United States Air and Radiation EPA420-R-03-008
Environmental Protection April 2003
Agency
&EPA Draft Regulatory Impact
Analysis: Control of
Emissions from Nonroad
Diesel Engines
> Printed on Recycled Paper
-------
EPA420-R-03-008
April 2003
Draft 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
NOTICE
This technical report does not necessarily represent final EPA decisions or positions.
It is intended to present technical analysis of issues using data that are currently available.
The purpose in the release of such reports is to facilitate the exchange of
technical information and to inform the public of technical developments which
may form the basis for a final EPA decision, position, or regulatory action.
-------
Table of Contents
Executive Summary
CHAPTER 1: Industry Characterization 1-1
1.1 Characterization of Engine Manufacturers 1-2
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) 1-7
1.2.3 Equipment Using Engines Rated between 56kW and 130 kW (75 and 175 hp) 1-8
1.2.4 Equipment Using Engines Rated between 130 and 560 kW (175 and 750
hp) 1-10
1.2.5 Equipment Using Engines Rated over 560 kW (750 hp) 1-12
1.3 Refinery Operations 1-13
1.3.1 The Supply-Side 1-13
1.3.2 The Demand Side 1-20
1.3.3 Industry Organization 1-27
1.3.4 Markets and Trends 1-31
1.4 Distribution and Storage Operations 1-36
1.4.1 The Supply-Side 1-36
1.4.2 The Demand-Side 1-38
1.4.3 Industry Organization 1-38
1.4.4 Markets and Trends 1-39
CHAPTER 2: Air Quality, Health, and Welfare Effects 2-1
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 2-12
2.1.2.1 CurrentPM Air Quality 2-12
2.1.2.2 Risk of Future Violations 2-22
2.1.3 Welfare Effects of Particulate Matter 2-33
2.1.3.1 Visibility Degradation 2-33
2.1.3.2 Other Effects 2-46
2.2 Air Toxics 2-50
2.2.1 Diesel Exhaust PM 2-50
2.2.1.1 Potential Cancer Effects of Diesel Exhaust 2-50
2.2.1.2 Other Health Effects of Diesel Exhaust 2-54
2.2.1.3 Diesel Exhaust PM Ambient Levels 2-55
iii
-------
2.2.1.4 Diesel Exhaust PM Exposures 2-65
2.2.2 Gaseous Air Toxics 2-69
2.2.2.1 Benzene 2-73
2.2.2.2 1,3-Butadiene 2-77
2.2.2.3 Formaldehyde 2-80
2.2.2.4 Acetaldehyde 2-83
2.2.2.5 Acrolein 2-85
2.2.2.6 Polycyclic Organic Matter 2-87
2.2.2.7 Dioxins 2-87
2.3 Ozone 2-87
2.3.1 Health Effects of Ozone 2-89
2.3.2 Attainment and Maintenance of the 1-Hour and 8-Hour Ozone NAAQS .... 2-91
2.3.2.1 1-Hour Ozone Nonattainment Areas and Concentrations 2-92
2.3.2.2 8-Hour Ozone Levels: Current and Future Concentrations 2-95
2.3.2.3 Potentially Counterproductive Impacts on Ozone Concentrations from NOx
Emissions Reductions 2-107
2.3.3 Welfare Effects Associated with Ozone and its Precursors 2-112
2.4 Carbon Monoxide 2-115
2.4.1 General Background 2-115
2.4.2 Health Effects of CO 2-116
2.4.3 CO Nonattainment 2-117
CHAPTER 3: Emissions Inventory 3-1
3.1 Nonroad Diesel Baseline Emissions Inventory Development 3-1
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 3-13
3.1.3 Commercial Marine Vessels and Locomotives 3-15
3.1.4 Recreational Marine Engines 3-20
3.1.5 Fuel Consumption for Nonroad Diesel Engines 3-23
3.2 Contribution of Nonroad Diesel Engines to National Emission Inventories 3-25
3.2.1 Baseline Emissions Inventory Development 3-25
3.2.2 PM25 Emissions 3-26
3.2.3 NOX Emissions 3-27
3.2.4 SO2 Emissions 3-28
3.2.5 VOC Emissions 3-28
3.2.6 CO Emissions 3-29
3.3 Contribution of Nonroad Diesel Engines to Selected Local Emission Inventories . . 3-36
3.3.1 PM25 Emissions 3-36
3.3.2 NOX Emissions 3-39
3.4 Nonroad Diesel Controlled Emissions Inventory Development 3-42
3.4.1 Land-Based Diesel Engines—PM25, NOX, SO2, VOC, and CO
Emissions 3-42
3.4.2 Land-Based Diesel Engines—Air Toxics Emissions 3-51
3.4.3 Commercial Marine Vessels and Locomotives 3-52
IV
-------
3.4.4 Recreational Marine Engines 3-54
3.5 Anticipated Emission Reductions With the Proposed Rule 3-58
3.5.1 PM25 Reductions 3-58
3.5.2 NOX Reductions 3-67
3.5.3 SO2 Reductions 3-69
3.5.4 VOC and Air Toxics Reductions 3-76
3.5.5 CO Reductions 3-79
3.6 Emission Inventories Used for Air Quality Modeling 3-80
CHAPTER 4: Technologies and Test Procedures for Low-Emission Engines 4-1
