)
Year
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2031
2032
2033
2034
2035
2036
NPVb
Engineering
Cost/Unit
—
—
$1,453
$1,452
$1,457
$1,455
$1,454
$1,380
$1,380
$1,351
$1,351
$1,351
$1,351
$1,351
$1,351
$1,351
$1,351
$1,351
$1,351
$1,351
$1,351
$1,351
$1,351
$1,351
$1,351
$1,351
$1,351

Absolute
Change in
Price
—
—
$1,375
$1,375
$1,350
$1,350
$1,350
$1,350
$1,350
$1,350
$1,350
$1,350
$1,350
$1,350
$1,350
$1,350
$1,350
$1,350
$1,350
$1,350
$1,350
$1,350
$1,350
$1,350
$1,350
$1,350
$1,350

Change in
Price (%)
0.0%
0.0%
0.0%
0.0%
0.0%
25.0%
25.0%
24.6%
24.6%
24.6%
24.6%
24.6%
24.6%
24.6%
24.6%
24.6%
24.6%
24.6%
24.6%
24.6%
24.6%
24.6%
24.6%
24.6%
24.6%
24.6%
24.6%
24.6%
24.6%
24.6%

Change in
Quantity
(%)
-0.003%
-0.004%
-0.004%
-0.007%
-0.013%
-0.017%
-0.018%
-0.019%
-0.020%
-0.020%
-0.020%
-0.020%
-0.020%
-0.020%
-0.020%
-0.020%
-0.020%
-0.020%
-0.020%
-0.020%
-0.020%
-0.020%
-0.020%
-0.020%
-0.020%
-0.020%
-0.020%
-0.020%
-0.020%
-0.020%

Total Change in Producer
Engineering Surplus for Engine
Costs (103) Manufacturers (103)
—
—
$90,913
$92,337
$94,162
$95,561
$96,960
$93,480
$94,879
$94,288
$95,687
$97,086
$98,485
$99,884
$101,283
$102,682
$104,081
$105,480
$106,879
$108,278
$109,677
$111,075
$112,474
$113,873
$115,272
$116,671
$118,070
$1.431.405
-$1
-$3
-$3
-$5
-$11
-$4,892
-$4,894
-$6,885
-$6,886
-$6,886
-$2,008
-$2,009
-$19
-$19
-$19
-$19
-$20
-$20
-$20
-$21
-$21
-$21
-$21
-$22
-$22
-$22
-$23
-$23
-$23
-$23
-$25.444
a  Figures are in 2002 dollars.
b  Net present values are calculated using a social discount rate of 3 percent over the 2004 to 2036 time period.
                                            10-97

-------
Final  Regulatory Impact Analysis
Table 10A-6. Impacts on the Engine Market and Engine Manufacturers:
(Average Price per Engine = $20,000)a
176-600hp
Engine (176hp to 600hp)
Year
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2031
2032
2033
2034
2035
2036
NPVb
Engineering
Cost/Unit
—
—
$2,517
$2,511
$2,012
$2,574
$2,567
$2,258
$2,255
$2,253
$2,133
$2,132
$2,132
$2,131
$2,130
$2,130
$2,129
$2,128
$2,128
$2,127
$2,127
$2,126
$2,126
$2,125
$2,124
$2,124
$2,123
$2,123

Absolute
Change in
Price
—
—
$2,191
$2,189
$1,696
$2,136
$2,135
$2,135
$2,134
$2,133
$2,132
$2,131
$2,131
$2,130
$2,129
$2,129
$2,128
$2,127
$2,127
$2,126
$2,126
$2,125
$2,124
$2,124
$2,123
$2,123
$2,122
$2,122

Change in
Price (%)
0.0%
0.0%
0.0%
0.0%
11.0%
10.9%
8.5%
10.7%
10.7%
10.7%
10.7%
10.7%
10.7%
10.7%
10.7%
10.7%
10.6%
10.6%
10.6%
10.6%
10.6%
10.6%
10.6%
10.6%
10.6%
10.6%
10.6%
10.6%
10.6%
10.6%

Change in
Quantity
(%)
-0.003%
-0.004%
-0.004%
-0.008%
-0.014%
-0.018%
-0.019%
-0.021%
-0.021%
-0.021%
-0.021%
-0.021%
-0.021%
-0.021%
-0.021%
-0.021%
-0.021%
-0.021%
-0.021%
-0.021%
-0.021%
-0.021%
-0.021%
-0.021%
-0.021%
-0.021%
-0.021%
-0.021%
-0.021%
-0.021%

Total Change in Producer
Engineering Surplus for Engine
Costs (103) Manufacturers (103)
—
—
$101,112
$102,473
$83,408
$108,339
$109,668
$97,915
$99,244
$100,573
$96,607
$97,936
$99,265
$100,594
$101,923
$103,253
$104,582
$105,911
$107,240
$108,570
$109,899
$111,228
$112,557
$113,887
$115,216
$116,545
$117,874
$119,203
$1.561.195
-$3
-$7
-$7
-$13
-$13,109
-$13,118
-$13,121
-$18,421
-$18,423
-$5,342
-$5,342
-$5,343
-$48
-$48
-$49
-$49
-$50
-$51
-$51
-$52
-$53
-$54
-$54
-$55
-$56
-$56
-$57
-$58
-$58
-$59
-$69.509
a  Figures are in 2002 dollars.
b  Net present values are calculated using a social discount rate of 3 percent over the 2004 to 2036 time period.
                                             10-98

-------
                                                          Economic Impact Analysis
Table 10A-7. Impacts on the Engine Market and Engine Manufacturers: >601hp
(Average Price per Engine = $80,500)a
Engine (>601hp)
Year
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2031
2032
2033
2034
2035
2036
NPVb
Engineering
Cost/Unit
—
—
$3,771
$3,758
$3,081
$3,817
$7,679
$6,857
$6,042
$6,032
$5,780
$5,347
$5,347
$5,347
$5,347
$5,347
$5,347
$5,347
$5,347
$5,347
$5,347
$5,347
$5,347
$5,347
$5,347
$5,347
$5,347
$5,347

Absolute
Change in
Price
i
-$i
$2,908
$2,907
$2,242
$2,730
$6,149
$6,149
$5,343
$5,343
$5,343
$5,343
$5,343
$5,343
$5,343
$5,343
$5,343
$5,343
$5,343
$5,343
$5,343
$5,343
$5,343
$5,343
$5,343
$5,343
$5,343
$5,343

Change in
Price (%)
0.0%
0.0%
0.0%
0.0%
3.6%
3.6%
2.8%
3.4%
7.6%
7.6%
6.6%
6.6%
6.6%
6.6%
6.6%
6.6%
6.6%
6.6%
6.6%
6.6%
6.6%
6.6%
6.6%
6.6%
6.6%
6.6%
6.6%
6.6%
6.6%
6.6%

Change in
Quantity
-0.002%
-0.004%
-0.004%
-0.007%
-0.013%
-0.017%
-0.017%
-0.019%
-0.020%
-0.020%
-0.020%
-0.020%
-0.020%
-0.020%
-0.020%
-0.020%
-0.020%
-0.020%
-0.020%
-0.020%
-0.020%
-0.020%
-0.020%
-0.020%
-0.020%
-0.020%
-0.020%
-0.020%
-0.020%
-0.020%

Total Change in Producer
Engineering Surplus for Engine
Costs (103) Manufacturers (103)
—
—
$6,156
$6,228
$5,182
$6,514
$13,296
$12,044
$10,761
$10,893
$10,582
$9,921
$10,054
$10,187
$10,319
$10,452
$10,584
$10,717
$10,850
$10,982
$11,115
$11,248
$11,380
$11,513
$11,646
$11,778
$11,911
$12,044
$150.134
i
-$2
-$1,409
-$1,410
-$1,411
-$1,856
-$2,649
-$1,244
-$1,244
-$1,244
-$800
-$7
-$7
-$7
-$8
-$8
-$8
-$8
-$8
-$8
-$8
-$8
-$8
-$8
-$9
-$9
-$9
-$9
-$9.762
a  Figures are in 2002 dollars.
b  Net present values are calculated using a social discount rate of 3 percent over the 2004 to 2036 time period.
                                            10-99

-------
Final Regulatory Impact Analysis
       APPENDIX 10B:  Impacts on Equipment Markets
   This appendix provides the time series of impacts from 2007 through 2036 for the equipment
markets.  The equipment markets are the markets associated with the production and
consumption of equipment that use nonroad diesel engines.  Seven equipment types were
modeled:

   •   agricultural
   •   construction
   •   pumps and compressors
   •   generators and welder sets
   •   refrigeration and air conditioning
   •   general industrial
   •   lawn and garden

   Forty-two equipment markets were modeled, representing 7 horsepower categories within 7
application categories. There are 7 horsepower/application categories that did not have sales in
2000 and are not included in the model, so the total number of diesel equipment markets is 42
rather than 49.R

   There are two sets of tables in this  appendix. Tables 10B-1 through 10B-7 provide a
summary of the time series of impacts  for the seven equipment markets included in the analysis.
Tables 10B-8 through 10B-49 provide the time series impacts for each equipment market by
horsepower grouping. Each table includes the following:

   •   average equipment price
   •   average engineering costs (variable and fixed) per piece of equipment
       -   Note that in the engineering cost analysis, fixed costs for equipment manufacturers
          are recovered in the first ten years (see Chapter 6)
   •   absolute change in the market price ($)
       -   Note that the estimated absolute change in market price is based on variable costs
          only; see Appendix 101 for a sensitivity analysis including fixed costs as well
   •   relative change in the market price (%)
   •   relative change in the market quantity (%)
   •   total engineering  (regulatory) costs associated with each market ($)
   •   change in producer surplus for  all manufacturers in the market

   As described in Section 10.3.3.1, approximately 65 percent of engines  are sold on the market
and these are referred to as "merchant" engines. The remaining 35 percent are consumed
RThese seven equipment categories that did not have sales in 2000 are: agricultural equipment >600 hp; gensets &
   welders > 600 hp; refrigeration & A/C > 71 hp (4 hp categories); and lawn & garden >600 hp.

                                        10-100

-------
                                                      Economic Impact Analysis
internally by integrated equipment manufacturers and are referred to as "captive" engines.  The
engineering costs and changes in producer surplus presented in this appendix include total
equipment costs as well as captive engine costs. Because captive engines never pass through the
engines markets, they therefore present an additional cost for integrated equipment producers.

   All prices and costs are presented in $2002, and real equipment prices are assumed to be
constant. The engineering cost per piece of equipment peak around 2014 as the fixed cost per
equipment are phased in and then are depreciated over the next several years.

   A greater percentage of the cost of the regulation is borne by the various equipment markets
than is borne by the engine market. However, a substantial percentage of the cost is still passed
along through increased equipment prices. For each equipment market as a whole, price
increases range from an average increase of 1.31 percent in the general industrial equipment
market to 5.4 percent in the pumps and compressors market. For specific types of equipment,
the price increases range from 0.7 percent for construction <25, 176-600 and >600 hp, and
general industrial equipment (<25 hp), to 9.4 percent for pumps and compressors 76-100 hp.

   Even though the cost per piece of equipment and market impacts (in terms of percentage
change in price and quantity) stabilize after the initial years of the regulation, the engineering
costs and produce surplus changes continue to gradually increase because the projected baseline
population of equipment increases over time.
                                         10-101

-------
Final Regulatory Impact Analysis
Table 10B-1
. Impacts on Agricultural Equipment Market and Manufacturers
(Average Price per Equipment = $24,200)a'b
Agricultural Equipment
Year
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2031
2032
2033
2034
2035
2036
NPVC
Absolute
Engineering Change in
Cost/Unit Price
—
$94
$93
$89
$836
$1,278
$1,432
$1,611
$1,529
$1,448
$1,423
$1,390
$1,349
$1,347
$1,263
$1,230
$1,218
$1,190
$1,189
$1,189
$1,189
$1,189
$1,189
$1,189
$1,188
$1,188
$1,188
$1,188
$1,188
$1,188

-$1
$67
$67
$62
$630
$1,021
$1,158
$1,268
$1,191
$1,191
$1,191
$1,191
$1,190
$1,190
$1,190
$1,190
$1,190
$1,190
$1,189
$1,189
$1,189
$1,189
$1,189
$1,189
$1,188
$1,188
$1,188
$1,188
$1,188
$1,188

Change in
Price (%)
0.0%
0.5%
0.5%
0.5%
0.9%
1.6%
3.1%
3.2%
2.7%
2.7%
2.7%
2.7%
2.7%
2.7%
2.7%
2.7%
2.7%
2.7%
2.7%
2.7%
2.7%
2.7%
2.7%
2.7%
2.7%
2.7%
2.7%
2.7%
2.7%
2.7%

Change in
Quantity
-0.004%
-0.006%
-0.006%
-0.010%
-0.019%
-0.024%
-0.025%
-0.027%
-0.027%
-0.027%
-0.027%
-0.027%
-0.027%
-0.027%
-0.027%
-0.027%
-0.027%
-0.027%
-0.027%
-0.027%
-0.027%
-0.027%
-0.027%
-0.027%
-0.027%
-0.027%
-0.027%
-0.027%
-0.027%
-0.027%

i
Total
Engineering
Costs (103) ]
—
$6,217
$6,304
$6,163
$136,011
$201,592
$205,681
$242,214
$238,948
$227,805
$227,549
$227,388
$223,284
$225,968
$209,555
$203,133
$203,137
$198,628
$201,312
$203,996
$206,680
$209,364
$212,048
$214,731
$217,415
$220,099
$222,783
$225,467
$228,151
$230,834
$3.203.099
Change in Producer
Surplus for
Equipment
Manufacturers (103)
-$114
-$2,359
-$2,364
-$2,578
-$36,021
-$48,332
-$51,844
-$65,974
-$65,991
-$52,188
-$49,273
-$46,453
-$39,690
-$39,703
-$20,621
-$11,540
-$8,884
-$1,716
-$1,740
-$1,764
-$1,788
-$1,813
-$1,837
-$1,861
-$1,885
-$1,909
-$1,933
-$1,957
-$1,982
-$2,006
-$396.969
a   Figures are in 2002 dollars.
b   Average price per equipment for the market is a weighted average of the price of equipment by hp.
0   Net present values are calculated using a social discount rate of 3 percent over the 2004 to 2036 time period.
                                             10-102

-------
                                                          Economic Impact Analysis
Table 10.B-2. Impacts on Construction Equipment Market and Manufacturers
(Average Price per Equipment = $128,100)a'b
Construction Equipment
Year
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2031
2032
2033
2034
2035
2036
NPVC
Absolute
Engineering Change in
Cost/Unit Price
—
$82
$81
$77
$771
$1,342
$1,455
$1,621
$1,658
$1,574
$1,523
$1,495
$1,452
$1,440
$1,359
$1,323
$1,313
$1,285
$1,272
$1,272
$1,272
$1,272
$1,272
$1,272
$1,271
$1,271
$1,271
$1,271
$1,271
$1,271

-$1
$58
$58
$53
$567
$1,073
$1,172
$1,268
$1,285
$1,285
$1,266
$1,266
$1,266
$1,266
$1,265
$1,265
$1,265
$1,265
$1,265
$1,265
$1,265
$1,264
$1,264
$1,264
$1,264
$1,264
$1,264
$1,264
$1,264
$1,263

Change in
Price (%)
0.0%
0.2%
0.2%
0.2%
0.4%
0.9%
1.6%
1.6%
1.4%
1.4%
1.4%
1.4%
1.4%
1.4%
1.4%
1.4%
1.4%
1.4%
1.4%
1.4%
1.4%
1.4%
1.4%
1.4%
1.4%
1.4%
1.4%
1.4%
1.4%
1.4%

Change in
Quantity
(%)
-0.004%
-0.006%
-0.006%
-0.011%
-0.021%
-0.027%
-0.028%
-0.031%
-0.032%
-0.032%
-0.032%
-0.032%
-0.032%
-0.032%
-0.032%
-0.032%
-0.032%
-0.032%
-0.032%
-0.032%
-0.032%
-0.032%
-0.032%
-0.032%
-0.032%
-0.032%
-0.032%
-0.032%
-0.032%
-0.032%

Total
Engineering
Costs (103)
—
$2,791
$2,819
$2,764
$129,258
$222,497
$215,758
$252,584
$277,706
$265,984
$260,346
$261,583
$257,237
$257,684
$237,148
$225,352
$225,367
$218,660
$217,689
$220,554
$223,419
$226,284
$229,149
$232,014
$234,880
$237,745
$240,610
$243,475
$246,340
$249,206
$3.510.842
Change in Producer
Surplus for
Equipment
Manufacturers (103)
-$227
-$1,822
-$1,831
-$2,307
-$41,345
-$60,765
-$64,049
-$81,136
-$87,572
-$72,975
-$68,895
-$67,318
-$60,158
-$57,783
-$34,427
-$19,817
-$17,019
-$7,497
-$3,712
-$3,763
-$3,814
-$3,865
-$3,915
-$3,966
-$4,017
-$4,068
-$4,119
-$4,170
-$4,221
-$4,272
-$545.099
a    Figures are in 2002 dollars.
b    Average price per equipment for the market is a weighted average of the price of equipment by hp.
0    Net present values are calculated using a social discount rate of 3 percent over the 2004 to 2036 time period.
                                            10-103

-------
Final Regulatory Impact Analysis
Table
10B-3. Impacts on Pumps and Compressor Equipment Market and
(Average Price per Equipment = $13,700)a'b
Pumps and Compressors
Year
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2031
2032
2033
2034
2035
2036
NPVC
Engineering
Cost/Unit
—
$135
$134
$128
$340
$682
$952
$1,006
$923
$899
$878
$842
$826
$824
$800
$793
$780
$773
$772
$772
$772
$772
$772
$772
$772
$772
$772
$772
$772
$772

Absolute
Change in
Price
—
$98
$98
$93
$255
$563
$817
$847
$766
$766
$765
$765
$765
$765
$765
$765
$765
$765
$764
$764
$764
$764
$764
$764
$764
$764
$764
$764
$764
$764

Change in
Price (%)
0.0%
1.1%
1.1%
1.1%
1.4%
3.6%
6.1%
6.1%
5.4%
5.4%
5.4%
5.4%
5.4%
5.4%
5.4%
5.4%
5.4%
5.4%
5.4%
5.4%
5.4%
5.4%
5.4%
5.4%
5.4%
5.4%
5.4%
5.4%
5.4%
5.4%

Change in
Quantity
(%)
0.000%
-0.001%
-0.001%
-0.001%
-0.002%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%

i
Total
Engineering
Costs (103) ]
—
$176
$176
$176
$1,011
$2,102
$2,685
$3,136
$3,115
$3,126
$3,137
$2,971
$2,982
$2,993
$2,306
$1,526
$1,155
$785
$784
$795
$805
$816
$827
$838
$849
$860
$871
$882
$893
$904
$27.665
Manufacturers
Change in Producer
Surplus for
Equipment
Manufacturers (103)
—
-$177
-$177
-$177
-$876
-$1,668
-$2,051
-$2,432
-$2,444
-$2,444
-$2,444
-$2,268
-$2,268
-$2,268
-$1,571
-$779
-$398
-$17
-$5
-$5
-$5
-$5
-$5
-$5
-$5
-$5
-$5
-$5
-$5
-$6
-$17.056
a   Figures are in 2002 dollars.
b   Average price per equipment for the market is a weighted average of the price of equipment by hp.
0   Net present values are calculated using a social discount rate of 3 percent over the 2004 to 2036 time period.
                                             10-104

-------
                                                          Economic Impact Analysis
Table 10.B-4. Impacts on Generator Sets and Welding Equipment Market
(Average Price per Equipment = $9,200)a'b
Generator Sets and
Year
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2031
2032
2033
2034
2035
2036
NPVC
Engineering
Cost/Unit
—
$169
$168
$161
$202
$354
$631
$644
$563
$557
$548
$512
$507
$507
$502
$493
$481
$478
$478
$478
$478
$478
$478
$478
$478
$478
$478
$478
$478
$478

Absolute
Change in
Price
—
$123
$123
$117
$149
$285
$553
$558
$479
$479
$479
$479
$479
$479
$479
$479
$479
$479
$479
$479
$479
$478
$478
$478
$478
$478
$478
$478
$478
$478

Change in
Price (%)
0.0%
1.6%
1.6%
1.5%
1.6%
2.3%
5.5%
5.5%
4.6%
4.6%
4.6%
4.6%
4.6%
4.6%
4.6%
4.6%
4.6%
4.6%
4.6%
4.6%
4.6%
4.6%
4.6%
4.6%
4.6%
4.6%
4.6%
4.6%
4.6%
4.6%

Welders
Change in
Quantity
(%)
0.000%
-0.001%
-0.001%
-0.001%
-0.002%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%


Total
Engineering
Costs (103)
—
$7,721
$7,832
$7,677
$11,511
$25,652
$41,613
$43,801
$40,244
$40,403
$40,314
$37,930
$38,054
$38,566
$38,247
$36,440
$34,816
$34,523
$35,035
$35,547
$36,058
$36,570
$37,082
$37,594
$38,106
$38,618
$39,130
$39,642
$40,154
$40,666
$563.662
and Manufacturers
Change in Producer
Surplus for
Equipment
Manufacturers (103)
—
-$2,899
-$2,899
-$2,902
-$4,090
-$7,014
-$9,151
-$10,345
-$10,345
-$9,992
-$9,391
-$6,496
-$6,109
-$6,109
-$5,278
-$2,959
-$824
-$19
-$19
-$19
-$20
-$20
-$20
-$21
-$21
-$21
-$22
-$22
-$22
-$23
-$69.507
a    Figures are in 2002 dollars.
b    Average price per equipment for the market is a weighted average of the price of equipment by hp.
0    Net present values are calculated using a social discount rate of 3 percent over the 2004 to 2036 time period.
                                            10-105

-------
Final Regulatory Impact Analysis
Table 10B-5. Impacts on Refrigeration and Air-Conditioning Equipment Market and
Manufacturers (Average Price per Equipment = $6,3 14)a'b
Refrigeration and Air
Year
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2031
2032
2033
2034
2035
2036
NPVC
Engineering
Cost/Unit
—
$208
$206
$197
$196
$195
$768
$766
$610
$609
$607
$546
$546
$545
$545
$545
$522
$522
$522
$522
$522
$522
$522
$522
$522
$522
$522
$522
$522
$522

Absolute
Change in
Price
—
$152
$152
$144
$143
$143
$676
$676
$521
$521
$521
$521
$521
$521
$521
$521
$521
$521
$521
$521
$521
$521
$521
$521
$521
$521
$521
$521
$521
$521

Conditioning
Change in
Change in Quantity
Price (%) (%)
0.0%
0.6%
0.6%
0.6%
0.6%
0.6%
2.1%
2.1%
1.7%
1.7%
1.7%
1.7%
1.7%
1.7%
1.7%
1.7%
1.7%
1.7%
1.7%
1.7%
1.7%
1.7%
1.7%
1.7%
1.7%
1.7%
1.7%
1.7%
1.7%
1.7%

0.000%
-0.001%
-0.001%
-0.001%
-0.002%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%


Total
Engineering
Costs (103)
—
$447
$447
$447
$447
$447
$2,551
$2,565
$2,418
$2,429
$2,440
$2,005
$2,016
$2,027
$2,038
$2,049
$732
$743
$754
$765
$776
$787
$798
$810
$821
$832
$843
$854
$865
$876
$22.468
Change in Producer
Surplus for
Equipment
Manufacturers (103)
—
-$449
-$449
-$452
-$456
-$459
-$1,792
-$1,793
-$1,793
-$1,793
-$1,793
-$1,347
-$1,347
-$1,347
-$1,348
-$1,348
-$19
-$20
-$20
-$20
-$21
-$21
-$21
-$21
-$22
-$22
-$22
-$23
-$23
-$23
-$12.722
a   Figures are in 2002 dollars.
b   Average price per equipment for the market is a weighted average of the price of equipment by hp.
0   Net present values are calculated using a social discount rate of 3 percent over the 2004 to 2036 time period.
                                             10-106

-------
                                                          Economic Impact Analysis
Table 10.B-6. Impacts on General Industrial Equipment Market and Manufacturers
(Average Price per Equipment = $91,200)a'b
General Industrial
Year
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2031
2032
2033
2034
2035
2036
NPVC
Absolute
Engineering Change in
Cost/Unit Price
—
$64
$63
$60
$516
$1,320
$1,429
$1,549
$1,537
$1,483
$1,431
$1,409
$1,372
$1,366
$1,313
$1,268
$1,260
$1,236
$1,231
$1,231
$1,231
$1,231
$1,231
$1,231
$1,230
$1,230
$1,230
$1,230
$1,230
$1,230

—
$46
$46
$44
$387
$1,101
$1,200
$1,260
$1,242
$1,242
$1,234
$1,234
$1,234
$1,234
$1,234
$1,234
$1,233
$1,233
$1,233
$1,233
$1,233
$1,233
$1,233
$1,233
$1,233
$1,233
$1,233
$1,232
$1,232
$1,232

Change in
Price (%)
0.0%
0.1%
0.1%
0.1%
0.3%
1.1%
1.4%
1.4%
1.3%
1.3%
1.3%
1.3%
1.3%
1.3%
1.3%
1.3%
1.3%
1.3%
1.3%
1.3%
1.3%
1.3%
1.3%
1.3%
1.3%
1.3%
1.3%
1.3%
1.3%
1.3%

Change in
Quantity
(%)
0.000%
-0.001%
-0.001%
-0.001%
-0.002%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%

Total
Engineering
Costs (103)
—
$557
$563
$552
$7,656
$27,925
$29,960
$33,740
$34,239
$34,263
$33,767
$33,729
$33,618
$33,896
$29,901
$24,474
$24,119
$21,873
$21,724
$22,021
$22,319
$22,616
$22,914
$23,212
$23,509
$23,807
$24,104
$24,402
$24,700
$24,997
$401.039
Change in Producer
Surplus for
Equipment
Manufacturers (103)
$1
-$287
-$287
-$294
-$4,870
-$11,353
-$12,069
-$15,024
-$15,489
-$15,216
-$14,467
-$14,131
-$13,723
-$13,705
-$9,412
-$3,688
-$3,036
-$493
-$47
-$47
-$48
-$48
-$49
-$50
-$50
-$51
-$52
-$52
-$53
-$54
-$102.642
a    Figures are in 2002 dollars.
b    Average price per equipment for the market is a weighted average of the price of equipment by hp.
0    Net present values are calculated using a social discount rate of 3 percent over the 2004 to 2036 time period.
                                            10-107

-------
Final Regulatory Impact Analysis
Table 10.B-7. Impacts on Lawn and Garden Equipment Market and Manufacturers
(Average Price per Equipment = $17,700)a'b
Lawn and Garden
Year
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2031
2032
2033
2034
2035
2036
NPVb
Engineering
Cost/Unit
—
$164
$163
$156
$195
$361
$604
$616
$544
$539
$529
$496
$491
$491
$486
$479
$469
$467
$467
$467
$467
$467
$467
$467
$466
$466
$466
$466
$466
$466

Absolute
Change in
Price
—
$119
$119
$113
$144
$292
$530
$535
$465
$465
$465
$465
$465
$465
$465
$465
$465
$465
$465
$465
$465
$465
$465
$465
$465
$465
$465
$465
$465
$465

Change in
Price (%)
0.0%
1.0%
1.0%
0.9%
1.0%
1.4%
2.5%
2.5%
2.2%
2.2%
2.2%
2.2%
2.2%
2.2%
2.2%
2.2%
2.2%
2.2%
2.2%
2.2%
2.2%
2.2%
2.2%
2.2%
2.2%
2.2%
2.2%
2.2%
2.2%
2.2%

Change in
Quantity
(%)
0.000%
-0.001%
-0.001%
-0.001%
-0.002%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%

i
Total
Engineering
Costs (103) ]
—
$2,293
$2,331
$2,289
$2,604
$3,590
$5,759
$6,106
$5,667
$5,734
$5,801
$5,266
$5,333
$5,400
$5,234
$4,596
$4,113
$3,940
$4,007
$4,075
$4,142
$4,209
$4,276
$4,343
$4,410
$4,477
$4,544
$4,611
$4,678
$4,745
$76.592
Change in Producer
Surplus for
Equipment
Manufacturers (103)
—
-$838
-$838
-$839
-$1,074
-$1,780
-$2,097
-$2,338
-$2,338
-$2,338
-$2,338
-$1,736
-$1,736
-$1,736
-$1,503
-$799
-$249
-$9
-$9
-$9
-$10
-$10
-$10
-$10
-$10
-$10
-$10
-$11
-$11
-$11
-$17.642
a   Figures are in 2002 dollars.
b   Average price per equipment for the market is a weighted average of the price of equipment by hp.
0   Net present values are calculated using a social discount rate of 3 percent over the 2004 to 2036 time period.
                                             10-108

-------
           Economic Impact Analysis
Table 10B-8. Impacts on Agricultural Equipment Market and Manufacturers (<25 hp)
(Average Price per Equipment = $6,900)a
Agricultural Equipment
Engineering
Year Cost/Unit
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2031
2032
2033
2034
2035
2036
NPVb
a Figures are
b Net present
—
$177
$176
$168
$167
$166
$136
$136
$135
$135
$135
$123
$123
$123
$123
$123
$123
$123
$123
$123
$123
$123
$123
$123
$123
$123
$123
$123
$123
$123

Absolute
Change in
Price
—
$129
$129
$122
$122
$122
$122
$122
$122
$122
$122
$122
$122
$122
$122
$122
$122
$122
$122
$122
$122
$122
$122
$122
$122
$122
$122
$122
$122
$122

in 2002 dollars.
values are calculated using
Change in
Price (%)
0.0%
1.9%
1.9%
1.8%
1.8%
1.8%
1.8%
1.8%
1.8%
1.8%
1.8%
1.8%
1.8%
1.8%
1.8%
1.8%
1.8%
1.8%
1.8%
1.8%
1.8%
1.8%
1.8%
1.8%
1.8%
1.8%
1.8%
1.8%
1.8%
1.8%

a social discount
(<25hp)
Change in
Quantity
(%)
-0.004%
-0.006%
-0.006%
-0.010%
-0.019%
-0.024%
-0.025%
-0.027%
-0.027%
-0.027%
-0.027%
-0.027%
-0.027%
-0.027%
-0.027%
-0.027%
-0.027%
-0.027%
-0.027%
-0.027%
-0.027%
-0.027%
-0.027%
-0.027%
-0.027%
-0.027%
-0.027%
-0.027%
-0.027%
-0.027%


Total
Engineering
Costs (103)
—
$666
$675
$666
$674
$683
$608
$617
$625
$634
$642
$395
$404
$412
$421
$429
$437
$446
$454
$463
$471
$479
$488
$496
$505
$513
$522
$530
$538
$547
$9.600
rate of 3 percent over the 2004
Change in Producer
Surplus for
Equipment
Manufacturers (103)
-$1
-$341
-$341
-$343
-$348
-$351
-$269
-$271
-$271
-$272
-$272
-$17
-$18
-$18
-$18
-$19
-$19
-$19
-$20
-$20
-$21
-$21
-$21
-$22
-$22
-$22
-$23
-$23
-$24
-$24
-$2.622
to 2036 time period.
10-109

-------
Final Regulatory Impact Analysis
Table 10B-9. Impacts on Agricultural Equipment Market and Manufacturers (26-50 hp)
(Average Price per Equipment = $14,400)a
Agricultural Equipment (25^hp<50)
Absolute
Engineering Change in
Year Cost/Unit Price
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2031
2032
2033
2034
2035
2036
NPVb
a Figures are
b Net present
—
$204
$203
$194
$193
$192
$986
$984
$773
$771
$769
$693
$692
$692
$691
$691
$661
$661
$661
$661
$661
$661
$661
$661
$661
$661
$661
$661
$661
$661

—
$147
$147
$139
$138
$138
$868
$868
$660
$660
$660
$660
$660
$660
$660
$660
$660
$660
$660
$660
$660
$660
$660
$660
$660
$660
$660
$660
$660
$660

in 2002 dollars.
values are calculated using
Change in
Price (%)
0.0%
1.0%
1.0%
1.0%
1.0%
1.0%
6.0%
6.0%
4.6%
4.6%
4.6%
4.6%
4.6%
4.6%
4.6%
4.6%
4.6%
4.6%
4.6%
4.6%
4.6%
4.6%
4.6%
4.6%
4.6%
4.6%
4.6%
4.6%
4.6%
4.6%

a social discount
Change in Total
Quantity Engineering
(%) Costs (103)
-0.004%
-0.006%
-0.006%
-0.010%
-0.019%
-0.024%
-0.025%
-0.027%
-0.027%
-0.027%
-0.027%
-0.027%
-0.027%
-0.027%
-0.027%
-0.027%
-0.027%
-0.027%
-0.027%
-0.027%
-0.027%
-0.027%
-0.027%
-0.027%
-0.027%
-0.027%
-0.027%
-0.027%
-0.027%
-0.027%

rate of 3 percent
—
$3,707
$3,762
$3,679
$3,731
$3,782
$20,616
$20,951
$17,064
$17,319
$17,575
$16,061
$16,316
$16,571
$16,826
$17,081
$15,546
$15,801
$16,057
$16,312
$16,567
$16,822
$17,077
$17,332
$17,587
$17,842
$18,097
$18,353
$18,608
$18,863
$248.449
over the 2004
Change in Producer
Surplus for
Equipment
Manufacturers (103)
-$7
-$1,225
-$1,225
-$1,238
-$1,268
-$1,284
-$3,639
-$3,648
-$3,649
-$3,651
-$3,653
-$1,886
-$1,887
-$1,888
-$1,890
-$1,891
-$103
-$105
-$107
-$108
-$110
-$112
-$114
-$115
-$117
-$119
-$121
-$122
-$124
-$126
-$25.062
to 2036 time period.
                                   10-110

-------
           Economic Impact Analysis
Table 10B-10. Impacts on Agricultural Equipment Market and Manufacturers (51-75 hp)
(Average Price per Equipment = $22,600)a
Agricultural Equipment (50^hp<75)
Absolute
Engineering Change in
Year Cost/Unit Price
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2031
2032
2033
2034
2035
2036
NPVb
a Figures are
b Net present
—
$226
$225
$214
$213
$212
$978
$976
$769
$767
$765
$687
$686
$686
$685
$685
$653
$653
$653
$653
$653
$653
$653
$653
$653
$653
$653
$653
$653
$653

—
$167
$167
$157
$156
$155
$856
$856
$651
$651
$651
$651
$651
$651
$651
$651
$651
$651
$651
$651
$651
$651
$651
$651
$651
$651
$651
$651
$651
$651

in 2002 dollars.
values are calculated using
Change in
Price (%)
0.0%
0.7%
0.7%
0.7%
0.7%
0.7%
3.8%
3.8%
2.9%
2.9%
2.9%
2.9%
2.9%
2.9%
2.9%
2.9%
2.9%
2.9%
2.9%
2.9%
2.9%
2.9%
2.9%
2.9%
2.9%
2.9%
2.9%
2.9%
2.9%
2.9%

a social discount
Change in Total
Quantity Engineering
(%) Costs (103)
-0.004%
-0.006%
-0.006%
-0.010%
-0.019%
-0.024%
-0.025%
-0.027%
-0.027%
-0.027%
-0.027%
-0.027%
-0.027%
-0.027%
-0.027%
-0.027%
-0.027%
-0.027%
-0.027%
-0.027%
-0.027%
-0.027%
-0.027%
-0.027%
-0.027%
-0.027%
-0.027%
-0.027%
-0.027%
-0.027%

rate of 3 percent
—
$1,844
$1,867
$1,818
$1,840
$1,863
$9,199
$9,326
$7,616
$7,713
$7,810
$7,086
$7,183
$7,280
$7,377
$7,474
$6,681
$6,777
$6,874
$6,971
$7,068
$7,165
$7,262
$7,359
$7,456
$7,553
$7,650
$7,747
$7,844
$7,941
$108.842
over the 2004
Change in Producer
Surplus for
Equipment
Manufacturers (103)
-$5
-$582
-$583
-$592
-$615
-$627
-$1,771
-$1,778
-$1,778
-$1,780
-$1,781
-$961
-$962
-$963
-$964
-$965
-$76
-$77
-$78
-$79
-$80
-$81
-$82
-$84
-$85
-$86
-$87
-$88
-$89
-$90
-$12.491
to 2036 time period.
10-111

-------
Final Regulatory Impact Analysis
Table 10B-1 1. Impacts on Agricultural Equipment Market and Manufacturers (76-100 hp)
(Average Price per Equipment = $22,400)a


Engineering
Year Cost/Unit
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2031
2032
2033
2034
2035
2036
NPVb
a Figures are
b Net present
—
—
—
—
—
$1,303
$1,302
$1,325
$1,324
$1,322
$1,247
$1,246
$1,218
$1,218
$1,218
$1,218
$1,218
$1,218
$1,218
$1,218
$1,218
$1,218
$1,218
$1,218
$1,218
$1,218
$1,218
$1,218
$1,218
$1,218

Agricultural
Absolute
Change in
Price
—
-$1
-$1
-$1
-$3
$1,175
$1,175
$1,166
$1,166
$1,166
$1,166
$1,166
$1,166
$1,166
$1,166
$1,166
$1,166
$1,166
$1,166
$1,166
$1,166
$1,166
$1,166
$1,166
$1,166
$1,166
$1,166
$1,166
$1,166
$1,166

in 2002 dollars.
values are calculated using a
Equipment
Change in
Price (%)
0.0%
0.0%
0.0%
0.0%
0.0%
3.5%
3.5%
3.5%
3.5%
3.5%
3.5%
3.5%
3.5%
3.5%
3.5%
3.5%
3.5%
3.5%
3.5%
3.5%
3.5%
3.5%
3.5%
3.5%
3.5%
3.5%
3.5%
3.5%
3.5%
3.5%

(70shp<100)
Change in
Quantity
(%)
-0.004%
-0.006%
-0.006%
-0.010%
-0.019%
-0.024%
-0.025%
-0.027%
-0.027%
-0.027%
-0.027%
-0.027%
-0.027%
-0.027%
-0.027%
-0.027%
-0.027%
-0.027%
-0.027%
-0.027%
-0.027%
-0.027%
-0.027%
-0.027%
-0.027%
-0.027%
-0.027%
-0.027%
-0.027%
-0.027%


Total
Engineering
Costs (103)
—
—
—
—
—
$13,727
$13,923
$14,767
$14,962
$15,157
$14,600
$14,796
$14,695
$14,890
$15,085
$13,661
$13,857
$13,635
$13,830
$14,026
$14,221
$14,416
$14,612
$14,807
$15,002
$15,198
$15,393
$15,588
$15,784
$15,979
$206.738
social discount rate of 3 percent over the 2004
Change in Producer
Surplus for
Equipment
Manufacturers (103)
-$5
-$10
-$10
-$18
-$39
-$2,422
-$2,426
-$3,146
-$3,146
-$3,147
-$2,396
-$2,397
-$2,102
-$2,102
-$2,103
-$485
-$486
-$70
-$71
-$72
-$73
-$74
-$75
-$76
-$77
-$78
-$79
-$80
-$81
-$82
-$18.829
to 2036 time period.
                                   10-112

-------
           Economic Impact Analysis
Table 10B-12. Impacts on Agricultural Equipment Market and Manufacturers (101-175 hp)
(Average Price per Equipment = $69,100)a
Agricultural
Absolute
Engineering Change in
Year Cost/Unit Price
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2031
2032
2033
2034
2035
2036
NPVb
a Figures are
b Net present
—
—
—
—
$1,623
$1,619
$1,664
$1,659
$1,654
$1,577
$1,574
$1,542
$1,539
$1,537
$1,388
$1,387
$1,351
$1,351
$1,351
$1,351
$1,351
$1,351
$1,351
$1,351
$1,351
$1,351
$1,351
$1,351
$1,351

in 2002 dollars.
values are calculated
-$!
-$i
-$3
-$6
$1,414
$1,414
$1,391
$1,391
$1,391
$1,391
$1,391
$1,391
$1,391
$1,391
$1,391
$1,391
$1,391
$1,391
$1,391
$1,391
$1,391
$1,391
$1,391
$1,391
$1,391
$1,391
$1,391
$1,391
$1,391

using a
Equipment (10(khp<175)
Change in
Price (%)
0.0%
0.0%
0.0%
0.0%
0.0%
2.0%
2.0%
2.0%
2.0%
2.0%
2.0%
2.0%
2.0%
2.0%
2.0%
2.0%
2.0%
2.0%
2.0%
2.0%
2.0%
2.0%
2.0%
2.0%
2.0%
2.0%
2.0%
2.0%
2.0%
2.0%

social discount
Change in Total
Quantity Engineering
(%) Costs (103)
-0.004%
-0.006%
-0.006%
-0.010%
-0.019%
-0.024%
-0.025%
-0.027%
-0.027%
-0.027%
-0.027%
-0.027%
-0.027%
-0.027%
-0.027%
-0.027%
-0.027%
-0.027%
-0.027%
-0.027%
-0.027%
-0.027%
-0.027%
-0.027%
-0.027%
-0.027%
-0.027%
-0.027%
-0.027%
-0.027%

rate of 3 percent
—
—
—
—
$50,277
$50,949
$53,852
$54,515
$55,178
$53,654
$54,317
$54,087
$54,750
$55,413
$48,590
$49,253
$48,004
$48,667
$49,330
$49,993
$50,656
$51,319
$51,982
$52,645
$53,308
$53,971
$54,634
$55,298
$55,961
$741.939
over the 2004
Change in Producer
Surplus for
Equipment
Manufacturers (103)
-$28
-$59
-$60
-$113
-$241
-$9,980
-$10,007
-$12,849
-$12,853
-$12,859
-$10,677
-$10,684
-$9,797
-$9,800
-$9,804
-$2,324
-$2,330
-$424
-$430
-$436
-$442
-$448
-$454
-$460
-$466
-$472
-$478
-$484
-$491
-$497
-$81.965
to 2036 time period.
10-113