4.1 Feasibility of Emission Standards 4-1
4.1.1 PM Control Technologies 4-1
4.1.2 NOx Control Technologies 4-17
4.1.3 Can These Technologies Be Applied to Nonroad Engines and
Equipment? 4-68
4.1.4 Are the Standards Proposed for Engines >25 hp and <75 hp Feasible? 4-80
4.1.5 Are the Standards Proposed for Engines <25 hp Feasible? 4-89
4.1.6 Meeting the Crankcase Emissions Requirements 4-93
4.1.7 Why Do We Need 15ppm Sulfur Diesel Fuel? 4-94
4.2. Supplemental Transient Emission Testing 4-103
4.2.1. Background and Justification 4-103
4.2.2. Data Collection and Cycle Generation 4-106
4.2.3 Composite Cycle Construction 4-119
4.2.4 Cycle Characterization Statistics 4-121
4.2.5 Cycle Normalization / Denormalization Procedure 4-122
4.2.6 Cycle Performance Regression Statistics 4-123
4.2.7 Constant-Speed Variable-Load Transient Test Procedure 4-124
4.2.8 Cycle Harmonization 4-127
4.2.9 Supplemental Cold Start Transient Test Procedure 4-137
4.2.10 Applicability of Component Cycles to Nonroad Diesel Market 4-140
4.2.11 Final Certification Cycle Selection Process 4-143
4.3 Feasibility of Not-to-Exceed Standards 4-144
4.3.1 What EPA concerns do all NTE standards address? 4-144
4.3.2 How does EPA characterize the highway NTE test procedures? 4-145
4.3.3 How does EPA characterize the alternate NTE test procedures mentioned
above? 4-145
4.3.4 What limits might be placed on NTE compliance under the alternate test
procedures? 4-146
4.3.5 How does the "constant-work" moving average work, and what does it
do? 4-150
4.3.6 What data would need to be collected in order to calculate emissions results using
the alternate NTE? 4-153
4.3.7 Could data from a vehicle's on-board electronics be used to calculate
emissions? 4-154
4.3.8 How would anyone test engines in the field? 4-154
iii
-------
4.3.9 How might in-use crankcase emissions be evaluated? 4-155
4.3.10 How might the agency characterize the technological feasability for
manufacturers to comply with NTE standards? 4-155
CHAPTER 5: Fuel Standard Feasibility 5-1
5.1 Blendstock 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 5-6
5.2.1 Introduction to Diesel Fuel Sulfur Control 5-6
5.2.2 Conventional Hydrotreating 5-7
5.2.3 Phillips S-Zorb Sulfur Adsorption 5-20
5.2.4 Linde Isotherming 5-23
5.2.5 Chemical Oxidation and Extraction 5-26
5.2.6 FCC Feed Hydrotreating 5-26
5.3 Feasibility of Producing 500 ppm Sulfur Nonroad Diesel Fuel in 2007 5-27
5.3.1 Expected use of Desulfurization Technologies for 2007 5-27
5.3.2 Leadtime Evaluation 5-29
5.4 Feasibility of Distributing 500 ppm Sulfur Non-Highway Diesel Fuel in 2007 and 500
ppm Locomotive and Marine Diesel Fuel in 2010 5-37
5.4.1 The Diesel Fuel Distribution System Prior to the Implementation of the
Proposed 500 ppm Sulfur Program: 5-37
5.4.2 Summary of the Proposed 500 ppm Sulfur Standards 5-38
5.4.3 Limiting Sulfur Contamination 5-40
5.4.4 Potential Need for Additional Product Segregation 5-40
5.5 Feasibility of Producing 15 ppm Sulfur Nonroad Diesel Fuel in 2010 5-46
5.5.1 Expected use of Desulfurization Technologies for 2010 5-46
5.5.2 Leadtime Evaluation 5-49
5.6 Feasibility of Distributing 15 ppm Sulfur Nonroad Diesel Fuel in 2010 5-50
5.6.1 The Diesel Fuel Distribution System Prior to the Implementation of the
Proposed 15 ppm Nonroad Diesel Sulfur Program 5-50
5.6.2 Summary of the Proposed 15 ppm Nonroad Diesel Sulfur Standard 5-50
5.6.3 Limiting Sulfur Contamination 5-51
5.6.4 Potential need for Additional Product Segregation Due to the Implementation of
the Proposed 15 ppm Sulfur Specification for Nonroad Diesel Fuel 5-52
5.7 Impacts on the Engineering and Construction Industry 5-54
5.7.1 Design and Construction Resources Related to Desulfurization
Equipment 5-54
5.7.2 Number and Timing of Revamped and New Desulfurization Units 5-55
5.7.3 Timing of Desulfurization Projects Starting up in the Same Year 5-63
5.7.4 Timing of Design and Construction Resources Within a Project 5-64
5.7.5 Projected Levels of Design and Construction Resources 5-65
5.8 Supply of Nonroad, Locomotive, and Marine Diesel Fuel (NRLM) 5-70