-------
Final Regulatory Impact Analysis
Table 10B-13. Impacts on Agricultural Equipment Market and Manufacturers (176-600 hp)
(Average Price per Equipment = $143,700)a
Agricultural Equipment (175^hp<600)
Absolute
Engineering Change in
Year Cost/Unit Price
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2031
2032
2033
2034
2035
2036
NPVb
a Figures are
b Net present
—
—
—
—
$2,970
$2,958
$2,439
$3,107
$3,092
$2,777
$2,768
$2,759
$2,634
$2,627
$2,294
$2,292
$2,291
$2,209
$2,208
$2,208
$2,207
$2,206
$2,206
$2,205
$2,204
$2,204
$2,203
$2,203
$2,202
$2,202
-$1
-$3
-$3
-$6
$2,255
$2,251
$1,741
$2,200
$2,199
$2,198
$2,197
$2,197
$2,196
$2,195
$2,194
$2,194
$2,193
$2,192
$2,191
$2,191
$2,190
$2,189
$2,189
$2,188
$2,187
$2,187
$2,186
$2,186
$2,185
$2,185
Change in
Price (%)
0.0%
0.0%
0.0%
0.0%
1.6%
1.6%
1.2%
1.5%
1.5%
1.5%
1.5%
1.5%
1.5%
1.5%
1.5%
1.5%
1.5%
1.5%
1.5%
1.5%
1.5%
1.5%
1.5%
1.5%
1.5%
1.5%
1.5%
1.5%
1.5%
1.5%
Change in Total
Quantity Engineering
(%) Costs (103)
-0.004%
-0.006%
-0.006%
-0.010%
-0.019%
-0.024%
-0.025%
-0.027%
-0.027%
-0.027%
-0.027%
-0.027%
-0.027%
-0.027%
-0.027%
-0.027%
-0.027%
-0.027%
-0.027%
-0.027%
-0.027%
-0.027%
-0.027%
-0.027%
-0.027%
-0.027%
-0.027%
-0.027%
-0.027%
-0.027%
—
—
—
—
$129,766
$131,260
$110,384
$142,701
$144,166
$131,803
$133,268
$134,733
$130,600
$132,065
$114,433
$115,898
$117,363
$113,965
$115,430
$116,895
$118,360
$119,824
$121,289
$122,754
$124,219
$125,684
$127,149
$128,614
$130,079
$131,544
$1.887.531
in 2002 dollars.
values are calculated using
a social discount
rate of 3 percent
over the 2004
Change in Producer
Surplus for
Equipment
Manufacturers (103)
-$68
-$143
-$146
-$274
-$33,510
-$33,668
-$33,733
-$44,283
-$44,293
-$30,479
-$30,494
-$30,508
-$24,924
-$24,931
-$5,842
-$5,856
-$5,870
-$1,021
-$1,035
-$1,048
-$1,062
-$1,076
-$1,090
-$1,104
-$1,118
-$1,132
-$1,145
-$1,159
-$1,173
-$1,187
-$256.000
to 2036 time period.
                                   10-114

-------
           Economic Impact Analysis
Table 10.
B-14. Impacts on Construction Equipment Market and Manufacturers (<25 hp)
(Average Price per Equipment = $18,000)a
Construction Equipment (<25hp)
Engineering
Year Cost/Unit
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2031
2032
2033
2034
2035
2036
NPVb
a Figures are
b Net present
—
$177
$176
$168
$167
$166
$136
$136
$135
$135
$135
$123
$123
$123
$123
$123
$123
$123
$123
$123
$123
$123
$123
$123
$123
$123
$123
$123
$123
$123

Absolute
Change in
Price
—
$129
$129
$122
$122
$121
$121
$121
$121
$121
$121
$121
$121
$121
$121
$121
$121
$121
$121
$121
$121
$121
$121
$121
$121
$121
$121
$121
$121
$121

in 2002 dollars.
values are calculated using
Change in
Price (%)
0.0%
0.7%
0.7%
0.7%
0.7%
0.7%
0.7%
0.7%
0.7%
0.7%
0.7%
0.7%
0.7%
0.7%
0.7%
0.7%
0.7%
0.7%
0.7%
0.7%
0.7%
0.7%
0.7%
0.7%
0.7%
0.7%
0.7%
0.7%
0.7%
0.7%

a social discount
Change in
Quantity
(%)
-0.004%
-0.006%
-0.006%
-0.011%
-0.021%
-0.027%
-0.028%
-0.031%
-0.032%
-0.032%
-0.032%
-0.032%
-0.032%
-0.032%
-0.032%
-0.032%
-0.032%
-0.032%
-0.032%
-0.032%
-0.032%
-0.032%
-0.032%
-0.032%
-0.032%
-0.032%
-0.032%
-0.032%
-0.032%
-0.032%

Total
Engineering
Costs (103)
—
$370
$371
$370
$371
$372
$364
$365
$366
$367
$368
$39
$40
$41
$42
$42
$43
$44
$45
$46
$47
$47
$48
$49
$50
$51
$52
$52
$53
$54
$3.325
rate of 3 percent over the 2004
Change in Producer
Surplus for
Equipment
Manufacturers (103)
-$3
-$343
-$344
-$350
-$363
-$371
-$365
-$370
-$372
-$373
-$374
-$46
-$47
-$48
-$48
-$49
-$50
-$51
-$52
-$53
-$54
-$55
-$56
-$57
-$58
-$59
-$60
-$61
-$62
-$63
-$3.348
to 2036 time period.
10-115

-------
Final Regulatory Impact Analysis
Table 10.B-15. Impacts on Construction Equipment Market and Manufacturers (26-50 hp)
(Average Price per Equipment = $29,700)a
Construction Equipment (25^hp<50)
Absolute
Engineering Change in
Year Cost/Unit Price
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2031
2032
2033
2034
2035
2036
NPVb
a Figures are
b Net present
—
$204
$203
$194
$193
$192
$986
$984
$773
$771
$769
$693
$692
$692
$691
$691
$661
$661
$661
$661
$661
$661
$661
$661
$661
$661
$661
$661
$661
$661

—
$146
$146
$138
$137
$137
$867
$867
$659
$659
$659
$659
$659
$659
$659
$659
$659
$659
$659
$659
$659
$659
$659
$659
$659
$659
$659
$659
$659
$659

in 2002 dollars.
values are calculated using
Change in
Price (%)
0.0%
0.5%
0.5%
0.5%
0.5%
0.5%
2.9%
2.9%
2.2%
2.2%
2.2%
2.2%
2.2%
2.2%
2.2%
2.2%
2.2%
2.2%
2.2%
2.2%
2.2%
2.2%
2.2%
2.2%
2.2%
2.2%
2.2%
2.2%
2.2%
2.2%

a social discount
Change in Total
Quantity Engineering
(%) Costs (103)
-0.004%
-0.006%
-0.006%
-0.011%
-0.021%
-0.027%
-0.028%
-0.031%
-0.032%
-0.032%
-0.032%
-0.032%
-0.032%
-0.032%
-0.032%
-0.032%
-0.032%
-0.032%
-0.032%
-0.032%
-0.032%
-0.032%
-0.032%
-0.032%
-0.032%
-0.032%
-0.032%
-0.032%
-0.032%
-0.032%

rate of 3 percent
—
$438
$440
$437
$439
$441
$3,293
$3,323
$3,006
$3,030
$3,053
$2,723
$2,747
$2,770
$2,794
$2,817
$1,428
$1,451
$1,475
$1,498
$1,521
$1,545
$1,568
$1,592
$1,615
$1,639
$1,662
$1,685
$1,709
$1,732
$32.256
over the 2004
Change in Producer
Surplus for
Equipment
Manufacturers (103)
-$8
-$345
-$345
-$362
-$397
-$420
-$1,864
-$1,875
-$1,882
-$1,884
-$1,885
-$1,534
-$1,536
-$1,538
-$1,540
-$1,543
-$132
-$134
-$137
-$139
-$141
-$143
-$145
-$148
-$150
-$152
-$154
-$156
-$159
-$161
-$14.120
to 2036 time period.
                                   10-116

-------
           Economic Impact Analysis
Table 10.B-16. Impacts on Construction Equipment Market and Manufacturers (51-75 hp)
(Average Price per Equipment = $3 l,600)a
Construction Equipment
Engineering
Year Cost/Unit
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2031
2032
2033
2034
2035
2036
NPVb
a Figures are
b Net present
—
$226
$225
$214
$213
$212
$978
$976
$769
$767
$765
$687
$686
$686
$685
$685
$653
$653
$653
$653
$653
$653
$653
$653
$653
$653
$653
$653
$653
$653

Absolute
Change in
Price
—
$167
$167
$157
$156
$155
$856
$856
$650
$650
$650
$650
$650
$650
$650
$650
$650
$650
$650
$650
$650
$650
$650
$650
$650
$650
$650
$650
$650
$650

in 2002 dollars.
values are calculated using
Change in
Price (%)
0.0%
0.5%
0.5%
0.5%
0.5%
0.5%
2.7%
2.7%
2.1%
2.1%
2.1%
2.1%
2.1%
2.1%
2.1%
2.1%
2.1%
2.1%
2.1%
2.1%
2.1%
2.1%
2.1%
2.1%
2.1%
2.1%
2.1%
2.1%
2.1%
2.1%

(50shp<70)
Change in
Quantity
(%)
-0.004%
-0.006%
-0.006%
-0.011%
-0.021%
-0.027%
-0.028%
-0.031%
-0.032%
-0.032%
-0.032%
-0.032%
-0.032%
-0.032%
-0.032%
-0.032%
-0.032%
-0.032%
-0.032%
-0.032%
-0.032%
-0.032%
-0.032%
-0.032%
-0.032%
-0.032%
-0.032%
-0.032%
-0.032%
-0.032%


Total
Engineering
Costs (103)
—
$1,983
$2,007
$1,957
$1,980
$2,002
$10,288
$10,422
$8,629
$8,731
$8,834
$7,991
$8,093
$8,196
$8,298
$8,401
$7,067
$7,169
$7,272
$7,374
$7,477
$7,580
$7,682
$7,785
$7,887
$7,990
$8,092
$8,195
$8,297
$8,400
$118.863
a social discount rate of 3 percent over the 2004
Change in Producer
Surplus for
Equipment
Manufacturers (103)
-$8
-$710
-$711
-$728
-$764
-$788
-$2,484
-$2,495
-$2,502
-$2,504
-$2,505
-$1,561
-$1,563
-$1,565
-$1,567
-$1,569
-$134
-$136
-$138
-$140
-$142
-$144
-$146
-$148
-$150
-$152
-$154
-$156
-$158
-$160
-$17.987
to 2036 time period.
10-117

-------
Final Regulatory Impact Analysis
Table 10.B-17 Impacts on Construction Equipment Market and Manufacturers (76-100 hp)
(Average Price per Equipment = $57,900)a


Engineering
Year Cost/Unit
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2031
2032
2033
2034
2035
2036
NPVb
a Figures are
b Net present
—
—
—
—
—
$1,303
$1,302
$1,325
$1,324
$1,322
$1,247
$1,246
$1,218
$1,218
$1,218
$1,218
$1,218
$1,218
$1,218
$1,218
$1,218
$1,218
$1,218
$1,218
$1,218
$1,218
$1,218
$1,218
$1,218
$1,218

Construction
Absolute
Change in
Price
—
-$1
-$1
-$2
-$3
$1,174
$1,174
$1,165
$1,164
$1,164
$1,164
$1,164
$1,164
$1,164
$1,164
$1,164
$1,164
$1,164
$1,164
$1,164
$1,164
$1,164
$1,164
$1,164
$1,164
$1,164
$1,164
$1,164
$1,164
$1,164

Equipment
Change in
Price (%)
0.0%
0.0%
0.0%
0.0%
0.0%
2.0%
2.0%
2.0%
2.0%
2.0%
2.0%
2.0%
2.0%
2.0%
2.0%
2.0%
2.0%
2.0%
2.0%
2.0%
2.0%
2.0%
2.0%
2.0%
2.0%
2.0%
2.0%
2.0%
2.0%
2.0%

(7(khp<100)
Change in
Quantity
(%)
-0.004%
-0.006%
-0.006%
-0.011%
-0.021%
-0.027%
-0.028%
-0.031%
-0.032%
-0.032%
-0.032%
-0.032%
-0.032%
-0.032%
-0.032%
-0.032%
-0.032%
-0.032%
-0.032%
-0.032%
-0.032%
-0.032%
-0.032%
-0.032%
-0.032%
-0.032%
-0.032%
-0.032%
-0.032%
-0.032%


Total
Engineering
Costs (103)
—
—
—
—
—
$23,156
$23,465
$25,237
$25,545
$25,854
$25,024
$25,333
$25,192
$25,501
$25,809
$21,977
$22,285
$21,527
$21,836
$22,144
$22,452
$22,761
$23,069
$23,377
$23,686
$23,994
$24,303
$24,611
$24,919
$25,228
$339.723
in 2002 dollars.
values are calculated using a social discount rate of 3 percent over the 2004
Change in Producer
Surplus for
Equipment
Manufacturers (103)
-$15
-$30
-$31
-$62
-$127
-$5,449
-$5,460
-$6,995
-$7,007
-$7,011
-$5,875
-$5,879
-$5,434
-$5,437
-$5,440
-$1,303
-$1,306
-$244
-$247
-$251
-$254
-$258
-$262
-$265
-$269
-$272
-$276
-$279
-$283
-$287
-$45.057
to 2036 time period.
                                   10-118

-------
           Economic Impact Analysis
Table 10.B-18. Impacts on Construction Equipment Market and Manufacturers (101-175 hp)
(Average Price per Equipment = $122,700)a
Construction Equipment (10(khp<175)
Absolute
Engineering Change in
Year Cost/Unit Price
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2031
2032
2033
2034
2035
2036
NPVb
a Figures are
b Net present
—
—
—
—
—
$1,623
$1,619
$1,664
$1,659
$1,654
$1,577
$1,574
$1,542
$1,539
$1,537
$1,388
$1,387
$1,351
$1,351
$1,351
$1,351
$1,351
$1,351
$1,351
$1,351
$1,351
$1,351
$1,351
$1,351
$1,351

-$1
-$2
-$2
-$4
-$7
$1,412
$1,411
$1,389
$1,388
$1,388
$1,388
$1,388
$1,388
$1,388
$1,388
$1,388
$1,388
$1,388
$1,388
$1,388
$1,388
$1,388
$1,388
$1,388
$1,388
$1,388
$1,388
$1,388
$1,388
$1,388

in 2002 dollars.
values are calculated using
Change in
Price (%)
0.0%
0.0%
0.0%
0.0%
0.0%
1.2%
1.2%
1.1%
1.1%
1.1%
1.1%
1.1%
1.1%
1.1%
1.1%
1.1%
1.1%
1.1%
1.1%
1.1%
1.1%
1.1%
1.1%
1.1%
1.1%
1.1%
1.1%
1.1%
1.1%
1.1%

a social discount
Change in Total
Quantity Engineering
(%) Costs (103)
-0.004%
-0.006%
-0.006%
-0.011%
-0.021%
-0.027%
-0.028%
-0.031%
-0.032%
-0.032%
-0.032%
-0.032%
-0.032%
-0.032%
-0.032%
-0.032%
-0.032%
-0.032%
-0.032%
-0.032%
-0.032%
-0.032%
-0.032%
-0.032%
-0.032%
-0.032%
-0.032%
-0.032%
-0.032%
-0.032%
$1
—
—
—
—
—
$68,698
$69,612
$73,652
$74,553
$75,455
$73,387
$74,289
$73,979
$74,881
$75,783
$66,164
$67,065
$65,280
$66,182
$67,083
$67,985
$68,887
$69,788
$70,690
$71,592
$72,493
$73,395
$74,297
$75,198
$76,100
.011.838
rate of 3 percent over the 2004
Change in Producer
Surplus for
Equipment
Manufacturers (103)
-$51
-$105
-$107
-$215
-$438
-$14,076
-$14,114
-$18,081
-$18,122
-$18,134
-$15,171
-$15,183
-$13,984
-$13,994
-$14,004
-$3,496
-$3,508
-$833
-$844
-$856
-$868
-$880
-$891
-$903
-$915
-$927
-$939
-$950
-$962
-$974
-$118.002
to 2036 time period.
10-119

-------
Final Regulatory Impact Analysis
Table 10.B-19. Impacts on Construction Equipment Market and Manufacturers (176-600 hp)
(Average Price per Equipment = $3 12,900)a
Construction
Absolute
Engineering Change in
Year Cost/Unit Price
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2031
2032
2033
2034
2035
2036
NPVb
a Figures are
b Net present
—
—
—
—
$2,970
$2,958
$2,439
$3,107
$3,092
$2,777
$2,768
$2,759
$2,634
$2,627
$2,294
$2,292
$2,291
$2,209
$2,208
$2,208
$2,207
$2,206
$2,206
$2,205
$2,204
$2,204
$2,203
$2,203
$2,202
$2,202
-$2
-$5
-$5
-$9
$2,248
$2,241
$1,731
$2,189
$2,187
$2,186
$2,185
$2,184
$2,184
$2,183
$2,182
$2,181
$2,181
$2,180
$2,179
$2,178
$2,178
$2,177
$2,176
$2,176
$2,175
$2,175
$2,174
$2,173
$2,173
$2,172
Equipment (175^hp<600)
Change in
Price (%)
0.0%
0.0%
0.0%
0.0%
0.7%
0.7%
0.6%
0.7%
0.7%
0.7%
0.7%
0.7%
0.7%
0.7%
0.7%
0.7%
0.7%
0.7%
0.7%
0.7%
0.7%
0.7%
0.7%
0.7%
0.7%
0.7%
0.7%
0.7%
0.7%
0.7%
Change in Total
Quantity Engineering
(%) Costs (103)
-0.004%
-0.006%
-0.006%
-0.011%
-0.021%
-0.027%
-0.028%
-0.031%
-0.032%
-0.032%
-0.032%
-0.032%
-0.032%
-0.032%
-0.032%
-0.032%
-0.032%
-0.032%
-0.032%
-0.032%
-0.032%
-0.032%
-0.032%
-0.032%
-0.032%
-0.032%
-0.032%
-0.032%
-0.032%
-0.032%
—
—
—
—
$103,262
$104,397
$88,557
$114,342
$115,456
$106,203
$107,317
$108,431
$105,349
$106,462
$88,274
$89,388
$90,502
$86,700
$87,814
$88,928
$90,042
$91,156
$92,270
$93,384
$94,498
$95,612
$96,726
$97,839
$98,953
$100,067
$1.477.053
in 2002 dollars.
values are calculated
using a
social discount
rate of 3 percent
over the 2004
Change in Producer
Surplus for
Equipment
Manufacturers (103)
-$110
-$225
-$229
-$461
-$30,609
-$30,925
-$31,005
-$40,265
-$40,352
-$30,010
-$30,022
-$30,046
-$25,874
-$25,894
-$6,612
-$6,637
-$6,661
-$1,769
-$1,793
-$1,817
-$1,841
-$1,865
-$1,889
-$1,913
-$1,936
-$1,960
-$1,984
-$2,008
-$2,032
-$2,056
-$250.397
to 2036 time period.
                                   10-120

-------
           Economic Impact Analysis
Table 10.B-20. Impacts on Construction Equipment Market and Manufacturers (>600 hp)
(Average Price per Equipment = $847,400)a
Construction Equipment
Absolute
Engineering Change in
Year Cost/Unit Price
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2031
2032
2033
2034
2035
2036
NPVb
a Figures are
b Net present
—
—
—
—
$4,519
$4,496
$3,797
$4,684
$9,206
$8,364
$7,517
$7,489
$7,218
$6,767
$6,151
$6,142
$6,133
$5,997
$5,458
$5,458
$5,458
$5,458
$5,458
$5,458
$5,458
$5,458
$5,458
$5,458
$5,458
$5,458

-$6
-$11
-$11
-$22
$2,923
$2,909
$2,230
$2,727
$6,205
$6,205
$5,387
$5,387
$5,387
$5,388
$5,388
$5,388
$5,388
$5,388
$5,388
$5,388
$5,388
$5,388
$5,388
$5,388
$5,388
$5,388
$5,388
$5,388
$5,388
$5,388

in 2002 dollars.
values are calculated using
Change in
Price (%)
0.0%
0.0%
0.0%
0.0%
0.4%
0.4%
0.3%
0.4%
0.8%
0.8%
0.7%
0.7%
0.7%
0.7%
0.7%
0.7%
0.7%
0.7%
0.7%
0.7%
0.7%
0.7%
0.7%
0.7%
0.7%
0.7%
0.7%
0.7%
0.7%
0.7%

a social discount
0600hp)

Change in Total
Quantity Engineering
(%) Costs (103)
-0.004%
-0.006%
-0.006%
-0.011%
-0.021%
-0.027%
-0.028%
-0.031%
-0.032%
-0.032%
-0.032%
-0.032%
-0.032%
-0.032%
-0.032%
-0.032%
-0.032%
-0.032%
-0.032%
-0.032%
-0.032%
-0.032%
-0.032%
-0.032%
-0.032%
-0.032%
-0.032%
-0.032%
-0.032%
-0.032%

rate of 3 percent
—
—
—
—
$23,207
$23,431
$20,179
$25,243
$50,150
$46,344
$42,363
$42,777
$41,837
$39,833
$36,149
$36,563
$36,978
$36,488
$33,066
$33,480
$33,895
$34,309
$34,724
$35,138
$35,552
$35,967
$36,381
$36,795
$37,210
$37,624
$527.785
over the 2004
Change in Producer
Surplus for
Equipment
Manufacturers (103)
-$31
-$63
-$65
-$130
-$8,646
-$8,735
-$8,757
-$11,056
-$17,335
-$13,058
-$13,061
-$13,068
-$11,720
-$9,307
-$5,214
-$5,221
-$5,227
-$4,330
-$500
-$506
-$513
-$519
-$526
-$532
-$539
-$545
-$551
-$558
-$564
-$571
-$96.188
to 2036 time period.
10-121

-------
Final Regulatory Impact Analysis
Table 10B-21. Impacts on Pumps and Compressor Equipment Market and
Manufacturers (<25 hp)
(Average Price per Equipment = $6,000)a
Pumps and Compressor Equipment (<25hp)
Absolute
Engineering Change in
Year Cost/Unit Price
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2031
2032
2033
2034
2035
2036
NPVb
a Figures are
b Net present
—
$177
$176
$168
$167
$166
$136
$136
$135
$135
$135
$123
$123
$123
$123
$123
$123
$123
$123
$123
$123
$123
$123
$123
$123
$123
$123
$123
$123
$123

—
$129
$129
$123
$123
$123
$123
$123
$123
$123
$123
$123
$123
$123
$123
$123
$123
$123
$123
$123
$123
$123
$123
$123
$123
$123
$123
$123
$123
$123

in 2002 dollars.
values are calculated using
Change in
Price (%)
0.0%
2.2%
2.2%
2.0%
2.0%
2.0%
2.0%
2.0%
2.0%
2.0%
2.0%
2.0%
2.0%
2.0%
2.0%
2.0%
2.0%
2.0%
2.0%
2.0%
2.0%
2.0%
2.0%
2.0%
2.0%
2.0%
2.0%
2.0%
2.0%
2.0%

a social discount
Change in Total
Quantity Engineering
(%) Costs (103)
0.000%
-0.001%
-0.001%
-0.001%
-0.002%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%

rate of 3 percent
—
$96
$96
$96
$96
$96
$96
$96
$96
$96
$96
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
$752
over the 2004
Change in Producer
Surplus for
Equipment
Manufacturers (103)
—
-$96
-$96
-$96
-$97
-$97
-$97
-$97
-$97
-$97
-$97
-$1
-$1
-$1
-$1
-$1
-$1
-$1
-$1
-$1
-$1
-$1
-$1
-$1
-$1
-$1
-$1
-$1
-$1
-$1
-$760
to 2036 time period.
                                   10-122

-------
           Economic Impact Analysis
Table 10B-22. Impacts on Pumps and Compressor Equipment Market and
Manufacturers (26-50 hp)
(Average Price per Equipment = $12,200)a
Pumps and Compressor Equipment (25^hp<50)
Engineering
Year Cost/Unit
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2031
2032
2033
2034
2035
2036
NPVb
a Figures are
b Net present
—
$204
$203
$194
$193
$192
$986
$984
$773
$771
$769
$693
$692
$692
$691
$691
$661
$661
$661
$661
$661
$661
$661
$661
$661
$661
$661
$661
$661
$661

Absolute
Change in
Price
—
$147
$147
$139
$139
$139
$870
$870
$661
$661
$661
$661
$661
$661
$661
$661
$661
$661
$661
$661
$661
$661
$661
$661
$661
$661
$661
$661
$661
$661

in 2002 dollars.
values are calculated using
Change in
Price (%)
0.0%
1.2%
1.2%
1.1%
1.1%
1.1%
7.1%
7.1%
5.4%
5.4%
5.4%
5.4%
5.4%
5.4%
5.4%
5.4%
5.4%
5.4%
5.4%
5.4%
5.4%
5.4%
5.4%
5.4%
5.4%
5.4%
5.4%
5.4%
5.4%
5.4%

a social discount
Change in
Quantity
(%)
0.000%
-0.001%
-0.001%
-0.001%
-0.002%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%

Total
Engineering
Costs (103)
—
$41
$41
$41
$41
$41
$356
$359
$337
$339
$340
$301
$303
$305
$307
$309
$112
$113
$115
$117
$119
$121
$123
$124
$126
$128
$130
$132
$134
$135
$3.189
rate of 3 percent over the 2004
Change in Producer
Surplus for
Equipment
Manufacturers (103)
—
-$41
-$41
-$41
-$42
-$42
-$241
-$241
-$241
-$241
-$241
-$200
-$200
-$200
-$200
-$200
-$1
-$1
-$1
-$1
-$1
-$1
-$1
-$1
-$1
-$1
-$1
-$1
-$1
-$1
-$1.673
to 2036 time period.
10-123

-------
Final Regulatory Impact Analysis
Table 10B-23. Impacts on Pumps and Compressor Equipment Market and
Manufacturers (51-75 hp)
(Average Price per Equipment = $10,600)a
Pumps and Compressor Equipment (50^hp<70)
Engineering
Year Cost/Unit
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2031
2032
2033
2034
2035
2036
NPVb
a Figures are
b Net present
—
$226
$225
$214
$213
$212
$978
$976
$769
$767
$765
$687
$686
$686
$685
$685
$653
$653
$653
$653
$653
$653
$653
$653
$653
$653
$653
$653
$653
$653

Absolute
Change in
Price
—
$167
$167
$158
$158
$158
$858
$858
$653
$653
$653
$653
$653
$653
$653
$653
$653
$653
$653
$653
$653
$653
$653
$653
$653
$653
$653
$653
$653
$653

in 2002 dollars.
values are calculated using
Change in
Price (%)
0.0%
1.6%
1.6%
1.5%
1.5%
1.5%
8.1%
8.1%
6.2%
6.2%
6.2%
6.2%
6.2%
6.2%
6.2%
6.2%
6.2%
6.2%
6.2%
6.2%
6.2%
6.2%
6.2%
6.2%
6.2%
6.2%
6.2%
6.2%
6.2%
6.2%

a social discount
Change in
Quantity
(%)
0.000%
-0.001%
-0.001%
-0.001%
-0.002%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%

Total
Engineering
Costs (103)
—
$39
$39
$39
$39
$39
$328
$329
$309
$311
$312
$275
$276
$278
$279
$281
$99
$101
$102
$104
$105
$107
$108
$110
$111
$112
$114
$115
$117
$118
$2.896
rate of 3 percent over the 2004
Change in Producer
Surplus for
Equipment
Manufacturers (103)
—
-$39
-$39
-$39
-$39
-$39
-$222
-$222
-$222
-$222
-$222
-$183
-$183
-$183
-$183
-$183
-$1
-$1
-$1
-$1
-$1
-$1
-$1
-$1
-$1
-$1
-$1
-$1
-$1
-$1
-$1.542
to 2036 time period.
                                   10-124

-------
           Economic Impact Analysis
Table 10B-24. Impacts on Pumps and Compressor Equipment Market and
Manufacturers (76-100 hp)
(Average Price per Equipment = $12,500)a
Pumps and Compressor Equipment (7(khp<100)
Engineering
Year Cost/Unit
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2031
2032
2033
2034
2035
2036
NPVb
a Figures are
b Net present
—
—
—
—
—
$1,303
$1,302
$1,325
$1,324
$1,322
$1,247
$1,246
$1,218
$1,218
$1,218
$1,218
$1,218
$1,218
$1,218
$1,218
$1,218
$1,218
$1,218
$1,218
$1,218
$1,218
$1,218
$1,218
$1,218
$1,218

Absolute
Change in
Price
—
—
—
—
—
$1,178
$1,178
$1,169
$1,169
$1,169
$1,169
$1,169
$1,169
$1,169
$1,169
$1,169
$1,169
$1,169
$1,169
$1,169
$1,169
$1,169
$1,169
$1,169
$1,169
$1,169
$1,169
$1,169
$1,169
$1,169

in 2002 dollars.
values are calculated using
Change in
Price (%)
0.0%
0.0%
0.0%
0.0%
0.0%
9.4%
9.4%
9.4%
9.4%
9.4%
9.4%
9.4%
9.4%
9.4%
9.4%
9.4%
9.4%
9.4%
9.4%
9.4%
9.4%
9.4%
9.4%
9.4%
9.4%
9.4%
9.4%
9.4%
9.4%
9.4%

a social discount
Change in
Quantity
(%)
0.000%
-0.001%
-0.001%
-0.001%
-0.002%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%

Total
Engineering
Costs (103)
—
—
—
—
—
$823
$827
$998
$1,003
$1,007
$1,011
$1,016
$1,020
$1,025
$1,029
$452
$456
$311
$315
$320
$324
$328
$333
$337
$342
$346
$351
$355
$360
$364
$9.294
rate of 3 percent over the 2004
Change in Producer
Surplus for
Equipment
Manufacturers (103)
—
—
—
—
—
-$583
-$583
-$733
-$733
-$733
-$733
-$733
-$733
-$733
-$733
-$151
-$151
-$1
-$1
-$1
-$1
-$1
-$1
-$1
-$1
-$1
-$1
-$1
-$1
-$1
-$5.030
to 2036 time period.
10-125

-------
Final Regulatory Impact Analysis
Table 10B-25. Impacts on Pumps and Compressor Equipment Market and
Manufacturers (101-175 hp)
(Average Price per Equipment = $23,800)a
Pumps and Compressor Equipment (100^hp<175)
Engineering
Year Cost/Unit
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2031
2032
2033
2034
2035
2036
NPVb
a Figures are
b Net present
—
—
—
—
—
$1,623
$1,619
$1,664
$1,659
$1,654
$1,577
$1,574
$1,542
$1,539
$1,537
$1,388
$1,387
$1,351
$1,351
$1,351
$1,351
$1,351
$1,351
$1,351
$1,351
$1,351
$1,351
$1,351
$1,351
$1,351

Absolute
Change in
Price
—
—
—
—
—
$1,421
$1,421
$1,399
$1,399
$1,399
$1,399
$1,399
$1,399
$1,399
$1,399
$1,399
$1,399
$1,399
$1,399
$1,399
$1,399
$1,399
$1,399
$1,399
$1,399
$1,399
$1,399
$1,399
$1,399
$1,399

in 2002 dollars.
values are calculated using
Change in
Price (%)
0.0%
0.0%
0.0%
0.0%
0.0%
6.0%
6.0%
5.9%
5.9%
5.9%
5.9%
5.9%
5.9%
5.9%
5.9%
5.9%
5.9%
5.9%
5.9%
5.9%
5.9%
5.9%
5.9%
5.9%
5.9%
5.9%
5.9%
5.9%
5.9%
5.9%

a social discount
Change in
Quantity
(%)
0.000%
-0.001%
-0.001%
-0.001%
-0.002%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%

Total
Engineering
Costs (103)
—
—
—
—
—
$266
$267
$325
$326
$327
$328
$329
$o o r\
330
$OO1
331
$332
$124
$125
$72
$73
$74
$75
$76
$77
$78
$79
$80
$81
$82
$83
$84
$2.796
rate of 3 percent over the 2004
Change in Producer
Surplus for
Equipment
Manufacturers (103)
—
—
—
—
—
-$210
-$210
-$263
-$263
-$263
-$263
-$263
-$263
-$263
-$263
-$54
-$54
—
—
—
—
—
—
—
—
—
—
—
—
—
-$1.807
to 2036 time period.
                                   10-126

-------
           Economic Impact Analysis
Table 10B-26. Impacts on Pumps and Compressor Equipment Market and
Manufacturers (176-600 hp)
(Average Price per Equipment = $53,000)a
Pumps and Compressor Equipment (175^hp<600)
Engineering
Year Cost/Unit
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2031
2032
2033
2034
2035
2036
NPVb
a Figures are
b Net present
—
—
—
—
$2,970
$2,958
$2,439
$3,107
$3,092
$2,777
$2,768
$2,759
$2,634
$2,627
$2,294
$2,292
$2,291
$2,209
$2,208
$2,208
$2,207
$2,206
$2,206
$2,205
$2,204
$2,204
$2,203
$2,203
$2,202
$2,202

Absolute
Change in
Price
—
—
—
-$1
$2,265
$2,264
$1,755
$2,216
$2,215
$2,214
$2,213
$2,212
$2,211
$2,210
$2,210
$2,209
$2,208
$2,207
$2,207
$2,206
$2,205
$2,205
$2,204
$2,203
$2,203
$2,202
$2,202
$2,201
$2,200
$2,200

in 2002 dollars.
values are calculated using
Change in
Price (%)
0.0%
0.0%
0.0%
0.0%
4.3%
4.3%
3.3%
4.2%
4.2%
4.2%
4.2%
4.2%
4.2%
4.2%
4.2%
4.2%
4.2%
4.2%
4.2%
4.2%
4.2%
4.2%
4.2%
4.2%
4.2%
4.2%
4.2%
4.2%
4.2%
4.2%

a social discount
Change in
Quantity
(%)
0.000%
-0.001%
-0.001%
-0.001%
-0.002%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%

Total
Engineering
Costs (103)
—
—
—
—
$821
$823
$797
$1,010
$1,012
$1,015
$1,017
$1,019
$1,021
$1,023
$341
$343
$345
$173
$175
$177
$180
$182
$184
$186
$188
$190
$192
$195
$197
$199
$8.508
rate of 3 percent over the 2004
Change in Producer
Surplus for
Equipment
Manufacturers (103)
—
—
—
—
-$685
-$685
-$686
-$860
-$860
-$860
-$860
-$860
-$860
-$860
-$176
-$176
-$176
-$1
-$1
-$1
-$1
-$1
-$1
-$1
-$1
-$1
-$1
-$1
-$1
-$1
-$6.048
to 2036 time period.
10-127

-------
Final Regulatory Impact Analysis
Table 10B-27. Impacts on Pumps and Compressor Equipment Market and
Manufacturers (>600 hp)
(Average Price per Equipment = $88,000)a
Pumps and Compressor Equipment (^600hp)
Engineering
Year Cost/Unit
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2031
2032
2033
2034
2035
2036
NPVb
a Figures are
b Net present
—
—
—
—
$4,519
$4,496
$3,797
$4,684
$9,206
$8,364
$7,517
$7,489
$7,218
$6,767
$6,151
$6,142
$6,133
$5,997
$5,458
$5,458
$5,458
$5,458
$5,458
$5,458
$5,458
$5,458
$5,458
$5,458
$5,458
$5,458

Absolute
Change in
Price
—
-$1
-$1
-$2
$2,965
$2,964
$2,287
$2,790
$6,271
$6,271
$5,453
$5,453
$5,453
$5,453
$5,453
$5,453
$5,453
$5,453
$5,453
$5,453
$5,453
$5,453
$5,453
$5,453
$5,453
$5,453
$5,453
$5,453
$5,453
$5,453

in 2002 dollars.
values are calculated using
Change in
Price (%)
0.0%
0.0%
0.0%
0.0%
3.4%
3.4%
2.6%
3.2%
7.1%
7.1%
6.2%
6.2%
6.2%
6.2%
6.2%
6.2%
6.2%
6.2%
6.2%
6.2%
6.2%
6.2%
6.2%
6.2%
6.2%
6.2%
6.2%
6.2%
6.2%
6.2%

a social discount
Change in
Quantity
(%)
0.000%
-0.001%
-0.001%
-0.001%
-0.002%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%

Total
Engineering
Costs (103)
—
—
—
—
$15
$15
$14
$18
$32
$32
$31
$31
$32
$32
$18
$18
$18
$15
$3
$3
$o
3
$o
3
$3
$3
$3
$3
$3
$o
3
$o
3
$3
$231
rate of 3 percent over the 2004
Change in Producer
Surplus for
Equipment
Manufacturers (103)
—
—
—
—
-$14
-$14
-$14
-$16
-$29
-$29
-$29
-$29
-$29
-$29
-$16
-$16
-$16
-$13
—
—
—
—
—
—
—
—
—
—
—
—
-$196
to 2036 time period.
                                   10-128

-------
           Economic Impact Analysis
Table 10.B-28. Impacts on Generator Sets and Welding Equipment
Manufacturers (<25 hp)
(Average Price per Equipment = $6,800)a
Generator Sets and Welding Equipment (<25hp)
Absolute
Engineering Change in
Year Cost/Unit Price
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2031
2032
2033
2034
2035
2036
NPVb
a Figures are
b Net present
—
$177
$176
$168
$167
$166
$136
$136
$135
$135
$135
$123
$123
$123
$123
$123
$123
$123
$123
$123
$123
$123
$123
$123
$123
$123
$123
$123
$123
$123

—
$129
$129
$123
$123
$123
$123
$123
$123
$123
$123
$123
$123
$123
$123
$123
$123
$123
$123
$123
$123
$123
$123
$123
$123
$123
$123
$123
$123
$123

in 2002 dollars.
values are calculated using
Change in
Price (%)
0.0%
1.9%
1.9%
1.8%
1.8%
1.8%
1.8%
1.8%
1.8%
1.8%
1.8%
1.8%
1.8%
1.8%
1.8%
1.8%
1.8%
1.8%
1.8%
1.8%
1.8%
1.8%
1.8%
1.8%
1.8%
1.8%
1.8%
1.8%
1.8%
1.8%

a social discount
Change in Total
Quantity Engineering
(%) Costs (103)
0.000%
-0.001%
-0.001%
-0.001%
-0.002%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%

rate of 3 percent
—
$3,795
$3,854
$3,794
$3,850
$3,906
$3,410
$3,466
$3,522
$3,578
$3,634
$2,627
$2,683
$2,739
$2,795
$2,851
$2,907
$2,963
$3,019
$3,075
$3,131
$3,187
$3,243
$3,299
$3,355
$3,411
$3,467
$3,523
$3,579
$3,634
$58.866
over the 2004
Market and
Change in Producer
Surplus for
Equipment
Manufacturers (103)
—
-$1,615
-$1,615
-$1,616
-$1,618
-$1,619
-$1,068
-$1,069
-$1,069
-$1,069
-$1,069
-$6
-$7
-$7
-$7
-$7
-$7
-$7
-$7
-$7
-$8
-$8
-$8
-$8
-$8
-$8
-$8
-$9
-$9
-$9
-$10.712
to 2036 time period.
10-129

-------
Final  Regulatory Impact Analysis
Table 10.B-29.
Impacts on Generator Sets and Welding Equipment Market and
Manufacturers (26-50 hp)
(Average Price per Equipment = $8,700)a
Generator Sets and Welding Equipment (25^hp<50) (
Year
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2031
2032
2033
2034
2035
2036
NPVb
Engineering
Cost/Unit
—
$204
$203
$194
$193
$192
$986
$984
$773
$771
$769
$693
$692
$692
$691
$691
$661
$661
$661
$661
$661
$661
$661
$661
$661
$661
$661
$661
$661
$661

Absolute
Change in
Price
—
$147
$147
$139
$139
$139
$870
$870
$661
$661
$661
$661
$661
$661
$661
$661
$661
$661
$661
$661
$661
$661
$661
$661
$661
$661
$661
$661
$661
$661

Change in
Price (%)
0.0%
1.7%
1.7%
1.6%
1.6%
1.6%
10.0%
10.0%
7.6%
7.6%
7.6%
7.6%
7.6%
7.6%
7.6%
7.6%
7.6%
7.6%
7.6%
7.6%
7.6%
7.6%
7.6%
7.6%
7.6%
7.6%
7.6%
7.6%
7.6%
7.6%

Change in
Quantity
(%)
0.000%
-0.001%
-0.001%
-0.001%
-0.002%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%

Total
Engineering
Costs (io3) n
—
$1,896
$1,922
$1,883
$1,907
$1,932
$10,977
$11,143
$9,227
$9,354
$9,481
$8,631
$8,758
$8,885
$9,012
$9,139
$7,746
$7,873
$8,000
$8,127
$8,254
$8,381
$8,508
$8,635
$8,762
$8,889
$9,017
$9,144
$9,271
$9,398
$128.538
Change in Producer
Surplus for
Equipment
Manufacturers (IO3)
—
-$713
-$713
-$714
-$715
-$716
-$2,502
-$2,502
-$2,502
-$2,502
-$2,502
-$1,525
-$1,525
-$1,525
-$1,525
-$1,525
-$5
-$5
-$5
-$5
-$5
-$5
-$5
-$5
-$5
-$5
-$6
-$6
-$6
-$6
-$16.831
a    Figures are in 2002 dollars.
b    Net present values are calculated using a social discount rate of 3 percent over the 2004 to 2036 time period.
                                            10-130

-------
           Economic Impact Analysis
Table 10.B-30. Impacts on Generator Sets and Welding Equipment
Manufacturers (51-75 hp)
(Average Price per Equipment = $8,300)a
Generator Sets and Welding Equipment (5(khp<70)
Absolute
Engineering Change in
Year Cost/Unit Price
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2031
2032
2033
2034
2035
2036
NPVb
a Figures are
b Net present
—
$226
$225
$214
$213
$212
$978
$976
$769
$767
$765
$687
$686
$686
$685
$685
$653
$653
$653
$653
$653
$653
$653
$653
$653
$653
$653
$653
$653
$653

—
$167
$167
$158
$158
$158
$858
$858
$653
$653
$653
$653
$653
$653
$653
$653
$653
$653
$653
$653
$653
$653
$653
$653
$653
$653
$653
$653
$653
$653

in 2002 dollars.
values are calculated using
Change in
Price (%)
0.0%
2.0%
2.0%
1.9%
1.9%
1.9%
10.3%
10.3%
7.9%
7.9%
7.9%
7.9%
7.9%
7.9%
7.9%
7.9%
7.9%
7.9%
7.9%
7.9%
7.9%
7.9%
7.9%
7.9%
7.9%
7.9%
7.9%
7.9%
7.9%
7.9%

a social discount
Change in Total
Quantity Engineering
(%) Costs (103)
0.000%
-0.001%
-0.001%
-0.001%
-0.002%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%

rate of 3 percent
—
$2,029
$2,056
$2,000
$2,025
$2,051
$9,825
$9,966
$8,049
$8,157
$8,265
$7,518
$7,626
$7,734
$7,842
$7,950
$7,443
$7,551
$7,659
$7,767
$7,875
$7,983
$8,091
$8,199
$8,307
$8,415
$8,523
$8,631
$8,739
$8,847
$118.426
over the 2004
Market and
Change in Producer
Surplus for
Equipment
Manufacturers (103)
—
-$570
-$570
-$570
-$571
-$571
-$1,472
-$1,472
-$1,472
-$1,472
-$1,472
-$617
-$617
-$617
-$617
-$617
-$2
-$2
-$2
-$2
-$2
-$2
-$2
-$2
-$2
-$2
-$2
-$2
-$2
-$2
-$9.648
to 2036 time period.
10-131

-------
Final Regulatory Impact Analysis
Table 10.B-3 1 . Impacts on Generator Sets and Welding Equipment
Manufacturers (76-100 hp)
(Average Price per Equipment = $18,000)a
Generator Sets and Welding Equipment (7(khp<100)
Absolute
Engineering Change in
Year Cost/Unit Price
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2031
2032
2033
2034
2035
2036
NPVb
a Figures are
b Net present
—
—
$1,303
$1,302
$1,325
$1,324
$1,322
$1,247
$1,246
$1,218
$1,218
$1,218
$1,218
$1,218
$1,218
$1,218
$1,218
$1,218
$1,218
$1,218
$1,218
$1,218
$1,218
$1,218
$1,218
$1,218
$1,218

—
—
$1,178
$1,178
$1,169
$1,169
$1,169
$1,169
$1,169
$1,169
$1,169
$1,169
$1,169
$1,169
$1,169
$1,169
$1,169
$1,169
$1,169
$1,169
$1,169
$1,169
$1,169
$1,169
$1,169
$1,169
$1,169

in 2002 dollars.
values are calculated using
Change in
Price (%)
0.0%
0.0%
0.0%
0.0%
0.0%
6.5%
6.5%
6.5%
6.5%
6.5%
6.5%
6.5%
6.5%
6.5%
6.5%
6.5%
6.5%
6.5%
6.5%
6.5%
6.5%
6.5%
6.5%
6.5%
6.5%
6.5%
6.5%
6.5%
6.5%
6.5%

a social discount
Change in Total
Quantity Engineering
(%) Costs (103)
0.000%
-0.001%
-0.001%
-0.001%
-0.002%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%

rate of 3 percent
—
—
$2,241
$2,265
$2,527
$2,552
$2,576
$2,524
$2,548
$2,543
$2,567
$2,592
$1,851
$1,876
$1,703
$1,727
$1,752
$1,776
$1,801
$1,825
$1,849
$1,874
$1,898
$1,923
$1,947
$1,971
$1,996
$30.552
over the 2004
Market and
Change in Producer
Surplus for
Equipment
Manufacturers (103)
—
-$1
-$842
-$842
-$1,069
-$1,069
-$1,069
-$993
-$993
-$963
-$963
-$963
-$198
-$199
-$2
-$2
-$2
-$2
-$2
-$2
-$2
-$2
-$2
-$2
-$2
-$2
-$2
-$7.004
to 2036 time period.
                                   10-132

-------
           Economic Impact Analysis
Table 10.B-32. Impacts on Generator Sets and Welding Equipment
Manufacturers (101-175 hp)
(Average Price per Equipment = $21,400)a
Generator Sets and Welding Equipment (10(khp<175)
Absolute
Engineering Change in
Year Cost/Unit Price
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2031
2032
2033
2034
2035
2036
NPVb
a Figures are
b Net present
—
—
$1,623
$1,619
$1,664
$1,659
$1,654
$1,577
$1,574
$1,542
$1,539
$1,537
$1,388
$1,387
$1,351
$1,351
$1,351
$1,351
$1,351
$1,351
$1,351
$1,351
$1,351
$1,351
$1,351
$1,351
$1,351

—
—
$1,421
$1,421
$1,399
$1,399
$1,399
$1,399
$1,399
$1,399
$1,399
$1,399
$1,399
$1,399
$1,399
$1,399
$1,399
$1,399
$1,399
$1,399
$1,399
$1,399
$1,399
$1,399
$1,399
$1,399
$1,399

in 2002 dollars.
values are calculated using
Change in
Price (%)
0.0%
0.0%
0.0%
0.0%
0.0%
6.6%
6.6%
6.5%
6.5%
6.5%
6.5%
6.5%
6.5%
6.5%
6.5%
6.5%
6.5%
6.5%
6.5%
6.5%
6.5%
6.5%
6.5%
6.5%
6.5%
6.5%
6.5%
6.5%
6.5%
6.5%

a social discount
Change in Total
Quantity Engineering
(%) Costs (103)
0.000%
-0.001%
-0.001%
-0.001%
-0.002%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%

rate of 3 percent
—
—
$11,755
$11,915
$12,544
$12,702
$12,860
$12,493
$12,651
$12,595
$12,753
$12,911
$11,515
$11,673
$11,434
$11,591
$11,749
$11,907
$12,065
$12,223
$12,381
$12,539
$12,697
$12,855
$13,013
$13,171
$13,329
$174.772
over the 2004
Market and
Change in Producer
Surplus for
Equipment
Manufacturers (103)
-$1
-$1
-$2,081
-$2,081
-$2,692
-$2,692
-$2,692
-$2,167
-$2,168
-$1,953
-$1,953
-$1,953
-$399
-$399
-$2
-$2
-$3
-$3
-$3
-$3
-$3
-$3
-$3
-$3
-$3
-$3
-$3
-$16.116
to 2036 time period.
10-133

-------
Final Regulatory Impact Analysis
Table 10.B-33. Impacts on Generator Sets and Welding Equipment
Manufacturers (176-600 hp)
(Average Price per Equipment = $21,400)a
Generator Sets and Welding Equipment (175^hp<600)
Absolute
Engineering Change in
Year Cost/Unit Price
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2031
2032
2033
2034
2035
2036
NPVb
a Figures are
b Net present
—
—
$2,970
$2,958
$2,439
$3,107
$3,092
$2,777
$2,768
$2,759
$2,634
$2,627
$2,294
$2,292
$2,291
$2,209
$2,208
$2,208
$2,207
$2,206
$2,206
$2,205
$2,204
$2,204
$2,203
$2,203
$2,202
$2,202

—
—
$2,266
$2,265
$1,756
$2,216
$2,215
$2,214
$2,214
$2,213
$2,212
$2,211
$2,210
$2,210
$2,209
$2,208
$2,207
$2,207
$2,206
$2,205
$2,205
$2,204
$2,203
$2,203
$2,202
$2,202
$2,201
$2,201

in 2002 dollars.
values are calculated using
Change in
Price (%)
0.0%
0.0%
0.0%
0.0%
6.3%
6.3%
4.9%
6.2%
6.2%
6.2%
6.2%
6.2%
6.2%
6.2%
6.2%
6.2%
6.2%
6.2%
6.2%
6.2%
6.2%
6.2%
6.2%
6.2%
6.2%
6.2%
6.2%
6.2%
6.2%
6.2%

a social discount
Change in Total
Quantity Engineering
(%) Costs (103)
0.000%
-0.001%
-0.001%
-0.001%
-0.002%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%

rate of 3 percent
—
—
$3,728
$3,767
$3,221
$4,154
$4,192
$3,877
$3,916
$3,954
$3,850
$3,888
$3,096
$3,134
$3,173
$3,000
$3,038
$3,077
$3,115
$3,154
$3,192
$3,231
$3,269
$3,308
$3,346
$3,385
$3,423
$3,462
$52.508
over the 2004
Market and
Change in Producer
Surplus for
Equipment
Manufacturers (103)
—
—
-$1,185
-$1,185
-$1,186
-$1,540
-$1,540
-$1,187
-$1,187
-$1,187
-$1,044
-$1,044
-$213
-$213
-$213
-$1
-$1
-$1
-$1
-$1
-$1
-$1
-$1
-$1
-$1
-$1
-$1
-$1
-$9.195
to 2036 time period.
                                   10-134

-------
           Economic Impact Analysis
Table 10B-34.
Impacts on Refrigeration and Air-Conditioning Equipment Market and
Manufacturers (<25 hp)
(Average Price per Equipment = $12,500)a
Refrigeration and Air-Conditioning
Absolute
Engineering Change in
Year Cost/Unit Price
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2031
2032
2033
2034
2035
2036
NPVb
a Figures are
b Net present
—
$177
$176
$168
$167
$166
$136
$136
$135
$135
$135
$123
$123
$123
$123
$123
$123
$123
$123
$123
$123
$123
$123
$123
$123
$123
$123
$123
$123
$123

—
$129
$129
$123
$123
$123
$123
$123
$123
$123
$123
$123
$123
$123
$123
$123
$123
$123
$123
$123
$123
$123
$123
$123
$123
$123
$123
$123
$123
$123

in 2002 dollars.
values are calculated using
Change in
Price (%)
0.0%
1.0%
1.0%
1.0%
1.0%
1.0%
1.0%
1.0%
1.0%
1.0%
1.0%
1.0%
1.0%
1.0%
1.0%
1.0%
1.0%
1.0%
1.0%
1.0%
1.0%
1.0%
1.0%
1.0%
1.0%
1.0%
1.0%
1.0%
1.0%
1.0%

a social discount
Equipment (<25hp)
Change in Total
Quantity Engineering
(%) Costs (103)
0.000%
-0.001%
-0.001%
-0.001%
-0.002%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%

rate of 3 percent
—
$168
$168
$168
$168
$168
$168
$168
$168
$168
$168
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
$1.310
over the 2004
Change in Producer
Surplus for
Equipment
Manufacturers (103)
—
-$168
-$168
-$168
-$169
-$169
-$169
-$170
-$170
-$170
-$170
-$2
-$2
-$2
-$2
-$2
-$2
-$2
-$2
-$2
-$2
-$2
-$2
-$2
-$2
-$2
-$3
-$3
$o
3
-$3
-$1.340
to 2036 time period.
10-135

-------
Final Regulatory Impact Analysis
Table 10B-35. Impacts on Refrigeration and Air-Conditioning Equipment Market and
Manufacturers (26-50 hp)
(Average Price per Equipment = $27,000)a
Refrigeration and Air-Conditioning Equipment (25^hp<50)
Engineering
Year Cost/Unit
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2031
2032
2033
2034
2035
2036
NPVb
a Figures are
b Net present
—
$204
$203
$194
$193
$192
$986
$984
$773
$771
$769
$693
$692
$692
$691
$691
$661
$661
$661
$661
$661
$661
$661
$661
$661
$661
$661
$661
$661
$661

Absolute
Change in
Price
—
$147
$147
$139
$139
$139
$869
$869
$661
$661
$661
$661
$661
$661
$661
$661
$661
$661
$661
$661
$661
$661
$661
$661
$661
$661
$661
$661
$661
$661

in 2002 dollars.
values are calculated using
Change in
Price (%)
0.0%
0.5%
0.5%
0.5%
0.5%
0.5%
3.2%
3.2%
2.4%
2.4%
2.4%
2.4%
2.4%
2.4%
2.4%
2.4%
2.4%
2.4%
2.4%
2.4%
2.4%
2.4%
2.4%
2.4%
2.4%
2.4%
2.4%
2.4%
2.4%
2.4%

a social discount
Change in
Quantity
(%)
0.000%
-0.001%
-0.001%
-0.001%
-0.002%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%

Total
Engineering
Costs (103)
—
$100
$100
$100
$100
$100
$871
$876
$823
$827
$832
$736
$740
$745
$749
$754
$273
$277
$281
$286
$290
$295
$299
$304
$308
$313
$317
$322
$326
$331
$7.790
rate of 3 percent over the 2004
Change in Producer
Surplus for
Equipment
Manufacturers (103)
—
-$101
-$101
-$102
-$103
-$104
-$590
-$590
-$590
-$591
-$591
-$490
-$490
-$490
-$490
-$490
-$5
-$5
-$5
-$5
-$5
-$5
-$5
-$5
-$6
-$6
-$6
-$6
-$6
-$6
-$4.126
to 2036 time period.
                                   10-136

-------
           Economic Impact Analysis
Table 10B-36.
Impacts on Refrigeration and Air-Conditioning Equipment Market and
Manufacturers (51-75 hp)
(Average Price per Equipment = $42,100)a
Refrigeration and Air-Conditioning Equipment (5(khp<70)
Absolute
Engineering Change in
Year Cost/Unit Price
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2031
2032
2033
2034
2035
2036
NPVb
a Figures are
b Net present
—
$226
$225
$214
$213
$212
$978
$976
$769
$767
$765
$687
$686
$686
$685
$685
$653
$653
$653
$653
$653
$653
$653
$653
$653
$653
$653
$653
$653
$653

—
$167
$167
$158
$157
$157
$858
$858
$653
$653
$653
$653
$653
$653
$653
$653
$653
$653
$653
$653
$653
$653
$653
$653
$653
$653
$653
$653
$653
$653

in 2002 dollars.
values are calculated using
Change in
Price (%)
0.0%
0.4%
0.4%
0.4%
0.4%
0.4%
2.0%
2.0%
1.6%
1.6%
1.6%
1.6%
1.6%
1.6%
1.6%
1.6%
1.6%
1.6%
1.6%
1.6%
1.6%
1.6%
1.6%
1.6%
1.6%
1.6%
1.6%
1.6%
1.6%
1.6%

a social discount
Change in Total
Quantity Engineering
(%) Costs (103)
0.000%
-0.001%
-0.001%
-0.001%
-0.002%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%

rate of 3 percent
—
$179
$179
$179
$179
$179
$1,512
$1,521
$1,428
$1,434
$1,441
$1,269
$1,276
$1,282
$1,289
$1,295
$459
$466
$472
$479
$486
$492
$499
$506
$512
$519
$526
$532
$539
$546
$13.368
over the 2004
Change in Producer
Surplus for
Equipment
Manufacturers (103)
—
-$180
-$180
-$182
-$184
-$187
-$1,032
-$1,033
-$1,033
-$1,033
-$1,033
-$855
-$855
-$855
-$855
-$855
-$12
-$13
-$13
-$13
-$13
-$13
-$13
-$14
-$14
-$14
-$14
-$14
-$15
-$15
-$7.255
to 2036 time period.
10-137

-------
Final Regulatory Impact Analysis
Table 10
.B-37.
Impacts on General Industrial Equipment Market and Manufacturers (<25 hp)
(Average Price per Equipment = $17,300)a
General Industrial Equipment (<25hp)
Absolute
Engineering Change in
Year Cost/Unit Price
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2031
2032
2033
2034
2035
2036
NPVb
a Figures are
b Net present
—
$177
$176
$168
$167
$166
$136
$136
$135
$135
$135
$123
$123
$123
$123
$123
$123
$123
$123
$123
$123
$123
$123
$123
$123
$123
$123
$123
$123
$123

—
$129
$129
$123
$123
$123
$123
$123
$123
$123
$123
$123
$123
$123
$123
$123
$123
$123
$123
$123
$123
$123
$123
$123
$123
$123
$123
$123
$123
$123

in 2002 dollars.
values are calculated using
Change in
Price (%)
0.0%
0.7%
0.7%
0.7%
0.7%
0.7%
0.7%
0.7%
0.7%
0.7%
0.7%
0.7%
0.7%
0.7%
0.7%
0.7%
0.7%
0.7%
0.7%
0.7%
0.7%
0.7%
0.7%
0.7%
0.7%
0.7%
0.7%
0.7%
0.7%
0.7%

a social discount
Change in Total
Quantity Engineering
(%) Costs (103)
0.000%
-0.001%
-0.001%
-0.001%
-0.002%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%

rate of 3 percent
—
$61
$61
$61
$61
$61
$61
$61
$61
$61
$61
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
$479
over the 2004
Change in Producer
Surplus for
Equipment
Manufacturers (103)
—
-$61
-$61
-$61
-$62
-$62
-$62
-$62
-$62
-$62
-$62
-$1
-$1
-$1
-$1
-$1
-$1
-$1
-$1
-$1
-$1
-$1
-$1
-$1
-$1
-$1
-$1
-$1
-$1
-$1
-$487
to 2036 time period.
                                   10-138

-------
           Economic Impact Analysis
Table 10.B-38. Impacts on General Industrial Equipment Market and Manufacturers (26-50 hp)
(Average Price per Equipment = $42,3 00)a
General Industrial Equipment (25^hp<50)
Engineering
Year Cost/Unit
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2031
2032
2033
2034
2035
2036
NPVb
a Figures are
b Net present
—
$204
$203
$194
$193
$192
$986
$984
$773
$771
$769
$693
$692
$692
$691
$691
$661
$661
$661
$661
$661
$661
$661
$661
$661
$661
$661
$661
$661
$661

Absolute
Change in
Price
—
$147
$147
$139
$139
$139
$870
$869
$661
$661
$661
$661
$661
$661
$661
$661
$661
$661
$661
$661
$661
$661
$661
$661
$661
$661
$661
$661
$661
$661

in 2002 dollars.
values are calculated using
Change in
Price (%)
0.0%
0.3%
0.3%
0.3%
0.3%
0.3%
2.1%
2.1%
1.6%
1.6%
1.6%
1.6%
1.6%
1.6%
1.6%
1.6%
1.6%
1.6%
1.6%
1.6%
1.6%
1.6%
1.6%
1.6%
1.6%
1.6%
1.6%
1.6%
1.6%
1.6%

a social discount
Change in
Quantity
(%)
0.000%
-0.001%
-0.001%
-0.001%
-0.002%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%

Total
Engineering
Costs (103)
—
$83
$83
$83
$83
$83
$664
$670
$616
$620
$624
$555
$559
$563
$567
$571
$251
$256
$260
$264
$268
$272
$276
$280
$284
$289
$293
$297
$301
$305
$6.249
rate of 3 percent over the 2004
Change in Producer
Surplus for
Equipment
Manufacturers (103)
—
-$71
-$71
-$72
-$72
-$73
-$400
-$400
-$400
-$400
-$400
-$326
-$327
-$327
-$327
-$327
-$3
-$3
-$3
-$3
-$3
$o
3
$o
3
-$3
-$3
-$3
-$3
-$3
$o
3
$o
3
-$2.785
to 2036 time period.
10-139

-------
Final Regulatory Impact Analysis
Table 10.B-39. Impacts on General Industrial Equipment Market and Manufacturers (51-75 hp)
(Average Price per Equipment = $56,400)a
General Industrial Equipment (50^hp<70)
Engineering
Year Cost/Unit
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2031
2032
2033
2034
2035
2036
NPVb
a Figures are
b Net present
—
$226
$225
$214
$213
$212
$978
$976
$769
$767
$765
$687
$686
$686
$685
$685
$653
$653
$653
$653
$653
$653
$653
$653
$653
$653
$653
$653
$653
$653

Absolute
Change in
Price
—
$167
$167
$158
$158
$157
$858
$858
$653
$653
$653
$653
$653
$653
$653
$653
$653
$653
$653
$653
$653
$653
$653
$653
$653
$653
$653
$653
$653
$653

in 2002 dollars.
values are calculated using
Change in
Price (%)
0.0%
0.3%
0.3%
0.3%
0.3%
0.3%
1.5%
1.5%
1.2%
1.2%
1.2%
1.2%
1.2%
1.2%
1.2%
1.2%
1.2%
1.2%
1.2%
1.2%
1.2%
1.2%
1.2%
1.2%
1.2%
1.2%
1.2%
1.2%
1.2%
1.2%

a social discount
Change in
Quantity
(%)
0.000%
-0.001%
-0.001%
-0.001%
-0.002%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%

Total
Engineering
Costs (103)
—
$413
$418
$408
$412
$417
$2,167
$2,195
$1,824
$1,845
$1,867
$1,687
$1,708
$1,730
$1,751
$1,772
$1,465
$1,486
$1,507
$1,529
$1,550
$1,571
$1,592
$1,614
$1,635
$1,656
$1,677
$1,699
$1,720
$1,741
$24.870
rate of 3 percent over the 2004
Change in Producer
Surplus for
Equipment
Manufacturers (103)
—
-$150
-$150
-$150
-$151
-$151
-$532
-$533
-$533
-$533
-$533
-$332
-$332
-$332
-$332
-$332
-$3
-$3
-$4
-$4
-$4
-$4
-$4
-$4
-$4
-$4
-$4
-$4
-$4
-$4
-$3.615
to 2036 time period.
                                   10-140

-------
           Economic Impact Analysis
Table 10.B-40. Impacts on General Industrial Equipment Market and Manufacturers (76-100
hp)
(Average Price per Equipment = $74,3 00)a
General Industrial Equipment (75^hp<100)
Engineering
Year Cost/Unit
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2031
2032
2033
2034
2035
2036
NPVb
a Figures are
b Net present
—
—
—
—
—
$1,303
$1,302
$1,325
$1,324
$1,322
$1,247
$1,246
$1,218
$1,218
$1,218
$1,218
$1,218
$1,218
$1,218
$1,218
$1,218
$1,218
$1,218
$1,218
$1,218
$1,218
$1,218
$1,218
$1,218
$1,218

Absolute
Change in
Price
—
—
—
—
—
$1,178
$1,178
$1,169
$1,169
$1,169
$1,169
$1,169
$1,169
$1,169
$1,169
$1,169
$1,169
$1,169
$1,169
$1,169
$1,169
$1,169
$1,169
$1,169
$1,169
$1,169
$1,169
$1,169
$1,169
$1,169

in 2002 dollars.
values are calculated using
Change in
Price (%)
0.0%
0.0%
0.0%
0.0%
0.0%
1.6%
1.6%
1.6%
1.6%
1.6%
1.6%
1.6%
1.6%
1.6%
1.6%
1.6%
1.6%
1.6%
1.6%
1.6%
1.6%
1.6%
1.6%
1.6%
1.6%
1.6%
1.6%
1.6%
1.6%
1.6%

a social discount
Change in
Quantity
(%)
0.000%
-0.001%
-0.001%
-0.001%
-0.002%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%

Total
Engineering
Costs (103)
—
—
—
—
—
$8,518
$8,625
$9,382
$9,489
$9,596
$9,325
$9,432
$9,390
$9,497
$9,604
$7,760
$7,867
$7,471
$7,578
$7,685
$7,792
$7,899
$8,006
$8,113
$8,220
$8,327
$8,434
$8,541
$8,648
$8,756
$122.225
rate of 3 percent over the 2004
Change in Producer
Surplus for
Equipment
Manufacturers (103)
—
-$1
-$1
-$2
-$4
-$2,336
-$2,337
-$2,990
-$2,990
-$2,990
-$2,611
-$2,611
-$2,462
-$2,462
-$2,462
-$511
-$511
-$9
-$9
-$9
-$10
-$10
-$10
-$10
-$10
-$10
-$10
-$10
-$11
-$11
-$18.884
to 2036 time period.
10-141

-------
Final Regulatory Impact Analysis

Table
10.B-41. Impacts on General Industrial Equipment Market and
Manufacturers (101-175 hp)
(Average Price per Equipment = $1 16,900)a
General Industrial Equipment
Absolute
Engineering Change in
Year Cost/Unit Price
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2031
2032
2033
2034
2035
2036
NPVb
a Figures are
b Net present
—
—
$1,623
$1,619
$1,664
$1,659
$1,654
$1,577
$1,574
$1,542
$1,539
$1,537
$1,388
$1,387
$1,351
$1,351
$1,351
$1,351
$1,351
$1,351
$1,351
$1,351
$1,351
$1,351
$1,351
$1,351
$1,351

—
-$1
$1,420
$1,420
$1,399
$1,399
$1,399
$1,399
$1,399
$1,399
$1,399
$1,399
$1,399
$1,399
$1,399
$1,399
$1,399
$1,399
$1,399
$1,399
$1,399
$1,399
$1,399
$1,399
$1,399
$1,399
$1,399

in 2002 dollars.
values are calculated using
Change in
Price (%)
0.0%
0.0%
0.0%
0.0%
0.0%
1.2%
1.2%
1.2%
1.2%
1.2%
1.2%
1.2%
1.2%
1.2%
1.2%
1.2%
1.2%
1.2%
1.2%
1.2%
1.2%
1.2%
1.2%
1.2%
1.2%
1.2%
1.2%
1.2%
1.2%
1.2%

a social discount
(10(khp<175)

Change in Total
Quantity Engineering
(%) Costs (103)
0.000%
-0.001%
-0.001%
-0.001%
-0.002%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%

rate of 3 percent
—
—
$11,708
$11,833
$13,023
$13,147
$13,272
$13,025
$13,150
$13,123
$13,247
$13,371
$9,722
$9,846
$9,007
$9,131
$9,256
$9,380
$9,504
$9,629
$9,753
$9,878
$10,002
$10,127
$10,251
$10,375
$10,500
$159.307
over the 2004
Change in Producer
Surplus for
Equipment
Manufacturers (103)
-$2
-$2
-$5
-$8
-$4,156
-$4,160
-$5,276
-$5,276
-$5,276
-$4,905
-$4,905
-$4,754
-$4,754
-$4,755
-$981
-$981
-$18
-$18
-$18
-$18
-$19
-$19
-$19
-$19
-$20
-$20
-$20
-$20
-$21
-$34.647
to 2036 time period.
                                   10-142

-------
           Economic Impact Analysis

Table
10.B-42. Impacts on General Industrial Equipment Market and
Manufacturers (176-600 hp)
(Average Price per Equipment = $154,200)a
General Industrial Equipment
Absolute
Engineering Change in
Year Cost/Unit Price
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2031
2032
2033
2034
2035
2036
NPVb
a Figures are
b Net present
—
—
$2,970
$2,958
$2,439
$3,107
$3,092
$2,777
$2,768
$2,759
$2,634
$2,627
$2,294
$2,292
$2,291
$2,209
$2,208
$2,208
$2,207
$2,206
$2,206
$2,205
$2,204
$2,204
$2,203
$2,203
$2,202
$2,202

—
-$1
$2,265
$2,264
$1,755
$2,215
$2,214
$2,213
$2,213
$2,212
$2,211
$2,210
$2,209
$2,209
$2,208
$2,207
$2,206
$2,206
$2,205
$2,204
$2,204
$2,203
$2,203
$2,202
$2,201
$2,201
$2,200
$2,200

in 2002 dollars.
values are calculated using
Change in
Price (%)
0.0%
0.0%
0.0%
0.0%
1.5%
1.5%
1.1%
1.4%
1.4%
1.4%
1.4%
1.4%
1.4%
1.4%
1.4%
1.4%
1.4%
1.4%
1.4%
1.4%
1.4%
1.4%
1.4%
1.4%
1.4%
1.4%
1.4%
1.4%
1.4%
1.4%

a social discount
(175shp<600)

Change in Total
Quantity Engineering
(%) Costs (103)
0.000%
-0.001%
-0.001%
-0.001%
-0.002%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%

rate of 3 percent
—
—
$6,434
$6,470
$5,975
$7,625
$7,662
$7,457
$7,494
$7,530
$7,469
$7,506
$3,727
$3,763
$3,799
$2,864
$2,901
$2,937
$2,974
$3,010
$3,046
$3,083
$3,119
$3,156
$3,192
$3,229
$3,265
$3,302
$76.149
over the 2004
Change in Producer
Surplus for
Equipment
Manufacturers (103)
-
$o
3
-$4,061
-$4,063
-$4,065
-$5,135
-$5,135
-$4,894
-$4,895
-$4,895
-$4,797
-$4,798
-$982
-$982
-$982
-$11
-$11
-$11
-$11
-$11
-$11
-$12
-$12
-$12
-$12
-$12
-$12
-$12
-$35.032
to 2036 time period.
10-143

-------
Final Regulatory Impact Analysis
Table 10.B-43. Impacts on General Industrial Equipment Market and Manufacturers (>600 hp)
(Average Price per Equipment = $345,700)a
General Industrial Equipment (^600hp)
Engineering
Year Cost/Unit
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2031
2032
2033
2034
2035
2036
NPVb
a Figures are
b Net present
—
—
—
—
$4,519
$4,496
$3,797
$4,684
$9,206
$8,364
$7,517
$7,489
$7,218
$6,767
$6,151
$6,142
$6,133
$5,997
$5,458
$5,458
$5,458
$5,458
$5,458
$5,458
$5,458
$5,458
$5,458
$5,458
$5,458
$5,458

Absolute
Change in
Price
—
-$1
-$1
-$2
$2,964
$2,963
$2,287
$2,789
$6,270
$6,270
$5,452
$5,452
$5,452
$5,452
$5,452
$5,452
$5,452
$5,452
$5,452
$5,452
$5,452
$5,452
$5,452
$5,452
$5,452
$5,452
$5,452
$5,452
$5,452
$5,452

in 2002 dollars.
values are calculated using
Change in
Price (%)
0.0%
0.0%
0.0%
0.0%
0.9%
0.9%
0.7%
0.8%
1.8%
1.8%
1.6%
1.6%
1.6%
1.6%
1.6%
1.6%
1.6%
1.6%
1.6%
1.6%
1.6%
1.6%
1.6%
1.6%
1.6%
1.6%
1.6%
1.6%
1.6%
1.6%

a social discount
Change in
Quantity
(%)
0.000%
-0.001%
-0.001%
-0.001%
-0.002%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%

Total
Engineering
Costs (103)
—
—
—
—
$665
$667
$634
$783
$1,439
$1,410
$1,371
$1,375
$1,369
$1,355
$881
$886
$890
$789
$346
$351
$355
$359
$364
$368
$372
$377
$381
$385
$390
$394
$11.760
rate of 3 percent over the 2004
Change in Producer
Surplus for
Equipment
Manufacturers (103)
—
—
—
—
-$512
-$512
-$513
-$629
-$1,095
-$1,061
-$1,061
-$1,061
-$1,050
-$1,031
-$554
-$554
-$554
-$449
-$2
-$2
-$2
-$2
-$2
-$2
-$2
-$2
-$2
-$2
-$2
-$2
-$7.192
to 2036 time period.
                                   10-144

-------
           Economic Impact Analysis
Table 10.B-44. Impacts on Lawn and Garden Equipment Market and Manufacturers (<25 hp)
(Average Price per Equipment = $9,300)a
Lawn and Garden Equipment (<25hp)
Absolute
Engineering Change in
Year Cost/Unit Price
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2031
2032
2033
2034
2035
2036
NPVb
a Figures are
b Net present
—
$177
$176
$168
$167
$166
$136
$136
$135
$135
$135
$123
$123
$123
$123
$123
$123
$123
$123
$123
$123
$123
$123
$123
$123
$123
$123
$123
$123
$123

—
$129
$129
$123
$123
$123
$123
$123
$123
$123
$123
$123
$123
$123
$123
$123
$123
$123
$123
$123
$123
$123
$123
$123
$123
$123
$123
$123
$123
$123

in 2002 dollars.
values are calculated using
Change in
Price (%)
0.0%
1.4%
1.4%
1.3%
1.3%
1.3%
1.3%
1.3%
1.3%
1.3%
1.3%
1.3%
1.3%
1.3%
1.3%
1.3%
1.3%
1.3%
1.3%
1.3%
1.3%
1.3%
1.3%
1.3%
1.3%
1.3%
1.3%
1.3%
1.3%
1.3%

a social discount
Change in Total
Quantity Engineering
(%) Costs (103)
0.000%
-0.001%
-0.001%
-0.001%
-0.002%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%

rate of 3 percent
—
$1,805
$1,836
$1,804
$1,834
$1,864
$1,597
$1,627
$1,657
$1,687
$1,717
$1,417
$1,447
$1,477
$1,507
$1,537
$1,568
$1,598
$1,628
$1,658
$1,688
$1,718
$1,749
$1,779
$1,809
$1,839
$1,869
$1,900
$1,930
$1,960
$29.853
over the 2004
Change in Producer
Surplus for
Equipment
Manufacturers (103)
—
-$629
-$629
-$629
-$630
-$630
-$333
-$333
-$333
-$333
-$333
-$2
-$2
-$2
$o
3
-$3
-$3
-$3
-$3
-$3
$o
3
$o
3
$o
3
-$3
-$3
-$3
-$3
-$3
$o
3
$o
3
-$3.868
to 2036 time period.
10-145

-------
Final Regulatory Impact Analysis
Table 10.B-45. Impacts on Lawn and Garden Equipment Market and Manufacturers (26-50 hp)
(Average Price per Equipment = $21,500)a
Lawn and Garden Equipment (25^hp<50)
Absolute
Engineering Change in
Year Cost/Unit Price
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2031
2032
2033
2034
2035
2036
NPVb
a Figures are
b Net present
—
$204
$203
$194
$193
$192
$986
$984
$773
$771
$769
$693
$692
$692
$691
$691
$661
$661
$661
$661
$661
$661
$661
$661
$661
$661
$661
$661
$661
$661

—
$147
$147
$139
$139
$139
$870
$870
$661
$661
$661
$661
$661
$661
$661
$661
$661
$661
$661
$661
$661
$661
$661
$661
$661
$661
$661
$661
$661
$661

in 2002 dollars.
values are calculated using
Change in
Price (%)
0.0%
0.7%
0.7%
0.6%
0.6%
0.6%
4.0%
4.0%
3.1%
3.1%
3.1%
3.1%
3.1%
3.1%
3.1%
3.1%
3.1%
3.1%
3.1%
3.1%
3.1%
3.1%
3.1%
3.1%
3.1%
3.1%
3.1%
3.1%
3.1%
3.1%

a social discount
Change in Total
Quantity Engineering
(%) Costs (103)
0.000%
-0.001%
-0.001%
-0.001%
-0.002%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%

rate of 3 percent
—
$474
$480
$471
$477
$482
$2,817
$2,858
$2,391
$2,422
$2,453
$2,228
$2,259
$2,290
$2,321
$2,353
$1,901
$1,933
$1,964
$1,995
$2,026
$2,057
$2,089
$2,120
$2,151
$2,182
$2,213
$2,245
$2,276
$2,307
$32.380
over the 2004
Change in Producer
Surplus for
Equipment
Manufacturers (103)
—
-$194
-$194
-$195
-$196
-$196
-$742
-$742
-$742
-$742
-$742
-$485
-$485
-$485
-$486
-$486
-$3
-$3
-$3
-$3
-$3
-$4
-$4
-$4
-$4
-$4
-$4
-$4
-$4
-$4
-$5.037
to 2036 time period.
                                   10-146

-------
           Economic Impact Analysis
Table 10.B-46. Impacts on Lawn and Garden Equipment Market and Manufacturers (51-75 hp)
(Average Price per Equipment = $33,100)a
Lawn and Garden Equipment (50^hp<75)
Absolute
Engineering Change in
Year Cost/Unit Price
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2031
2032
2033
2034
2035
2036
NPVb
a Figures are
b Net present
—
$226
$225
$214
$213
$212
$978
$976
$769
$767
$765
$687
$686
$686
$685
$685
$653
$653
$653
$653
$653
$653
$653
$653
$653
$653
$653
$653
$653
$653

—
$167
$167
$158
$158
$157
$858
$858
$653
$653
$653
$653
$653
$653
$653
$653
$653
$653
$653
$653
$653
$653
$653
$653
$653
$653
$653
$653
$653
$653

in 2002 dollars.
values are calculated using
Change in
Price (%)
0.0%
0.5%
0.5%
0.5%
0.5%
0.5%
2.6%
2.6%
2.0%
2.0%
2.0%
2.0%
2.0%
2.0%
2.0%
2.0%
2.0%
2.0%
2.0%
2.0%
2.0%
2.0%
2.0%
2.0%
2.0%
2.0%
2.0%
2.0%
2.0%
2.0%

a social discount
Chanw in Producer
Change in Total Surplus for
Quantity Engineering Equipment
(%) Costs (103) Manufacturers (103)
0.000%
-0.001%
-0.001%
-0.001%
-0.002%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%

rate of 3 percent
—
$14
$14
$14
$14
$14
$121
$122
$115
$115
$116
$102
$102
$103
$103
$104
$37
$37
$38
$38
$39
$40
$40
$41
$41
$42
$42
$43
$43
$44
$1.072
over the
—
-$14
-$14
-$15
-$15
-$15
-$83
-$83
-$83
-$83
-$83
-$68
-$68
-$68
-$68
-$68
-$1
-$1
-$1
-$1
-$1
-$1
-$1
-$1
-$1
-$1
-$1
-$1
-$1
-$1
-$577
2004 to 2036 time period.
10-147

-------
Final Regulatory Impact Analysis
Table 10.B-47. Impacts on Lawn and Garden Equipment Market and Manufacturers (76-100 hp)
(Average Price per Equipment = $38,500)a
Lawn and Garden Equipment (70shp<100)
Absolute
Engineering Change in
Year Cost/Unit Price
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2031
2032
2033
2034
2035
2036
NPVb
a Figures are
b Net present
—
—
$1,303
$1,302
$1,325
$1,324
$1,322
$1,247
$1,246
$1,218
$1,218
$1,218
$1,218
$1,218
$1,218
$1,218
$1,218
$1,218
$1,218
$1,218
$1,218
$1,218
$1,218
$1,218
$1,218
$1,218
$1,218

—
—
$1,178
$1,178
$1,169
$1,169
$1,169
$1,169
$1,169
$1,169
$1,169
$1,169
$1,169
$1,169
$1,169
$1,169
$1,169
$1,169
$1,169
$1,169
$1,169
$1,169
$1,169
$1,169
$1,169
$1,169
$1,169

in 2002 dollars.
values are calculated using
Change in
Price (%)
0.0%
0.0%
0.0%
0.0%
0.0%
3.1%
3.1%
3.0%
3.0%
3.0%
3.0%
3.0%
3.0%
3.0%
3.0%
3.0%
3.0%
3.0%
3.0%
3.0%
3.0%
3.0%
3.0%
3.0%
3.0%
3.0%
3.0%
3.0%
3.0%
3.0%

a social discount
Change in Total
Quantity Engineering
(%) Costs (103)
0.000%
-0.001%
-0.001%
-0.001%
-0.002%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%

rate of 3 percent
—
—
$529
$531
$641
$644
$647
$650
$653
$655
$658
$661
$290
$293
$200
$202
$205
$208
$211
$214
$217
$220
$222
$225
$228
$231
$234
$5.970
over the 2004
Change in Producer
Surplus for
Equipment
Manufacturers (103)
—
-$1
-$375
-$375
-$471
-$471
-$471
-$471
-$471
-$472
-$472
-$472
-$98
-$98
-$1
-$1
-$1
-$2
-$2
-$2
-$2
-$2
-$2
-$2
-$2
-$2
-$2
-$3.244
to 2036 time period.
                                   10-148

-------
           Economic Impact Analysis

Table 10
.B-48. Impacts on Lawn and Garden Equipment Market and
Manufacturers (101-175 hp)
(Average Price per Equipment = $29,200)a
Lawn and Garden Equipment
Engineering
Year Cost/Unit
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2031
2032
2033
2034
2035
2036
NPVb
a Figures are
b Net present
—
—
—
—
—
$1,623
$1,619
$1,664
$1,659
$1,654
$1,577
$1,574
$1,542
$1,539
$1,537
$1,388
$1,387
$1,351
$1,351
$1,351
$1,351
$1,351
$1,351
$1,351
$1,351
$1,351
$1,351
$1,351
$1,351
$1,351

Absolute
Change in
Price
—
—
—
—
—
$1,421
$1,421
$1,399
$1,399
$1,399
$1,399
$1,399
$1,399
$1,399
$1,399
$1,399
$1,399
$1,399
$1,399
$1,399
$1,399
$1,399
$1,399
$1,399
$1,399
$1,399
$1,399
$1,399
$1,399
$1,399

in 2002 dollars.
values are calculated using
Change in
Price (%)
0.0%
0.0%
0.0%
0.0%
0.0%
4.8%
4.8%
4.7%
4.7%
4.7%
4.7%
4.7%
4.7%
4.7%
4.7%
4.7%
4.7%
4.7%
4.7%
4.7%
4.7%
4.7%
4.7%
4.7%
4.7%
4.7%
4.7%
4.7%
4.7%
4.7%

a social discount
(10(khp<175)
Change in
Quantity
(%)
0.000%
-0.001%
-0.001%
-0.001%
-0.002%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%

Total
Engineering
Costs (103)
—
—
—
—
—
$420
$421
$514
$515
$517
$518
$520
$521
$523
$525
$195
$197
$114
$116
$117
$119
$120
$122
$124
$125
$127
$128
$130
$131
$133
$4.418
rate of 3 percent over the 2004
Change in Producer
Surplus for
Equipment
Manufacturers (103)
—
—
—
—
—
-$331
-$331
-$416
-$416
-$416
-$416
-$416
-$416
-$416
-$416
-$85
-$85
-$1
-$1
-$1
-$1
-$1
-$1
-$1
-$1
-$1
-$1
-$1
-$1
-$1
-$2.856
to 2036 time period.
10-149

-------
Final Regulatory Impact Analysis

Table 10
.B-49. Impacts on Lawn and Garden Equipment Market and
Manufacturers (176-600 hp)
(Average Price per Equipment = $64,300)a
Lawn and Garden Equipment
Engineering
Year Cost/Unit
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2031
2032
2033
2034
2035
2036
NPVb
a Figures are
b Net present
—
—
—
—
$2,970
$2,958
$2,439
$3,107
$3,092
$2,777
$2,768
$2,759
$2,634
$2,627
$2,294
$2,292
$2,291
$2,209
$2,208
$2,208
$2,207
$2,206
$2,206
$2,205
$2,204
$2,204
$2,203
$2,203
$2,202
$2,202

Absolute
Change in
Price
—
—
—
-$1
$2,265
$2,264
$1,755
$2,216
$2,215
$2,214
$2,213
$2,212
$2,211
$2,210
$2,210
$2,209
$2,208
$2,207
$2,207
$2,206
$2,205
$2,205
$2,204
$2,203
$2,203
$2,202
$2,202
$2,201
$2,200
$2,200

in 2002 dollars.
values are calculated using
Change in
Price (%)
0.0%
0.0%
0.0%
0.0%
3.5%
3.5%
2.7%
3.4%
3.4%
3.4%
3.4%
3.4%
3.4%
3.4%
3.4%
3.4%
3.4%
3.4%
3.4%
3.4%
3.4%
3.4%
3.4%
3.4%
3.4%
3.4%
3.4%
3.4%
3.4%
3.4%

a social discount
(175shp<600)
Change in
Quantity
(%)
0.000%
-0.001%
-0.001%
-0.001%
-0.002%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%

Total
Engineering
Costs (103)
—
—
—
—
$279
$280
$271
$344
$345
$346
$346
$347
$348
$349
$116
$117
$118
$59
$60
$60
$61
$62
$63
$63
$64
$65
$66
$66
$67
$68
$2.898
rate of 3 percent over the 2004
Change in Producer
Surplus for
Equipment
Manufacturers (103)
—
—
—
—
-$233
-$233
-$233
-$293
-$293
-$293
-$293
-$293
-$293
-$293
-$60
-$60
-$60
—
—
—
—
—
—
—
—
—
—
—
—
-$1
-$2.060
to 2036 time period.
                                   10-150

-------
                                                               Economic Impact Analysis
        APPENDIX IOC: Impacts on Application Markets


    This appendix provides the time series of impacts from 2007 through 2036 for the application markets and the
transportation service markets included in the model.