5.9 Desulfurization Effect on Other Non-Highway Diesel Fuel Properties 5-76
iv
-------
5.9.1 Fuel Lubricity 5-76
5.9.2 Volumetric Energy Content 5-79
5.9.3 Fuel Properties Related to Storage and Handling 5-81
5.9.4 Cetane Index and Aromatics 5-81
5.9.5 Other Fuel Properties 5-83
5.10 Feasibility of the Use of a Marker in Heating Oil from 2007-2010 and in Locomotive
and Marine Fuel from 2010-2014 5-84
Appendix 5 A: EPA's Legal Authority for Proposing Nonroad, Locomotive, and Marine
Diesel Fuel Sulfur Controls 5-90
CHAPTER 6: Estimated Engine and Equipment Costs 6-1
6.1 Methodology for Estimating Engine and Equipment Costs 6-1
6.2 Engine-Related Costs 6-4
6.2.1 Engine Fixed Costs 6-4
6.2.2 Engine Variable Costs 6-20
6.2.3 Engine Operating Costs 6-44
6.3 Equipment-Related Costs 6-55
6.3.1 Equipment Fixed Costs 6-56
6.3.2 Equipment Variable Costs 6-63
6.3.3 Potential Impact of the Transition Provisions for Equipment
Manufacturers 6-64
6.4 Summary of Engine and Equipment Costs 6-66
6.4.1 Engine Costs 6-66
6.4.2 Equipment Costs 6-68
6.5 Costs for Example Pieces of Equipment 6-69
6.5.1 Summary of Costs for Some Example Pieces of Equipment 6-69
6.5.2 Method of Generating Costs for Our Example Pieces of Equipment 6-70
CHAPTER 7: Estimated Costs of Low-Sulfur Fuels 7-2
7.1 Nonroad Fuel Volumes 7-2
7.1.1 Overview 7-2
7.1.2 Diesel Fuel Demand by PADD for 2000 7-3
7.1.3 Diesel Fuel Demand by PADD for 2008 7-21
7.1.4. Annual Diesel Fuel Demand (2000-2040) and Associated In-Use Sulfur
Levels 7-25
7.1.5 Refinery Supply Volumes 7-42
7.2 Refining Costs 7-49
7.2.1 Methodology 7-49
7.2.2 Refining Costs 7-106
7.3 Cost of Distributing Non-Highway Diesel Fuel 7-137
7.3.1 Distribution Costs Under the 500 ppm Sulfur Non-Highway Diesel Fuel
Program 7-137
7.3.2 Distribution Costs Under the 15 ppm Sulfur Nonroad Diesel Fuel
Program 7-139
7.3.3 Cost of Lubricity Additives 7-141
-------
7.3.4 Fuel Marker Costs 7-142
7.3.5 Distribution, Lubricity, and Marker Costs Under Alternative Sulfur Control
Options 7-143
7.4 Net Cost of the Two-Step Nonroad Diesel Fuel Program 7-145
7.5 Potential Fuel Price Impacts 7-146
Appendix 7A: Estimated Total Off-Highway Diesel Fuel Demand and Diesel Sulfur
Levels 7-151
CHAPTER 8: Estimated Aggregate Cost and Cost per Ton of Reduced Emissions 8-1
8.1 Projected Sales and Cost Allocations 8-1
8.2 Aggregate Engine Costs 8-4
8.2.1 Aggregate Engine Fixed Costs 8-4
8.2.2 Aggregate Engine Variable Costs 8-6
8.3 Aggregate Equipment Costs 8-9
8.3.1 Aggregate Equipment Fixed Costs 8-9
8.3.2 Aggregate Equipment Variable Costs 8-11
8.4 Aggregate Fuel Costs and Other Operating Costs 8-13
8.4.1 Aggregate Fuel Costs 8-14
8.4.2 Aggregate Oil Change Maintenance Savings 8-16
8.4.3 Aggregate CDPF & CCV Maintenance Costs and CDPF Regeneration
Costs 8-19
8.4.4 Summary of Aggregate Operating Costs 8-21
8.5 Summary of Total Aggregate Costs of the Proposed Program 8-23
8.6 Emission Reductions 8-26
8.7 Cost per Ton 8-28
8.7.1 Cost per Ton for the 500 ppm Fuel Program 8-28
8.7.2 Cost per Ton for the Proposed Program 8-30
CHAPTER 9: Cost-Benefit Analysis 9-1
9.1 Time Path of Emission Changes for the Proposed Standards 9-8
9.2 Development of Benefits Scaling Factors Based on Differences in Emission Impacts
Between Proposed and Modeled Preliminary Control Options 9-10
9.3 Summary of Modeled Benefits and Apportionment Method 9-11
9.3.1 Overview of Analytical Approach 9-12
9.3.2 Air Quality Modeling 9-13
9.3.3 Health Effect Concentration-Response Functions 9-15
9.3.4 Economic Values for Health Outcomes 9-18
9.3.5 Welfare Effects 9-19
9.3.6 Treatment of Uncertainty 9-23
9.3.7 Model Results 9-24
9.3.8 Apportionment of Benefits to NOx, SO2, and PM Emissions Reductions . . . 9-37
9.4 Estimated Benefits of Proposed Nonroad Diesel Engine Standards in 2020 and 2030 9-39
9.5 Development of Intertemporal Scaling Factors and Calculation of Benefits Over Tin$fe42
9.6 Comparison of Costs and Benefits 9-47
APPENDIX 9A: Benefits Analysis of Modeled Preliminary Control Option 9-65
vi
-------
APPENDIX 9B: Sensitivity Analyses of Key Parameters in the Benefits Analysis . . . 9-187
APPENDIX 9C: Visibility Benefits Estimates for Individual Class I Areas 9-206
CHAPTER 10: Economic Impact Analysis 10-1
10.1 Overview of Results 10-1
10.1.1 What is an Economic Impact Analysis? 10-1
10.1.2 What is EPA's Economic Analysis Approach for this Proposal? 10-1
10.1.3 What are the key features of the NDEEVI? 10-4
10.1.4 Summary of Economic Analysis 10-8
10.2 Economic Methodology 10-19
10.2.1 Behavioral Economic Models 10-20
10.2.2 Conceptual Economic Approach 10-20
10.2.3 Key Modeling Elements 10-28
10.3 Economic Impact Modeling 10-39
10.3.1 Operational Economic Model 10-39
10.3.2 Baseline Economic Data 10-40
10.3.3 Market Linkages 10-44
10.3.4 Compliance Costs 10-50
10.3.5 Supply and Demand Elasticity Estimates 10-59
10.3.6 Model Solution Algorithm 10-62
APPENDIX 10A: Impacts on the Engine Market and Engine Manufacturers 10-66
APPENDIX 1 OB: Impacts on Equipment Market and Equipment Manufacturers .... 10-75
APPENDIX IOC: Impacts on Application Market Producers and Consumers 10-84
APPENDIX 10D: Impacts on the Nonroad Fuel Market 10-88
APPENDIX 10E: Time Series of Social Cost 10-93
APPENDIX 10F: Model Equations 10-96
APPENDIX 10G: Elasticity Parameters for Economic Impact Modeling 10-102
APPENDIX 10H: Derivation of Supply Elasticity 10-118
APPENDIX 101: Sensitivity Analysis 10-119
CHAPTER 11: Small-Business Flexibility Analysis 11-1
11.1 Overview of the Regulatory Flexibility Act 11-1
11.2 Need for the Rulemaking and Rulemaking Objectives 11-2
11.3 Definition and Description of Small Entities 11-2
11.3.1 Description of Nonroad Diesel Engine and Equipment Manufacturers 11-3
11.3.2 Description of the Nonroad Diesel Fuel Industry 11-3
11.4 Summary of Small Entities to Which the Rulemaking Will Apply 11-4
11.4.1 Nonroad Diesel Engine Manufacturers 11-4
11.4.2 Nonroad Diesel Equipment Manufacturers 11-5
11.4.3 Nonroad Diesel Fuel Refiners 11-5
11.4.4 Nonroad Diesel Fuel Distributors and Marketers 11-6
11.5 Related Federal Rules 11-6
11.6 Projected Reporting, Recordkeeping, and Other Compliance Requirements 11-6
11.7 Projected Economic Effects of the Proposed Rulemaking 11-7
11.8 Regulatory Alternatives 11-8
vii
-------
11.8.1 Small Engine Manufacturers 11-8
11.8.2NonroadDiesel Equipment Manufacturers 11-13
11.8.3 Nonroad Diesel Fuel Refiners 11-16
11.8.4 Nonroad Diesel Fuel Distributors and Marketers 11-21
CHAPTER 12: Regulatory Alternatives 12-1
12.1 Range of Options Considered 12-1
12.1.1 One-Step Options 12-1
12.1.2 Two-Step Options 12-6
12.2 Emission Inventory Impacts Comparison 12-18
12.2.1 Assumptions Regarding Fuel Sulfur Content 12-19
12.2.2 Emission Inventories for Alternative Program Options 12-23
12.2.3 Cumulative Emission Reductions for Alternative Program Options 12-35
12.3 Benefits Comparison 12-37
12.4 Cost Analysis for Alternative Options 12-53
12.4.1 One Step Options 12-53
12.4.2 Two Step Options 12-60
12.4.3 Other Options 12-69
12.5 Costs per Ton 12-74
12.5.3 Incremental Cost per Ton for Option 2c 12-78
12.5.4 Incremental Cost per Ton for Option 2e 12-78
12.5.5 Incremental Cost per Ton for Option 3 12-79
12.5.6 Incremental Cost per Ton for Option 4 12-80
12.5.7 Incremental Cost per Ton for Option 5a 12-80
12.5.8 Incremental Cost per Ton for Option 5b 12-81
12.6 Summary and Assessment of Alternative Program Options 12-82
12.6.1 Summary of Results of Options Analysis 12-82
12.6.2 Discussion of Rationale, Issues, and Feasibility Assessment of
Options 12-85
Appendix 12A: Certification Fuel Sulfur Levels 12-106
Appendix 12B: Incremental Cost, Emission Reductions, Benefits, and Cost
Effectiveness 12-114
vin
-------
List of Acronyms
ABT
AEO
AGME
AT
BSFC
CCV
CDPF
CFR
CI
CMV
CO
DF
DI
DOC
EF
EGR
EIA
EIA
FR
FTC
GDP
HC
HD2007
hp
IDI
IRFA
kW
L&M
MPP
Averaging, Banking, and Trading
Annual Energy Outlook
Above-ground mining equipment
Aftertreatment
Brake Specific Fuel Consumption
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
Gross domestic product
Hydrocarbons
Heavy-duty 2007 refers to the final rule setting emission standards for 2007 and later engines
used in heavy-duty highway vehicles.