    There are 3 application markets: construction, agriculture, and manufacturing.

    There are 2 transportation service markets: locomotive and marine.

    Tables 10C-1 through 10C-5 provide the time series of impacts for these markets. Each table includes the
following:
    •   relative change in market price (%)
    •   relative change in market quantity (%)
    •   change in producer and consumer surplus for each application market

    For the three application markets, prices are expected to increase 0.02 percent in the manufacturing sector, 0.1
percent in the agricultural sector, and 0.5 percent in the construction sector.  Price increase are highest in about 2015,
and decrease thereafter.  Quantity decreases stabilize in about 2015 as well.

    For the transportation service markets, prices are expected to increase 0.03 percent in the locomotive sector and
0.006 percent in the marine sector. Price increases and quantity decreases stabilize in about 2015.
                                                10-151

-------
Final Regulatory Impact Analysis
                        Table 10C-1. Impacts on Agricultural Application Market3
Agriculture
Year
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2031
2032
2033
2034
2035
2036
NPVb
Change in Price (%)
0.030%
0.050%
0.050%
0.104%
0.142%
0.139%
0.136%
0.147%
0.154%
0.152%
0.150%
0.148%
0.146%
0.143%
0.140%
0.138%
0.136%
0.134%
0.132%
0.130%
0.128%
0.127%
0.125%
0.123%
0.121%
0.119%
0.118%
0.116%
0.114%
0.113%

Change in Quantity (%)
0.000%
-0.001%
-0.001%
-0.002%
-0.003%
-0.004%
-0.005%
-0.005%
-0.005%
-0.005%
-0.005%
-0.005%
-0.005%
-0.005%
-0.005%
-0.005%
-0.005%
-0.005%
-0.005%
-0.005%
-0.005%
-0.005%
-0.005%
-0.005%
-0.005%
-0.005%
-0.005%
-0.005%
-0.005%
-0.005%

Change in Producer and
Consumer Surplus ($103)
-$35,860
-$75,265
-$76,967
-$144,827
-$309,684
-$394,695
-$429,981
-$478,692
-$484,874
-$493,522
-$502,205
-$510,901
-$519,570
-$524,291
-$530,035
-$538,585
-$547,123
-$555,669
-$564,198
-$572,713
-$581,228
-$589,742
-$598,257
-$606,770
-$615,284
-$623,797
-$632,309
-$640,821
-$649,333
-$657,844
-$8.180.632
a   Figures are in 2002 dollars.
b   Net present values are calculated using a social discount rate of 3 percent over the 2004 to 2036 time period.
                                             10-152

-------
                                                          Economic Impact Analysis

Table 10C-2. Impacts
on Construction Application
Construction
Year
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2031
2032
2033
2034
2035
2036
NPVb
Change in Price (%)
0.105%
0.176%
0.174%
0.382%
0.526%
0.517%
0.508%
0.553%
0.587%
0.579%
0.573%
0.568%
0.565%
0.559%
0.554%
0.550%
0.544%
0.539%
0.533%
0.527%
0.522%
0.517%
0.512%
0.507%
0.502%
0.497%
0.492%
0.487%
0.482%
0.478%

Change in Quantity (%)
-0.001%
-0.001%
-0.001%
-0.002%
-0.004%
-0.005%
-0.006%
-0.006%
-0.006%
-0.006%
-0.006%
-0.006%
-0.006%
-0.006%
-0.006%
-0.006%
-0.006%
-0.006%
-0.006%
-0.006%
-0.006%
-0.006%
-0.006%
-0.006%
-0.006%
-0.006%
-0.006%
-0.006%
-0.006%
-0.006%

Market3
Change in Producer and
Consumer Surplus ($103)
-$47,524
-$97,113
-$99,303
-$199,991
-$409,111
-$548,053
-$584,333
-$650,082
-$689,966
-$702,193
-$709,196
-$721,412
-$733,610
-$744,027
-$754,910
-$767,057
-$779,171
-$791,302
-$803,409
-$815,495
-$827,581
-$839,668
-$851,754
-$863,841
-$875,929
-$888,016
-$900,104
-$912,193
-$924,281
-$936,370
-$11.525.673
a    Figures are in 2002 dollars.
b    Net present values are calculated using a social discount rate of 3 percent over the 2004 to 2036 time period.
                                            10-153

-------
Final  Regulatory Impact Analysis

Table 10C-3. Impacts on
Manufacturing Application Market"
Manufacturing
Year
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2031
2032
2033
2034
2035
2036
NPVb
Change in Price (%)
0.007%
0.015%
0.015%
0.028%
0.059%
0.074%
0.079%
0.086%
0.086%
0.086%
0.086%
0.086%
0.086%
0.086%
0.085%
0.085%
0.085%
0.085%
0.085%
0.085%
0.085%
0.085%
0.085%
0.085%
0.086%
0.086%
0.086%
0.086%
0.086%
0.086%

Change in Quantity (%)
-0.003%
-0.004%
-0.004%
-0.007%
-0.013%
-0.016%
-0.017%
-0.019%
-0.019%
-0.019%
-0.019%
-0.019%
-0.019%
-0.019%
-0.018%
-0.019%
-0.019%
-0.019%
-0.019%
-0.019%
-0.019%
-0.019%
-0.019%
-0.019%
-0.019%
-0.019%
-0.019%
-0.019%
-0.019%
-0.019%

Change in Producer and
Consumer Surplus ($103)
-$40,523
-$104,885
-$106,956
-$190,735
-$289,933
-$382,352
-$482,357
-$519,105
-$517,361
-$525,764
-$533,562
-$542,061
-$550,840
-$557,759
-$564,953
-$573,644
-$582,045
-$590,571
-$599,072
-$607,560
-$616,061
-$624,576
-$633,104
-$641,646
-$650,201
-$658,771
-$667,355
-$675,953
-$684,566
-$693,194
-$8.722.570
a    Figures are in 2002 dollars.
b    Net present values are calculated using a social discount rate of 3 percent over the 2004 to 2036 time period.
                                            10-154

-------
                                                          Economic Impact Analysis
Table 10C-4. Impacts on the Locomotive Transportation
Year
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2031
2032
2033
2034
2035
2036
NPVb
Manufacturing
Change in Price (%) Change
0.003%
0.005%
0.005%
0.010%
0.020%
0.027%
0.028%
0.031%
0.032%
0.032%
0.032%
0.032%
0.032%
0.032%
0.032%
0.032%
0.032%
0.032%
0.032%
0.032%
0.032%
0.032%
0.032%
0.032%
0.032%
0.032%
0.032%
0.032%
0.032%
0.032%


in Quantity (%)
-0.004%
-0.006%
-0.006%
-0.011%
-0.021%
-0.027%
-0.028%
-0.031%
-0.032%
-0.032%
-0.032%
-0.032%
-0.032%
-0.032%
-0.032%
-0.032%
-0.032%
-0.032%
-0.032%
-0.032%
-0.032%
-0.032%
-0.032%
-0.032%
-0.032%
-0.032%
-0.032%
-0.032%
-0.032%
-0.032%

Market3
Change in Producer and
Consumer Surplus ($103)
-$44
-$234
-$240
-$519
-$970
-$1,314
-$1,579
-$1,739
-$1,773
-$1,813
-$1,850
-$1,892
-$1,936
-$1,973
-$2,013
-$2,059
-$2,106
-$2,155
-$2,204
-$2,255
-$2,306
-$2,359
-$2,413
-$2,469
-$2,525
-$2,583
-$2,643
-$2,704
-$2,766
-$2,829
-$31.271
a  Figures are in 2002 dollars.
b  Net present values are calculated using a social discount rate of 3 percent over the 2004 to 2036 time period.
                                            10-155

-------
Final  Regulatory Impact Analysis
Table 10C-3. Impacts on the Marine Transportation
Year
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2031
2032
2033
2034
2035
2036
NPVb
Manufacturing
Change in Price (%) Change
0.001%
0.001%
0.001%
0.002%
0.004%
0.005%
0.006%
0.006%
0.006%
0.006%
0.006%
0.006%
0.006%
0.006%
0.006%
0.006%
0.006%
0.006%
0.006%
0.006%
0.006%
0.006%
0.006%
0.006%
0.006%
0.006%
0.006%
0.006%
0.006%
0.006%


in Quantity (%)
0.000%
-0.001%
-0.001%
-0.001%
-0.002%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%
-0.003%

Market3
Change in Producer and
Consumer Surplus ($103)
-$32
-$132
-$135
-$289
-$549
-$744
-$876
-$967
-$996
-$1,019
-$1,038
-$1,062
-$1,087
-$1,108
-$1,131
-$1,157
-$1,184
-$1,211
-$1,239
-$1,267
-$1,296
-$1,326
-$1,357
-$1,388
-$1,420
-$1,452
-$1,486
-$1,520
-$1,555
-$1,591
-$17.569
a  Figures are in 2002 dollars.
b  Net present values are calculated using a social discount rate of 3 percent over the 2004 to 2036 time period.
                                            10-156

-------
                                                     Economic Impact Analysis
            APPENDIX 10D: Impacts on the Nonroad Fuel Market
   This appendix provides the time series of impacts from 2007 through 2036 for the nonroad
diesel fuel market. Eight nonroad diesel fuel markets were modeled: 2 sulfur content levels (15
ppm and 500 ppm) for each of 4 PADDs (PADDs 1&3, PADD 2, PADD 4, and PADD 5). Note
that PADD 5 includes Alaska and Hawaii but excludes California fuel volumes that are not
affected by the program because they are covered by separate California nonroad diesel fuel
standards.

   Tables 10D-1 through 10D-4 provide the time series of impacts for the diesel fuel market for
the four regional fuel markets. Each table  includes the following:

   •   average price per gallon
   •   average engineering costs (variable and fixed) per gallon
   •   absolute change in the PADDs' nonroad diesel price ($)
       -  Note that the estimated absolute change in market price is based on average variable
          and fixed costs; see Appendix 101 for  sensitivity analyses reflecting maximum total
          costs and maximum variable costs
   •   relative change in the PADDs' nonroad diesel price (%)
   •   relative change in the PADDs' nonroad diesel quantity (%)
   •   total engineering (regulatory) costs associated with each PADD's fuel market ($)
   •   change in producer surplus for all fuel producers

   In 2001, about 68 percent of high-sulfur diesel fuel was consumed in nonroad diesel
equipment and about 32 percent was consumed in marine equipment and locomotive engines.8
The engineering costs and changes in producer surplus presented in this appendix include both of
these diesel fuel segments.

   All prices and costs are presented in $2002, and the real per-gallon prices are assumed to be
constant within each regional fuel market.  For each regional fuel market, the majority of the cost
of the regulation is passed along through increased fuel prices.
sThese percentages exclude heating oil; if high-sulfur heating oil is included, then about 35
   percent of high-sulfur fuel was consumed in nonroad diesel equipment and about 17 percent
   was consumed in marine equipment and locomotive engines.

                                        10-157

-------
Final  Regulatory Impact Analysis

Year
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2031
2032
2033
2034
2035
2036
NPVb
Table
Engineering
Cost/Unit
15ppm
—
—
—
$0.06
$0.06
$0.06
$0.06
$0.06
$0.06
$0.06
$0.06
$0.06
$0.06
$0.06
$0.06
$0.06
$0.06
$0.06
$0.06
$0.06
$0.06
$0.06
$0.06
$0.06
$0.06
$0.06
$0.06
$0.06
$0.06
$0.06

10D-1 . Impacts on the Nonroad Fuel Market in PADD 1&3
(Average Price per Gallon = $0.91)a
Engineering
Cost/Unit
SOOppm
$0.02
$0.02
$0.02
$0.02
$0.03
$0.03
$0.03
$0.03
$0.03
$0.03
$0.03
$0.03
$0.03
$0.03
$0.03
$0.03
$0.03
$0.03
$0.03
$0.03
$0.03
$0.03
$0.03
$0.03
$0.03
$0.03
$0.03
$0.03
$0.03
$0.03

Absolute
Change
in Price
$0.01
$0.02
$0.02
$0.04
$0.05
$0.05
$0.05
$0.06
$0.06
$0.06
$0.06
$0.06
$0.06
$0.06
$0.06
$0.06
$0.06
$0.06
$0.06
$0.06
$0.06
$0.06
$0.06
$0.06
$0.06
$0.06
$0.06
$0.06
$0.06
$0.06

Change
in Price
1.0%
1.8%
1.8%
4.1%
5.7%
5.7%
5.6%
6.1%
6.5%
6.5%
6.5%
6.5%
6.5%
6.5%
6.6%
6.6%
6.6%
6.6%
6.6%
6.6%
6.6%
6.6%
6.6%
6.6%
6.6%
6.6%
6.6%
6.6%
6.6%
6.6%

Change in
Quantity
-0.002%
-0.004%
-0.004%
-0.007%
-0.013%
-0.017%
-0.018%
-0.019%
-0.020%
-0.020%
-0.020%
-0.020%
-0.020%
-0.020%
-0.020%
-0.020%
-0.020%
-0.020%
-0.020%
-0.020%
-0.020%
-0.020%
-0.020%
-0.020%
-0.020%
-0.020%
-0.020%
-0.020%
-0.020%
-0.020%

Total
Engineering
Costs ($103)
$56,985
$99,743
$101,806
$236,629
$339,851
$346,465
$352,867
$390,537
$421,492
$429,036
$436,616
$444,324
$452,220
$462,196
$471,507
$479,447
$487,125
$494,924
$502,671
$510,413
$518,166
$525,932
$533,710
$541,500
$549,303
$557,119
$564,948
$572,789
$580,644
$588,512
$7,422,281
Change in
Producer
Surplus for Fuel
Producers ($103)
-$54
-$613
-$629
$65
-$2,313
-$3,292
-$3,624
-$4,187
-$4,532
-$4,625
-$4,689
-$4,783
-$4,877
-$5,027
-$5,164
-$5,259
-$5,353
-$5,448
-$5,542
-$5,636
-$5,730
-$5,824
-$5,918
-$6,012
-$6,106
-$6,200
-$6,294
-$6,388
-$6,482
-$6,576
-$76,083
a    Figures are in 2002 dollars.
b    Net present values are calculated using a social discount rate of 3 percent over the 2004 to 2036 time period.
                                            10-158

-------
                                                            Economic Impact Analysis
Table 10D-2. Impacts on the Nonroad Fuel Market in PADD 2
(Average Price per Gallon = $0.94)a
Year
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2031
2032
2033
2034
2035
2036
NPVb
Engineering
Cost/Unit
15ppm
—
—
—
$0.07
$0.07
$0.07
$0.07
$0.08
$0.08
$0.08
$0.08
$0.08
$0.08
$0.08
$0.08
$0.08
$0.08
$0.08
$0.08
$0.08
$0.08
$0.08
$0.08
$0.08
$0.08
$0.08
$0.08
$0.08
$0.08
$0.08

Engineering
Cost/Unit
SOOppm
$0.02
$0.02
$0.02
$0.03
$0.03
$0.03
$0.03
$0.03
$0.03
$0.03
$0.03
$0.03
$0.03
$0.03
$0.03
$0.03
$0.03
$0.03
$0.03
$0.03
$0.03
$0.03
$0.03
$0.03
$0.03
$0.03
$0.03
$0.03
$0.03
$0.03

Absolute
Change
in Price
$0.01
$0.02
$0.02
$0.05
$0.06
$0.06
$0.06
$0.07
$0.07
$0.07
$0.07
$0.07
$0.07
$0.07
$0.07
$0.07
$0.07
$0.07
$0.07
$0.07
$0.07
$0.07
$0.07
$0.07
$0.07
$0.07
$0.07
$0.07
$0.07
$0.07

Change
in Price
1.5%
2.6%
2.6%
5.0%
6.7%
6.7%
6.7%
7.3%
7.7%
7.7%
7.7%
7.7%
7.7%
7.5%
7.4%
7.4%
7.4%
7.4%
7.4%
7.4%
7.4%
7.4%
7.4%
7.4%
7.4%
7.5%
7.5%
7.5%
7.5%
7.5%

Change in
Quantity
-0.003%
-0.005%
-0.005%
-0.008%
-0.015%
-0.019%
-0.021%
-0.022%
-0.023%
-0.023%
-0.023%
-0.023%
-0.023%
-0.023%
-0.023%
-0.023%
-0.023%
-0.023%
-0.023%
-0.023%
-0.023%
-0.023%
-0.023%
-0.023%
-0.023%
-0.023%
-0.023%
-0.023%
-0.023%
-0.023%

Total
Engineering
Costs ($103)
$57,852
$101,359
$103,564
$204,945
$281,683
$287,389
$293,011
$323,985
$349,620
$356,353
$363,096
$369,869
$376,682
$374,491
$374,573
$381,107
$387,586
$394,090
$400,582
$407,040
$413,500
$419,963
$426,429
$432,896
$439,367
$445,840
$452,315
$458,794
$465,275
$471,758
$6,075,867
Change in
Producer
Surplus for Fuel
Producers ($103)
$64
-$544
-$558
$578
-$932
-$1,649
-$1,903
-$2,523
-$2,889
-$2,957
-$3,012
-$3,083
-$3,151
-$2,895
-$2,733
-$2,791
-$2,849
-$2,907
-$2,964
-$3,021
-$3,079
-$3,136
-$3,194
-$3,251
-$3,308
-$3,366
-$3,423
-$3,480
-$3,537
-$3,594
-$42,383
a   Figures are in 2001 dollars.
b   Net present values are calculated using a social discount rate of 3 percent over the 2004 to 2030 time period.
                                              10-159

-------
Final Regulatory Impact Analysis
Table 10D-3. Impacts on the Nonroad Fuel Market in PADD 4
(Average Price per Gallon = $0.91)a
Engineering
Cost/Unit
Year 15ppm
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2031
2032
2033
2034
2035
2036
NPVb
—
—
—
$0.13
$0.13
$0.13
$0.13
$0.13
$0.13
$0.13
$0.13
$0.13
$0.13
$0.13
$0.13
$0.13
$0.13
$0.13
$0.13
$0.13
$0.13
$0.13
$0.13
$0.13
$0.13
$0.13
$0.13
$0.13
$0.13
$0.13

Engineering
Cost/Unit
SOOppm
$0.04
$0.04
$0.04
$0.04
$0.09
$0.09
$0.09
$0.09
$0.09
$0.09
$0.09
$0.09
$0.09
$0.09
$0.09
$0.09
$0.09
$0.09
$0.09
$0.09
$0.09
$0.09
$0.09
$0.09
$0.09
$0.09
$0.09
$0.09
$0.09
$0.09

Absolute
Change
in Price
$0.02
$0.03
$0.03
$0.07
$0.09
$0.09
$0.09
$0.09
$0.10
$0.10
$0.10
$0.10
$0.10
$0.10
$0.10
$0.10
$0.10
$0.10
$0.10
$0.10
$0.10
$0.10
$0.10
$0.10
$0.10
$0.10
$0.10
$0.10
$0.10
$0.10

Change
in Price
2.0%
3.4%
3.4%
6.8%
9.1%
9.1%
9.1%
9.9%
10.6%
10.6%
10.6%
10.6%
10.6%
10.4%
10.3%
10.3%
10.3%
10.3%
10.4%
10.4%
10.4%
10.4%
10.4%
10.4%
10.4%
10.4%
10.4%
10.4%
10.4%
10.4%

Change in
Quantity
-0.003%
-0.005%
-0.005%
-0.009%
-0.016%
-0.020%
-0.021%
-0.023%
-0.024%
-0.024%
-0.024%
-0.024%
-0.024%
-0.024%
-0.024%
-0.024%
-0.024%
-0.024%
-0.024%
-0.024%
-0.024%
-0.024%
-0.024%
-0.024%
-0.024%
-0.024%
-0.024%
-0.024%
-0.024%
-0.024%

Change in
Total Producer
Engineering Surplus for Fuel
Costs ($103) Producers ($103)
$6,826
$11,955
$12,214
$24,781
$33,824
$34,500
$35,166
$39,254
$42,621
$43,461
$44,301
$45,142
$45,982
$45,886
$46,029
$46,840
$47,652
$48,463
$49,275
$50,081
$50,886
$51,692
$52,498
$53,304
$54,109
$54,915
$55,721
$56,527
$57,333
$58,138
$742,250
$34
-$34
-$35
$432
$459
$401
$390
$324
$273
$276
$280
$281
$284
$322
$349
$352
$356
$359
$363
$366
$369
$373
$376
$379
$383
$386
$390
$393
$397
$400
$5,626
a   Figures are in 2001 dollars.
b   Net present values are calculated using a social discount rate of 3 percent over the 2004 to 2030 time period.
                                              10-160

-------
                                                         Economic Impact Analysis
Table 10D-4. Impacts on the Nonroad Fuel Market in PADD 5
(Average Price per Gallon = $0.87)a
Engineering
Cost/Unit
Year 15ppm
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2031
2032
2033
2034
2035
2036
NPVb
—
—
—
$0.05
$0.05
$0.05
$0.05
$0.06
$0.07
$0.07
$0.07
$0.07
$0.07
$0.07
$0.07
$0.07
$0.07
$0.07
$0.07
$0.07
$0.07
$0.07
$0.07
$0.07
$0.07
$0.07
$0.07
$0.07
$0.07
$0.07

Engineering
Cost/Unit
SOOppm
$0.01
$0.01
$0.01
$0.02
$0.04
$0.04
$0.04
$0.04
$0.04
$0.04
$0.04
$0.04
$0.04
$0.04
$0.04
$0.04
$0.04
$0.04
$0.04
$0.04
$0.04
$0.04
$0.04
$0.04
$0.04
$0.04
$0.04
$0.04
$0.04
$0.04

Absolute
Change
in Price
$0.01
$0.01
$0.01
$0.02
$0.03
$0.03
$0.03
$0.04
$0.06
$0.06
$0.06
$0.06
$0.06
$0.06
$0.06
$0.06
$0.06
$0.06
$0.06
$0.06
$0.06
$0.06
$0.06
$0.06
$0.06
$0.06
$0.06
$0.06
$0.06
$0.06

Change
in Price
0.5%
0.9%
0.9%
1.8%
2.8%
2.8%
2.8%
4.4%
5.9%
5.9%
5.9%
5.9%
5.9%
5.9%
5.9%
5.9%
5.9%
5.9%
5.9%
5.9%
5.9%
5.9%
5.9%
5.9%
5.9%
5.9%
5.9%
6.0%
6.0%
6.0%

Change in
Quantity
-0.003%
-0.005%
-0.005%
-0.008%
-0.015%
-0.019%
-0.020%
-0.022%
-0.023%
-0.023%
-0.023%
-0.023%
-0.023%
-0.023%
-0.023%
-0.023%
-0.023%
-0.023%
-0.023%
-0.023%
-0.023%
-0.023%
-0.023%
-0.023%
-0.023%
-0.023%
-0.023%
-0.023%
-0.023%
-0.023%

Total
Engineering
Costs ($103)
$3,004
$5,266
$5,382
$11,146
$17,727
$18,083
$18,428
$29,541
$40,159
$40,915
$41,678
$42,453
$43,236
$44,001
$44,768
$45,551
$46,317
$47,090
$47,859
$48,627
$49,396
$50,166
$50,936
$51,707
$52,478
$53,251
$54,024
$54,797
$55,572
$56,347
$647,478
Change in
Producer
Surplus for Fuel
Producers ($103)
-$24
-$68
-$70
-$44
-$171
-$287
-$322
-$321
-$377
-$385
-$390
-$398
-$406
-$413
-$420
-$428
-$436
-$444
-$452
-$460
-$468
-$476
-$485
-$493
-$501
-$509
-$517
-$525
-$533
-$541
-$6,343
a    Figures are in 2001 dollars.
b    Net present values are calculated using a social discount rate of 3 percent over the 2004 to 2030 time period.
                                           10-161

-------
Final Regulatory Impact Analysis
                       APPENDIX 10E: Time Series of Social Cost
   This appendix provides a time series of the rule's estimated social costs from 2007 through 2036. Costs are
presented in 2002 dollars.
                                         10-162

-------
Table 10E-1.  Time Series of Market Impacts

Engine Producers Total
Equipment Producers Total
Construction Equipment
Agricultural Equipment
Industrial Equipment
Application Producers & Consumers Total
Total Producer-
Total Consumer
Construction
Agriculture
Manufacturing
Fuel Producers Total
PADD 1 & 3
PADD2
PADD 4
PADD 5
Transportation Services, Total
Locomotive
Marine
Application Markets Not Included in
Operating Savings
Total
2007
$0.0
$0.3
$0.2
$0.1
$0.0
$123.9
$45.5
$78.4
$47.5
$35.9
$40.5
$0.2
$0.1
$0.0
$0.0
$0.0
$18.9
$0.0
$0.0
$18.9
-$160.9
-$17.6
2008
$14.9
$8.8
$1.8
$2.4
$4.6
$277.3
$108.4
$168.8
$97.1
$75.3
$104.9
$1.7
$0.7
$0.8
$0.1
$0.1
$33.1
$0.2
$0.1
$32.7
-$281.9
$53.9
2009
$14.9
$8.8
$1.8
$2.4
$4.7
$283.2
$110.8
$172.4
$99.3
$77.0
$107.0
$1.7
$0.7
$0.8
$0.1
$0.1
$33.5
$0.2
$0.1
$33.1
-$288.0
$54.2
2010
$14.9
$9.6
$2.3
$2.6
$4.7
$535.6
$216.5
$319.1
$200.0
$144.8
$190.7
-$0.2
$0.1
-$0.1
-$0.3
$0.1
$71.5
$0.5
$0.3
$70.7
-$304.6
$326.7
2011
$29.4
$88.7
$41.3
$36.0
$11.4
$1,008.7
$418.5
$590.2
$409.1
$309.7
$289.9
$4.7
$2.6
$1.9
-$0.2
$0.4
$102.0
$1.0
$0.5
$100.5
-$311.4
$922.3
2012
$38.9
$131.4
$60.8
$48.3
$22.3
$1,325.1
$553.0
$772.1
$548.1
$394.7
$382.4
$7.2
$3.7
$2.9
$0.0
$0.6
$103.6
$1.3
$0.7
$101.6
-$302.2
$1,304.0
2013
$42.0
$143.1
$64.0
$51.8
$27.2
$1,496.7
$620.9
$875.7
$584.3
$430.0
$482.4
$8.0
$4.1
$3.3
$0.0
$0.6
$104.9
$1.6
$0.9
$102.4
-$284.7
$1,510.0
2014
$51.6
$179.0
$81.1
$66.0
$31.9
$1,647.9
$685.2
$962.7
$650.1
$478.7
$519.1
$9.6
$4.7
$4.0
$0.1
$0.7
$95.5
$1.7
$1.0
$92.8
-$293.0
$1,690.5
2015
$52
$186
$87
$66
$32

.4
.0
.6
.0
.4
$1,692.2
$706.4
$985
$690
$484
$517
$10
$5
$4
$0
$0
$88
$1
$1
$85
-$288
$1,741
'.8
.0
.9
.4
.5
.1
.4
.2
.7
.3
.8
.0
.5
.0
.3
2016
$37.9
$156.9
$73.0
$52.2
$31.8
$1,721.5
$718.6
$1,002.8
$702.2
$493.5
$525.8
$10.7
$5.2
$4.5
$0.2
$0.8
$89.2
$1.8
$1.0
$86.4
-$273.6
$1,742.6
(continued)
                10-163

-------
Table 10E-1. Time Series of Market Impacts (continued)

Engine Producers Total
Equipment Producers Total
Construction Equipment
Agricultural Equipment
Industrial Equipment
Application Producers & Consumers Total
Total Producer-
Total Consumer
Construction
Agriculture
Manufacturing
Fuel Producers Total
PADD 1 & 3
PADD2
PADD 4
PADD 5
Transportation Services, Total
Locomotive
Marine
Application Markets Not Included in
Operating Savings
Total
2017
$28
$148
$68
$49
$30
$1,745

.4
.6
.9
.3
.4
.0
$728.2
$1,016.8
$709
$502
$533
$10
$5
$4
$0
$0
$90
$1
$1
$87
-$260
$1,762
.2
.2
.6
.9
.3
.6
.2
.8
.2
.8
.0
.3
.8
.2
2018
$10.4
$139.7
$67.3
$46.5
$26.0
$1,774.4
$740.5
$1,033.9
$721.4
$510.9
$542.1
$11.1
$5.4
$4.7
$0.2
$0.8
$91.3
$1.9
$1.1
$88.3
-$249.4
$1,777.6
2019
$0.9
$125.0
$60.2
$39.7
$25.2
$1,804.0
$752.9
$1,051.1
$733.6
$519.6
$550.8
$11.3
$5.5
$4.8
$0.2
$0.8
$92.6
$1.9
$1.1
$89.6
-$239.3
$1,794.6
2020
$0.1
$122.7
$57.8
$39.7
$25.2
$1,826.1
$762.2
$1,063.8
$744.0
$524.3
$557.8
$11.2
$5.6
$4.6
$0.2
$0.8
$95.6
$2.0
$1.1
$92.6
-$227.4
$1,828.3
2021
$0.1
$74.2
$34.4
$20.6
$19.1
$1,849.9
$772.3
$1,077.6
$754.9
$530.0
$565.0
$11.2
$5.8
$4.5
$0.2
$0.8
$98.1
$2.0
$1.1
$95.0
-$218.2
$1,815.3
2022
$0.1
$40.9
$19.8
$11.5
$9.6
$1,879.3
$784.6
$1,094.7
$767.1
$538.6
$573.6
$11.5
$5.9
$4.5
$0.2
$0.8
$99.5
$2.1
$1.2
$96.2
-$212.8
$1,818.5
2023
$0.1
$30.4
$17.0
$8.9
$4.5
$1,908.3
$796.8
$1,111.6
$779.2
$547.1
$582.0
$11.7
$6.0
$4.6
$0.2
$0.8
$100.5
$2.1
$1.2
$97.2
-$208.1
$1,843.0
2024
$0.1
$9.8
$7.5
$1.7
$0.6
$1,937.5
$809.0
$1,128.5
$791.3
$555.7
$590.6
$11.9
$6.1
$4.7
$0.2
$0.9
$101.7
$2.2
$1.2
$98.4
-$204.2
$1,856.9
2025
$0.1
$5.6
$3.7
$1.7
$0.1
$1,966.7
$821.2
$1,145.5
$803.4
$564.2
$599.1
$12.1
$6.2
$4.8
$0.2
$0.9
$102.9
$2.2
$1.2
$99.4
-$200.7
$1,886.6
2026
$0.1
$5.6
$3.8
$1.8
$0.1
$1,995.8
$833.4
$1,162.4
$815.5
$572.7
$607.6
$12.3
$6.3
$4.9
$0.2
$0.9
$104.1
$2.3
$1.3
$100.6
-$198.0
$1,919.9
                  (continued)

-------
                                           Table 10E-1. Time Series of Market Impacts (continued)

Engine Producers Total
Equipment Producers Total
Construction Equipment
Agricultural Equipment
Industrial Equipment
Application Producers & Consumers Total
Total Producer
Total Consumer
Construction
Agriculture
Manufacturing
Fuel Producers Total
PADD 1 & 3
PADD2
PADD 4
PADD 5
Transportation Services, Total
Locomotive
Marine
Application Markets Not Included in
Operating Savings
Total
2027
$0.1
$5.7
$3.8
$1.8
$0.1
$2,024.9
$845.6
$1,179.3
$827.6
$581.2
$616.1
$12.5
$6.4
$5.0
$0.2
$0.9
$105.3
$2.3
$1.3
$101.7
-$196.0
$1,952.5
2028
$0.1
$5.8
$3.9
$1.8
$0.1
$2,054.0
$857.8
$1,196.2
$839.7
$589.7
$624.6
$12.7
$6.5
$5.1
$0.2
$0.9
$106.5
$2.4
$1.3
$102.8
-$194.9
$1,984.2
2029
$0.1
$5.9
$3.9
$1.8
$0.1
$2,083.1
$870.0
$1,213.1
$851.8
$598.3
$633.1
$13.0
$6.6
$5.2
$0.2
$0.9
$107.8
$2.4
$1.4
$104.0
-$194.3
$2,015.5
2030
$0.1
$5.9
$4.0
$1.9
$0.1
$2,112.3
$882.2
$1,230.1
$863.8
$606.8
$641.6
$13.2
$6.7
$5.2
$0.3
$1.0
$109.0
$2.5
$1.4
$105.2
-$194.1
$2,046.4
2031
$0.
$6.
$4.
$1.
$0.
$2,141.
$894.
$1,247.
$875.
$615.
$650.
$13.
$6.
$5.
$0.
$1.
$110.
$2.
$1.
$106.
-$194.

1
0
0
9
1
4
4
0
9
o
J
2
4
8
o
J
o
J
0
3
5
4
o
J
3
$2,076.9
2032
$0.1
$6.1
$4.1
$1.9
$0.1
$2,170.6
$906.6
$1,264.0
$888.0
$623.8
$658.8
$13.6
$6.9
$5.4
$0.3
$1.0
$111.6
$2.6
$1.5
$107.5
-$194.8
$2,107.2
2033
$0.1
$6.2
$4.1
$1.9
$0.1
$2,199.8
$918.8
$1,280.9
$900.1
$632.3
$667.4
$13.8
$7.0
$5.5
$0.3
$1.0
$112.9
$2.6
$1.5
$108.8
-$195.4
$2,137.4
2034
$0.2
$6.2
$4.2
$2.0
$0.1
$2,229.0
$931.1
$1,297.9
$912.2
$640.8
$676.0
$14.0
$7.1
$5.6
$0.3
$1.0
$114.2
$2.7
$1.5
$110.0
-$196.1
$2,167.5
2035
$0.2
$6.3
$4.2
$2.0
$0.1
$2,258.2
$943.3
$1,314.9
$924.3
$649.3
$684.6
$14.2
$7.2
$5.7
$0.3
$1.0
$115.6
$2.8
$1.6
$111.2
-$197.1
$2,197.3
2036
$0.2
$6.4
$4.3
$2.0
$0.1
$2,287.4
$955.5
$1,331.9
$936.4
$657.8
$693.2
$14.5
$7.3
$5.8
$0.3
$1.0
$116.9
$2.8
$1.6
$112.5
-$198.4
$2,227.0
a   Figures are in 2002 dollars.
b   Net present values are calculated using a social discount rate of 3 percent over the 2004 to 2036 time period.

-------
Final Regulatory Impact Analysis
                                APPENDIX 10F: Model Equations

    To enhance understanding of the economic model EPA used in this report, additional details about the model's
structure are provided in this appendix.  The equations describing supply, final demand, and intermediate (i.e.,
derived) demand relationships are presented below along with a brief description of the solution algorithm.

10F.1 Model Equations

    A constant-elasticity functional form was selected for all supply and final demand functions. The general form
and description of these equations are presented below:

                                  Supply Equation: Qx = a(Px - Ac - Ac/                           (10F.1)

                                    Final Demand Equation: Qx = aPxn                             (10F.2)

where
        x     =   production output,
        y     =   production input,
        Qx    =   quantity of output (x) supplied or demanded,
        Px    =   market price for output (x),
        a     =   constant,
        Ac    =   direct supply shift ($/Qx),
        Acy   =   indirect supply shift resulting from change in the price of input y, and
        e,r|    =   these parameters can be interpreted as the own-price elasticity of supply/demand for the
                  economic agent (see Tables 10.3-12 and  10.3-13 for values of these parameters).

    With this choice of functional form, the supply and demand elasticities are assumed to remain constant over the
range of output affected by the regulation. This can be demonstrated by applying the definition of own-price
elasticity of demand:

                                *i  . 1  =   Eap(1-£)  • £±1 = e.                        (10F.3)
                                dp      q                      a


    The intermediate input (Qy) demands is  specified within the supply chain as a function of output (QJ. The
subscript "0" denotes baseline and the subscript "1" denotes with regulation.

            Derived Demand Equation:    Qy = f(Qx)                                               (10F.4a)

                                        Qyl = Qy0(l+AQx/Qx)                                     (10F.4b)

    Computing Supply/Demand Function Constants. Using the baseline price, quantity, and elasticity parameter,
the value of the constants can be computed.  For example, supply function constants can be calculated as follows:

                                    Constant Calibration: a=    x°                                (10F.5)
                                                            (Pxo)£
    Direct Supply Shift (Dc). The  direct upward shift in the  supply function is calculated by using the annualized
compliance cost estimates provided by the engineering cost analysis.  Computing the supply shift in this manner
treats the compliance costs as the conceptual equivalent of a unit tax on output.

    Indirect Supply Shift (Dcy).  The indirect upward shift in the supply function is calculated by using the change in
input (y) prices (i.e., engines, equipment, and/or fuel) that result from the direct compliance costs introduced into the
model. Only two types of suppliers are  affected by these changes: equipment producers that use diesel engines and
application markets that use equipment with diesel  engines and diesel fuel. The term Dcy is computed as follows:


                                                 10-166

-------
                                                             Economic Impact Analysis
                                                 QyO
                                                  Y  .                                      (10F.6)
10F.2 Engine Markets

    As described in Section 10.3.3.1, seven separate engine markets were modeled segmented by engine size in
horsepower (the EIA includes more horsepower categories than the standards, allowing more efficient use of the
engine compliance cost estimates developed for this rule):

    •    less than 25 hp
    •    26 to 50 hp
        51 to 75 hp
        76 to 100 hp
    •    101 to 175 hp
        176 to 600 hp
    •    greater than 601 hp

    In each of these engine markets, there are three types of suppliers: captive suppliers (engines are consumed
internally by integrated equipment manufacturers), merchant suppliers (engines are sold on the open market), and
foreign suppliers.  These supply segments are represented by upward-sloping supply functions.  On the demand side,
consumers of engines include  integrated and nonintegrated equipment manufacturers1 and are represented by derived
demand functions (Eqs. [10F-4a] and [10F.4b]).

    Captive Domestic Supply Equation:        Sengcap = a^p—c)                       (10F.7)

    Merchant Domestic Supply Equation:       S engmer  = a2 (p — c)e                     (1 OF. 8)

    Import Supply Equation:              Meng = a(p-c)£                         (10F.9)

    Integrated Demand Equation:         D, = S (Seqmp)                                           (10F. 10)

    Nonintegrated Demand Equation:     DNI=S(Seqmp)                                          (10F.11)

    Market Clearing Condition:          Sengcap + Sengmer + Meng = D, + DNI                          (10F. 12)


10F.3 Equipment Markets

    As described in Section 10.3.3.2, integrated and nonintegrated equipment manufacturers supply their products
into a series of 42 equipment markets (7 horsepower categories within 7 application categories; there are 7
horsepower/application categories that did not have sales in 2000 and are not included in the model, so the total
number of diesel equipment markets is 42, not 49).u  The equipment types are:
TNote that engines sold to foreign equipment manufacturers are not included in the domestic
   engine market because they are subject to different (foreign) environmental regulations and
   hence are considered different products.

u These are: agricultural equipment >600 hp; gensets & welders > 600 hp; refrigeration & A/C >
   71 hp (4 hp categories); and lawn & garden >600 hp.

                                              10-167

-------
Final Regulatory Impact Analysis
    •   agricultural
    •   construction
    •   refrigeration
    •   generators and welder sets
    •   lawn and garden
    •   pumps and compressors
    •   general industrial

    Each individual equipment market is comprised of two aggregate suppliers groups: (1) domestic integrated
suppliers that produce and consume their own engines (captive engines) and (2) domestic nonintegrated suppliers
that purchase engines from the open market to be used in their equipment (merchant engines).

    On the demand side, each of the 42 equipment markets is linked to one of three application markets
(agricultural, construction, and manufacturers) is represented by derived demand functions (Eq. [10F.4a and
10F.4b])

    Domestic Integrated Supply Equation:      Seql  = a(p-c)£                          (10F.13)

    Domestic Nonintegrated Supply Equation:   SeqNI = a(p — c — cy)e                          (10F.14)

    Domestic Demand Equation:           Deq = Y  Qeq  1 H	—              (10F.15)
                                                        I    Qqppo J
    Market Clearing Condition:            Seql + SeqNI  = Deq                                         (10F. 16)

10F.4 Application  Markets

    As described in Section 10.3.3.3, there are three application markets that supply products and services to
consumers:

    •   agricultural
    •   construction
    •   manufacturing

    The supply in each of these three application markets is the sum of a domestic supply and an foreign (import)
supply. The consumers in the application markets are represented by a domestic demand and a foreign (export)
demand function.


    Supply Equation:                 Sapp  =  &(p       -   C  - /&p)Eks           (10F.17)

    Foreign (Import) Supply Equation:  Sapp =  apapp                                    (10F.18)

    Domestic Demand Equation:       D app = apn                                        (1 OF. 19)

    Foreign (Export) Demand Equation:    Xapp = ap11                                        (10F.20)

    Market Clearing Condition:        Sapp + Mapp = Dapp+Xapp                                       (10F.21)

    Po, Pi, and P2 are the baseline input shares of equipment, fuel, and transportation services.


1 OF.5 Fuel Markets

    As described in Section 10.3.3.4, eight nonroad diesel fuel markets were modeled: two distinct nonroad diesel
fuel commodities in four regional markets. The two fuels are:

                                               10-168

-------
                                                               Economic Impact Analysis
    •   500 ppm nonroad diesel fuel, and
    •   15 ppm nonroad diesel fuel.

    The four regional nonroad diesel fuel markets are

    •   PADD 1 and 3
    •   PADD2
    •   PADD4
    •   PADD 5 (includes Alaska and Hawaii; California fuel volumes that are not affected by the program because
        they are covered by separate California nonroad diesel fuel standards are not included in the analysis)

    The supply and demand for nonroad diesel fuel is specified for the model for four regional diesel fuel markets.
Derived demand of diesel fuel comes from three application markets. The equations for PADD district j are
specified below:

        Supply Equation:                Sj = a(Pj-Ac)e                                           (10F.22)
                                                               AQ
                                                                   app
                                                               Q,
                                                                                               (10F.23)
        Derived Demand Equation:               Dj = S Q -0 1 +
                                                               ''appO
        Market Clearing Condition:               SJ = DJ                                          (10F.24)


10F.6  Locomotive and Marine Transportation Markets

    There are two transportation service markets that supply services to the application markets:

    •   locomotive
    •   marine

    The supply in each of these three application markets is the sum of a domestic supply

    Supply Equation:                 S,OT=  a(PtTans  -  c  -/ftpftel)E-                          (10F.25)

    Market Clearing Condition:       S^ =0^                                                 (10F.26)


    P is the baseline input share of fuel '


10F.7  Market-Clearing Process and Equations

    Supply responses and market adjustments can be conceptualized as an interactive process. Producers facing
increased production costs due to compliance with the control program are willing to supply smaller quantities at the
baseline price.  This reduction in market supply leads to an increase in the market price that all producers and
consumers face, which leads to further responses by producers and consumers and thus new market prices, and so
on. The new with-regulation equilibrium is the result of a series of iterations in which price is adjusted and
producers and consumers respond, until a set of stable market prices arises where total market supply equals market
demand.