Horsepower
Indirect injection
Initial Regulatory Flexibility Analysis
kilowatt
Locomotive and marine
marginal physical product
IX
-------
NDEIM
NMHC
NPV
NR
NRLM
O&M
OMB
PM
ppm
PSR
RIA
SBA
SBAR
SBREFA
SER
SIC
stds
TAP
TPEM
VMP
VOC
ZHL
Nonroad Diesel Economic Impact Model
Non-methane hydrocarbons
Net present value
Nonroad
Nonroad, Locomotive, and Marine diesel fuel
operating and maintenance
Office of Management and Budget
Particulate matter
Parts per million
Power Systems Research
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
1039.625)
value of marginal product
Volatile organic compounds
Zero-Hour Emission Level
-------
Executive Summary
The Environmental Protection Agency (EPA) is proposing requirements to reduce emissions
of particulate matter (PM) and oxides of nitrogen (NOx) from nonroad diesel engines. This
proposal includes emission standards for new nonroad diesel engines. The proposal also
addresses the quality of the fuel used in nonroad engines, as well as locomotive and marine
engines, by specifying reduced sulfur levels.
This executive summary first highlights the proposed emission standards and fuel
requirements, then gives an overview of the analyses in the rest of this document.
Emission Standards and Engine Technologies
Tables 1 and 2 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 to
ensure compliance. As shown in the table, we are phasing in many of the proposed standards
over a two- or three-year period to address lead time, workload, and feasibility considerations.
Table 1
Proposed PM Standards (g/bhp-hr) and Schedule
Engine Power
hp<25 (kW<19)
25 750 (kW>560)
Model Year
2008
0.30
0.22
2009
2010
2011
0.01
0.01
2012
0.01
2013
0.02
ES-1
-------
Table 2
Proposed NOx and NMHC Standards and Schedule
Engine Power
25 750 (kW>560)
25 750 (kW>560)
Standard (g/bhp-hr)
NOx
NMHC
3.5NMHC+NOx
0.30
0.30
0.30
0.14
0.14
0.14
Phase-in Schedule
2011
50%
50%
2012
50%
50%
50%
2013
100%
50%
50%
50%
2014
100%
100%
100%
The proposal includes new provisions to help ensure that emission-control systems perform
as well when operating in actual service conditions as in the laboratory. These procedures will
also allow for testing an engine's emission levels while the machinery operates in normal service.
Controls on In-use Diesel Fuel
Just as lead was phased out of gasoline because it damages catalytic converters in cars, sulfur
can contaminate high-efficiency emission-control systems used on diesel engines. Nonroad
diesel fuel currently has sulfur levels up to 3,400 parts per million (ppm). This proposal would
reduce these levels by 99 percent, which is an essential step in achieving the emission reductions
anticipated under the proposal.
Starting in 2007, fuel sulfur levels in nonroad diesel fuel would be limited to a maximum of
500 ppm, the same as for current highway diesel fuel. This limit also covers fuels used in
locomotive and marine applications (though not to the marine residual fuel used by very large
engines on ocean-going vessels). Reducing fuel sulfur levels to 500 ppm or lower will provide
immediate public health benefits by reducing particulate emissions from engines in the existing
fleet of nonroad equipment, with the added benefit of reducing the cost of maintaining engines.
The proposal includes a second step of fuel controls to a 15-ppm limit on sulfur content that
would apply in 2010. This additional reduction in sulfur levels will further reduce PM emissions
from existing engines. More importantly, the ultra-low sulfur levels will make it possible for
engine manufacturers to use advanced emission-control systems that will achieve dramatic
reductions in both PM and NOx emissions. In addition, we are seriously considering whether to
ES-2
-------
establish new emissions standards in the future that would reduce the emissions from locomotive
and marine engines by more than 90 percent with the same advanced emission-control
technologies included in this proposal.
Estimated Costs of the Proposal
There are approximately 600 nonroad equipment manufacturers using diesel engines in
several thousand different equipment models. 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 3. Also shown in Table 3 are typical prices for
each piece of equipment for reference. See Chapter 6 for additional detailed information related
to cost analyses related to engines and equipment.
Table 3
Long-Term Costs for Several Example Pieces of Equipment"
Horsepower
Displacement (L)
Incremental Engine &
Equipment Cost
Long Term
Near Term
Estimated Equipment
Pnceb
GenSet
9hp
0.4
$120
$170
$3,500
Skid/Steer
Loader
33 hp
1.5
$760
$1,100
$13,500
Backhoe
76 hp
3.9
$1,210
$1,680
$50,000
Dozer
175 hp
10.5
$2,590
$3,710
$235,000
Agricultural
Tractor
250 hp
7.6
$2,000
$2,950
$130,000
Dozer
503 hp
18
$4,210
$6,120
$575,000
Off-
Highway
Truck
1000 hp
28
$6,780
$10,100
$700,000
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 upgrading to low-sulfur fuel takes into account all the
necessary changes in both refining and distribution practices. We have estimated the cost of
producing 500-ppm fuel to be on average 2.5 cents per gallon. Average costs for 15-ppm fuel are
estimated to be an additional 2.3 cents per gallon for a combined cost of 4.8 cents per gallon, as
shown in Table 4. 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 3.3 cents per gallon.