                          Market-Clearing Equation:  Total Supply = Total Demand.                   (10F.27)

The algorithm for determining with-regulation equilibria can be summarized by six recursive steps:

    1.  Impose the control costs on affected supply  segments, thereby affecting their supply decisions.


                                                10-169

-------
Final Regulatory Impact Analysis
    2.   Recalculate the market supply in each market.  Excess demand currently exists.
    3.   Determine the new prices via a price revision rule.  A rule similar to the factor price revision rule described
        by Kimbell and Harrison (1986) is used.  P; is the market price at iteration i, qd is the quantity demanded,
        and qs is the quantity supplied.  The parameter z influences the magnitude of the price revision and speed of
        convergence.  The revision rule increases the price when excess demand exists, lowers the price when
        excess supply exists, and leaves the price unchanged when market demand equals market supply. The price
        adjustment is expressed as follows:
                                                     (   Y
                                            P1+1 = PI '  —                                       (10F.26)
                                                     UJ
    4.   Recalculate market supply with new prices.
    5.   Compute market demand in each market.
    6.   Compare supply and demand in each market. If equilibrium conditions are not satisfied, go to Step 3,
        resulting in a new set of market prices. Repeat until equilibrium conditions are satisfied (i.e., the ratio of
        supply and demand is arbitrarily close to one).
                                                10-170

-------
                                                                Economic Impact Analysis
           APPENDIX 10G: Elasticity Parameters for Economic Impact Modeling

    The Nonroad Diesel Economic Impact Model (NDEIM) relies on elasticity parameters to estimate the
behavioral response of consumers and producers to the regulation and its associated costs. To operationalize the
market model, supply and demand elasticities are needed to represent the behavioral adjustments that are likely to be
made by market participants.  The following parameters are needed:

    •   supply and demand elasticities for application markets (agriculture, construction, and manufacturing)
    •   supply elasticities for equipment markets
    •   supply elasticities for engine markets
    •   supply elasticities for diesel fuel markets
    •   supply elasticities for locomotive and marine transportation markets

    Note that demand elasticities for the equipment, engine, diesel fuel, and transportation markets are not estimated
because they are derived internally in the model. They are a function of changes in output levels in the applications
markets.

    Tables 10G-1 and 10G-2 contain the demand and supply elasticities used to estimate the economic impact of the
rule. Two methods were used to obtain the supply and demand elasticities used in the NDEIM.  First, the
professional literature was surveyed to identify elasticity estimates used in published studies.  Second, when
literature estimates were not available for specific markets, established econometric techniques were used to estimate
supply and demand elasticity parameters directly.  Specifically, the supply elasticities for the agricultural and
construction application markets and the  supply elasticity for the diesel fuel market were obtained from the
literature.  The supply elasticity for the manufacturing market is assumed to be the same as for the construction
market. The supply elasticities for all of the  application markets and for equipment and engine markets were
estimated econometrically.

    This appendix discusses the literature for elasticities based on existing studies and presents the data sources and
estimation methodology and regression results for the econometric estimation.

    Finally, it should be noted that these elasticities reflect intermediate run behavioral changes.  In the long run,
supply and demand are expected to be more elastic since more substitutes may become available.
                                                10-171

-------
Final  Regulatory Impact Analysis
                                            Table 10G-1
                       Summary of Market Demand Elasticities Used in the NDEIM
Market
Applications
Agriculture
Estimate
-0.20
Source
EPA econometric
Method
Productivity shift
Input Data Summary
Annual time series from
   Construction
   Manufacturing
                              estimate
-0.96     EPA econometric
          estimate
-0.58     EPA econometric
          estimate
                            approach (Morgenstern,
                            Pizer, and Shih, 2002)
Simultaneous equation
(log-log) approach
Simultaneous equation
(log-log) approach.
1958- 1995 developed by
       Jorgenson et al.
(Jorgenson, 1990; Jorgenson,
Gollop, andFraumeni, 1987)

Annual time series from
1958- 1995 developed by
Jorgenson et al. (Jorgenson,
1990; Jorgenson, Gollop, and
Fraumeni, 1987)

Annual time series from
1958- 1995 developed by
Jorgenson et al. (Jorgenson,
1990; Jorgenson, Gollop, and
Fraumeni, 1987)
Transportation
Services
Locomotive
Marine
Equipment
Agriculture
Construction
Pumps/
compressors
Generators and
Welders
Refrigeration
Industrial
Lawn and
Garden
Engines
Diesel fuel

Derived demand
Derived demand
Derived demand
Derived demand
Derived demand
Derived demand

Derived demand
Derived demand
Derived demand

Derived demand
Derived demand

In the derived demand approach,


• compliance costs increase prices and decrease demand
for products and services in the application markets;
• this in turn leads to reduced demand for diesel

equipment, engines and fuel, which are inputs into the
production of products and services in the application
markets
















                                              10-172

-------
                                                                       Economic  Impact Analysis
Table 10G-2
Summary of Market Supply Elasticities Used in the NDEIM
Markets
Applications
Agriculture


Construction
Manufacturing


Transportation
Services
Locomotive

Marine


Equipment
Agriculture
Construction
Pumps/
compressors
Generators/
Welder Sets
Refrigeration
Industrial
Lawn and
Garden
Engines
Diesel fuel
Estimate

0.32


1
1



0.6

0.6



2.14
3.31
2.83
2.91
2.83
5.37
3.37
3.81
0.24
Source

Literature-based
estimate


Literature-based
estimate
Literature-based
estimate



Literature-based
estimate
Literature-based
estimate



EPA econometric
estimate
EPA econometric
estimate
EPA econometric
estimate
EPA econometric
estimate
EPA econometric
estimate
EPA econometric
estimate
EPA econometric
estimate
EPA econometric
estimate
Literature based
estimate
Method

Production-weighted
average of individual
crop estimates ranging
from 0.27 to 0.55.
(Lin etal, 2000)
Based on Topel and
Rosen, (1988).a
Literature estimates are
not available so assumed
same value as for
Construction market

Method based on Ivaldi
and McCollough (2001)
Literature estimates not
available so assumed
same value as for
locomotive market

Cobb-Douglas
production function
Cobb-Douglas
production function
Cobb-Douglas
production function
Cobb-Douglas
production function

Cobb-Douglas
production function
Cobb-Douglas
production function
Cobb-Douglas
production function
Based on Considine
(20021b
Input Data Summary

Agricultural Census data
1991 - 1995


Census data, 1963 - 1983
Not applicable



Association of American
Railroads 1978-1997
Not applicable



Census data 1958-1996; SIC
3523
Census data 1958-1996; SIC
3531
Census data 1958-1996; SIC
3561 and 3563
Census data 1958-1996; SIC
3548
Assumed same as
pumps/compressors
Census data 1958-1996; SIC
3537
Census data 1958-1996; SIC
3524
Census data 1958-1996; SIC
3519
From Energy Intelligence
Grourj (EIG): 1987-2000°
Most other studies estimate ranges that encompass 1.0, including DiPasquale (1997) and DiPasquale and Wheaton (1994).
Other estimates range from 0.02 to 1.0 (Greene and Tishchishyna, 2000). However, Considine (2002) is one of the few studies that estimates a
supply elasticity for refinery operations.  Most petroleum supply elasticities also include extraction.
This source refers to the data used by Considine in his 2002 study.
                                                     10-173

-------
Final Regulatory Impact Analysis
10G.1 Application Markets - Demand Elasticities

   There are three application markets in the NDEIM: agricultural, construction, and
manufacturing.  Demand elasticities for the construction and manufacturing application markets
were estimated using a simultaneous equation (two-stage least squares) method. This approach
was also investigated for the agricultural application market; however, the estimated demand
elasticity parameter for that market was not statistically significant. For this reason, a production
function approach (Morgenstern, Pizer and Shih, 2002) was employed for the agricultural
application market. Publicly available data developed by Dale Jorgenson and his associates
(Jorgenson, 1990; Jorgenson, Gollop, and Fraumeni, 1987) were used in the regression analysis.
A time series of 38 observations, from 1958 to 1995, was used to estimate the demand
elasticities in both the two-stage least squares and production function approach. Both of these
techniques are described below.

10G.1.1 Construction and Manufacturing Demand Elasticities

   10G.1.1.1 Description of Simultaneous Equation Method

   The demand elasticities for the construction and manufacturing application markets were
estimated using a simultaneous equation (two-stage least squares) approach.  The methodology
is described below and the individual regression results are presented in Appendix 10F.

   In a partial equilibrium model, supply and demand are represented by a series of
simultaneous interdependent equations, in which the price and quantity produced of a product
are simultaneously determined by the interaction of producers and consumers in the market. In
simultaneous equations models, where one variable feeds back in to the other equations, the error
terms are correlated with the endogenous variable. As a result, estimating parameter values
using the ordinary least squares (OLS) regression method for each individual equation yields
biased and inconsistent parameter estimates. Therefore, OLS is not an appropriate estimation
technique.
   Instead, a simultaneous equations approach is used.  In the simultaneous equations approach
both the supply and demand equations for the market are specified and parameters for the two-
equation system are estimated simultaneously.

   The log-log version of the model is specified as follows:

          Supply: Qts = a0 + a^, + a2PLt + a3PKt + a4PMt + et              (1OG. 1 a)

          Demand:  Qtd= b0 + bjPt + b2HHt + b3It + v,                      (lOG.lb)

where
   Qt = log of quantity of the market product in year t
   Pt  = log of price of the market product in year t
   PL,   = log of cost of labor inputs in year t
                                        10-174

-------
                                                      Economic Impact Analysis
   PKt   = log of cost of capital inputs in production in year t
   PMt   = log of cost of material inputs in production in year t
   HHt   = log of number of households in year t
   I,  = average income per household in year t
   et, vt = error terms in year t

                             /*.
The parameter estimates ax and bl are the estimated price elasticity of supply and price elasticity
of demand, respectively.

   The first equation defines quantity supplied in each year as a function of the product price
and the cost of inputs: labor, capital and materials.  The second equation defines the quantity
demanded in each year as a function of the production price, the number of households, and the
average income per household. The equilibrium condition is that supply equals demand

                                  equilibrium: Qts  = Qtd

   Application of this two-stage least square regression approach was successful for estimating
the demand elasticity parameters for use here but was unsuccessful for estimating the supply
elasticities. The supply elasticity estimates were negative and not statistically significant.
Therefore, as noted above, literature estimates were used for the supply elasticities for the three
application markets in the NDEIM.

   To estimate the demand elasticities using this two-stage least squares approach, it is
necessary to first estimate the reduced-form equation for price using OLS. The reduced-form
equation expresses price as a function of all exogenous variables in the system:
                              P, = fn(PL, , PKj , PMt , HHt , It)

The results of this regression are used to develop fitted values of the dependent price variable Pt
(this is a new instrumental variable for price).  The fitted values by construction will be
independent of error terms in the demand equation.  In the second stage regression, the fitted
price variable Pt (the instrumental variable) is used as a replacement for Pt, in the demand
equation.  An OLS is performed on this equation, which produces a consistent, unbiased estimate
of the demand elasticity bj.

   10G.1.1.2 Construction Application Market Demand Elasticity

   The results of the simultaneous equation method for the construction demand elasticity are
presented in Table 10G-3. The estimated demand elasticity is -0.96 and is statistically
significant with a t-statistic of -3.83.  This inelastic estimate implies that a 1 percent increase in
price will lead to a 0.96 percent decrease in demand for construction, and means that the quantity
of goods and services demanded is expected to be fairly insensitive to price changes.

                       Table 10G-3. Construction Demand Elasticity
                                         10-175

-------
Final Regulatory Impact Analysis
 Number of Observations =   29
 R squared =   0.78
 Adjusted R squared = 0.75
 Variable                        Estimated Coefficients              t-statistic
intercept
In price
In number of households
In average income per
18.83
-0.96
-1.73
-1.67
5.19
-3.83
-3.37
5.34
 household
   10G.1.1.3 Manufacturing Application Market Demand Elasticity

   The results of the simultaneous equation method for the manufacturing market are presented
in Table 10G-4. The estimated demand elasticity is -0.58 and is statistically significant with a t-
statistic of-2.24. This inelastic estimate implies that a 1 percent increase in price will lead to a
0.58 percent decrease in the demand for manufactured products, and means that the quantity of
goods and services demanded is expected to be fairly insensitive to price changes.

                      Table 10G-4.  Manufacturing Demand Elasticity
 Number of Observations =   29
 R squared =   0.83
 Adjusted R squared =0.81
Variable
intercept
In price
In number of households
In average income per
household
Estimated Coefficients
6.16
-0.58
0.19
0.62
t-statistic
0.84
-2.24
0.23
1.49
10G.1.2 Agricultural Application Market Demand Elasticity

   10G.1.2.1: Description of Productivity Shift Approach

   When the simultaneous equation method was attempted for the agricultural application
market, the resulting demand elasticity parameter estimate was not statistically significant.
Thus, the demand elasticity for the agricultural market was estimated using the productivity shift
approach.  This is a technique that regresses historical data for aggregate output on industry
productivity (Morgenstern, Pizer, and Shih, 2002).

                                        10-176

-------
                                                     Economic Impact Analysis
   As shown in Figure 10G-1, changes in industry productivity represent shifts in the supply
curve. The supply curve shifts in conjunction with the known output values trace-out the
demand curve and enables the estimation of the demand elasticity. Because the agricultural
sector is relatively small compared to the entire economy, it is reasonable to assume that the
productivity changes do not shift the demand curve through income effects.

                                       Figure 10G-1
                         Productivity Shifts Trace-Out Demand Curve
                           \
                            \
                              \
                                                                Q

   The demand elasticity (£d) is estimated through a simple regression of the annual change in
the natural log of outputs on change in the natural log of productivity:

                              A In output,  = £d A In prod, + e,

where
   output, = output t is the industry output in year t
   prod, = industry productivity in year t
   e, = random error term

   The change in the natural log of productivity is computed as the log difference between the
annual change in input price and the annual change in output price:
           Aln prod, = £sh K
                                                                    (10.G-2)
where
   P   =
   PO
   i)   =
          input prices
          =  output prices
          input shares
                                        10-177

-------
Final Regulatory Impact Analysis
Eq. (10G.2) is a similar to a standard quantity-based definition of productivity (output divided by
input), but expressed in terms of input and output prices.  Under a competitive market with zero-
profit assumptions, revenue equals cost, and the price of output must equal the price of input
divided by the standard definition of productivity:
Thus,

                                    PI/PO=QO/QI

where
       Qo = quantity of output
       Qj = quantity of input

Since Q0/ Q: is a quantity based productivity, Pj / P0 is an equivalent measure of productivity
according to the above equation. The difference in logged changes in Pj and P0 is a valid
measure of productivity growth (Pizer, 2002).

10G.1.2.2 Agricultural Application Market Demand Elasticity

   The results of the estimated agricultural model are presented in Table 10G-5. The demand
elasticity estimate is -0.20 and is statistically significant with a t-statistic of 2.31.  This implies
that a 1 percent increase in price will lead to  a 0.2 percent decrease in demand, and means that
the quantity of goods and services demanded is expected to be fairly insensitive to price changes.

                       Table 10G-5.  Agricultural Demand Elasticity
 Number of Observations =   38
 R squared =  0.13
 Adjusted R squared =0.11
Variable
intercept
In productivity t
Estimated Coefficients
0.02
-0.20
t-statistic
3.49
2.31
10G.2 Application Market - Supply Elasticities

   Professional literature sources were used to obtain supply elasticity estimates for the
applications markets. These literature sources used are described below.

   It should be noted that both of the econometric estimation methods described above, the
simultaneous equation approach and the production function approach, were also attempted for
the supply elasticities.  However, because of the great variety of the production processes in

                                        10-178

-------
                                                       Economic Impact Analysis
these aggregate industry sectors (heterogeneity), parameter estimates were either not statistically
significant or did not conform with standard microeconomic theory (i.e., estimates were not
upward sloping).

10G.2.1 Agricultural Application Market Supply Elasticity

   Obtaining reasonable estimates of supply response in agriculture has been a persistent
problem since the inception of farm price support programs in the 1930s. The nonrecourse
marketing loans, deficiency payments, and conservation set-asides that make up the current farm
price support system distort equilibrium prices to the point that any econometric estimates are
difficult to formulate or support.

   A recent study by economists at the USDA's Economic Research Service provides an
approach to estimating agricultural demand elasticities (Lin et al., 2000). Taking into account
recent changes in the 1996 Farm Bill, the authors measure nationwide acreage price elasticity
values for the seven major agricultural crops, obtaining values ranging from 0.269 for soybeans
to 0.550 for sorghum. Although a composite number for all farm output is not reported, an
average value of 0.32 can be obtained by weighting the reported values by the acreage planted
for each crop. This value was used for the supply elasticity in the agriculture application market.
This estimated elasticity is inelastic, which means that the quantity of goods and services
supplied is expected to be fairly insensitive to price changes.

   Although the literature estimates vary, this estimate conforms closely to historical evidence
and economic theory of small but positive supply elasticities. This determination of price having
little impact on supply (referred to as inelastic supply) is consistent with a historical observation
that total acreage cultivated varies little from year to year. Between 1986 and 2001, for instance,
U.S. cropland harvested has ranged from 289 to 318 million acres, with an average of 305
million acres  over that 15-year period. A low supply elasticity is also supported by the fact that
there are few alternative uses (except in the very long run) for cropland,  capital, and labor
employed in farming. Abandonment or redeployment of farm assets is an often irreversible
decision, and one not greatly affected by annual price swings.

10G.2.2 Construction Application Market Supply Elasticity

   Although the construction market does not suffer from government-induced distortions to
prices and quantities, the evidence on supply elasticity is even more varied than that for
agriculture. Estimates of supply elasticity ranging from near zero to infinity have been reported
in credible papers on housing construction published during the past 20 to 30 years. A literature
survey paper  by DiPasquale (1997) describes the methodological issues that have led to this
variety of responses.  A key issue is the conceptual problem of distinguishing between increases
in the stock the of housing (or other structures) through new construction and changes in the
flow of housing services, which can also include renovation, apartment or condominium
conversion, and abandonment.

   DiPasquale cites a number of published studies that suggest that a value of 1.0 for supply
elasticity is appropriate. In the study that most closely matches the analysis for this regulation,

                                          10-179

-------
Final Regulatory Impact Analysis
Poterba (1984) estimated elasticity of new construction with respect to real house prices ranging
from 0.5 to 2.3, depending on the specification. A study by Topel and Rosen investigating
asset-markets and also found a short-run elasticity value of 1.0 (Topel and Rosen, 1988).
Finally, DiPasquale cites one of her own papers that estimated values of 1.0 to 1.2 for the price
elasticity of construction (DiPasquale and Wheaton, 1994). Based on these studies, a value of
1.0 was used for the supply elasticity in the construction application market.  This unit elastic
elasticity means that the quantity supplied is expected to vary directly with changes in prices.

   Estimates of supply response for other portions of the construction market, namely
nonresidential buildings and nonbuilding (roads and bridges, water and  sewer systems, etc.), are
not available in the literature.  However, the similarity between technologies employed in
construction of residential and other nonindustrial buildings suggests that supply elasticities
should be comparable.  In addition, residential construction accounts for a significant portion of
construction activity. According to the Census Bureau's most recent Annual  Value of
Construction Put in Place report, residential and nonindustrial buildings accounted for about 77
percent of the $842 billion in construction spending in 2001, with new residential housing
making up about 33 percent (U.S. Census Bureau, 2002).

10G.2.3 Manufacturing Application Market Supply Elasticity

   No supply elasticity estimates were available in the professional literature for the aggregate
manufacturing sector. For this reason, a unitary supply elasticity of 1.0 was used in the model.
This unit elastic elasticity means that the quantity supplied is expected to vary directly with
changes in prices. A sensitivity analysis for this assumed elasticity is presented in Appendix 101.

10G.3 Engine and  Equipment  Markets Supply Elasticity

   Published sources for the price elasticity of supply for diesel engine and diesel equipment
markets were not available. Therefore, the  supply elasticities used in the model were estimated
econometrically using a production function cost minimization approach.

10G.3.1 Production Function Cost Minimization Approach

   The production function cost minimization approach for econometrically estimating  the
supply elasticities is based on the cost-minimizing behavior of the firm subject to production
function constraints.  The production function describes the relationship between output and
inputs.  For this analysis, a Cobb-Douglas, or multiplicative form, was used as the functional
form of the production function:

                    Qt = A ktak LtaL Mtak tA                     (10G-3)

where
   Qt = output in year t
                                         10-180

-------
                                                      Economic Impact Analysis
   Kt  = real capital consumed in production in year tv
   Lt  = quantify of labor used in year t
   Mt = material inputs in year t
   t   = a time trend variable to reflect technology changes

This equation can be written in linear form by taking the natural logarithms of each side of the
equation. The parameters of this model, OCK, OCL, OCM, can then be estimated using linear regression
techniques:

                     In Qt = In A + ak In kt + a In Lt + am In Mt + A In t.

Under the assumptions of a competitive market and perfect competition, the elasticity of supply
with respect to the price  of the final product can be expressed in terms of the parameters of the
production function:

               Supply Elasticity = (c^ + aj I  (I- a{ - aj             (10G-4)

   This underlying relationship is derived from the technical production function and the
behavioral profit maximization conditions. The derivation for equation (10G-4) is provided in
Appendix 10H.

   In a competitive market, a firm will supply output as long as the marginal cost (MC) of
producing the next unit does not exceed the marginal revenue (MR, i.e., the price).  In a short-
run analysis, where capital stock is assumed to be fixed (or a sunk cost of production), the firm
will adjust its variable inputs of labor and material to minimize the total cost of producing a
given level  of output.

   The supply function is estimated by minimization, subject to the technical constraints of the
production function, and then setting the MC = P to determine the quantity produced as a
function of market price.  To maintain the desired properties of the Cobb-Douglas production
function, it is necessary to place restrictions on the estimated coefficients. For example, if OCL +
OCM = 1, then the supply elasticity will be undefined.  Alternatively, if OCL + OCM > 1, this yields a
negative supply elasticity.  Thus, a common assumption is that OCK + OCL + OCM = 1.  This implies
constant returns to scale, which is consistent with most empirical studies.

10G.3.2 Data for Estimating Engine and Equipment Supply Elasticities

   The data for the supply elasticity estimation were obtained from the National  Bureau of
Economic Research-Center for Economic Studies (NBER-CES). All nominal values were
deflated into $1987, using the appropriate price index. The following variables were used:

   •   value of shipments
^Capital consumed is defined as the value added minus labor expenditures, divided by the price
   index for capital.
                                         10-181

-------
Final Regulatory Impact Analysis
   •   price index of value shipments
   •   production worker wages
   •   implicit GDP deflators
   •   cost of materials
   •   price index for materials
   •   real capital stock
   •   investment
   •   price index for investment
   •   value added
   •   price index for capital

The capital (k) variable used in the Cobb-Douglas regression analysis is calculated as:

                 K = (Value Added - Labor Costs) / Price Index for Capital

This provides a measure of capital consumed as opposed to using a measure of total capital stock
in place at the firm.

10G.3.3 Engine Supply Elasticity Regression Results

   The results of the estimated production function is presented in Table 10G-6. All parameter
estimates are statistically significant at the 95 percent confidence level and the supply elasticity
is calculated to be 3.81.  This elastic elasticity estimate means that the quantities supplied in this
market are expected to be very responsive to price changes.

                          Table 10G-6. Engine Supply Elasticity
  Supply Elasticity = 3.81
 Number of Observations = 33
 R-squared = 0.9978
 Goldfeld-Quandt F = 1.88
 Note:  F(14,14)  = 2.46.
Variable
Intercept
InK
InT
InM
InL
Estimated Coefficients
0.954
0.2081
0.0215
0.5909
0.201
t-statistic
24.76
4.77
2.37
13.4
5.55
10G.3.4 Equipment Supply Elasticity Regression Results

   The results of the estimated production functions are presented in Tables 10G-7 through
10G-12.  The supply elasticities are calculated from the estimated coefficients for InM and InL as

                                         10-182

-------
                                                      Economic Impact Analysis
described in Equation G10-4.  The supply elasticities range from approximately 1.0 for
refrigeration to 5.4 for general industrial equipment. The average supply elasticity is 3.6. These
elastic elasticity estimates means that the quantities supplied in this market are expected to be
responsive to price changes.
                        Table 10G-7.  Agricultural Supply Elasticity
 Supply Elasticity = 2.14
 Number of Observations = 33
 R-squared = 0.9969
 Goldfeld-QuandtF = 2.01
 Note:  F(14,14) = 2.46
Variable
Intercept
InK
InT
InM
InL
Estimated Coefficients
1.1289
0.3189
-0.0241
0.4952
0.1858
t-statistic
20.81
11.12
-3.10
10.29
4.64
                       Table 10G-8.  Construction Supply Elasticity
 Supply Elasticity = 3.31
 Number of Observations = 33
 R-squared = 0.9926
 Goldfeld-Quandt F = 1.76
 Note:  F(14,14) = 2.46
 Variable
Estimated Coefficients
t-statistic
 Intercept
 InK
 InT
 InM
 InL
        1.172
        0.2318
       -0.0617
        0.1511
        0.6172
  28.54
   5.83
  -7.08
   4.54
  13.97
                                         10-183

-------
Final Regulatory Impact Analysis
                        Table 10G-9.  Industrial Supply Elasticity
 Supply Elasticity = 5.37
 Number of Observations = 33
 R-squared = 0.9949
 Goldfeld-Quandt F = 1.23
 Note:   F(14,14) = 2.46
Variable
Intercept
InK
InT
InM
InL
Estimated Coefficients
0.6927
0.157
-0.00739
0.0412
0.8018
t-statistic
18.29
3.47
-0.76
0.96
21.9
                                Table 10G-10.  Garden
 Supply Elasticity = 3.37
 Number of Observations = 33
 R-squared = 0.9963
 Goldfeld-Quandt F = 1.18
 Note:   F(14,14) = 2.46
Variable
Intercept
InK
InT
InM
InL
Estimated Coefficients
0.6574
0.2287
0.0413
0.0644
0.7069
t-statistic
13.34
3.75
2.78
1.72
11.23
                                       10-184

-------
                                                     Economic Impact Analysis
                                 Table 10G-11.  Gensets
 Supply Elasticity = 2.91
 Number of Observations = 33
 R-squared = 0.9909
 Goldfeld-Quandt F = 1.16
 Note:  F(14,14) = 2.46
 Variable
Estimated Coefficients
t-statistic
 Intercept
 InK
 InT
 InM
 InL
        1.1304
        0.2557
        0.0325
        0.3797
        0.3646
  11.09
   3.6
  2.73
  4.67
  4.51
                                 Table 10G-12. Pumps
 Supply Elasticity = 2.83
 Number of Observations = 33
 R-squared = 0.9979
 Goldfeld-Quandt F = 1.40
 Note:  F(14,14) = 2.46
Variable
Intercept
InK
InT
InM
InL
Estimated Coefficients
0.9367
0.2608
-0.207
0.0891
0.6501
t-statistic
19.01
4.45
-1.74
1.57
14.48
10G.4 Diesel Fuel Supply Elasticity:  Literature Estimate
   Very few studies have attempted to quantify supply responsiveness for individual refined
products, such as diesel fuel.  For example, a study for the California Energy Commission stated
"There do not seem to be credible estimates of gasoline supply elasticity" (Finizza, 2002).
However, sources agree that refineries have little or no ability to change output in response to
price: high fixed costs compel them to operate as close to their capacity limit as possible. The
Federal Trade Commission (FTC) analysis made this point explicitly (FTC, 2001).
   Greene and Tishchishyna (2000) reviewed supply elasticity estimates available in the
literature. The supply elasticity values cited in most of these studies were for "petroleum" or
"oil" production in the United States, which includes exploration, distribution and refining
                                        10-185

-------
Final Regulatory Impact Analysis
activities. The lowest short-term numbers cited were 0.02 to 0.05, with long-run values ranging
from 0.4 to 1.0. It seems likely that these extremely low numbers are influenced by the limited
domestic supply of crude petroleum and the difficulty of extraction.

   A recent paper by Considine (2002) provides one of the few supply elasticity estimates for
refining production (excluding extraction and distribution) based on historical price and quantity
data.  In this study, Considine estimates a refining production supply elasticity of 0.24.  This
estimate is for aggregate refinery production and includes distillate and nondistillate fuels.
Because petroleum products are made in  strict proportion and refineries have limited ability to
adjust output mix in the short to medium  run, it is reasonable to assume that supply is relatively
inelastic and similar across refinery products.  This value of 0.24 was used for the supply
elasticity for this market. This estimated elasticity is inelastic, which means that the quantity of
goods and services supplied is expected to be fairly insensitive to price changes.
10G.4 Locomotive and Marine Supply Elasticities:  Literature Estimate

   Over the past three decades, several studies have empirically estimated railroad cost
functions (see for example Braeutigam, 1999). One of the most recent studies by Ivald and
McCullough (2001) estimated a multi-product cost function for railroad services using data from
the Association of American Railroads (1978 to 1997). They report cost elasticities for which
we can derive a supply elasticity parameter for rail transportation servicesw. The supply
parameters are slightly elastic (1.6), suggesting a one percent change in the market price of the
services would induce producers increase service supply more than one percent.

   Similar studies for marine transportation services are generally restricted to the study of the
liner shipping industry (see for example Klein and Kyle, 1997). However, these ocean carrier
services are not directly comparable to commercial marine services in the Great Lakes and
Inland River Ports in the United States. Instead, they are more likely to be consistent with on-
land transportation services provided by the railroad sector. As a result, we have assumed the
supply elasticity parameter for best characterizes the supply responses of the marine
transportation market included in NDEIM.
wUnder the assumption of perfect competition, supply elasticities can be derived by taking the
   inverse of the reported cost elasticities. Therefore, Invalid and McCullough's cost elasticity of
   0.6 is used to compute a supply elasticity of 1/0.6 = 1.6.
                                         10-186

-------
                                                  Economic Impact Analysis
              APPENDIX 10H: Derivation of Supply Elasticity

   This appendix derives the underlying relationship for the supply elasticity used in the
production function approach described in Appendix 10G.

Cobb-Douglas:
   Q = Lakla   where  Q = output
                      L = labor input
                      k = capital input
Cost Minimization:
   Marginal Revenue Product of Labor = Wage Rate
   MRPL = P • MPL = w
   MPL=   3Q=aLa-lkl-a
           dL
           w
         Pak
    i-«_ Pak1
            l-a
   L =
       I  w

Substitute Back into Cobb-Douglas:
   y
                  l-a
   y=    *«k=«k
       _  a ,  _.   a  ,  cn\  . ,
       ~     lnP+ - In — +lnk
         l-a     l-a  vwy
    31ny    a
    31nP   l-a
              = Supply Elasticity
                                      10-187

-------
Final Regulatory Impact Analysis
                       APPENDIX 101: Sensitivity Analysis
   The Economic Impact Analysis presented in this Chapter 10 is based on the Nonroad Diesel
Economic Impact Model (NDEIM) developed for this analysis.  The NDEIM reflects certain
assumptions about behavioral responses (modeled by supply and demand elasticities) and how
costs are treated by producers.  This appendix presents a sensitivity analysis for several model
components by varying how they are treated. Five model components are examined:

   •      Scenario 1: alternative market supply and demand elasticity parameters
   •      Scenario 2: alternative ways to treat fuel market costs
   •      Scenario 3: alternative way to treat operating costs
   •      Scenario 4: alternatives way to treat engine and equipment fixed costs
   •      Scenario 5: alternative discount rates

   The results of these sensitivity analyses are presented below. All of the results are presented
for 2013 only. The results for the application and transportation service markets do not include
the operating savings. Instead, operating savings are added into the total social costs as a
separate item.

   In general, varying the model parameters does not significantly change the results of the
economic impact assessment analysis presented above. Total social costs are about the same
across all sensitivity analysis scenarios, $1,510 million.  In addition, varying these model
parameters does not significantly affect the way the social costs are borne.  In all cases, the
application markets bear the majority of the burden (about 83 percent), although there are small
differences in the way the costs are  borne among the scenarios. The exception is Scenario 2, the
fuel cost scenario. In the maximum total cost scenario, the share of the social costs borne by the
application market exceeds the social costs of the rule ($2,029 million versus $1,510.9 million
for the rule), indicating that refiners will gain from the rule (about  $526 million).  In the
maximum variable cost scenario, the share of the social costs borne by the application market
also exceeds the social costs of the rule ($1,584 million versus $1,510.9 million for the rule),
indicating that refiners would gain from the rule in this scenario as well (about $79 million).
There are also differences in the way the application market costs are shared among producers
and consumers in  that market, especially for Scenario 1.

   With regard to the market analysis, expected percentage changes for price and price and
quantity for each market are about the same as in the base case.  Prices are expected to increase
about 2.14. 2.9, and  6 percent for the engine, equipment, and fuel markets respectively, while
quantities. These  engine and equipment percentage price increases are stable across scenarios
except in Scenario 4, in which engine and equipment fixed costs and included in the model. In
this case, the  expected engine price increase goes up from about 21.4 percent to 23.0 percent and
the expected equipment price increase goes up from about 2.9 percent to 3.4 percent.  The fuel
percentage price increases are also stable across scenarios, with the exception of Scenario 2, in
which a price increase of 11 percent is expected in the maximum total cost scenario and a 7
                                         10-188

-------
                                                      Economic Impact Analysis
percent increase is expected in the maximum variable cost scenario.

   Percentage decreases in the quantities produced in the markets are also relatively stable
across the scenarios with decreases of 0.01, 0.02, and 0.02 percent expected for the engine,
equipment, and fuel markets respectively.  There is some variation in absolute quantities across
the scenarios, but these are negligible when compared to the total output of each market. The
largest change in absolute quantity of output is associated with Scenario 1, when supply
elasticities are varied.  The largest decline is 107 engines, 189 equipment units, and 3.25 million
gallons of fuel; the smallest is 44 engines, 74 equipment units, and 1.29 million gallons of fuel.
This is in comparison to 79 engines,  139 equipment units, and 2.38 million gallons of fuel in the
base case.

   For the application market, the expected price increase remains stable across the scenarios at
about 0.1 percent, and the expected quantity decrease at about 0.02. Prices in the transportation
service markets are expected to increase about 0.0.01 percent and quantity to decrease about 0.01
percent.

101.1 Model Elasticity Parameters

   Key model parameters include supply and demand elasticity estimates used by the model to
characterize behavioral responses of producers and consumers in each market.

   Consumer demand and producer supply responsiveness to changes in the commodity prices
are referred to by economists as "elasticity." The measure is typically expressed as the
percentage change in quantity (demanded or supplied) brought about by a percent change in own
price. A detailed discussion regarding the estimation and selection of the elasticities used in the
NDEIM are discussed in Appendix 10G. This component of the sensitivity analysis examines
the impact of changes in selected elasticity values, holding other parameters constant.  The goal
is to determine whether alternative elasticity values significantly alter conclusions in this report.

101.1.1 Application Markets (Supply and Demand Elasticity Parameters)

   The choice of supply and demand elasticities for the application market is important because
changes in quantities in the application markets are the key drivers in the derived demand
functions used to link impacts in the  engine, equipment, and fuel markets. In addition, the
distribution of regulatory costs depends on the relative supply and demand elasticities used in
the analysis. For example, consumers will bear less  of the regulatory burden if they are more
responsive to price changes than producers.

   Table  101-1 reports the upper- and lower-bound values of the application market elasticity
parameters (supply and demand) used in the sensitivity analysis. The variation in  estimates
reported in the literature were used for supply elasticity ranges. For the manufacturing market,
an assumed elasticity of 1.0 was used. For the purpose of this sensitivity analysis, the  same
upper and lower bounds were  used as for the construction market. For demand elasticity values,
a 90 percent confidence interval was computed using the coefficient and standard error values
                                         10-189

-------
Final Regulatory Impact Analysis
reported in the econometric analysis (see Appendix 10G).

           Table 101-1.  Sensitivity Analysis of the Supply and Demand Elasticities
                               for the Application Markets
Parameter/Market
Supply elasticity
Agriculture
Construction

Manufacturing
Demand elasticity
Agriculture
Construction
Manufacturing
Elasticity
Source

Literature
estimate
Literature
estimate
Assumed value

EPA estimate
EPA estimate
EPA estimate
Upper Bound

0.55
2.3

2.3

-0.35
-1.39
-1.02
Base Case

0.32
1

1

-0.20
-0.96
-0.58
Lower Bound

0.027
0.5

0.5

-0.054
-0.534
-0.140
Note:     For literature estimates, the variations in estimates reported were used to develop
          elasticity ranges. In contrast, EPA computed upper- and lower-bound estimates using
          the coefficient and standard error values associated with its econometric analysis and
          reflect a 90 percent confidence interval.
   The results of the NDEIM using these alternative elasticity values for the application markets
are reported in Tables 101-2 and 101-3. As can be seen in those tables, market prices are stable
across the upper- and lower-bound sensitivity scenarios.  Absolute quantities vary but the
percentage changes in output are negligible for the two scenarios.

   The change in total social surplus for 2013 also remains nearly unchanged across all
scenarios and is approximately the same as for the rule ($1,510 million). However, consumers in
the application market bear a smaller share of the social costs when they are more responsive to
price changes relative to producers (supply lower bound and demand upper bound scenarios).
As shown, consumers bear approximately 34.5 and 46.5 percent, respectively, in these scenarios
compared to 58.5 percent in the base case. In contrast, they bear a higher share (up to 78.5
percent) when they are less responsive to price changes relative to producers (supply upper
bound and demand lower bound scenarios). While the burden of the fuel market changes
slightly, it always remain below 1 percent of the social costs.
                                         10-190

-------
                                                      Economic Impact Analysis
        Table 101-2.  Application Market Sensitivity Analysis for Supply Elasticitiesa'b
Scenario
Application Markets
Price ($/q)
Quantity (q/yr)
Change in Consumer Surplus
($106/yr)
Change in Producer Surplus
($106/yr)
Change in Total Surplus
($106/yr)
Equipment Markets
Price ($/q)
Quantity (gal/yr)
Change in Producer Surplus
($106/yr)
Engine Markets
Price ($/q)
Quantity (gal/yr)
Change in Producer Surplus
($106/yr)
Fuel Markets
Price ($/q)
Quantity (q/yr)
Change in Producer Surplus
($106/yr)
Transportation Services
Price ($/q)
Quantity (q/yr)
Change in Producer Surplus
($106/yr)
Applications Not Included in
NDEIM ($106/yr)
Operating Savings ($106/yr)
Social Costs ($106/yr)
Base Case
Absolute

NA
NA
$876
$621
$1,497

$975
-139
$143

$821
-79
$42

$0.06
-2.38
$8

NA
NA
$2
$102.4
-$284.7
$1,510.0
Relative

0.10%
-0.02%
NA
NA
NA

2.9%
-0.02%
NA

21.4%
-0.01%
NA

6.0%
-0.02%
NA

0.01%
-0.01%
NA
NA
NA
NA
Supply Upper Bound
Absolute

NA
NA
$1,113
$377
$1,490

$973
-189
$145

$821
-107
$42

$0.06
-3.25
$12

NA
NA
$3
$102.4
-$284.7
$1,509.9
Relative

0.11%
-0.02%
NA
NA
NA

2.9%
-0.02%
NA

21.4%
-0.02%
NA

6.0%
-0.03%
NA

0.01%
-0.01%
NA
NA
NA
NA
Supply Lower Bound
Absolute

NA
NA
$520
$985
$1,505

$977
-74
$141

$821
-44
$42

$0.06
-1.29
$3

NA
NA
$2
$102.4
-$284.7
$1,510.1
Relative

0.05%
-0.01%
NA
NA
NA

2.9%
-0.01%
NA

21.4%
-0.01%
NA

6.0%
-0.01%
NA

0.01%
-0.01%
NA
NA
NA
NA
a  Sensitivity analysis is presented for 2013.
b  Figures are in 2002 dollars.
                                         10-191

-------
        Table 101-3. Application Market Sensitivity Analysis for Demand Elasticities"'11
Scenario
Application Markets
Price ($/q)
Quantity (q/yr)
Change in Consumer Surplus
($106/yr)
Change in Producer Surplus
($106/yr)
Change in Total Surplus
($106/yr)
Equipment Markets
Price ($/q)
Quantity (gal/yr)
Change in Producer Surplus
($106/yr)
Engine Markets
Price ($/q)
Quantity (gal/yr)
Change in Producer Surplus
($106/yr)
Fuel Markets
Price ($/q)
Quantity (q/yr)
Change in Producer Surplus
($106/yr)
Transportation Services
Price ($/q)
Quantity (q/yr)
Change in Producer Surplus
($106/yr)
Applications Not Included in
NDEIM ($106/yr)
Operating Savings ($106/yr)
Social Costs ($106/yr)
Base Case
Absolute

NA
NA
$876
$621
$1,497

$975
-139
$143

$821
-79
$42

$0.06
-2.38
$8

NA
NA
$2
$102.4
-$284.7
$1,510.0
Relative

0.10%
-0.02%
NA
NA
NA

2.9%
-0.02%
NA

21.4%
-0.01%
NA

6.0%
-0.02%
NA

0.01%
-0.01%
NA
NA
NA
NA
Demand
Absolute

NA
NA
$695
$798
$1,493

$974
-170
$144

$821
-96
$42

$0.06
-2.89
$10

NA
NA
$3
$102.4
-$284.7
$1,509.9
Upper Bound
Relative

0.08%
-0.02%
NA
NA
NA

2.9%
-0.02%
NA

21.4%
-0.02%
NA

6.0%
-0.02%
NA

0.01%
-0.01%
NA
NA
NA
NA
Demand
Absolute

NA
NA
$1,181
$323
$1,503

$977
-88
$142

$821
-50
$42

$0.06
-1.54
$4

NA
NA
$1
$102.4
-$284.7
$1,510.0
Lower Bound
Relative

0.12%
-0.01%
NA
NA
NA

2.9%
-0.01%
NA

21.4%
-0.01%
NA

6.0%
-0.01%
NA

0.01%
0.00%
NA
NA
NA
NA
a  Sensitivity analysis is presented for 2013.
b  Figures are in 2002 dollars.
                                           10-192

-------
                                                       Economic Impact Analysis
101.1.2  Equipment, Engine and Diesel Fuel Markets (Supply Elasticity Parameters)

    Sensitivity analysis was also conducted for the engine, equipment, and diesel fuel market
supply elasticities. The range of supply elasticity values evaluated for each market are provided
in Table 101-4.  The engine and equipment market supply elasticities are derived
econometrically.  Therefore, the upper and lower bound values were computed using the
coefficient and standard error values associated with the econometric analysis and reflect a 90
percent  confidence interval (see Appendix 10G).

    The fuel market supply elasticity was obtained from the literature. The value for the lower
bound for the sensitivity analysis is based on the range of available estimates. The value for the
upper bound was derived from a set of regulatory studies of the petroleum refining industry that
were conducted using a techno-economic method to estimate supply costs at the individual
refinery level (EPA, 2000; CRA/BOB, 2000; MathPro, 2002).  Synthetic industry supply curves
(i.e., marginal cost curves) were developed from these studies and yielded supply elasticities
ranging from 0.2 to 2.0. Therefore, the sensitivity analysis uses 2.0 as an upper bound for the
supply elasticity of nonroad diesel fuel.

    Three sets of sensitivity results are presented in Tables 101-5, 101-6, and 101-7, where supply
elasticities are changed in the equipment, engines, and fuel markets, respectively.

                                       Table 101-4
     Engine, Equipment, and Diesel Fuel Market Sensitivity Analysis for Supply Elasticity
                                       Parameters
Market
Supply
Engines
Equipment
Agriculture
Construction
Refrigeration
Industrial
Garden
Generator
Pumps
Diesel fuel
Elasticity Source

EPA Estimate

EPA Estimate
EPA Estimate
EPA Estimate
EPA Estimate
EPA Estimate
EPA Estimate
EPA Estimate
Literature Estimate
Upper
Bound

7.64

3.72
6.06
5.62
12.93
7.96
12.14
5.62
2
Base Case

3.81

2.14
3.31
2.83
5.37
3.37
2.91
2.83
0.2
Lower
Bound

2.33

1.31
2.09
1.62
2.9
1.82
1.12
1.62
0.04
Note:  For literature estimates, the variations in estimates reported were used to develop
       elasticity ranges. In contrast, EPA computed upper- and lower-bound estimates using the
       coefficient and standard error values associated with its econometric analysis and reflect
       a 90 percent confidence interval.

                                         10-193

-------
Final Regulatory Impact Analysis
   Tables 101-5 and 101-6 contain the results of varying the engine and equipment supply
elasticities. When these elasticities are allowed to vary, all quantitative estimates for both
market impacts (price and quantity changes) and social impacts (how the burden is shared across
markets) remain nearly unchanged when compared with the rule, across both the upper and
lower bound supply elasticity scenarios for equipment and engines. These results imply that the
results presented in Section 10.1 are not sensitive to the supply elasticity values used in the
engine and equipment markets, because the derived demand for engines and  equipment is highly
inelastic (it is a function of the inelastic demand and supply in the application markets), and so
almost all of the compliance costs are passed on to the application markets through price
increases.

   Table 101-7 contains the results of varying the fuel supply elasticity.  The results for the
upper bound is nearly identical to the base case. However, in the case of the lower bound
(producers are less sensitive to price changes), the expected percentage change in the price of
fuel decreases from 6 percent in the base case to 5.6 percent.  There is a reallocation of surplus
loss from the application markets to the fuel markets. In the base case, the application markets
are expected to bear about 83 percent of the social costs ($1,497 million), while the fuel market
is expected to bear about 0.5 percent ($8  million). When the lower bound of the supply elasticity
for the fuel market is used, the share of the application markets decreases  to 80 percent ($1,436
million) while the share of the fuel markets increases to about 4 percent ($70 million). The total
welfare  losses are stable, however, at $1,510.

   The demand elasticities for the equipment and engine diesel fuel markets are derived as part
of the model, and therefore sensitivity analysis was not conducted on  those parameters.x  In other
words, the change in the application market quantities determines the demand responsiveness in
the engine, equipment,  and diesel fuel markets. As a result, the demand sensitivity analysis for
these markets is indirectly shown in Table 101-2.  Nonroad diesel equipment and fuel
expenditures  are relatively small shares of total production costs for the application markets.
Therefore changes in these input prices do not significantly alter input demand (i.e., demand in
these markets is highly  inelastic).
xFor a discussion of the concept of derived demand, see Section 10.2.2.3 Incorporating
   Multimarket Interactions.

                                         10-194

-------
                                                     Economic Impact Analysis
           Table 101-5.  Equipment Market Supply Elasticity Sensitivity Analysis
                                                                         a,b
Scenario
Application Markets
Price ($/q)
Quantity (q/yr)
Change in Consumer Surplus
($106/yr)
Change in Producer Surplus
($106/yr)
Change in Total Surplus
($106/yr)
Equipment Markets
Price ($/q)
Quantity (q/yr)
Change in Producer Surplus
($106/yr)
Engine Markets
Price ($/q)
Quantity (q/yr)
Change in Producer Surplus
($106/yr)
Fuel Markets
Price ($/q)
Quantity (q/yr)
Change in Producer Surplus
($106/yr)
Transportation Services
Price ($/q)
Quantity (q/yr)
Change in Producer Surplus
($106/yr)
Applications Not Included in
NDEIM ($106/yr)
Operating Savings ($106/yr)
Social Costs ($106/yr)
Base Case
Absolute

NA
NA
$876
$621
$1,497

$975
-139
$143

$821
-79
$42

$0.06
-2.38
$8

NA
NA
$2
$102.4
-$284.7
$1,510.0
Relative

0.10%
-0.02%
NA
NA
NA

2.9%
-0.02%
NA

21.4%
-0.01%
NA

6.0%
-0.02%
NA

0.01%
-0.01%
NA
NA
NA
NA
Supply
Absolute

NA
NA
$877
$622
$1,499

$977
-139
$141

$821
-76
$42

$0.06
-2.39
$8

NA
NA
$2
$102.4
-$284.7
$1,510.0
Upper Bound
Relative

0.10%
-0.02%
NA
NA
NA

2.9%
-0.02%
NA

21.4%
-0.01%
NA

6.0%
-0.02%
NA

0.01%
-0.01%
NA
NA
NA
NA
Supply
Absolute

NA
NA
$874
$620
$1,494

$972
-139
$146

$821
-79
$42

$0.06
-2.38
$8

NA
NA
$2
$102.4
-$284.7
$1,510.0
Lower Bound
Relative

0.10%
-0.02%
NA
NA
NA

2.9%
-0.02%
NA

21.4%
-0.01%
NA

6.0%
-0.02%
NA

0.01%
-0.01%
NA
NA
NA
NA
a  Sensitivity analysis is presented for 2013.
b  Figures are in 2002 dollars.
                                        10-195

-------
Final Regulatory Impact Analysis
             Table 101-6. Engine Market Supply Elasticity Sensitivity Analysis'
                                                                         a,b
Scenario
Application Markets
Price ($/q)
Quantity (q/yr)
Change in Consumer Surplus
($106/yr)
Change in Producer Surplus
($106/yr)
Change in Total Surplus
($106/yr)
Equipment Markets
Price ($/q)
Quantity (q/yr)
Change in Producer Surplus
($106/yr)
Engine Markets
Price ($/q)
Quantity (q/yr)
Change in Producer Surplus
($106/yr)
Fuel Markets
Price ($/q)
Quantity (q/yr)
Change in Producer Surplus
($106/yr)
Transportation Services
Price ($/q)
Quantity (q/yr)
Change in Producer Surplus
($106/yr)
Applications Not Included in
NDEIM ($106/yr)
Operating Savings ($106/yr)
Social Costs ($106/yr)
Base Case
Absolute

NA
NA
$876
$621
$1,497

$975
-139
$143

$821
-79
$42

$0.06
-2.38
$8

NA
NA
$2
$102.4
-$284.7
$1,510.0
Relative

0.10%
-0.02%
NA
NA
NA

2.9%
-0.02%
NA

21.4%
-0.01%
NA

6.0%
-0.02%
NA

0.01%
-0.01%
NA
NA
NA
NA
Supply
Absolute

NA
NA
$876
$621
$1,497

$975
-139
$143

$821
-79
$42

$0.06
-2.38
$8

NA
NA
$2
$102.4
-$284.7
$1,510.0
Upper Bound
Relative

0.10%
-0.02%
NA
NA
NA

2.9%
-0.02%
NA

21.4%
-0.01%
NA

6.0%
-0.02%
NA

0.01%
-0.01%
NA
NA
NA
NA
Supply
Absolute

NA
NA
$876
$621
$1,497

$975
-139
$143

$821
-77
$42

$0.06
-2.38
$8

NA
NA
$2
$102.4
-$284.7
$1,510.0
Lower Bound
Relative

0.10%
-0.02%
NA
NA
NA

2.9%
-0.02%
NA

21.4%
-0.01%
NA

6.0%
-0.02%
NA

0.01%
-0.01%
NA
NA
NA
NA
a  Sensitivity analysis is presented for 2013.
b  Figures are in 2002 dollars.
                                         10-196

-------
                                                   Economic Impact Analysis
Table 101-7. Fuel
Scenario
Application Markets
Price ($/q)
Quantity (q/yr)
Change in Consumer Surplus
($106/yr)
Change in Producer Surplus
($106/yr)
Change in Total Surplus
($106/yr)
Equipment Markets
Price ($/q)
Quantity (q/yr)
Change in Producer Surplus
($106/yr)
Engine Markets
Price ($/q)
Quantity (q/yr)
Change in Producer Surplus
($106/yr)
Fuel Markets
Price ($/q)
Quantity (q/yr)
Change in Producer Surplus
($106/yr)
Transportation Services
Price ($/q)
Quantity (q/yr)
Change in Producer Surplus
($106/yr)
Applications Not Included in
NDEIM ($106/yr)
Operating Savings ($106/yr)
Social Costs ($106/yr)
Market Supply Elasticity
Base Case
Absolute

NA
NA
$876
$621
$1,497

$975
-139
$143

$821
-79
$42

$0.06
-2.38
$8

NA
NA
$2
$102.4
-$284.7
$1,510.0
Relative

0.10%
-0.02%
NA
NA
NA

2.9%
-0.02%
NA

21.4%
-0.01%
NA

6.0%
-0.02%
NA

0.01%
-0.01%
NA
NA
NA
NA
Sensitivity Analysisa'b
Supply Upper Bound
Absolute

NA
NA
$878
$623
$1,501

$975
-140
$143

$821
-78
$42

$0.06
-2.39
-$2

NA
NA
$2
$102.4
-$284.7
$1,510.6
Relative

0.10%
-0.02%
NA
NA
NA

2.9%
-0.02%
NA

21.4%
-0.01%
NA

6.0%
-0.02%
NA

0.01%
-0.01%
NA
NA
NA
NA
Supply
Absolute

NA
NA
$839
$597
$1,436

$975
-134
$143

$821
-75
$42

$0.05
-2.31
$70

NA
NA
$3
$102.4
-$284.7
$1,510.6

Lower Bound
Relative

0.09%
-0.01%
NA
NA
NA

2.9%
-0.02%
NA

21.4%
-0.01%
NA

5.6%
-0.02%
NA

0.01%
-0.01%
NA
NA
NA
NA
a  Sensitivity analysis is presented for 2013.
b  Figures are in 2002 dollars.
                                       10-197

-------
Final Regulatory Impact Analysis
10.1.2  Fuel Market Supply Shift Alternatives

    Section 10.2 discusses alternative approaches to shifting the supply curve in the market
model. Three alternatives for the fuel market supply shift are investigated in this sensitivity
analysis:

    •   Total average (variable + fixed) cost shift—the results presented in Section 10.1 and the
       appendices are generated using this cost shift.
    •   Total maximum (variable + fixed) cost shift
    •   Variable maximum cost shift

    To model the total and variable maximum cost scenarios, the high-cost producer is
represented by a separate supply curve as shown in Figure 101-1. The remainder of the market is
represented as a single aggregate supplier.  The high-cost producer's supply curve is then shifted
by Cmax (either total or variable), and the  aggregate supply curve is shifted by Cagg.  Using this
structure, the high-cost producer will determine price as long as

    •   the decrease in market quantity does not shut down the high-cost producer,  and
    •   the supply from aggregate producers is highly inelastic (i.e., remaining producers are
       operating close to capacity); thus, the aggregate producers cannot expand output in
       response to the price increase.

                                       Figure 101-1
                         High Cost Producer Drives Price Increases
                                        agg
                                             Q
                                               agg
          High Cost Supplier
Aggregate Remaining
     Suppliers
Fuel Market
   Note that the aggregate supply curve is no longer shifted by the average compliance costs but
slightly less than the average because the high-cost producer has been removed.  The adjusted
average aggregate cost shift (Cagg) is calculated from the following:

                   Cave*Q,0, = Cmax * Qmax + Cagg * Qagg            (101.2)

where Cave is the average control cost for the total population;  Qmax, Cmax, and Qagg, Cagg are the
                                         10-198

-------
                                                       Economic Impact Analysis
baseline output and cost shift for the maximum cost producer; and the baseline output and cost
shift for the remaining aggregate producers, respectively.

   The results of this sensitivity analysis are reported in Table 101-8.

                                       Table 101-8
                 Sensitivity Analysis to Cost Shifts in the Diesel Fuel Market

Scenario
Application Markets
Price ($/q)
Quantity (q/yr)
Change in Consumer Surplus ($106/yr)
Change in Producer Surplus ($106/yr)
Change in Total Surplus ($106/yr)
Equipment Markets
Price ($/q)
Quantity (q/yr)
Change in Producer Surplus ($106/yr)
Engine Markets
Price ($/q)
Quantity (q/yr)
Change in Producer Surplus ($106/yr)
Fuel Markets
Price ($/q)
Quantity (q/yr)
Change in Producer Surplus ($106/yr)
Transportation Services
Price ($/q)
Quantity (q/yr)
Change in Producer Surplus ($106/yr)
Applications Not Included in NDEIM
Operating Savings ($106/yr)
Social Costs ($106/yr)
Average Total Scenario
Absolute
Change

NA
NA
$876
$621
$1,497

$975
-139
$143

$821
-79
$42

$0.06
-2.38
$8

NA
NA
$2
$102.4
-$284.7
$1,510.0
Relative
Change (%)

0.10%
-0.02%
NA
NA
NA

2.9%
-0.02%
NA

21.4%
-0.01%
NA

6.0%
-0.02%
NA

0.01%
-0.01%
NA
NA
NA
NA
Maximum Total Scenario
Absolute
Change

NA
NA
$1,176
852
$2,029

$973
-177
$145

$821
-100
$42

$0.10
-3.02
-$526

NA
NA
$4
$102.4
-$284.7
$1,510.9
Relative
Change (%)

0.14%
-0.02%
NA
NA
NA

2.9%
-0.02%
NA

21.0%
-0.02%
NA

11.0%
-0.02%
NA

0.01%
-0.01%
NA
NA
NA
NA
Maximum Variable Scenario
Absolute
Change

NA
NA
$919
665
$1,584

$975
-138
$143

$821
-78
$42

$0.06
-2.36
-$79

NA
NA
$3
$102.4
-$284.7
$1,510.9
Relative
Change (%)

0.10%
-0.02%
NA
NA
NA

2.9%
-0.02%
NA

21.0%
-0.01%
NA

7.0%
-0.02%
NA

0.01%
-0.01%
NA
NA
NA
NA
a  Sensitivity analysis is presented for 2013.
b  Figures are in 2002 dollars.
   The total and variable maximum cost shift scenarios lead to different conclusions for two
important variables: the estimated market price increase for diesel fuel and the estimated welfare
impact for affected refineries. Under the base case (total average cost scenario), refiners pass
most of the average compliance costs on to the application markets, and the net decrease in
                                         10-199

-------
Final Regulatory Impact Analysis
producer surplus for refiners is relatively small ( about $7.8 million, or 0.6 percent of total social
costs), and prices are expected to increase about 6.0 percent. Note that these are industry
averages, and individual refiners will gain or lose because compliance costs vary across
individual refineries.

   In the total maximum cost scenario, the highest operating cost refinery determines the new
market price through the impacts on both fixed and variable costs. This refinery has the highest
per-unit supply shift, which leads to a higher price increase relative to the average cost scenario.
As a result, all refiners except the highest cost refiner are expected to benefit from the rule, by
about $526 million, because the change in market price exceeds the additional per-unit
compliance costs for most of the refineries (i.e., most refiners have costs less than the costs for
the highest operating cost refinery).  Consequently, in this scenario the producers and consumers
in the application market are expected to bear a larger share of the total cost of the program:
$2,029 million compared to $1,497 million,  out of total social costs of about $1,510 million for
the welfare costs of the rule without considering the operating savings.

   The variable maximum cost scenario is similar to the total maximum cost scenario because
the highest cost refinery determines the with-regulation market price.  However, the variable
maximum cost scenario leads to an expected price increase that is smaller than the total
maximum cost scenario because the refiner supply shift includes only variable compliance costs.
In other words, the refiners do not pass along any fixed costs; they absorb the fixed costs.
However, the refinery industry still experiences a small net surplus gain ($79 million) because
the change in market price (driven by the maximum variable cost) exceeds the additional
per-unit compliance costs for many of the refineries (i.e., many refiners still have total costs less
than the costs for the highest operating cost refinery in this scenario)/ The net surplus gain for
refiners is smaller than the total maximum scenario ($79 million  compared to $596 million)
because refiners absorb fixed costs, and the projected market price increase is smaller.  Again,
consumers and producers in the application markets are expected to bear a larger share of the
total cost of the program, about $1,584 million.

   The results of this sensitivity analysis suggest that the expected impacts on producers and
consumers in the application markets and on refiners is affected by how refinery costs are
modeled.  The NDEEVI models these costs based on the average (variable + fixed) cost scenario,
reflecting a competitive market situation in all regional markets.  However, if the highest cost
refinery drives the new market price, then prices are expected to  increase more, with a larger
contraction in output. In this case, consumers and producers in the application market are
expected to bear more than the cost of the rule. When the highest cost refinery's variable costs
drive the new market price, then prices will increase  slightly more that the base case (from 6
percent to 7 percent), producers and consumer will again bear more of the burden of the rule, and
refiners bear less than in the base case.
YAlso, see Table 7.6-1 and related text in Chapter 7 regarding the possible diesel fuel price
   increases for the maximum operating cost scenario

                                          10-200

-------
                                                       Economic Impact Analysis
101.3 Operating Cost Scenario

   In the base case analysis presented in Chapter 10, operating savings are not included in the
market analysis. As explained in Section 10.3.5.3, this approach is used because these operating
savings are not expected to affect consumer decisions with respect to new engines and
equipment. However, these operating savings accrue to society and so they are added to social
costs after changes in price and quantity are estimated.  In the analysis for 2013, $284.7 million
in operating savings are applied to the application markets; these savings are expected to accrue
to producers in these markets. Specifically, $265.5 million are applied to the social costs for the
three application markets and for the transportation services providers ($243.2 million and $22.3
million, respectively) and $19.2 million are applied to the social costs for those markets not
included  in NDEIM.2 The results of this base case analysis are set out in Table 10.1-4. In the
summary presented in Table 101-9, all of the operating  savings are presented as a separate item.

   In this sensitivity analysis, we modify the analysis to include operating savings in the market
analysis.  This scenario considers the possibility that some portion of the operating savings
realized by users of nonroad engines, equipment, and fuel can be transmitted to consumers
through the market relationships specified in the model, thereby affecting prices and output.  The
operating savings are modeled as a cost reduction (benefit) for producers in the application
markets and service providers in the locomotive and marine sectors.AA  Specifically, they are
treated as negative supply shift for the supply curves in these markets.  Treating operating
savings like this reduces the size of the supply shift and illustrates how operating savings may be
shared among producers and consumers in  these markets.

   The results  of this sensitivity analysis are included in Table 1-9.  In this scenario, the price
increase and quantity decrease in the application markets are expected to be smaller (0.08
percent compared to 0.10 percent for price, and -0.01 percent compared to -0.02 percent for
quantity). This is a direct result of the smaller supply shift.  Although the estimated total social
costs associated with the rule are comparable for both scenarios, $1,510.1 million compared to
$1.510.0  million in the base case, there are two important distributional consequences associated
with including operating savings in the market analysis. First,  almost all of the locomotive and
marine savings ($22 million) are now directly passed to the application markets in the form of
lower prices.  As a result, the application markets benefit from operating savings in
transportation services and they bear 80.6 percent of the total social costs instead of 83.4 percent
(the change in total  application market surplus decreases from $1,254 to $1,234 million).
Second, a portion of the operating savings is now distributed to consumers in application
markets.  In 2013, the change in consumer  surplus  in the application markets decreases from
$876 million to $709 million. The change in producer surplus is smaller, and decreases from
z See Section 10.3.5.3 for a description of how the operating savings are estimated.

AAWe only consider cost savings for market included in NDEIM (the three application markets
   and the transportation service markets). This amounts to $265 million, or 93 percent of the
   operating savings.  The remaining $19 million is added as a line item to the social costs for
   application markets not included in NDEIM.

                                         10-201

-------
Final Regulatory Impact Analysis
$621 to $525 million.
                                      Table 101-9
                    Operating Savings Included in the Market Analysisa'b
Scenario
Application Markets
Price ($/q)
Quantity (q/yr)
Change in Consumer Surplus
Change in Producer Surplus
Change in Total Surplus ($106/yr)
Equipment Markets
Price ($/q)
Quantity (q/yr)
Change in Producer Surplus
Engine Markets
Price ($/q)
Quantity (q/yr)
Change in Producer Surplus
Fuel Markets
Price ($/q)
Quantity (q/yr)
Change in Producer Surplus
Transportation Services
Price ($/q)
Quantity (q/yr)
Change in Producer Surplus
Applications Not Included in
Operating Savings ($106/yr)
Total Social Cost
Base
Absolute
Change

NA
NA
$876
$621
$1,497

$975
-139
$143

$821
-79
$42

$0.06
-2.38
$8

NA
NA
$2
$102.4
-$284.7
$1,510.0
Case (2013)
Relative Change
(%)

0.10%
-0.02%
NA
NA
NA

2.9%
-0.02%
NA

21.4%
-0.01%
NA

6.0%
-0.02%
NA

0.01%
-0.01%
NA
NA
NA
NA
Adding Operating
Absolute
Change

NA
NA
$709
$525
$1,234

$976
-93
$142

$821
-53
$42

$0.06
-1.57
$6

NA
NA
$2
$102.4
-$19.2
$1,510.1
Savings To App
Relative Change
(%)

0.08%
-0.01%
NA
NA
NA

2.9%
-0.01%
NA

21.4%
-0.01%
NA

6.0%
-0.01%
NA

0.01%
-0.01%
NA
NA
NA
NA
a  Sensitivity analysis is presented for 2013.
b  Figures are in 2002 dollars.
                                        10-202

-------
                                                     Economic Impact Analysis
101.4 Engine and Equipment Fixed Cost Shift Scenario

   As discussed in Section 10.3 only the variable costs are used to shift the supply curve in the
engines and equipment markets. Fixed costs are assumed to be R&D costs that are absorbed by
engine and equipment markets over a 5-year period and hence do not affect market prices or
quantities.  As a result, producers are not able to pass any of these costs on and bear all fixed
costs as a decrease in producer surplus.

   In this scenario, the supply shift for engine producers includes the fixed and variable
compliance costs.  The results are presented in Table 101-10.  In this scenario, engine producers
are able to pass along the majority of the fixed compliance costs to the downstream markets
rather than absorb them as a one-to-one reduction in profits. As expected, this scenario leads to a
higher projected price increases for the engine and equipment markets (from 2.9 percent in the
baseline case to 3.4 percent for equipment markets and from 21.4 percent in the baseline case to
23.0 percent for engine markets), and the share of the social costs borne by these markets
decreases from 9.5 percent to 0.2 percent for the equipment markets, and from 2.8 percent to 0
percent for the engine markets.  These costs are passed on to the application markets, and their
expected share of the compliance burden increases from 83 percent to 93 percent.  However, the
total social costs of the regulation are not expected to change measurably as the higher prices
lead to almost no change in the demand for equipment and engines.
                                        10-203

-------
Final Regulatory Impact Analysis
     Table 101-10 Fixed Costs Added to Supply Shift in Engine and Equipment Marketsa>b
Scenario
Application Markets
Price ($/q)
Quantity (q/yr)
Change in Consumer Surplus
Change in Producer Surplus
Change in Total Surplus
Equipment Markets
Price ($/q)
Quantity (q/yr)
Change in Producer Surplus
Engine Markets
Price ($/q)
Quantity (q/yr)
Change in Producer Surplus
Fuel Markets
Price ($/q)
Quantity (q/yr)
Change in Producer Surplus
Transportation Services
Price ($/q)
Quantity (q/yr)
Change in Producer Surplus
Applications Not Included in
Operating Savings ($106/yr)
Social Costs ($106/yr)
Base
Absolute
Change

NA
NA
$876
$621
$1,497

$975
-139
$143

$821
-79
$42

$0.06
-2.38
$8

NA
NA
$2
$102.4
-$284.7
$1,510.0
Case (2013)
Relative Change
(%)

0.10%
-0.02%
NA
NA
NA

2.9%
-0.02%
NA

21.4%
-0.01%
NA

6.0%
-0.02%
NA

0.01%
-0.01%
NA
NA
NA
NA
Shocking Engine
Markets by
Absolute
Change

NA
NA
$978
$697
$1,675

$1,192
-156
$5

$898
-87
$0

$0.06
-2.67
$9

NA
NA
$3
$102.4
-$284.7
$1,509.9
and Equipment
Total Costs
Relative Change
(%)

0.11%
-0.02%
NA
NA
NA

3.4%
-0.02%
NA

23.0%
-0.02%
NA

6.0%
-0.02%
NA

0.01%
-0.01%
NA
NA
NA
NA
a   Sensitivity analysis is presented for 2013.
b   Figures are in 2002 dollars.
101.5 Alternative Social Discount Rates

   Future benefits and costs are commonly discounted to account for the time value of money.
                                       10-204

-------
                                                       Economic Impact Analysis
The market and economic impact estimates presented in Section 10.1 calculate the present value
of economic impacts using a social discount rate of 3 percent, yielding a total social cost of
$27.2 billion. The 3 percent discount rate reflects the commonly used substitution rate of
consumption over time.  An alternative is the OMB-recommended discount rate of 7 percent that
reflects the commonly used real private rate of investment.  Table 101-11 shows the present
value calculated over 2004 to 2030 using both the 3 and 7 percent social discount rates. With the
7 percent social discount rate, the present value of total social costs decreases to $13.9 billion.
                             Table 101-11. Net Present Values3

Engine Producers Total
Equipment Producers Total
Construction Equipment
Agricultural Equipment
Industrial Equipment
Application Producers &
Consumers Total
Total Producer
Total Consumer
Construction
Agriculture
Manufacturing
Fuel Producers Total
PADD 1 & 3
PADD2
PADD 4
PADD 5
Transportation Services Total
Locomotive
Marine
Application Markets Not
Included in NDEIM
Total

Market
Surplus
(106)
$256
$1,162
$545
$397
$220
$28,429
$11,838
$16,591
$11,526
$8,181
$8,723
$169
$85
$69
$3
$12
$1,653
$31
$18
$1,604
$31,669
NPV (3%)
Operating Cost
Savings
(io6)





-$3,757


-$1,779
-$1,208
-$770






-$160
-$204
-$315
-$4,437

Total
$256
$1,162
$545
$397
$220
$24,672


$9,746
$6,973
$7,953
$169
$85
$69
$3
$12
$973
-$129
-$187
$1,288
$27,232
NPV (7%)
Market
Surplus
(IO6)
$180
$740
$343
$255
$141
$14,663
$6,096
$8,567
$5,922
$4,222
$4,519
$86
$43
$35
$1
$6
$900
$16
$9
$875
$16,569
Operating
Cost Savings
(io6)





-$2,309


-$1,093
-$742
-$473






-$97
-$113
-$182
-$2,701
Total
$180
$740
$343
$255
$141
$12,354


$4,829
$3,480
$4,046
$86
$43
$35
$1
$6
$508
-$82
-$104
$693
$13,868
a  Figures are in 2001 dollars.
b  Figures are in 2002 dollars.
                                         10-205

-------
CHAPTER 11: Small-Business Flexibility Analysis
   11.1 Overview of the Regulatory Flexibility Act  	11-1
   11.2 Need for the Rulemaking and Rulemaking Objectives	11-2
   11.3 Issues Raised by Public Comments	11-2
       11.3.1 Comments Regarding Small Business Engine and Equipment Manufacturers  11-3
       11.3.2 Comments Regarding Small Fuel Refiners, Distributors, and Marketers .... 11-3
          11.3.2.1 General Comments on Small Refiner Flexibility	11-3
          11.3.2.2 Comments on the Small Refiner Definition	11-4
          11.3.2.3 Comments on the Baseline Approach 	11-4
          11.3.2.4 Comments on Small Refiner 'Option 4'  	11-4
          11.3.2.5 Comments on Emission Impacts of the Small Refiner Provisions 	11-5
          11.3.2.6 Comments on Inclusion of a Crude Capacity Limit for Small Refiners  . 11-5
          11.3.2.7 Comments on Leadtime Afforded for Mergers and Acquisitions	11-6
          11.3.2.8 Necessity of Small Refiner Program	11-6
          11.3.2.9 Comments on Fuel Marker  	11-6
   11.4 Description of Affected Entities 	11-6
       11.4.1 Description of Nonroad Diesel Engine and Equipment Manufacturers	11-7
          11.4.1.1 Nonroad Diesel Engine Manufacturers	11-8
          11.4.1.2 Nonroad Diesel Equipment Manufacturers	11-8
       11.4.2 Description of the Nonroad Diesel Fuel Industry	11-9
          11.4.2.1 Nonroad Diesel Fuel Refiners	11-9
          11.4.2.2 Nonroad Diesel Fuel Distributors and Marketers 	11-10
   11.5 Projected Reporting, Recordkeeping, and Other Compliance Requirements of the
       Regulation  	11-10
   11.6 Steps to Minimize Significant Economic Impact on Small Entities 	11-11
       11.6.1 Transition and Hardship Provisions for Small Engine  Manufacturers	11-12
          11.6.1.1 Panel Recommendations  	11-12
          11.6.1.2 What We Proposed   	11-13
          11.6.1.3 Provisions Being Finalized in This Rule	11-14
       11.6.2 Transition and Hardship Provisions for Nonroad Diesel Equipment Manufacturers
           	 11-16
          11.6.2.1 Panel Recommendations  	11-16
          11.6.2.2 What We Proposed   	11-17
          11.6.2.3 Provisions in the Final Rule	11-19
       11.6.3 Transition and Hardship Provisions for Nonroad Diesel Fuel Small Refiners 11-20
          11.6.3.1 Panel Recommendations  	11-20
          11.6.3.2 What We Proposed   	11-22
          11.6.3.3 Provisions in the Final Rule	11-23
       11.6.4 Transition and Hardship Provisions for Nonroad Diesel Fuel Small Distributors
          and Marketers 	11-26
          11.6.4.1 Panel Recommendations  	11-26
          11.6.4.2 What We Proposed   	11-26
          11.6.4.3 Provisions in the Final Rule	11-26
   11.7 Conclusion	11-27

-------
                                                 Small-Business Flexibility Analysis
       CHAPTER  11:  Small-Business Flexibility Analysis
   This chapter discusses our Final Regulatory Flexibility Analysis, which evaluates the
potential impacts of new standards on small entities. Pursuant to the requirements of the
Regulatory Flexibility Act, as amended by the Small Business Regulatory Enforcement Fairness
Act of 1996 (SBREFA), which generally requires an agency to prepare a regulatory flexibility
analysis of any rule subject to notice-and-comment rulemaking requirements under the
Administrative Procedure Act or any other statute unless the agency certifies that the rule will
not have a significant economic impact on a substantial number of small entities.  Prior to
issuing a proposal for this rulemaking, we analyzed the potential impacts of these regulations on
small entities. As a part of this analysis, we convened a Small Business Advocacy Review Panel
(SBAR Panel, or 'the Panel'). During the Panel process, we gathered information and
recommendations from Small Entity Representatives (SERs) on how to reduce the impact of the
rule on small entities, and those comments are detailed in the Final Panel Report which is located
in the public record for this rulemaking (Docket A-2001-28, Document No. II-A-172).

11.1 Overview of the Regulatory Flexibility Act

   In accordance with section 609(b) of the Regulatory Flexibility Act, we convened an SBAR
Panel before conducting the Regulatory Flexibility Analysis.  A summary of the Panel's
recommendations can be found in our proposal. Further, the Final Panel Report contains a
detailed discussion of the Panel's advice  and recommendations (as well as the SER
recommendations). The regulatory alternatives that are being adopted in this final rule are
described below.

   Section 609(b) of the Regulatory Flexibility Act further directs the Panel to report on the
comments of small entity representatives and make findings on issues related to identified
elements of the Regulatory Flexibility Analysis under  section 603 of the Regulatory Flexibility
Act.  Key elements of a Regulatory Flexibility Analysis are:
   a description and, where feasible, an estimate of the number of small entities to which the
   proposed rule applies;
-  projected reporting, record keeping, and other compliance requirements of the proposed rule,
   including an estimate of the classes of small entities that would be subject to the rule and the
   type of professional skills necessary to prepare reports or other records;
-  an identification, to the extent practicable, of all other relevant federal rules that may
   duplicate, overlap, or conflict with the proposed rule;
   any significant alternatives to the proposed rule that accomplish the stated objectives of
   applicable statutes and that minimize any  significant economic impact of the proposed rule
   on small entities.

   The Regulatory Flexibility Act was amended by SBREFA to ensure that concerns regarding
small entities are adequately considered during the development of new regulations that affect

                                          11-1

-------
Final Regulatory Support Document
those entities. Although we are not required by the Clean Air Act to provide special treatment to
small businesses, the Regulatory Flexibility Act requires us to carefully consider the economic
impacts that our rules will have on small entities.  The recommendations made by the Panel may
serve to help lessen these economic impacts on small entities when consistent with Clean Air
Act requirements.

11.2 Need for the Rulemaking and Rulemaking Objectives

   A detailed discussion on the need for and objectives of this rule are in the preamble to the
final rule.  Controlling emissions from nonroad engines and equipment, in conjunction with
diesel fuel controls, has important public health and welfare benefits.  With the advent of more
stringent controls on highway vehicles and their fuels, emissions from nonroad sources, unless
controlled, will contribute significantly more harmful pollution than those from highway
sources.

   Section 213(a)(3) of the Clean Air Act requires EPA to regulate NOx emissions from
nonroad engines and vehicles upon an EPA determination that nonroad engines contribute to
emissions in a nonattainment area.  In part, section 213(a)(3) authorizes EPA to promulgate
standards for designated pollutants (including NOx) that require the greatest degree of emission
reduction achievable from application of technology to nonroad engines (or vehicles) while
giving "appropriate consideration to the cost of applying such technology within the period of
time available to  manufacturers and to noise, energy, and safety factors associated with the
application of such technology." Section 213(a)(4) applies to all pollutants not specifically
identified in section 213(a)(3), and authorizes EPA to promulgate "appropriate" standards for
such pollutants, taking into account "costs, noise, safety, and energy factors associated with the
application of technology which the Administrator determines will be available" for those
engines (or vehicles). Controls on PM implement this provision.

   Similarly, section 21 l(c)(l) authorizes EPA to regulate fuels if any emission product of the
fuel causes or contributes to air pollution that may endanger public health or welfare, or that may
impair the performance of emission-control technology on engines and vehicles. We believe
there is an opportunity for cost-effective emission reductions on a large scale.

11.3 Issues Raised by Public Comments

   During the public comment period we received numerous comments regarding various
aspects of the NPRM, including our proposed small business provisions.  The following section
provides a summary of the comments that we received on our proposed provisions.  More
information on these comments can be found in the Final Summary and Analysis of Comments,
which is a part of the rulemaking record.
                                          11-2

-------
                                                 Small-Business Flexibility Analysis
11.3.1 Comments Regarding Small Business Engine and Equipment Manufacturers

   One small business engine manufacturer commented that the proposed provisions for small
manufacturers are appropriate and strongly supported their inclusion in the final rule. The
manufacturer raised many concerns of why it believes that it is necessary to include such
provisions, such as: larger/higher-volume manufacturers will have priority in supply of new
technologies and will thus have more R&D time to complete development of these systems
before they are available to smaller manufacturers; and, smaller manufacturers do not command
the same amount of attention from potential suppliers of critical technologies for T4 controls,
and are thus concerned that they may not be able to attract a manufacturer to work with them on
the development of compliant technologies. The small manufacturer believes that the additional
three-year time period proposed for small business engine manufacturers in the NPRM is
necessary for its company, and is the company's estimate  of the time that it will take for these
technologies to be available to small engine manufacturers.

   The Small Business Administration's Office of Advocacy ("Advocacy") raised the concern
that the rule would impose significant burdens on a substantial number of small entities with
little corresponding environmental benefit.  Advocacy commented that we should exclude
smaller engines (those under 75 hp) from further regulation  in order to comply with the
Regulatory Flexibility Act and fulfill the requirement of reducing the burden  on small engine
classes.  Advocacy recommended that PM standards for engines in the 25-75  hp powerband
should not be based on performance of aftertreatment technologies.  Advocacy believes that the
proposed flexibilities will not suffice on their own to appropriately minimize  the regulatory
burdens on small entities; and Advocacy noted that during the SBREFA process, some small
equipment manufacturers stated that although EPA would allow some  equipment to be sold
which would not require new emissions controls, engine manufacturers would not produce or
sell such equipment. Advocacy also commented that we have not shown that substantial
numbers of small businesses have taken advantage of previous small business flexibilities, or
that small businesses would be able to take advantage of the flexibilities under this rule. Lastly,
Advocacy commented that although full compliance with  the more stringent emissions controls
requirements would be delayed for small manufacturers, small business manufacturers
eventually will be required to produce equipment meeting the new requirements.

11.3.2 Comments Regarding Small Fuel Refiners, Distributors, and Marketers

   11.3.2.1 General Comments on Small Refiner Flexibility

   One small refiner commented that it is not plausible at this time to  evaluate the impact of the
three fuels regulations on the refining industry (and small  refiners), however it stated that we
should continue to evaluate the impacts and act quickly to avoid shortages and price spikes and
we should be prepared, if necessary, to act quickly in considering changes in  the regulations to
avoid these problems. We also received comment that some small refiners that produce
locomotive and marine fuels fear that future sulfur reductions to these markets could be very
damaging.

-------
Final Regulatory Support Document
    11.3.2.2 Comments on the Small Refiner Definition

    A small refiner commented that the proposed redefinition of a small refiner (to not
grandfather as small refiners those that were small for highway diesel) would both negate the
benefits afforded under the small refiner provisions in the Highway Diesel Sulfur rule and
disqualify its status as a small refiner.  The refiner suggested that we clarify the language and
include provisions for continuance of small refiner flexibility for refiners who qualified under
the Highway Diesel  Sulfur rule (and have not been disqualified as the result of a merger or
acquisition).

    11.3.2.3 Comments on the Baseline Approach

    A coalition of small refiners provided comments on a few aspects of concern.  The small
refiners believe that  the fuel segregation, and ensuing marking and dying, provisions are quite
complex.  One small refiner believes that mandating a minimum volume of NRLM production
would conflict with the purpose of maintaining adequate on-highway volumes of 15ppm sulfur
fuel and unnecessarily restricts small refiners,  and offered suggestions in their comments on how
to improve the language.

    11.3.2.4 Comments on Small Refiner 'Option 4'

    A coalition of small refiners commented that if the final rule is not issued before January 1,
2004, a provision should be made to accommodate those  small refiners planning to take
advantage of the proposed small refiner "Option 4" (the NRLM/Gasoline Compliance option). A
small refiner echoed the concerns of the small refiner coalition, commenting that delayed
fmalization of the final rule would undermine the benefits of small refiner flexibility Option 4.
The small refiner is concerned that a delay in issuing the rule, and subsequent delay in the
opportunity to apply the interim gasoline flexibility, would negate its opportunity to take full
advantage of the credits the refiner now has, as it would not be able to comply with the 300 ppm
cap. The small refiner suggested that we allow small refiners to apply for temporary relief and
operate under the Option 4 provision.

    A small refiner commented that, in the NPRM, it was unclear if a small refiner could elect to
use any or all of the  first three of the small refiner provisions if it did not elect to use Option 4.
Further, the refiner understood that if Option 4 was chosen, a small refiner could not use any of
the first three options. The refiner believes that it is important that a small refiner be able to use
Options 1, 2, and 3 in combination with each other, and stated that we need to clarify the intent
in the final rule. The small refiner also commented that the provisions in 40 CFR §§ 80.553 and
80.554 are not clear  and should be revised to clarify their intent.  Specifically, the refiner
questioned whether or not a small refiner who committed to producing ULSD by June 1, 2006 in
exchange for an extension of its interim gasoline sulfur standards (under 40 CFR 80.553) could
elect to exercise the  options allowed under 40  CFR 80.554.

    Another small refiner raised the concern that the small refiner Option 4  only provides an
adjustment to those small refiners whose small refiner gasoline sulfur standards were established

                                          11-4

-------
                                                  Small-Business Flexibility Analysis
through the hardship process of 40 CFR § 80.240. The small refiner suggested that we finalize a
compliance option that allows a 20% increase in small refiner gasoline sulfur standards be
extended to all small refiners, not just those with standards established pursuant 40 CFR §
80.240(a), and offers suggested language in its comments.

    11.3.2.5 Comments on Emission Impacts of the Small Refiner Provisions

    A state environmental group commented that the provisions for small refiners raise
substantial environmental concerns.  The group is concerned that these provisions will allow
small refiners the ability to produce gasoline with an unknown sulfur content for an unknown
length of time; this fuel may then be sold at the refiner's retail outlet, and may become the
primary fuel for some vehicles, which alters vehicle fleet emissions performance. This
environmental group also commented that the absence of any process of notification regarding
small business provisions to notify States of these provisions is troubling. The group's concern
is that any deviations from fuel content regulations that affect fuels consumed, can significantly
alter their inventories and can undermine the State planning process. The group suggested that in
the future there should be greater communication from us regarding decisions that impact the
quality of fuels consumed in a state, and thus impact the quality of that state's air.