ES-3
-------
Table 4
Increased Cost of Providing Nonroad,
Locomotive and Marine Diesel Fuel (cents per gallon of affected fuel)
Step One - 500 ppm NRLM diesel fuel
Step Two - 15 ppm Nonroad diesel fuel
Step Two - 500 ppm Locomotive and Marine diesel fuel
Refining
2.2
4.4
2.2
Distribution
0.3
0.4
0.2
Total
2.5
4.8
2.4
Cost per Ton of Reduced Emissions
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 5 compares
these aggregate costs with the corresponding estimated emission reductions to present cost-per-
ton figures for the various pollutants.
Table 5
Aggregate Cost per Ton for the Proposed Two-Step Fuel Program
and Engine Program—2004-2036 Net Present Values at 3% Discount Rate ($2001)
Pollutant
NOx+NMHC
PM
SOX
Aggregate Discounted Lifetime Cost per ton
$810
$8,700
$200
Estimated Emission Reductions, Air Quality Impacts and Benefits
Based on our most recent nationwide inventory used for this proposal (1996), we estimate
that the nonroad diesel engines affected by this proposal contribute about 44 percent of diesel PM
emissions and 12 percent of NOx emissions from mobile sources. By 2020, these engines will
emit over 60 percent of diesel PM and 20 percent of NOx from mobile sources. When fully
implemented, this proposal would reduce PM and NOx emissions from nonroad diesel engines
by more than 90 percent. It will also virtually eliminate emissions of sulfur oxides (SOx) from
these engines, which amounted to nearly 300,000 tons in 1996, and would otherwise grow to
approximately 380,000 tons by 2020. These dramatic emission reductions emissions are a
critical part of the effort by federal, state, local, and tribal governments to reduce the health-
related impacts of air pollution.
Reducing NOx and PM emissions from nonroad diesel engines by more than 90 percent
would provide a wide range of benefits for public health and the environment. We have
estimated that, by 2030, controlling these emissions would annually prevent 9,600 premature
ES-4
-------
deaths, over 8,300 hospitalizations, and almost a million work days lost. All told, the monetized
health benefits of this rule would be $81 billion annually once the program is fully phased in.
Costs for both the engine and fuel requirements would be significantly less, at approximately
$1.5 billion annually. See the fact sheet referenced below for further description of these
environmental benefits.
Economic Impact Analysis
An Economic Impact Analysis was prepared for this proposal to estimate its potential
economic impacts on producers and consumers of nonroad engines and equipment and fuels, and
related industries. The Economic Impact Analysis 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. A multi-market partial
equilibrium approach was used to track changes in price and quantity for 60 integrated product
markets. The model and data inputs are described in Chapter 10.
As shown in Table 6, the market analysis predicts that the overall economic impact of the
proposed emission control program on society is expected to be small, on average. According to
this analysis, the average prices of goods and services produced using equipment and fuel
affected by the proposal (the application markets) are expected to increase about 0.02 percent.
Engine prices are expected to increase, on average, about 22.9 percent in 2013, decreasing to
about 19.5 percent for 2020 and after. The average price increase for nonroad equipment is
expected to be about 5.2 percent in 2013, decreasing to 4.4 percent by 2020. The average price
increase for nonroad diesel fuel for all years is expected to be about 4.1 percent. Quantities of
products affected by this proposal are expected to decline negligibly, by less that 0.02 percent.
Table 6
Summary of Expected Market Impacts, 2013 and 2020
Market
Engines
Equipment
Application
markets*
Nonroad Fuel
Markets
2013
Average
engineering
cost per unit
$1,087
$1,021
—
$0.039
Price change
22.9%
5.2%
0.02%
4.1%
Quantity
change
-0.013%
-0.014%
-0.010%
-0.013%
2020
Average
engineering
cost per unit
$1,028
$1,018
—
$0.039
Price change
19.5%
4.4%
0.02%
4.1%
Quantity
change
-0.013%
-0.014%
-0.010%
-0.014%
"Commodities in the application markets are normalized; only percentage changes are presented
ES-5
-------
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 estimated to be about $1,202 million. About 82 percent of the total social costs are expected
to be borne by producers and consumers in the application markets, indicating a majority of the
costs are expected to be passed on in the form of higher prices. When these estimated impacts
are broken down, 58 percent are expected to be borne by consumers in the application markets
and 42 percent are expected to be borne by producers in the application markets. Equipment
manufacturers are expected to bear about 10 percent of the total social costs. These are primarily
the costs associated with equipment redesign. Engine manufacturers are expected to bear about
2.5 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 5 percent is accounted for by fuel
marker costs and the additional costs of 15 ppm fuel being sold in to markets such as marine
diesel, locomotive, and home heating fuel that do not require it.
In 2020, the social costs of the rule are expected to increase to about $1,510 million.
Producers and consumers in the applications markets are expected to bear nearly all of these
costs, about 94 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.