    Another state environmental group commented on the flexibility provisions for small
refiners; the group is concerned that the exemption will not have a minor effect on the nation's
fuel supply, as the state is an intermountain western state.  The group comments that the impact
of this exemption is concentrated in these states, namely Washington and Oregon- states which
are served primarily by refineries that will be allowed to delay compliance with the ULSD
standards until 2014.  Therefore, the group commented, residents of these areas are denied air
quality benefits equivalent to those promised the rest of the country. The group is concerned that
those seeking to purchase and use equipment in the West will be subject to the ULSD standard
regardless of fuel supply and availability in their area. Further, they would be faced with
problems such as misfueling, the need to defer the purchase of new equipment, or paying a
premium for a 'boutique' fuel.

    11.3.2.6 Comments on Inclusion of a Crude Capacity  Limit for Small Refiners

    Two non-small refiners supported the inclusion of the 155,000 bpcd limit; further, one refiner
commented that any refiner with the financial wherewithal to acquire additional refineries to
allow its crude capacity to exceed 155,000 bpcd should not be able to retain status as a small
refiner.  Another commenter stated that if we were to finalize the 155,000 bpcd limit, we should
not apply it in cases of a merger between two small refiners.  The  commenter further stated that a
merger of two small companies in a hardship condition does not imply improved financial health
in  the same way that an acquisition would.  A small refiner is commented that it supports the
addition of the capacity limit in the small refiner definition as it would correct the problem of the
inadvertent loop-hole in the two previous fuel rules.  Though the refiner  did raise the concern
that the wording of the proposed language may result in small refiners such as itself, who grew
by normal business practice, being disqualified as small refiners.
                                          11-5

-------
Final Regulatory Support Document
    11.3.2.7 Comments on Leadtime Afforded for Mergers and Acquisitions

    A non-small refiner suggested that we limit the provision of affording a two-year leadtime to
small refiners who lose their small status due to merger or acquisition to the case where a small
refiner merges with another small refiner. Further, the refiner commented that it would be
inappropriate to allow such small refiners to be able to generate credits for "early" production of
lower sulfur diesels during this two-year leadtime.  Lastly, the refiner commented that a small
refiner which acquires a non-small refiner, and thus loses its small refiner status,  should not be
eligible for hardship provisions. Another non-small refiner commented that it supports  the two-
year lead time for refineries that lose their status as a small refiner due to a merger or acquisition.

    11.3.2.8 Necessity of Small Refiner Program

    A non-small refiner provided comment on the NPRM stating the belief that the proposed
provisions for small refiners are not practical.  The refiner is concerned that having provisions
for small refiners adds a level of complication, results in emissions losses, increases the potential
for ULSD contamination, and create an unfair  situation in the marketplace.  Similarly, another
non-small refiner and a trade group representing many refiners and others in the fuels industry
commented that they oppose the extension of compliance deadlines for small refiners, as this can
result in inequitable situations that may affect the refining industry for some time and can put the
distribution system at risk for contamination of lower sulfur fuels. They further stated that all
refiners will face challenges in complying with the upcoming standards and would not
significantly alter the business decisions that small refiners would make.  They also stated that
non-small refiners face similar issues with their older and/or smaller refineries, but will  not have
the benefit of being able to postpone making these decisions as small refiners will.

    11.3.2.9 Comments on Fuel Marker

    We received comments from terminal operators stating that the proposed heating oil marker
requirements would force small terminal operators to install expensive  injection equipment and
that they would not be able to recoup the costs.

11.4 Description of Affected Entities

    Small entities include small businesses, small organizations, and small governmental
jurisdictions. For assessing the impacts of the  rule on small entities, a small entity is defined as:
(1) a small business that meets the definition for business based on the  Small Business
Administration's (SBA) size standards (see Table 11-1); (2) a small governmental jurisdiction
that is a government of a city, county, town, school district or special district with a population
of less than 50,000; and (3) a small organization that is any not-for-profit enterprise that is
independently owned and operated and is not dominant in its field.  Table 11-1 provides an
overview of the primary SBA small business categories potentially affected by this regulation.
                                          11-6

-------
                                                     Small-Business Flexibility Analysis
    The following sections discuss the small entities directly regulated by this final rule—namely
nonroad diesel engine manufacturers, nonroad diesel equipment manufacturers, and nonroad fuel
refiners and fuel marketers/distributors.  Also, Table 11-2 lists our assessment of the number of
small entities that will be directly affected by  this rulemaking.

                                         Table 11-1
                                 Small Business Definitions
Industry
Engine manufacturers
Equipment manufacturers:
- construction equipment
- industrial truck manufacturers (i.e., forklifts)
- all other nonroad equipment manufacturers
Fuel refiners
Fuel distributors
Defined as small
entity by SBA if:
Less than 1,000 employees
Less than 750 employees
Less than 750 employees
Less than 500 employees
Less than 1500 employees11
varies
Major SIC Codes3
Major Group 35
Major Group 35
Major Group 35
Major Group 35
2911
varies
a Standard Industrial Classification
b In previous rulemakings to set fuel requirements, we have included a provision that a refiner must also have a company-
wide crude refining capacity of no greater than 155,000 barrels per calendar day to qualify for the small-refiner
flexibilities,. We have included this criterion in the small-refiner definition for this final rule.
                                         Table 11-2
           Number of Small Entities To Which the Nonroad Diesel Rule Will Apply
Industry
Engine manufacturers
Equipment manufacturers
Fuel refiners
Fuel distributors
Defined as small entity by SBA if:
Less than 1,000 employees
(see criteria in Table 11-1)
Less than 1500 employees
varies
Number of Affected Entities
4a
335a
26
(see discussion in 1 1.4.2.2)
a The numbers of affected entities for these categories are taken from the total number of companies that were used in our
screening analysis (i.e., companies with publicly available employee and sales data).
11.4.1 Description of Nonroad Diesel Engine and Equipment Manufacturers

    To assess how many small engine and equipment manufacturers would be directly affected
by the rule, we first created a database consisting of firms listed in the Power Systems Research
(PSR) database and compared this with the list of companies from the analysis performed for the
1998 nonroad final rule and with membership lists from trade organizations. We then found
sales and employment data for the parent companies of these firms using databases such as the
Thomas Register and Dun and Bradstreet. Due to the wide variety in the types of equipment that
                                            11-7

-------
Final Regulatory Support Document
use nonroad diesel engines, there are numerous SIC codes in which the equipment manufacturers
report their sales, though the majority of the firms are listed under the SIC major group 35xx-
Industrial and Commercial Machinery and Computer Equipment.

    We conducted a preliminary industry profile to identify the engine and equipment
manufacturers that are in the nonroad diesel sector.  We identified more than 1,000 businesses
that fit this  description; however, due to a lack of publicly available sales or employment data,
some of these entities could not be confirmed for consideration in the analysis.

    11.4.1.1 Nonroad Diesel Engine Manufacturers

    Using information from the preliminary industry profile, we identified a total of 61 engine
manufacturers. The top 10 engine manufacturers comprise over 80 percent of the total market,
while the other 51 companies make up the remaining percentage.A  Of the 61 manufacturers, four
fit the SBA definition of a small entity.  These four manufacturers were Anadolu Motors,
Farymann Diesel GmbH, Lister-Fetter Group, and V & L Tools (parent company of Wisconsin
Motors LLC, formerly 'Wis-Con Total Power'). These businesses comprise approximately 8
percent of the total engine sales for the year 2000. Lister Fetter and V & L Tools were the only
two manufacturers which had certified engines for model year 2000.

    Wisconsin Motors produces diesel engines for a small niche market and served as a Small
Entity Representative (SER) during the Small Business Advocacy Review Panel process,
speaking to the needs of small engine manufacturers.

    11.4.1.2 Nonroad Diesel Equipment Manufacturers

    This rule will result in equipment manufacturers incurring some increased costs as a result of
the need to  make changes to their equipment to accommodate the addition of aftertreatment
technologies.  The vast majority of equipment manufacturers are not integrated companies,
meaning that they do not make the engines they install. Thus, most equipment manufacturers are
largely dependent on engine manufacturers for the availability of pre-production information
about the new engines and for a sufficient supply of the engines once production begins.
Equipment  manufacturers that are small businesses may, in general, face a disproportionate
degree of hardship in adapting to these types of changes in design and increased costs of new,
cleaner engines.

    To determine the number of equipment manufacturers, we also used the industry profile that
was conducted.  From this, we identified more than 700 manufacturers with sales and/or
employment data that could be included in the screening analysis.  These businesses included
manufacturers in the construction, agricultural, and outdoor power equipment (mainly, lawn and
garden equipment) sectors of the nonroad diesel market.  The equipment produced by these
manufacturers ranged from small (sub-25 hp walk-behind equipment) to large (in excess of 750
   A All sales information used for this analysis was 2000 data.

                                         11-8

-------
                                                  Small-Business Flexibility Analysis
hp, such as mining and construction equipment).  Of the manufacturers with available sales and
employment data (approximately 500 manufacturers), small equipment manufacturers represent
68 percent of total equipment manufacturers (and these manufacturers account for 11 percent of
nonroad diesel equipment industry sales).  Thus, the majority of the small entities that could
potentially experience a significant impact as a result of this rulemaking are in the nonroad
equipment manufacturing sector.

   While a few small equipment manufacturers did serve as SERs during the SBREFA Panel
process, a trade association representing many equipment manufacturers also served as a SER.
We believe that due to the large number of small equipment manufacturers, this SER was better
able to contact and disseminate information to the large universe of small entities in this category
and serve as a voice for some of the extremely small equipment manufacturers.

11.4.2 Description of the Nonroad Diesel Fuel Industry

   The analysis that we developed for the refining  industry is built on analyses that were
performed for the gasoline and highway diesel sulfur programs in recent years. Information
about the characteristics of refiners came from sources including the Energy Information
Administration within the U.S. Department of Energy, and from oil industry literature. Our
assessment was that the refining industry is located primarily in SIC 2911. In both the gasoline
sulfur and highway diesel sulfur rules, we applied specific small-refiner flexibilities to refiners
that have no more than 1500 employees and no greater than 155,000 barrels per calendar day
crude capacity. For transporters, distributors, and marketers of nonroad diesel fuel, trade groups
were our key sources for information about this industry.  We determined that this industry
sector includes several types of businesses that fall  into several different SB A small entity
criteria; our assessment was that the vast majority of these entities are small.

   11.4.2.1 Nonroad Diesel Fuel  Refiners

   Our assessment is that 26 high-sulfur (nonroad and locomotive and marine) refiners,
collectively owning 33 refineries, meet SBA's definition of a small business for the refining
industry.  The 33 refineries appear to meet both of the employee number and production volume
criteria mentioned above, out of a total of approximately 91 nonroad refineries. These small
refiners produce approximately 6 percent of the total high-sulfur diesel fuel. Note that because
of the dynamics in the refining industry (such as mergers and acquisitions), this figure could
likely change.

   A few small refiners, as well as representatives  of an ad-hoc coalition of some of the small
refiners participated in the SBREFA process. These small refiners, and those in which they
represented, provide high sulfur  diesel fuel for various non-highway markets and applications,
and own and operate refineries throughout the country.
                                          11-9

-------
Final Regulatory Support Document
    11.4.2.2 Nonroad Diesel Fuel Distributors and Marketers

    The industry that transports, distributes, and markets nonroad diesel fuel encompasses a wide
range of businesses, including bulk terminals, bulk plants, fuel oil dealers, and diesel fuel
trucking operations, and totals thousands of entities that have some role in this activity. More
than 90 percent of these entities meet small-entity criteria. Common carrier pipeline companies
are also a part of the distribution system; 10 of them are small businesses.

    Similar to the nonroad small business equipment sector, the universe of nonroad fuel
distributors and marketers is quite large, so representatives of fuel pipeline and fuel marketing
trade groups participated in the SBREFA process. We believe that these representatives were
very capable of speaking to the needs of their members that are small entities and were also
better able to disseminate SER outreach information to these markets.

11.5 Projected Reporting, Recordkeeping, and Other Compliance
Requirements of the Regulation

    For engine and equipment manufacturers, EPA is continuing many of the reporting,
recordkeeping, and compliance requirements prescribed for these categories in 40 CFR part 89.
These include, certification requirements and provisions related to reporting of production,
emissions information, use of transition provisions, etc. The types of professional skills required
to prepare reports and records is also similar to the types of skills set out in 40 CFR part 89. Key
differences in the requirements of this final rule, as compared to 40 CFR part 89, are the
reporting of emissions information and defect reporting — we are finalizing an increase in the
number of data points (i.e. transient testing) that will be required for reporting emissions
information, as well as adopting an increased reporting burden for Tier 3 and earlier engines for
defect reporting. In addition, we are requiring that manufacturers report to us if they learn that a
substantial number of their engines have emission-related defects. This is generally not an
affirmative requirement to collect information. However, if manufacturers learn that there are,
or might be, a substantial number of emission-related defects, then they must send us
information describing the defects.

    For any fuel control program, we must have the assurance that fuel produced by refiners
meets the applicable standard, and that the fuel continues to meet this standard as it passes
downstream through the distribution system to the ultimate end user. Which is of particular
importance in regards to diesel fuel, since the aftertreatment technologies expected to be used to
meet the engine standards are highly sensitive to sulfur. Many of the recordkeeping, reporting,
and compliance provisions we are finalizing are fairly consistent with those currently in place for
other fuel programs, including the current 15 ppm highway diesel regulation. For example,
recordkeeping involves the use of product transfer documents, which are already required under
the 15 ppm highway diesel sulfur rule (40 CFR 80.560). We  are finalizing additional
recordkeeping and reporting requirements for refiners, importers, and fuel distributors to
implement the designate and track provisions. Discussions with parties from all segments of the
distribution system indicated that the records necessary were  analogous to records already kept


                                          11-10

-------
                                                  Small-Business Flexibility Analysis
as a normal process of conducting business. Consequently, the only significant additional
burden would be associated with the reporting requirement.

   General requirements for reporting for refiners and importers include: registration (if the
refiner or importer is not registered under a previous fuel program), pre-compliance reports (on a
refiner or importer's progress towards meeting the nonroad diesel fuel requirements as specified
in this rule), quarterly designation reports, and annual reports. All parties, from the refiner to the
terminal, will be required to report volumes of designated fuels received and distributed, as well
as compliance with quarterly and annual limits. All parties in the distribution system will be
required to keep product transfer documents (PTDs), though refiners and importers are required
to initially generate and provide information on commercial PTDs that identify the diesel fuel
with meeting specific needs (i.e.  15 ppm highway diesel, 500 ppm highway diesel, etc.). Also,
small refiners in Alaska that choose to delay compliance must, at a minimum, report end users of
their fuel.  These end users must at a minimum also keep records of these fuel purchases. As
with previous fuel regulations, small refiners will be required to apply for small refiner status
and small refiner baselines.

   In general, we are requiring that all records be kept for at least five years. This
recordkeeping requirement should impose little additional burden, as five years is the applicable
statute of limitations for current fuel programs.

   Section X.B of the preamble to the final rule includes a discussion of the estimated burden
hours and costs  of the recordkeeping and reporting that will be required by this final rule.
Detailed information on the reporting and recordkeeping measures associated with this
rulemaking are described in the Information Collection Requests (ICRs), also located in the
preamble to this rulemaking-- 1897.05 for nonroad diesel engines, and 1718.05 for fuel-related
items.

11.6 Steps to Minimize Significant Economic Impact on Small Entities

   As a part of the SBREFA process, we conducted outreach to a number of small entities
representing the various sectors covered in this rulemaking and convened a Panel to gain
feedback and advice from these representatives.  Prior to convening the Panel, we held outreach
meetings with the SERs to learn the needs of small businesses and potential challenges that these
entities may face.  The outreach meetings also helped to provide the SERs an opportunity to gain
a better understanding of the upcoming standards.  The feedback that we received from SERs as
a result of these meetings was used during the Panel convening for developing regulatory
alternatives to mitigate the impacts of the rulemaking on small businesses. General concerns
raised by SERs  during the SBREFA process were potential difficulty and costs of compliance
with the upcoming standards.

   The Panel consisted of members from EPA, the Office of Management and Budget (OMB),
and the Small Business Administration's Office of Advocacy ('Advocacy').  Following the Panel
convening, a Final Panel Report detailing all of the alternatives that were recommended by the
                                         11-11

-------
Final Regulatory Support Document
Panel (as well as individual Panel members) was issued. We either proposed or requested
comment on the various recommendations put forth by the Panel. Below we discuss those
flexibility options recommended in the Panel Report, our proposed regulatory alternatives, and
those provisions which are being finalized. We are finalizing many of the provisions
recommended by the Panel, with exceptions noted below. We believe that the provisions that we
are finalizing will help to mitigate the burden imposed upon small entities in complying with this
rule.

11.6.1 Transition and Hardship Provisions for Small Engine Manufacturers

   11.6.1.1 Panel Recommendations

   The following provisions were recommended by the Panel for nonroad diesel  small business
engine manufacturers. During the SBREFA process and the development of the rule, we
considered both a one-step approach as well as the two-step approach in the final  rule.  To be
eligible for the recommended provisions set out below, a manufacturer would have to have
certified in model year 2002 or earlier and would be limited to 2500 units per year (to allow for
some market growth). The Panel recommended these qualifications to prevent misuse of the
transition and hardship provisions as a way to enter the nonroad diesel market or to gain unfair
market position relative to other manufacturers.

   For an approach that entails only one phase of standards, the Panel recommended that a
manufacturer could opt to delay compliance for a period of up to three years.  The Panel also
recommended that we take comment on whether this delay period should be two,  three, or four
years.  Each delay would be pollutant-specific (i.e., the delay would apply to each pollutant as it
is phased in).

   For an approach with two phases of standards the Panel recommended the following
transition provisions:
       •   an engine manufacturer could skip the first phase and comply on time  with the
          second; or,
          a manufacturer could  delay compliance with each phase of standards for up to two
          years.

   The Panel recommended that there should not be any PM aftertreatment-based standards for
engines between 25 and 75 hp. It was also recommended by the Panel that an emission-credit
program of averaging, banking, and trading (ABT) be included as part of the overall rulemaking
program.

   The Panel recommended that two types of hardship provisions be extended to small engine
manufacturers. These provisions are:
       •   for the case of a catastrophic event or other extreme unforseen circumstances beyond
   the control  of the manufacturer that could not have been avoided with reasonable discretion
   (such as fire, tornado, or supplier not fulfilling contract); and
                                         11-12

-------
                                                  Small-Business Flexibility Analysis
       •  for the case where a manufacturer has taken all reasonable business, technical, and
          economic steps to comply but cannot do so.

    The Panel recommended that either hardship relief provision could provide lead time for up
to 2 years- in addition to the transition provisions- and a manufacturer would have to
demonstrate to the Agency's satisfaction that failure to sell the noncompliant engines would
jeopardize the company's solvency.  The Panel further recommended that the Agency may
require that the manufacturer make up the lost environmental benefit through the use of
programs such as supplemental environmental projects.

    11.6.1.2 What We Proposed

    Due to the structure of the standards and their timing,  we proposed transition provisions, for
small engine manufacturers encompassing both approaches recommended by the Panel (with the
inclusion of the 2,500 unit limit for each manufacturer). Following the recommendations of the
Panel, we proposed the following transition provisions for small business engine manufacturers:

    •   for PM-
          small engine manufacturers could delay compliance with the standards for up to three
          years for engines under 25 hp,  and for engines between 75 and 175 hp (as these
          engines only have one standard)
          small engine manufacturers could have the option to delay compliance for one year if
          interim standards are met for engines between 50 and 75 hp (for this power category
          we would be treating the PM standard as a two phase standard) with the stipulation
          that small manufacturers could not use PM credits to meet the interim standard; also,
          if a small manufacturer elects the optional approach to the standard (elects to skip the
          interim standard), no further relief would be provided

    •   for NOxB-
          a three year delay in the program for engines in the 25-50 hp and the 75-175 hp
          categories, consistent with the one-phase approach recommendation above;
          a small engine manufacturer could be afforded up to two years of hardship (in
          addition to the transition flexibilities) upon demonstrating to EPA a significant
          hardship situation;
          small engine manufacturers would be able to participate in an averaging, banking,
          and trading (ABT) program (which we proposed as part of the overall rulemaking
          program for all manufacturers); and,
          no NOx aftertreatment-based standards for engines 75 hp and under.

    We did propose ABT provisions for all nonroad engine manufacturers to enhance the
flexibility offered to engine manufacturers as they make the transition to meet the more stringent
    B EPA did not propose a change in the NOx standard for engines under 25 hp and those between 50 and 75 hp.
For these two power bands, EPA would retain the Tier 3 standards.

                                          11-13

-------
Final Regulatory Support Document
standards. We proposed to retain the basic structure of the current nonroad diesel ABT program,
with some changes to accommodate implementation of the emission standards.  Though the
Panel recommended small engine manufacturer-specific ABT provisions, we did not believe it
would be appropriate to provide a different ABT program for small business engine
manufacturers. Discussions during the SBAR process indicated that small-volume
manufacturers would need extra time to comply due to cost and personnel constraints, and we
found little reason to believe that ABT provisions specific to small manufacturers would create
an incentive to accelerate compliance. Small manufacturers would, of course, always be able to
participate in  the general ABT program.

   We proposed the majority of the Panel's recommendations for small business engine
manufacturers, with noted specific provision elements for PM and NOx. As we disagreed
strongly with  the Panel's recommendation that there not be any PM aftertreatment-based
standards for  engines between 25 and 75 hp, we requested comment on this recommendation,
noting our strong reservations.  In addition, we proposed the Panel recommended hardship
provisions for small business engine manufacturers to provide a useful safety valve in the event
of unforeseen extreme hardship.

   11.6.1.3 Provisions Being Finalized in This Rule

   For nonroad diesel small business engine manufacturers, we are finalizing many of the
transition and hardship provisions that we proposed; we are finalizing some revisions to the
transition provisions, as described below, and we are finalizing all of the hardship provisions that
were proposed.  While we believe that emissions from nonroad engines have a significant effect
on emissions, we also believe that offering these transition provisions to small business engine
manufacturers will have a negligible effect on air quality and the emissions inventory, and
provide an appropriate measure of lead time for these small entities. Further, we continue to
believe that a  complete exemption from the upcoming standards (even assuming that such an
exemption could be justified legally)  would put small business engine manufacturers at a
competitive disadvantage as eventually the rest of the market will be producing engines that are
compliant with these new standards and the equipment produced will only be able to
accommodate these compliant engines. With the transition provisions, small business engine
manufacturers will be in compliance with the Tier 4 standards in the long run and the flexibility
options will give them appropriate lead time to comply.  Further, we received comments from a
small business engine manufacturer stating that such provisions are necessary and adequate to
ease the burden of compliance with the upcoming standards. As such, we believe that the
transition provisions we are adopting will be of significant help for small business engine
manufacturers, and is part of our consideration of appropriate costs in assessing lead time
pursuant to section 213 (b) of the Act.

   We are finalizing the following transition provisions for small business engine
manufacturers:

   For engines under 25 hp-
                                         11-14

-------
                                                   Small-Business Flexibility Analysis
   •   PM- a manufacturer may elect to delay compliance with the standard for up to three
       years.
   •   NOx- there is no change in the level of the existing NOx standard for engines in this
       category, so no special provisions are being provided.
   For engines in the 25-50 hp category-
   •   PM- manufacturers must comply with the interim standards (the Tier 4 requirements that
       begin in model year 2008) on time, and may elect to delay compliance with the 2013 Tier
       4 requirements (0.02 g/bhp-hr PM standard) for up to three years.
   •   NOx- a manufacturer may elect to delay compliance with the standard for up to three
       years.
   For engines in the 50-75 hp category-
   •   PM- A small business engine manufacturer may delay compliance with the 2013 Tier 4
       requirement of 0.02 g/bhp-hr PM for up to three years provided that it complies with the
       interim Tier 4 requirements that begin in model year 2008 on time, without the use of
       credits. Alternatively, a manufacturer may elect to skip the interim standard completely.
       Manufacturers choosing this option will receive only one additional year for compliance
       with the 0.02 g/bhp-hr standard (i.e. compliance in 2013, rather than 2012).  See Section
       III.C of the preamble to the  final rule  for a fuller explanation of these provisions.
   •   NOx- there is no change in the level of the NOx standard for engines in this category,
       therefore no special provisions are being provided.
   For engines in the 75 to 175 hp  category -
   •   PM- a manufacturer may elect to delay compliance with the standard for up to three
       years.
   •   NOx- a manufacturer may elect to delay compliance with the standard for up to three
       years.

   Regarding the Office of Advocacy's comments on the technical feasibility of PM and NOx
aftertreatment devices. As we proposed in the NPRM, we are not adopting standards based on
performance of NOx aftertreatment technologies for engines under 75  hp.  We believe the factual
record, as documented in the preamble, the Summary and Analysis of Comments (e.g., the
response to comment 3.1.4.3), and elsewhere in this RIA, does not support the claim that the PM
standards will  not be technically feasible in 2013 for the 25-75 hp engines.  As  set out at length
in Section 4.1.3, among  other places, performance of PM traps is not dependent on engine size.
Furthermore, as we discussed in the preamble to this final rule and earlier in Chapter 6, we
believe that such standards are feasible for these engines at reasonable cost0, and will help to
improve very significant air quality problems, especially by reducing exposure to diesel PM and
by aiding in attainment of the PM 2.5 and ozone National Ambient Air Quality Standards.
Indeed, given these facts, we do not believe that an alternative of no aftertreatment-based PM
standards for these engines would be appropriate under section 213(a)(4).  We believe the
transition and hardship provisions being finalized for small business engine manufacturers in this
   c As the cost issues raised in SBA's comments relate to all manufacturers (not just small business
manufacturers), further information on the costs of this technology as well as the benefits analysis, can be found in
Section VI of this preamble (and also Chapters 6 and 9, respectively).

                                          11-15

-------
Final Regulatory Support Document
rule are reasonable and are a factor in our ultimate finding that the PM standards for engines in
the 25-75 hp range are appropriate, and that the lead time provided for these standards is the
earliest possible after appropriate consideration of compliance costs.

11.6.2 Transition and Hardship Provisions for Nonroad Diesel Equipment Manufacturers

   11.6.2.1 Panel Recommendations

   For small business equipment manufacturers the Panel recommended that we propose to
continue the transition provisions offered for the Tier 1 and Tier 2 nonroad diesel emission
standards, as set out in 40 CFR 89.102, with some modifications. Those recommended transition
provisions were:
   •   Percent-of-Production Allowance: Over a period of seven model years, equipment
       manufacturers may install engines not certified to the new emission standards in an
       amount of equipment equivalent to 80 percent of one year's production. This would be
       implemented by power category with the average determined over the period in which
       the flexibility is used.

   •   Small-Volume Allowance: A manufacturer could exceed the 80 percent allowance in
       seven years as described above, provided that the previous Tier engine use does not
       exceed 700 total over seven years, and 200 in any given year.  This would be limited to
       one family per power category. Alternatively, the Panel recommended, at the
       manufacturer's choice by power category, a program that eliminates the "single family
       provision" restriction with revised total and annual sales limits as shown below:
       -   For power categories below 175 hp, a manufacturer could use 525 previous Tier
          engines (over  seven years) with an annual cap of 150 units (these engine numbers are
          separate for each of the three power categories defined in the regulations).
          For power categories above 175 hp, a manufacturer could use 350 previous Tier
          engines (over  seven years) with an annual cap of 100 units (these engine numbers are
          separate for each of the two power categories defined in the regulations).
       The Panel recommended that we seek comment on the total number of engines and
       annual cap values listed above. Advocacy believed the transition to the Tier 4
       technology will be more costly and technically difficult, and therefore suggested that
       small business equipment manufacturers may therefore need more liberal flexibility
       allowances especially for equipment using the lower hp engines.  SBA and OMB
       recommended that we seek comment on implementing the small-volume allowance (700
       engine provision) for small equipment manufacturers without a limit on the number of
       engine families that could be covered in any power category, as these Panel members
       were concerned that the Panel's recommended flexibility might not adequately address
       the approximately 50 percent of small business equipment models where the annual sales
       per model is less than 300 and the fixed costs are higher.

   •   An allowance for  small business equipment manufacturers to be able to borrow from the
       TierS/Tier 4 flexibilities for use in the Tier 2/Tier 3 time frame.
                                         11-16

-------
                                                  Small-Business Flexibility Analysis
    The Panel recommended that - similar to the application of flexibility options that are
currently in place - the three transition provisions listed above should be provided to all
equipment manufacturers to maximize the likelihood that the application of these flexibilities
would result in the availability of previous Tier engines for use by the small business equipment
manufacturers.  (See discussion on transition provisions for all equipment manufacturers in
Section III.B of the preamble to this final rule.)

   The Panel also recommended that we seek comment on the need for and value of special
"application-specific" alternative standards for small equipment manufacturers for equipment
configurations that present unusually challenging technical issues for compliance. Further,
Advocacy suggested that we include a technological review of the standards in the 2008
timeframe in the proposal, and the Panel recommended that we consider this.

   The Panel recommended that the following two types of hardship provisions be extended to
small equipment manufacturers:
   •   for the case of a catastrophic event or other extreme unforseen circumstances beyond the
       control of the manufacturer that could not have been avoided with reasonable discretion
       (such as fire, tornado, or supplier not fulfilling contract); and
   •   for the case where a manufacturer has taken all reasonable business, technical, and
       economic steps to comply but cannot.  In this case relief would have to be sought before
       there is imminent jeopardy that a manufacturer's equipment could not be sold and a
       manufacturer would  have to demonstrate to the Agency's satisfaction that failure to get
       permission to sell equipment with a previous Tier engine would create a serious
       economic hardship.  Hardship relief of this nature could not be sought by a manufacturer
       that also manufactures the engines for its own equipment.

   11.6.2.2 What We Proposed

   Following the Panel's recommendation, we proposed both the Percent-of-Production and
Small-Volume Allowances for all equipment manufacturers. Within limits, small  business
equipment manufacturers would be able to continue to use their current engine/equipment
configuration and avoid out-of-cycle equipment redesign until the allowances are exhausted or
the time limit passes. We did not propose the Panel's suggested exemption and annual cap
values; however, we did request comment on these items.  We also requested  comment on
implementing the small-volume allowance provision without the single family limit provision
using caps slightly lower than 700 units, with the provision being applied separately to each
engine power category subject to the proposed standards.

   We also proposed and requested comment on requirements associated with the use of
transition provisions by foreign importers. During the SBREFA Panel process,  the Panel
discussed the possible misuse of the transition provisions by using them as  a loophole to enter
the nonroad diesel equipment market or to gain unfair market position relative to other
manufacturers.  The Panel recognized that this was a possible problem, and believed that the
requirement for small business equipment manufacturers and importers to have  reported
equipment sales using certified engines in model year 2002 or earlier was a sufficient safeguard.

                                         11-17

-------
Final Regulatory Support Document
Upon further analysis, we found that importers of equipment from a foreign equipment
manufacturer could as a group import more excepted equipment from that foreign manufacturer
than 80 percent of that manufacturer's production for the U.S. market or more than the small-
volume allowances identified in the transition provisions. This would create a potentially
significant disparity between the treatment of foreign and domestic equipment manufacturers.
We did not intend this situation, and we believe it is not needed to provide reasonable lead time
for foreign equipment manufacturers.

   To ensure that the transition provisions meet the intended goal of alleviating the burden on
small business  equipment manufacturers, we requested comment on the additional requirement
that only the small business nonroad diesel equipment manufacturer that is most responsible for
the installing engines, and the designing, manufacturing, and assembling processes, would
qualify for the  allowances provided under the small equipment manufacturer transition
provisions.  For importers, only a small importer that produced or manufactured nonroad diesel
equipment would be eligible for these transition provisions. A small importer that does not
manufacture or produce equipment does not face a burden in meeting the standards, and
therefore would not receive any allowances under the transition provisions directly, but could
import exempt equipment if it is covered by an allowance or by transition provisions associated
with a foreign small business equipment manufacturer. We proposed this requirement to transfer
the flexibility offered in the transition provisions to the party with the burden. We would also
allow transition provisions and allowances to be used by foreign small business equipment
manufacturers  in the same way as domestic small business equipment manufacturers, while
avoiding the potential for misuse by importers of unnecessary allowances.

   We also proposed the Panel's recommendation that equipment manufacturers be allowed to
borrow from Tier 4 flexibilities in the Tier2/3 time frame. A more detailed discussion on this
issue, as well as the proposed recommendations for importers, can be found in Section VII.B of
the preamble to the proposed rule, and Section III.B of the preamble to the final rule.

   With regard to the Panel recommendation of a provision allowing small business
manufacturers  to request limited "application-specific" alternative standards for equipment
configurations  that present unusually challenging technical issues for compliance, we requested
comment on this recommendation (in Section VII. C of the preamble to the proposed rule);
however, we did not receive any public comments on this matter. We believed (and continue to
believe) that the transition provisions that we proposed would provide latitude, at least in the
near term, and  a properly structured emission credit program for the engine manufacturers. Even
if one were to assume that these flexibilities provide insufficient lead time (which may not be the
case), application-specific standards would still be cumbersome for both the small business
equipment manufacturers and for the Agency. Further, this provision could potentially have
provided more lead time than could be justified and undermine achievable emission reductions.
Moreover, no participant in the SBAR process offered any empirical support that such a problem
existed, nor have such issues been demonstrated (or raised)  by any equipment manufacturers in
implementing the current nonroad standards. We do note, however, that we are adopting a
Technical Hardship provision for all equipment manufacturers, which allows  a case-by-case
showing of extreme and unpreventable technical difficulty which can justify additional lead time

                                         11-18

-------
                                                 Small-Business Flexibility Analysis
for specific applications. See Section III.B.2.b to the preamble to the final rule.  We believe that
this provision meets some of the concerns voiced by the Panel.

   We proposed that the Panel's recommended hardship provisions be extended to small
business equipment manufacturers in addition to the transition provisions described above.  To
be eligible for these hardship provisions (as well as for the proposed transition provisions),
equipment manufacturers and importers must have reported equipment sales using certified
engines in model year 2002 or earlier.  As discussed earlier, we noted this requirement to thwart
misuse of the provisions as a loophole to enter the nonroad diesel equipment market or to gain
unfair market position relative to other manufacturers and we request comment on this
restriction.  Either relief provision would provide additional lead time for small business
equipment manufacturers for up to two model years based on the circumstances, and hardship
relief would not be available until other allowances have been exhausted.

   11.6.2.3 Provisions in the Final Rule

   We are finalizing many of the transition and hardship provisions that we proposed for small
business nonroad equipment manufacturers, with some modifications as noted below.  Adopting
an alternative on which we solicited comment, the final rule will allow all equipment
manufacturers the opportunity to choose between two options:
    •  manufacturers would be allowed to exempt 700 pieces of equipment over seven years,
       with one engine family; or,
   •   manufacturers using the small-volume allowance could exempt
          525 machines over seven years (with a maximum of 150 in any given year) for each
          of the three power categories below 175 horsepower, and
          350 machines over seven years (with a maximum of 100 in any given year) for the
          two power categories above 175 horsepower.
Concurrent with the revised caps, manufacturers could exempt engines from more than one
engine family under the small-volume allowance program. Based on sales information for small
businesses, we estimated that the alternative small-volume allowance program to include lower
caps and allow manufacturers to exempt more than one engine family would keep the total
number of engines eligible for the allowance at roughly the same overall level as the 700-unit
program.

   We believe that these provisions will afford small manufacturers the type of transition
leeway recommended by the Panel.  Further, these transition provisions could allow small
business equipment manufacturers to postpone any redesign needed on low sales volume or
difficult equipment packages, thus saving decreasing the strain on financial resources and- in
many cases, limited- engineering personnel. Within limits, small business equipment
manufacturers would be able to continue to use their current engine/equipment configuration and
avoid out-of-cycle equipment redesign until the allowances are exhausted or the time limit
passes.

   We are not finalizing the requirement that small equipment manufacturers and importers
have reported equipment sales using certified engines in model year 2002 or earlier. Please see

                                         11-19

-------
Final Regulatory Support Document
Section III.C.2.a.ii of the preamble for a detailed discussion on our decision to eliminate this
requirement from ths rule.

   We are also finalizing three additional provisions. Two of these provisions are being
finalized for all equipment manufacturers, and therefore small business equipment manufacturers
may also take advantage of them.  These are the Technical Hardship Provision and the Early Tier
4 Engine Incentive Program, and are discussed in greater detail in Sections III.B.2.b and e of the
preamble. The third provision is being finalized for small business equipment manufacturers
only, for the 20-50 hp category. This provision is discussed in greater detail in Section
III.C.2.b.ii of the preamble.

11.6.3  Transition and Hardship Provisions for Nonroad Diesel Fuel Small Refiners

   11.6.3.1 Panel Recommendations

   During the SBREF A process, the Panel considered a range of options and regulatory
alternatives for providing small refiners with flexibility in complying with new sulfur standards
for nonroad diesel fuel. Taking into consideration the comments received on these ideas during
the outreach meetings with SERs, as well as additional business and technical information
gathered about potentially affected small entities,  the Panel recommended that whether we
propose a one-step or a two-step approach, we should provide for delayed compliance for small
refiners as shown in Table 11-3 below.

                                        Table  11-3
                       SBREFA Panel Small-Refiner Options Under
                 Potential 1-Step and 2-Step Nonroad Diesel Base Programs
                 Recommended Sulfur Standards (in parts per million, ppm)

l-Step
Program
Non-small b
Small
2-Step
Program
Non-small °
Small
2006

-
—

-
-
2007

-
—

500
-
2008

15
—

500
-
2009

15
-

500
-
2010

15
-

15
500
2011

15
-

15
500
2012

15
15

15
500
2013

15
15

15
500
2014

15
15

15
15
2015+

15
15

15
15
a New standards are assumed to take effect June 1 of the applicable year.
b Assumes 500 ppm standard for marine + locomotive fuel for nonsmall refiners for 2008, and for small refiners for 2012
and later.
0 Assumes 500 ppm standard for marine + locomotive fuel for nonsmall refiners for 2007, and for small refiners for 2010
and later.
                                          11-20

-------
                                                  Small-Business Flexibility Analysis
   The Panel also recommended that we propose certain provisions to encourage early
compliance with lower sulfur standards. The Panel recommended that we propose that small
refiners be eligible to select one of the two following options (with the maximum per-gallon
sulfur cap for any small refiner remaining at 450 ppm):

   •   Credits for Early Desulfurization- The Panel recommended that we propose, as part of an
       overall trading program, a credit trading system that allows small refiners to generate and
       sell credits for nonroad diesel fuel that meets the small-refiner standards earlier than that
       required in the above table. Such credits could be used to offset higher sulfur fuel
       produced by that refiner or by another refiner that purchases the credits.

   •   Limited Relief on Small-Refiner Interim Gasoline Sulfur Standards- The Panel
       recommended that a small refiner producing its entire nonroad diesel fuel pool at 15 ppm
       sulfur by June 1, 2006, and that chooses not to generate nonroad credits for its early
       compliance, receive a 20 percent relaxation in its assigned small-refiner interim gasoline
       sulfur standards.

   The Panel recommended that we propose small refiner hardship provisions modeled after
those established under the gasoline sulfur and highway diesel fuel sulfur programs (see 40 CFR
80.270 and 80.560).  Specifically, it was recommended that we propose a process that, like the
hardship provisions of the gasoline and highway diesel rules, would allow small refiners to seek
case-by-case approval of applications for temporary waivers to the nonroad diesel sulfur
standards, based on a demonstration of extreme hardship circumstances. This provision was
recommended as it would allow domestic and foreign refiners, including small refiners, to
request additional flexibility based on a showing of unusual circumstances resulting in extreme
hardship and significantly affecting the ability of the small refiner to comply by the applicable
date, despite its  best efforts.

   11.6.3.2 What We Proposed

   We proposed the  small refiner transition provisions as recommended by the Panel for a two-
step program (as we chose to propose a two-step fuel implementation program), which are
shown in Table  11-4 below.
                                         11-21

-------
Final Regulatory Support Document
                                       Table 11-4
                Small-Refiner Options 2-Step Nonroad Diesel Base Programs
                Recommended Sulfur Standards (in parts per million (ppm))a
Under 2-Step
Program
Non-smallb
Small
2006
—
—
2007
500
—
2008
500
—
2009
500
—
2010
15
500
2011
15
500
2012
15
500
2013
15
500
2014
15
15
2015+
15
15
a New standards are assumed to take effect June 1 of the applicable year.
b Assumes 500 ppm standard for marine + locomotive fuel for nonsmall refiners for 2007 and later and for small refiners
   for 2010 and later.
   The proposed provisions were to address the concerns that small refiners raised during the
SBREFA process and during the development of the proposal, while still expeditiously
achieving air quality benefits and ensuring timely availability of 15 ppm nonroad diesel fuel for
the introduction of 2011 model year nonroad diesel engines and equipment.

   In accordance with the Panel recommendation of encouraging early compliance with the
standards, we proposed that small refiners be able to choose between the two Panel-
recommended options discussed above ('Credits for Early Desulfurization' and 'Limited Relief
on Small-Refiner Interim Gasoline Sulfur Standards') to provide incentives for such early
compliance. Following the Panel's recommendation, we proposed that the per-gallon cap for
either option could not exceed 450 ppm under any circumstances (this is also consistent with the
gasoline sulfur program).

   For the  'Credits for Early Desulfurization' option, we proposed that a small refiner would be
able to generate NRLM diesel sulfur credits for production of 500 ppm NRLM diesel fuel before
June 1, 2010, and for production of 15 ppm nonroad fuel from June 1,  2010 through May 31,
2012. During discussions with small refiners during the development  of the proposal, some
small refiners indicated that they might find it necessary to produce fuel meeting the nonroad
diesel sulfur standards earlier than required under the small-refiner program.  These small
refiners listed various reasons for this, including: a limited number of grades of diesel fuel that
their respective distribution  systems would carry; economically advantageous to make 500 ppm
or 15 ppm fuel earlier so as not to lose market share; and one small refiner indicated that it may
decide to desulfurize  its nonroad pool at the same time as its highway  diesel fuel, in June of 2006
(due to limitations in  its distribution system and to take advantage of economies of scale).

   For the  option of  'Limited Relief on Small-Refiner Interim Gasoline Sulfur Standards', we
proposed that a small refiner qualifying for this option  would receive a 20 percent revision in its
interim small-refiner  gasoline sulfur standards for the duration of the program (i.e., through
either 2007 or 2010, depending on whether the refiner had extended its participation in the
gasoline sulfur interim program by complying with the highway diesel standard at the beginning
of that program (June, 2006, as provided in 40  CFR 80.552(c))), beginning January 1, 2004. In
                                         11-22

-------
                                                   Small-Business Flexibility Analysis
addition, we proposed that a small refiner wishing to use this option would be required to
produce a minimum of 85 percent of the volume represented by its non-highway distillate
baseline percentage at 15 ppm by June 1, 2006. Further, if the refiner began to produce gasoline
in 2004 at the higher interim standard of this provision but then either failed to meet the 15 ppm
standard for its nonroad fuel or failed to meet the 85 percent requirement, the small refiner's
original interim gasoline sulfur standard would be reinstated.  The refiner would then need to
compensate for the higher gasoline levels that it had enjoyed by either purchasing gasoline sulfur
credits or producing an equivalent volume of gasoline below the required sulfur levels.

   We also requested comment on a slightly different compliance schedule which would have
required small refiners to produce 15 ppm nonroad diesel fuel beginning June 1,  2013,  one year
earlier than proposed above. Such a schedule would align the end of the interim small-refiner
provisions with the end of the proposed phase-in for nonroad engines and equipment and
eliminate higher sulfur nonroad fuel from the distribution system by the time all new engines
required 15 ppm fuel.

   We also proposed  small refiner hardship provisions, as recommended by the  Panel, which are
identical to those offered under the gasoline sulfur and highway diesel fuel sulfur programs.
These provisions would be evaluated on a case-by-case basis to provide short-term relief to those
refiners needing additional lead time due to extreme hardship  circumstances.

   11.6.3.3 Provisions in the Final Rule

   In addition to regulating nonroad diesel fuel to a 15 ppm sulfur limit, we are also finalizing a
15 ppm standard for locomotive and marine diesel fuel. As a result, we have modified  the
proposed provisions to also incorporate flexibility for small refiners in meeting the 15 ppm
locomotive and marine standard. Given the regulatory transition provisions that we are
finalizing for small refiners and small terminal operators, we are confident  about going forward
with the 500 ppm sulfur standard for NRLM diesel fuel in 2007, and the 15 ppm sulfur standard
for nonroad diesel fuel in 2010 and locomotive and marine diesel fuel in 2012, as part of our
general program.

   We are finalizing the Panel's recommendation of delayed compliance for small refiners
along with transition provisions to encourage early compliance with the new standards.  The
transition provisions that we are finalizing for small refiners are as follows:

   •   NRLMDelay Option- Small refiners will be required to comply with the standards set out
       in Table 11-5 below, meeting the 500 ppm sulfur standard in 2010 and the  15 ppm sulfur
       standard in 2014.°  This is identical to the relief proposed in the NPJAM (which  small
       refiners considered sufficient and supported) with the exception that it applies not only to
   D Since new engines with sulfur sensitive emission controls will begin to become widespread during this time,
small refiner fuel will need to be segregated and only supplied for use in pre-2011 nonroad equipment or in
locomotives or marine engines.

                                          11-23

-------
Final Regulatory Support Document
       nonroad fuel, but also to locomotive and marine fuel. However, this delay option is not
       being finalized for the Northeast and Mid-Atlantic areas due to the removal of the heating
       oil marker in these areas (see discussion in Section V of the preamble). Removal of the
       marker provision for heating oil in these areas will help to alleviate the concern raised by
       small terminal operators in their comments regarding the cost of adding a marker to
       heating oil.  At the same time, its removal is not expected to impact small refiners since
       we do not anticipate that they would have marketed fuel in this area. Further, this
       provision will be finalized in Alaska only if a refiner gets an approved compliance plan
       for segregating their fuel to the end user.

       The NRLM Credit Option- Some small refiners have indicated that they might need to
       produce fuel meeting the NRLM diesel fuel sulfur standards earlier than required under
       the small refiner program described above (distribution systems might limit the number
       of grades of diesel fuel that will be carried, it may be economically advantageous to make
       compliant NRLM diesel fuel earlier to prevent losing market share, etc.) This option
       allows small refiners to participate in the NRLM diesel fuel sulfur credit banking and
       trading program discussed in Section IV of the preamble.  Generating and selling credits
       could provide small refiners with funds to help defray the costs of early NRLM
       compliance.

       The NRLM/Gasoline Compliance Option- This option is available to small refiners that
       produce greater than 95 percent of their NRLM diesel fuel  at the 15 ppm sulfur standard
       by June 1, 2006 and elect not to use the provision described above to earn NRLM diesel
       fuel sulfur credits for this early compliance.E For small refiners choosing this option, the
       applicable small refiner annual average and per-gallon cap  gasoline sulfur standards will
       be increased by 20 percent for the duration of the interim program; however, in no case
       may the per-gallon gasoline sulfur cap exceed 450 ppm.
   E This is down from the 100% requirement proposed to allow for some contamination losses in the process of
delivering fuel from the refinery. Production volumes in the final rule are based upon actual delivered volumes. The
5% allowance for greater than 15 ppm fuel should provide adequate flexibility for any refiner's contamination
issues, while not providing any opportunity to significantly alter their compliance plans.

                                          11-24

-------
                                                  Small-Business Flexibility Analysis
                                       Table 11-5
              Sulfur Standards for the NRLM Diesel Fuel Small Refiner Program
                                (in parts per million (ppm))z

Non-Small- NR
Non-Small- LM
Small- all
NRLM
2006
-
-
-
2007
500
500
-
2008
500
500
-
2009
500
500
-
2010
15
500
500
2011
15
500
500
2012
15
15
500
2013
15
15
500
2014
15
15
15
2015+
15
15
15
 Notes:
 a New standards are assumed to take effect June 1 of the applicable year.
   A small refiner may choose to use the relaxed standards (the NRLM Delay option), the
NRLM Credit option, or both in combination.  Thus any fuel that it produces from crude at or
below the sulfur standards earlier than required will qualify for generating credits. However, the
NRLM/Gasoline Compliance option may not be used in combination with either the NRLM
Delay option or the NRLM Credit option, since a small refiner must produce at least 85 percent
of its NRLM diesel fuel at the 15 ppm sulfur standard under the NRLM/Gasoline Compliance
option.

   Combined with the transition provisions for small refiners, the  compliance schedule that we
are adopting will achieve the air quality benefits of the nonroad diesel program as soon as
possible, while helping to ensure that small refiners will have adequate time to raise capital for
new or upgraded fuel desulfurization equipment.  Most small refiners have limited additional
sources of income beyond refinery earnings for financing and typically do not have the financial
backing that larger and generally more integrated companies have. They can therefore benefit
from this additional time to accumulate capital internally or to secure capital financing from
lenders. This will help to offset the disproportionate financial burden facing small refiners.

   We recognize that while the sulfur levels in the proposed program can be achieved using
conventional refining technologies, new technologies are also  being developed that may reduce
the capital and/or operational costs of sulfur removal. We believe  that allowing small refiners
some additional time for newer technologies to be proven out by other refiners may have the
added benefit of reducing the risks faced by small refiners. Further, this additional time may
also  increase the availability of engineering and construction resources.  Some refiners will need
to install additional processing equipment to meet the nonroad diesel sulfur requirements.
Vendors will be more likely to contract their services with the larger refiners first, as their
projects will offer larger profits for the vendors.  Therefore, we anticipate that there may be
significant competition for technology services, engineering resources, and construction
management and labor. Temporarily delaying compliance for small refiners will allow for lower
costs of improvements in desulfurization technology and would spread out the demand for
construction and engineering resources, and likely reduce any  cost premiums  caused by limited
supply.
                                          11-25

-------
Final Regulatory Support Document
11.6.4 Transition and Hardship Provisions for Nonroad Diesel Fuel Small Distributors and
Marketers

    11.6.4.1 Panel Recommendations

    During the SBREFA process, we were considering both a one-step fuel approach, and the
two-step approach that we are finalizing.  The Panel recognized that a two-step fuel approach
would include the possibility of there being two grades of nonroad diesel fuel in the market place
for at least a transition period, the Panel recommended that we study the issue of multiple fuel
grades in the distribution system further during our development of the NPRM. In discussions
that took place during the SBREFA process, distributors supported a one-step  approach as it
would have no significant impact on their operations.  However, they did offer suggestions on
how they might deal with this issue, but indicated that there would be adverse  impacts in some
circumstances.  (A more complete discussion of costs and related issues relevant to fuel
distributors under the proposed program is located in Chapter 7 of the Draft Regulatory Impact
Analysis.)

    11.6.4.2 What We Proposed

    Our proposed fuel sulfur program was designed to minimize the need for additional product
segregation and the associated feasibility and cost issues for fuel distributors associated with it.
Beyond the accommodation of fuel distributor concerns during the overall design of the
proposed program, we did not believe it possible for us to provide special provisions for
particular (i.e., small) fuel distributors to further limit the potential impact of the proposed rule.
However, to allow for a smooth transition of diesel  fuel in the distribution system to 15 ppm, we
proposed that parties downstream of the refineries be allowed a small amount of additional time
to turnover their tanks to 15 ppm. Specifically, we  proposed that at the terminal level, nonroad
diesel fuel would be required to meet the 15 ppm standard beginning July  15, 2010. At bulk
plants, wholesale purchaser-consumers, and any retail stations carrying nonroad diesel, this fuel
would have to meet the  15 ppm standard by September 1, 2010. The proposed transition
schedule for compliance with the 15 ppm standard at refineries, terminals, and secondary
distributors would be the same as those allowed under the recently promulgated highway diesel
fuel program. Lastly, to avoid the costs associated with segregating 500 ppm NRLM diesel fuel
from 500 ppm highway fuel, we proposed that the existing requirement that NRLM diesel fuel
be dyed leaving the refinery would need to be made voluntary (this element of the proposed rule
is discussed in more detail in Section 11.7 of the proposed RIA).

    11.6.4.3 Provisions in the Final Rule

    We are finalizing provisions to alleviate the problems raised in the public comments on our
NPRM regarding small terminal operators (heating oil marker requirements would force small
terminal operators to install expensive injection equipment  and they would not be able to recoup
these costs). To decrease the burden on these  small operators, we are not requiring the addition
of a fuel marker to home heating oil for terminals in much of PADD 1 (Northeast/Mid-Atlantic
Area). This Northeast/Mid-Atlantic Area covers the vast majority of heating oil that will be

                                         11-26

-------
                                                   Small-Business Flexibility Analysis
marketed; however, we are not allowing small refiner or credit fuel to be sold in the
Northeast/Mid-Atlantic Area. Further, we expect that few terminals outside of
Northeast/Mid-Atlantic Area would need to put in injection equipment, since very little fuel
above 500 ppm will be marketed outside of this area except directly from the refinery gate.

11.7 Conclusion

    Throughout the entire rulemaking process, we conducted substantial outreach- including
convening a Panel during the SBREFA process as well as meetings with  other stakeholders- to
gather information about the effect of this final rule on small entities.  We also took into account
comments received during the public comment period and information from contractor studies in
developing regulatory transition provisions to ease the burden on small entities. From this
information (and performing a cost-to-sales ratio test- a ratio of the estimated annualized
compliance costs to the value of sales per company)F, we found that approximately 4 percent (13
companies) of small entities in the engine and equipment manufacturing industry were affected
between 1 and 3 percent of sales (i.e., the estimated costs of compliance with the final rule will
be greater than 1 percent, but less than 3 percent, of their sales).  One percent of small entities (4
companies) were affected at greater than 3 percent. In all, 17 of the 518 potentially affected
small engine and equipment manufacturers are estimated to have compliance costs that could
exceed 1 percent of their sales.

    Similarly, small refiners in general would likely experience a significant and disproportionate
financial hardship in complying with the fuel-sulfur requirements in this rule.  One indication of
this disproportionate hardship for small refiners is the relatively  high projected cost per gallon
for producing compliant nonroad diesel fuel. Refinery modeling (of all refineries) indicates that
without special provisions,  refining costs for small refiners on average would be about 2.3 cents
per gallon higher than the costs for non-small refiners. The majority of the cost for meeting the
fuel requirements in this final rule are related to refining, with only 15  to 25 percent of the
estimated costs being related to distribution.  Allowing highway and off-highway  diesel  fuel
meeting the same sulfur specification to be shipped fungibly until it leaves the terminal obviates
the need for additional storage tankage in this segment of the distribution system.0 The final rule
allows 500 ppm highway and 500 ppm NRLM fuel to be shipped fungibly as proposed.
However, it also allows high sulfur NRLM and heating oil to be shipped  fungibly. Furthermore,
the final  rule allows 500 ppm off-highway diesel engine  fuel to be mixed with high-sulfur diesel
fuel as long as its designation changes.
   F The cost-to-sales ratio test assumes that control costs are completely absorbed by each entity and does not
account for or consider interaction between manufacturers/producers and consumers in a market context.

   G Including the refinery, pipeline, marine tanker, and barge segments of the distribution system.

                                          11-27

-------
CHAPTER 12: Regulatory Alternatives
    12.1 Overview	12-1
    12.2 Description of Options	12-2
       12.2.1 One-Step Options 	12-3
       12.2.2 Two-Step Options	12-7
          12.2.2.1 Options Evaluated for Proposal	12-7
          12.2.2.2 Option 5c	12-18
             12.2.2.2.1 Emission Inventory Impacts 	12-19
             12.2.2.2.2 Cost Analysis	12-20
             12.2.2.2.3 Benefits Comparison	12-20
             12.2.2.2.4 Costs Per Ton	12-21

-------
                                                             Regulatory Alternatives
               CHAPTER 12: Regulatory Alternatives

   Our final program represents a combination of engine and fuel standards and their associated
timing that we believe to be superior to the alternatives considered given feasibility, cost, and
environmental impact. In this chapter we present the alternative program options that we
evaluated in order to make this determination.  These alternatives are cast as twelve specific
program options.

12.1  Overview

   In the Draft RIA supplementing our Notice of Proposed Rulemaking, we presented a detailed
analysis of twelve specific program options.  These options were used to illustrate variations in
both the timing and level of the engine and fuel standards, as well as the applicability of those
standards to different segments of off-highway engines and fuels.  To evaluate each option, we
conducted emission-inventory modeling, estimated costs and benefits, and calculated cost-
effectiveness. We then assessed the appropriateness of each option in comparison to our
proposed engine and fuel program, and presented our rationale for not proposing to implement
each of the options.

   Following release of the proposal, we received comments on some of the options that we
evaluated. Our detailed responses to those comments can be found in Section 8 of the Summary
and Analysis of Comments document. Our reasoning set forth in Chapter 12 of the Draft RIA
supporting the proposal  also still applies  as well for options we have not adopted.

   We examined the costs, inventory impacts, benefits, and cost-effectiveness of each of the
options as presented in the Draft RIA incrementally to our proposed program.  Given that the
final program includes some elements that differ from the proposed program, these same new
elements would need to  be included in each of the options in order to maintain the same
incremental differences  in program structure between the final program and each option. As a
result, we do not believe that a complete  revision to the calculated values for costs, inventory
impacts, benefits, and cost-effectiveness  is warranted, since we would expect them to be very
similar to those presented in the Draft RIA. Also, we would not expect recent modifications to
the NONROAD emissions model to change the incremental differences between the final
program and each of the options. We refer the reader to the detailed evaluations of the options
presented in the Draft RIA.

   The remainder of this section will present a description of the twelve options originally
evaluated in the context of the NPRM to  remind readers of the program issues we investigated.
However, during the course of reviewing comments on our proposed  program, we determined
that an additional evaluation of small  engine standards was warranted. This additional scenario
was labeled Option 5c, and the results of that evaluation are presented below In Section 12.2.2.2.
                                         12-1

-------
Final Regulatory Impact Analysis
12.2 Description of Options

   Our proposed emission-control program consisted of a two-step program to reduce the sulfur
content of nonroad diesel fuel in conjunction with the NOx and PM engine standards.  During
the development of our program, we also considered a one-step fuel program wherein all sulfur
reductions in the diesel fuel occur in a single step. Since the fuel provisions and timing dictate to
a large extent the possible engine standards, we structured this section to first discuss issues of
variations in the fuel program. Thus, the Program Options are divided into One-Step and Two-
Step options, to highlight the fuel sulfur program and its driving impact on the engine standards.
Within each of these fuel program approaches, we considered several variations and
combinations with engine standards.

   This Section  provides both text summaries of each Program Option as well as diagrams
showing how the engine and fuel standards would be implemented over time. For the diagrams,
previous standards  are labelled as Tier 1, Tier 2, or Tier 3 as appropriate.  For reference, Figure
12.2-1  shows the actual standards associated with Tier 1, Tier 2, and Tier 3 labels (40 CFR
89.112).
                                          12-2

-------
                                                               Regulatory Alternatives
                                      Figure 12.2-1
                            Existing Engine and Fuel Standards
hp group
2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015
Nonroad engine standards (g/bhp-hr)a
hp<25
25  750
Tier 2: 5.6 NOx+NMHC, 0.6 PM
Tier 2: 5.6 NOx+NMHC, 0.4 PM
Tier 2: Tier 3:
5.6 NOx+NMHC 3.5 NOx+NMHC
0.3 PM 0.3 PM
Tier 2: Tier 3:
4.9 NOx+NMHC 3 .0 NOx+NMHC
0.2 PM 0.2 PM
Tier 2: Tier 3:
4.8 3.0 NOx+NMHC
NOx+NMHC 0. 1 PM
0.1PM
Tierl: Tier 2:
6.9 NOx 4.8 NOx+NMHC
0.4PM 0.1PM
Fuel sulfur standard (ppm)
Loco &
marine
Nonroad
Uncontrolled
Uncontrolled
" Applies to model years.
12.2.1 One-Step Options

   One-step options were those in which the fuel sulfur standard was applied in a single step;
there were no phase-ins or step changes. In all one-step options, the transient test cycle was
required concurrently with the introduction of the transitional Tier 4 engine standards in any
horsepower group.

   Option  la differed from Options 1 and Ib in terms of the engine standards and their
associated timing.  Option Ib differed from Option 1 only in the timing of the fuel sulfur
standard, and was intended to generate additional early sulfate PM reductions. As a result, we
did not lower the certification fuel sulfur level to 15ppm in 2007 and 2008 when modeling this
Option, since doing so would permit manufacturers to take advantage of the lower sulfur and
thus reduce the PM reductions associated with their certified engines.
                                           12-2

-------
Final Regulatory Impact Analysis
   The one-step options are summarized in Table 12.2.1-1.  The specifics of the three one-step
options are shown in the standard charts in Figures 12.2.1-2, 3, and 4, while the previous Tier 1,
Tier 2, and Tier 3 standards were shown in Figure 12.2-1.  Only changes to the standards are
shown in these three figures, i.e. if no new standard for a given pollutant is indicated, the
previous standard applies.

                                     Table 12.2.1-1
                              Summary of One-Step Options
Option
Option 1
Option la
Option Ib
Summary Description
• Fuel sulfur < 15ppm in June 2008 for nonroad, < SOOppm for locomotives and
marine engines
• <50 hp: PM stds only in 2009
• 25-75 hp: PM aftertreatment -based standards and EGR or equivalent NOx
technology in 20 1 3 ; no NOx aftertreatment
• >75 hp: PM aftertreatment-based standards phasing in beginning in 2009; NOx
aftertreatment-based standards phasing in beginning in 20 1 1
See Figure 12.2.1-1
• Fuel sulfur < 15ppm in June 2008
• PM aftertreatment-based standards introduced in 2009-10
• NOx aftertreatment-based standards introduced in 201 1-12
See Figure 12.2.1-2
Same as Option la, except fuel sulfur standard required two years earlier
See Figure 12.2.1-3
                                          12-4

-------
                                                                      Regulatory Alternatives
                                         Figure 12.2.1-1
                           Engine and Fuel Standards Under Option 1
hp group
2005
2006
2007
2008
2009
2010
2011
2012
2013
2014 2015
Nonroad engine standards (g/bhp-hr)a
hp<25
25  750

Tierl
Tier 2




TierS




0.30
0.22

50%: 0
Tier 2
PM
PM

01PM

0.02PM, 3. 3Y NOx
0.01 PM
50%: 0.30 NOx
50%: 0.0 1PM, 0.30 NOx
0.30 NOx
Fuel sulfur standard (ppm/
Loco &
marine
Nonroad
Uncontrolled
Uncontrolled
500 ppm
15 ppm
" Applies to model years.  If no standard is shown for a given pollutant, the previous standard applies.
p Applies to calender years.  Begins in June.
Y Actual standard is 3.5g/bhp-hr NOx+NMHC, equivalent to the Tier 3 standard for 50-75hp. For modeling
purposes, NOx portion of this standard is assumed to be 3.3g/bhp-hr.
                                               12-5

-------
Final Regulatory Impact Analysis
                                  Figure 12.2.1-2
                      Engine and Fuel Standards Under Option la
hp group
2005

hp<25
25  750

Tierl
2006

Tier 2

2007
2008
Nonroad


2009
2010
2011
2012 2013 2014 2015
engine standards (g/bhp-hr)a

TierS
Tier 2


0.01
PM

0.30 NOx
Fuel sulfur standard (ppm/
Loco &
marine
Nonroad
Uncontrolled
Uncontrolled
15 ppm
15 ppm
" Applies to model years. If no standard is shown for a given pollutant, the previous standard applies.
p Applies to calender years. Begins in June.
                                      12-6

-------
                                                              Regulatory Alternatives
                                     Figure 12.2.1-3
                        Engine and Fuel Standards Under Option Ib
hp group
2005
2006

hp<25
25  750

Tierl
Tier 2

2007



2008
Nonroad
2009
2010
2011
2012 2013 2014 2015
engine standards (g/bhp-hr)a

TierS
Tier 2


0.01
PM

0.30 NOx
Fuel sulfur standard (ppm/
Loco &
marine
Nonroad
Uncont
rolled
Uncont
rolled
15 ppm
15 ppm
" Applies to model years. If no standard is shown for a given pollutant, the previous standard applies.
p Applies to calender years. Begins in June.
12.2.2 Two-Step Options

   Two-step options were those in which the fuel sulfur standard was set first at SOOppm for
several years, and then was lowered further to 15ppm.  The exact timing of the introduction of
the SOOppm and the 15ppm standards varied among each of the two-step options.  In addition,
we considered a variety of engine standards and phase-ins.  In the two-step options, the transient
test cycle was required concurrently with the introduction of the transitional  Tier 4 engine
standards. The one exception was Option 5b, under which the existing steady-state test applied
indefinitely for engines below 75 hp.

       12.2.2.1 Options Evaluated for Proposal

   The proposed program formed the basis for all of the two-step alternative program options.
The two-step options that we evaluated for the NPRM are summarized in Table 12.2.2-1.  The
specifics of these two-step options are shown in the standard charts in Figures 12.1.2-2 through
                                          12-7

-------
Final Regulatory Impact Analysis
11, while the previous Tier 1, Tier 2, and Tier 3 standards were shown in Figure 12.2-1. As for
the one-step standard charts, only changes to the standards are shown, i.e. if no new standard for
a given pollutant is indicated, the previous standard applies.
                                            Table 12.2.2-1
                                   Summary of Two-Step Options
 Option
Summary Description
 Proposed program
• 500 ppm in 2007; 15 ppm in 2010 for nonroad engines only
• >25 hp: PM aftertreatment-based standards introduced 2011-2013
• >75 hp: NOx aftertreatment-based standards introduced and phased-in 2011-2014
• <25 hp: PM standards in 2008
• 25-75 hp: PM standards in 2008 (optional for 50-75 hp)
• >750hp: PM and NOx standards phased-in 2011-2014
See Figure 12.2.2-1
 Option 2a
Same as our proposed program, except:
• Transitional sulfur standard of 500 ppm is introduced one year earlier
See Figure 12.2.2-2
 Option 2b
Same as our proposed program, except:
• Final sulfur standard of 15 ppm is introduced one year earlier
• Trap-based PM standards begin one year earlier for all engines
See Figure 12.2.2-3
 Option 2c
Same as our proposed program, except:
• Final sulfur standard of 15 ppm is introduced one year earlier
• Trap-based PM standards begin one year earlier for 175 - 750 hp engines
See Figure 12.2.2-4
 Option 2d
Same as our proposed program, except:
• Final NOx standard for 25 - 75 hp engines is lowered to 0.30 g/bhp-hr
• A phase-in for the NOx standard for this horsepower group is included
See Figure 12.2.2-5
 Option 2e
Same as our proposed program, except:
• No new Tier 4 NOx standards.
See Figure 12.2.2-6
 Option 3
Same as our proposed program, except:
• Above-ground mining equipment >750 hp remains at the Tier 2 standards
See Figure 12.2.2-7
 Option 4
Same as our proposed program, except:
• 15 ppm final sulfur standard applies to fuel used by locomotives and marine engines in
addition to all other nonroad engines
See Figure 12.2.2-8
 Option 5 a
Same as our proposed program, except:
• No new Tier 4 standards for <75 hp engines
See Figure 12.2.2-9
 Option 5b
Same as our proposed program, except:
• No trap-based PM standards for <75 hp engines
• No new Tier 4 NOx standards for <75 hp engines
See Figure 12.2.2-10
                                                 12-8

-------
                                                                       Regulatory Alternatives
                                          Figure 12.2.2-1
                     Engine and Fuel Standards under the Proposed Program
hp group
2005

hp<25
25  750

Tierl
2006
2007
2008
Nonroad
Tier 2



2009
2010
2011
2012
2013
2014 2015
engine standards (g/bhp-hr)a
0.30 PM
0.22 PM
TierS
Tier 2


0.02PM, 3. 3e NOx
100%Y : 0.01 PM
50%Y : 0.30 NOx
50%8 : 0.0 1PM, 0.30 NOx
0.01 PM
0.30 NOx
Fuel sulfur standard (ppm/
Loco &
marine
Nonroad
Uncontrolled
Uncontrolled

500 ppm
500 ppm
15 ppm
a Applies to model years. If no standard is shown for a given pollutant, the previous standard applies.
p Applies to calender years. Begins in June.
Y All engines must meet 0.01 PM, but only 50% of engines must meet the new NOx standard of 0.30. All engines
must use the transient test cycle.
8 Only 50% of engines must meet both the new PM and NOx standards on the transient test cycle. Remaining
engines meet Tier 2 standards on the steady-state test cycle.
e Actual standard is 3.5g/bhp-hr NOx+NMHC, equivalent to the Tier 3  standard for 50-75hp. For modeling
purposes, NOx portion of this standard is assumed to be 3.3g/bhp-hr.
                                                12-9

-------
Final Regulatory Impact Analysis
                                          Figure 12.2.2-2
                           Engine and Fuel Standards under Option 2a
hp group
2005
2006

hp<25
25  750

Tierl
Tier 2

2007
2008
Nonroad


2009
2010
2011
2012
2013
2014 2015
engine standards (g/bhp-hr)a
0.30 PM
0.22 PM
•
TierS

Tier 2


50%8 : 0

0.02PM, 3. 3e NOx
100%Y : 0.01 PM
50%Y : 0.30 NOx
01 PM, 0.
30 NOx
0.01 PM
0.30 NOx
Fuel sulfur standard (ppm/
Loco &
marine
Nonroad
Uncon-
trolled
Uncon-
trolled

500 ppm

500
ppm

15 ppm
a Applies to model years. If no standard is shown for a given pollutant, the previous standard applies.
p Applies to calender years. Begins in June.
Y All engines must meet 0.01 PM, but only 50% of engines must meet the new NOx standard of 0.30. All engines
must use the transient test cycle.
8 Only 50% of engines must meet both the new PM and NOx standards on the transient test cycle. Remaining
engines meet Tier 2 standards on the steady-state test cycle.
e Actual standard is 3.5g/bhp-hr NOx+NMHC, equivalent to the Tier 3 standard for 50-75hp. For modeling
purposes, NOx portion of this standard is assumed to be 3.3g/bhp-hr.
                                               12-10

-------
                                                                        Regulatory Alternatives
                                          Figure 12.2.2-3
                            Engine and Fuel Standards under Option 2b
hp group
2005
2006
2007
2008
Nonroad
hp<25
25  750

Tierl
Tier 2



2009
2010
2011
2012
2013
2014 2015
engine standards (g/bhp-hr)a
0.30 PM
0.22 PM
•
TierS
Tier 2

0.01
PM
50%:
0.01
PM
0.01
PM

0.02
PM

0.02PM, 3. 3e NOx
50%Y : 0.30 NOx
50%8 : 0.01 PM,
0.30 NOx
100%:
0.01
PM
0.01 PM
0.30 NOx
Fuel sulfur standard (ppm)p
Loco &
marine
Nonroad
Uncontrolled
Uncontrolled

500 ppm
500 ppm
15 ppm
a Applies to model years. If no standard is shown for a given pollutant, the previous standard applies.
p Applies to calender years. Begins in June.
Y All engines must meet 0.01 PM, but only 50% of engines must meet the new NOx standard of 0.30. All engines
must use the transient test cycle.
8 Only 50% of engines must meet both the new PM and NOx standards on the transient test cycle. Remaining
engines meet Tier 2 standards on the steady-state test cycle.
e Actual standard is 3.5g/bhp-hr NOx+NMHC, equivalent to the Tier 3 standard for 50-75hp. For modeling
purposes, NOx portion of this standard is assumed to be 3.3g/bhp-hr.
                                                12-11

-------
Final Regulatory Impact Analysis
                                          Figure 12.2.2-4
                           Engine and Fuel Standards under Option 2c
hp group
2005

hp<25
25  750

Tierl
2006
2007
2008
Nonroad
Tier 2



2009
2010
2011
2012
2013
2014 2015
engine standards (g/bhp-hr)a
0.30 PM
0.22 PM

TierS
Tier 2

0.01
PM



0.02PM, 3. 3e NOx
100%Y : 0.01 PM
50%Y : 0.30 NOx
50%8 : 0.0 1PM, 0.30 NOx
0.01 PM
0.30 NOx
Fuel sulfur standard (ppm/
Loco &
marine
Nonroad
Uncontrolled
Uncontrolled

500 ppm
500 ppm
15 ppm
a Applies to model years. If no standard is shown for a given pollutant, the previous standard applies.
p Applies to calender years. Begins in June.
Y All engines must meet 0.01 PM, but only 50% of engines must meet the new NOx standard of 0.30. All engines
must use the transient test cycle.
8 Only 50% of engines must meet both the new PM and NOx standards on the transient test cycle. Remaining
engines meet Tier 2 standards on the steady-state test cycle.
e Actual standard is 3.5g/bhp-hr NOx+NMHC, equivalent to the Tier 3 standard for 50-75hp. For modeling
purposes, NOx portion of this standard is assumed to be 3.3g/bhp-hr.
                                               12-12

-------
                                                                       Regulatory Alternatives
                                          Figure 12.2.2-5
                           Engine and Fuel Standards under Option 2d
hp group
2005
2006
2007
2008
2009
2010
2011
2012
2013
2014 2015 2016
Nonroad engine standards (g/bhp-hr)a
hp<25
25  750

Tierl
Tier 2



0.30 PM
0.22 PM
TierS
Tier 2


100%Y :
50%Y :
NOx
50%8 : 0.01 PM,
NOx
0.30
0.02 PM NOx
50%: 0.30 NOx
0.01 PM
0.30
0.30
0.01 PM
0.30 NOx
Fuel sulfur standard (ppm/
Loco &
marine
Nonroad
Uncontrolled
Uncontrolled
500 ppm
500 ppm
15 ppm
" Applies to model years. If no standard is shown for a given pollutant, the previous standard applies.
p Applies to calender years.  Begins in June.
Y All engines must meet 0.01 PM, but only 50% of engines must meet the new NOx standard of 0.30. All engines
must use the transient test cycle.
8 Only 50% of engines must meet both the new PM and NOx standards on the transient test cycle. Remaining
engines meet Tier 2 standards on the steady-state test cycle.
                                               12-13

-------
Final Regulatory Impact Analysis
                                        Figure 12.2.2-6
                          Engine and Fuel Standards under Option 2e
hp group

hp<25
25  750

Loco &
marine
Nonroad
2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015
Nonroad engine standards (g/bhp-hr)a
0.30 PM
0.22 PM 0.02 PM
Tier 2

Tier3 0.01PM

Tierl Tier 2 50%8 : 0.01 PM 0.01PM
Fuel sulfur standard (ppm/
Uncontrolled cnn
500 ppm
Uncontrolled 500 ppm 15 ppm
" Applies to model years. If no standard is shown for a given pollutant, the previous standard applies.
p Applies to calender years. Begins in June.
8 Only 50% of engines must meet the new PM standard on the transient test cycle. Remaining engines meet Tier 2
standards on the steady-state test cycle.
                                              12-14

-------
                                                                        Regulatory Alternatives
                                          Figure 12.2.2-7
                            Engine and Fuel Standards under Option 3
hp group
2005
2006
2007
2008
Nonroad
hp<25
25  750

Tierl
Tier 2



2009
2010
2011
2012
2013
2014 2015
engine standards (g/bhp-hr)a
0.30 PM
0.22 PM
•
TierS

Tier 2



0.02PM, 3. 3e NOx
100%Y : 0.01 PM
50%Y : 0.30 NOx
50%8 : 0.01 PM, 0.
Mining equipment
at Tier 2
30 NOx
remains
0.01 PM
0.30 NOx
0.01 PM
0.30 NOx
Mining
equipment
at Tier 2
Fuel sulfur standard (ppm/
Loco &
marine
Nonroad
Uncontrolled
Uncontrolled

500 ppm
500
ppm

15 ppm
" Applies to model years. If no standard is shown for a given pollutant, the previous standard applies.
p Applies to calender years. Begins in June.
Y All engines must meet 0.01 PM, but only 50% of engines must meet the new NOx standard of 0.30. All engines
must use the transient test cycle.
8 Only 50% of engines not used in mining equipment must meet both the new PM and NOx standards on the
transient test cycle.  Remaining engines meet Tier 2 standards on the steady-state test cycle.
e Actual standard is 3.5g/bhp-hr NOx+NMHC, equivalent to the Tier 3 standard for 50-75hp. For modeling
purposes, NOx portion of this standard is assumed to be 3.3g/bhp-hr.
                                                12-15

-------
Final Regulatory Impact Analysis
                                          Figure 12.2.2-8
                            Engine and Fuel Standards under Option 4
hp group
2005

hp<25
25  750

Tierl
2006
2007
2008
Nonroad
Tier 2



2009
2010
2011
2012
2013
2014 2015
engine standards (g/bhp-hr)a
0.30 PM
0.22 PM

TierS

Tier 2


0.02PM, 3. 3e NOx
100%Y : 0.01 PM
50%Y : 0.30 NOx
50%8 : 0.0 1PM, 0.30 NOx
0.01 PM
0.30 NOx
Fuel sulfur standard (ppm/
Loco &
marine
Nonroad
Uncontrolled
Uncontrolled
500 ppm
500 ppm
15 ppm
15 ppm
a Applies to model years. If no standard is shown for a given pollutant, the previous standard applies.
p Applies to calender years. Begins in June.
Y All engines must meet 0.01 PM, but only 50% of engines must meet the new NOx standard of 0.30. All engines
must use the transient test cycle.
8 Only 50% of engines must meet both the new PM and NOx standards on the transient test cycle. Remaining
engines meet Tier 2 standards on the steady-state test cycle.
e Actual standard is 3.5g/bhp-hr NOx+NMHC, equivalent to the Tier 3 standard for 50-75hp. For modeling
purposes, NOx portion of this standard is assumed to be 3.3g/bhp-hr.
                                               12-16

-------
                                                                       Regulatory Alternatives
                                          Figure 12.2.2-9
                           Engine and Fuel Standards under Option 5a
hp group
2005

hp<25
25  750

Tierl
2006
2007
2008
Nonroad
Tier 2



2009 | 2010
2011
2012
2013
2014 2015
engine standards (g/bhp-hr)a

TierS
Tier 2



100%Y : 0.01 PM
50%Y : 0.30 NOx
50%8 : 0.01 PM, 0.
30 NOx
0.01 PM
0.30 NOx
Fuel sulfur standard (ppm/
Loco &
marine
Nonroad
Uncontrolled
Uncontrolled
500 ppm
500 ppm
| 15 ppm
a Applies to model years. If no standard is shown for a given pollutant, the previous standard applies.
p Applies to calender years.  Begins in June.
Y All engines must meet 0.01 PM, but only 50% of engines must meet the new NOx standard of 0.30. All engines
must use the transient test cycle.
8 Only 50% of engines must meet both the new PM and NOx standards on the transient test cycle. Remaining
engines meet Tier 2 standards on the steady-state test cycle.
                                               12-17

-------
Final Regulatory Impact Analysis
                                      Figure 12.2.2-10
                         Engine and Fuel Standards under Option 5b
hp group
2005

hp<25
25  750

Tierl
2006
2007
2008
Nonroad
Tier 2



2009
2010
2011
2012
2013
2014 2015
engine standards (g/bhp-hr)a
0.30 PM
0.22 PM

TierS

Tier 2


100%Y : 0.01 PM
50%Y : 0.30 NOx
50%8 : 0.01 PM, 0.
30 NOx
0.01 PM
0.30 NOx
Fuel sulfur standard (ppm/
Loco &
marine
Nonroad
Uncontrolled
Uncontrolled

500 ppm
500 ppm
15 ppm
a Applies to model years. If no standard is shown for a given pollutant, the previous standard applies.
p Applies to calender years. Begins in June.
Y All engines must meet 0.01 PM, but only 50% of engines must meet the new NOx standard of 0.30. All engines
must use the transient test cycle.
8 Only 50% of engines must meet both the new PM and NOx standards on the transient test cycle. Remaining
engines meet Tier 2 standards on the steady-state test cycle.
       12.2.2.2 Option 5c

    As described in Section 12.2.2.1, Option 5b represented an alternative program in which we
would not apply trap-based PM standards or new NOx standards to engines under 75hp.  As
described in Sections II. A and II.B of the preamble, we continue to believe that the application
of PM filters to small engines is both feasible and is an important element of our efforts to
address air quality concerns associated with nonroad engines.  Therefore, we have not finalized
Option 5b and the proposed Tier 4 PM and NOx standards for <75hp engines are included in the
program we are finalizing.

    Some of the original concerns raised about <75hp engines were again raised in response to
the NPRM for a smaller group of engines with rated horsepower between 25 and  50 hp. In the
process of considering this issue, we evaluated a new Option 5c in which the trap-based PM
                                           12-18

-------
                                                               Regulatory Alternatives
standard and the Tier 4 NOx standard would not be applied to 25 - 50 hp engines, but would
continue to apply to above 50 hp engines.  This specific option is a refinement of Option 5b, but
was not evaluated for the NPRM. We evaluated this Option 5c as part of our overall evaluation
of a wide variety of alternative options.  We are presenting the results of our analysis here.

   As described above, we did not repeat the analyses for Options 1  through 5b for this final
rule, but instead refer the reader to the draft RIA for those analyses. The draft RIA presented the
inventory impacts, benefits, costs, and cost-effectiveness of each of the options in comparison to
the proposed program. For Option 5c, however, we evaluated the inventory impacts, benefits,
costs, and cost-effectiveness in comparison to the final program.
       12.2.2.2.1 Emission Inventory Impacts

   Option 5c is identical to our final program, except that it would not require 25-50hp engines
to meet the trap-based PM standards that are in our final program, nor would it require these
engines to meet the Tier 4 NOx standards. As a result, the PM and NOx emission reductions for
Option 5c would be lower than those for our final program.  However, under this option
pollutants other than PM and NOx would also be affected.  For instance, the reductions in
hydrocarbons and CO that will occur for our final program are generated primarily through the
presence of catalyzed diesel particulate traps, so the removal of the trap-based PM standards for
25-50 hp engines will also produce a corresponding reduction in the HC and CO benefits.

   In evaluating the inventory impacts of Option 5c, we assumed that the 2008 PM standards for
25-50 hp engines were met using a steady-state test cycle for both our final program and Option
5c.  Whether these engines should  be required to meet standards under a transient test procedure
is a separate issue from the use of after-treatment. Our analysis was designed to focus in the
impacts of requiring the use of aftertreatment.

   Thus Option 5c produces fewer benefits for all pollutants starting in 2013 in comparison to
our final program.  Table 12.2.2.2.1-1 shows the net impact of Option 5c on the 30-year net
present value inventory estimates.

                                    Table 12.2.2.2.1-1
                  50-State 30-Year Net Present Value Emission Increases
                    For Option 5c In Comparison to Final Program (tons)

PM
NOx + NMHC
3% discount rate
56,833
381,459
7% discount rate
25,238
170,819
                                         12-19

-------
Final Regulatory Impact Analysis
       12.2.2.2.2 Cost Analysis

   Option 5c would reduce the overall costs of the program since 25-50 hp engines would not
need to install PM traps nor make engine modifications to comply with more stringent NOx
standards. We calculated the total nationwide cost savings by summing the per-engine savings
across all engines for each year starting in 2013.  Table 12.2.2.2.2-1 shows the resulting 30-year
net present value cost savings for Option 5c. Costs were allocated to the various pollutants
according to the methodology described in Chapter 8 of the RIA.

                                    Table 12.2.2.2.2-1
                     50-State 30-Year Net Present Value Cost Savings
                  For Option 5c In Comparison to Final Program (Smillion)

All pollutants
PM
NOx + NMHC
3% discount rate
2,041
1,514
527
7% discount rate
997
735
263
       12.2.2.2.3 Benefits Comparison

   We were able to estimate the benefits of Option 5c using the benefit-transfer methodology
developed in Chapter 9 for estimating the monetized benefits of the final program. The specific
methodology is described in Section 9.5 "Development of Intertemporal Scaling Factors and
Calculation of Benefits Over Time" and will not be repeated here.  To use that methodology
requires input of 48-state emission reductions for NOx, PM2.5 and SO2 associated with Option
5c.  We cannot estimate 50-state benefits due to the fact that our air quality modeling work
covers only 48 states, and we are unable to extrapolate those results to Alaska or Hawaii. PM2.5
is used for these calculations rather than PM10 because the underlying health effect studies rely
on PM2.5 data.

   Accounting for the reduction in monetised health and welfare benefits from the net emission
inventory impacts of Option 5c in comparison to our final program produces 30-year net present
value of loss in benefits of $36.6 billion at a 3 percent discount rate, and $14.8 billion at a 7
percent discount rate. This loss in benefits is much larger than the costs savings associated with
not applying trap-based PM standards to 25-50-hp engines as shown in Table 12.2.2.2.2-1,
highlighting the fact that there is a substantial net benefit to society of applying the trap-based
PM standards to 25-50 hp engines.
                                         12-20

-------
                                                               Regulatory Alternatives
       12.2.2.2.4 Costs Per Ton

   The cost-effectiveness of the final standards for 25-50 hp engines can be calculated from the
values in Tables Table 12.2.2.2.1-1 and Table 12.2.2.2.2-1.  The results are given in Table
12.2.2.2.4-1.

                                    Table 12.2.2.2.4-1
                   50-State 30-Year Net Present Value Cost-Effectiveness
                   For Option 5c In Comparison to Final Program ($/ton)

PM
NOx + NMHC
3% discount rate
26,600
1,400
7% discount rate
29,100
1,500
                                          12-21

-------