Alternative program options
In the course of designing our proposed 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 evaluation of twelve alternative program options, including
an assessment of technical feasibility, cost, cost-effectiveness, inventory impact, and health and
welfare benefits for each alternative. Table 12.6.1-1 summarizes the alternative program options,
while the accompanying text in Section 12.6 presents our rationale for choosing the proposed
program rather than one of the alternatives.
ES-6
-------
CHAPTER 1: Industry Characterization
1.1 Characterization of Engine Manufacturers 1-2
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
1.1.3EnginesRatedbetween56andl30kW(75andl75hp) 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) 1-7
1.2.3 Equipment Using Engines Rated between 56kW and 130 kW (75 and 175 hp) 1-8
1.2.4 Equipment Using Engines Rated between 130 and 560 kW (175 and 750 hp) 1-10
1.2.5 Equipment Using Engines Rated over 560 kW (750 hp) 1-12
1.3 Refinery Operations 1-13
1.3.1 The Supply-Side 1-13
1.3.2 The Demand Side 1-20
1.3.3 Industry Organization 1-27
1.3.4 Markets and Trends 1-31
1.4 Distribution and Storage Operations 1-36
1.4.1 The Supply-Side 1-36
1.4.2 The Demand-Side 1-38
1.4.3 Industry Organization 1-38
1.4.4 Markets and Trends 1-39
-------
Industry Characterization
CHAPTER 1: Industry Characterization
In understanding the impact of emissions 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. In the remainder of this draft RIA, to the
extent data regarding engine/equipment populations, sales or other industry specific data has
been used, that data is explained and referenced in the relevant section of the draft RIA. The
information presented in this chapter will be most helpful for the reader who is 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 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 (e.g., 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 and 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
horsepower 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
-------
Draft Regulatory Support Document
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
recognize that the PSR database is not comprehensive, but have not identified a better source to
provide consistent data for identifying additional companies.
1.1.1 Engines Rated between 0-19 kW (0 and 25 hp)
In year 2000, sales of engines in this category comprised 18% (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% of engines with approximately 281,157 units
sold in year 2000. Direct Injection (DI) engines account for 59% of this category with 165,427
units. Yanmar has approximately 19% 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% 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% of the DI market, followed by Caterpillar/Perkins (22%),
Cummins (21%), Case/New Holland (8%), Volvo (4%), and then by Komatsu and Detroit Diesel
1-2
-------
Industry Characterization
(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).
1.1.5 Engines Rated over 560 kW (750 hp)
This is the smallest nonroad category with approximately 5,633 engines comprising 1% 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 proposed
rulemaking. Table 1.2-1 below contains examples of the types of nonroad equipment which
would be impacted by this proposal, 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
Forklift
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
1-3
-------
Draft Regulatory Support Document
Based on horsepower rating of the engine it uses, a fraction of applications such as air
compressors, generator sets, hydropower units, irrigation sets, pumps and welders is considered to
be stationary and hence not subject to EPA's proposed.standards. However, the tables in sections
1.2.1 to 1.2.5 account for all equipment manufactured, whether stationary or mobile within an engine
horsepower 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%
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 which 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
Manufacturer
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 Sains
13,394
11,042
9,970
5,559
5,191
5,117
Percentage
of Market
12%
10%
9%
5%
5%
5%
Engine
Charap.tnriyation*
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.
For these top six OEMs, their sales 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% of the equipment in this horsepower 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
Snips
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
Snips
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
-------
Draft 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 % of total sales are by Ingersoll- Rand. For the 19 to 56 kW range, the
equipments use either direct or indirect injection engines that are water or oil-cooled and are either
naturally aspirated or turbo-charged. The six leading manufacturers produce 53% of the equipment
in this category. These manufacturers are listed in Table 1.2-4. They manufacture equipment typical
of the market e.g. agricultural tractors, generator sets, skid-steer loaders and refrigeration/AC. These
top selling applications represent about 70% of the market as seen in Table 1.2-5. The top 90% 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 Sains
40,199
23,194
19,090
17,752
14,391
12,484
Percentage of
Market
17%
10%
8%
7%
6%
5%
Engine
Charap.tnriyation*
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%
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 which has a five year total sales
volume of 905,000. These manufacturers are shown in Table 1.2-6.
1-7
-------
Draft 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.
Of these top six OEMs, their sales 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% of
the equipment in this horsepower 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 Snips
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 Snips
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 turbo charged . A few are
naturally aspirated. The six leading manufacturers produce 56% 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% of the market. The top 90% of this market is supplied by 60 OEMs.
1-9
-------
Draft 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
Charap.tnriyation*
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 996.7000s!
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
Sains
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 Sains
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 only produced for the nonroad market
segments of construction equipment and welders and generators. As much as 35% 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% 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 hence not impacted by the proposed rule. 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
EComatsu
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
-------
Draft 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 EPA's proposed regulations. Section 1.4 discusses
distribution of refined petroleum products through pipelines from refineries, as well as storage
operations for these products. Both Sections 1.3 and 1.4 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 the new EPA 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 that the proposed regulations will mandate.
For each of the major products, several product streams from the refinery will be blended into
a finished mixture. For example, diesel fuel will typically contain 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
-------
Draft Regulatory Support Document
Figure 1.3-1
The Modern Refinery
i and
Light Gasoline
It
Ends
Plant
LPG
L.ght Str.-ryht Run Gawtine
Isomettotion teomwnte
Hurt
Reformer
Rrformale
Dtosel
FuH
Asphatt
Source: Chevron. 2002. Diesel Fuel Refining and Chemistry. As accessed on August 19, 2002.
.
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 EPA Regulation. Over the next few
years, EPA regulations will come into effect that require much lower levels of residual sulfur for
both gasoline and highway diesel fuel. 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.
1-15
-------
Draft Regulatory Support Document
Types of Products. The major products made at petroleum refineries are unbranded
commodities, which must meet established specifications for fuel value, density, vapor pressure,
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 prior to final sale, whereas
nonroad fuel is not. To prevent accidental or deliberate misuse, nonroad diesel fuel must be dyed
prior to 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 U.S. 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.
1-16
-------
Industry Characterization
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%
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 U.S. 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.
1-17
-------
Draft Regulatory Support Document
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 shows 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
Description of Petroleum Refineries—Census Bureau Data
NAICS324110—
Petroleum Refineries
2000
1999
1998
1997
1992 (reported as SIC 29 11)
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 US Census Bureau, 1997 Economic Census - Manufacturing, Industry Series EC97M-3241A, Table 1.
1998-2000 data from US Census Bureau, Annual Survey of Manufactures-2000, 2000, Statistics for Industry Groups and
Industries MOO(AS)-1, Table 2.
1-18
-------
Industry Characterization
Table 1.3-4
Petroleum Refinery Costs of Production, 1997-2000
Petroleum Refinery
Costs of Production
Cost of Materials (106)
as % 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 prior to 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).
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
1-19
-------
Draft Regulatory Support Document
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 HI, 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 to
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 (HSD), low-sulfur diesel (LSD), and fuel oil for
space heating;
• industrial—LSD for highway use, HSD for nonroad fuels, 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 HSD;
• electric utility—residual fuel and distillate fuel oil for power generation;
• railroad—HSD and LSD used for locomotives;
• vessel bunking—combination of fuel oil and residual fuel for marine engines;
• on-highway diesel—LSD for highway trucks and automobiles;
• military—HSD sales to the Armed Forces; and
• off-highway diesel—HSD and LSD used in construction and other industries.
1-20
-------
Industry Characterization
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.
1-21
-------
Draft Regulatory Support Document
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 (EIA). 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
On-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 (EIA). 2001b. Fuel Oil and Kerosene
Sales, 2000, Tables 7-12. Washington, DC.
1-22
-------
Industry Characterization
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
On-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 LSD fuel. PADD I (the East Coast) and PADD H (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 HI 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.
1-23
-------
Draft Regulatory Support Document
Table 1.3-8
Sales for On-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%
1-24
-------
Industry Characterization
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 LSD, HSD, 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 (ULSD) to highway
ULSD 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 EPA regulation, it is important to have a good
understanding of the overall refinery industry structure. The degree of industry concentration,
1-25
-------
Draft Regulatory Support Document
regional patterns of 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.
There is no convincing evidence in the literature that markets should be modeled as
imperfectly competitive, however. Although 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, 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 CR of 48.6 percent, and an HHI of 422. Merger guidelines
followed by the FTC and Department of Justice consider that there is little potential for pricing
power in an industry with an FtHI 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
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 FtHI figures. The results are presented in Table 1.3-13.
1-26
-------
Industry Characterization
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/ re finery _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 in, 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.
1-27
-------
Draft Regulatory Support Document
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, USA
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 (EIA). 2002a. Petroleum Supply Annual 2001.
Tables 16, 17, and 20. Washington, DC. Table 20.
• Markets in PADDs n and HI, which are not overly concentrated or geographically
isolated, should be expected to 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 that would be
a concern for competitive reasons.
Market Structure—Firms and Facilities. PADD HI has the greatest number of refineries
affected by the EPA nonroad regulation and will account for the largest volume of new ULSD
nonroadfuel. 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 HI 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 contains 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.
1-28
-------
Industry Characterization
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. It should be noted that PADD in 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 U.S. 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).
1-29
-------
Draft Regulatory Support Document
Many of the smallest refineries are certified as small businesses by EPA. 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 ULSD is likely to require significant
flexibility on the part of EPA.
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.
1-30
-------
Table 1.3-17
Top Refineries in Each PADD by Total Crude Distillation Capacity
Name
of Company
Sunoco Inc. (R&M)
PADD T Phillips 66 Co.
Phillips 66 Co.
Motiva Enterprises LLC
Sunoco Inc.
TOTAL
BP Products North America, Inc.
r> A r^n TT Phillips 66 Co.
IrYLJlJ 11
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)
-------
Figure 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)
-------
Figure 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.
. Washington, DC.
-------
Table 1.3-18
Sales of Distillate Fuel Oils to End Users 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.
-------
Industry Characterization
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 U.S. (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
only add value 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.
1-35
-------
Draft Regulatory Support Document
Figure 1.4-1
Typical Sequence in which Products are Batched While in Transit on Colonial System
Reformulated Regular Gasoline
• Low Sulfur Diesel Fuel
OKerosine/J?t 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, LSD 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 ULSD
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 ULSD for nonroad applications will not add additional
complexity to their operations. EPA expects that there will 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
1-36
-------
Industry Characterization
material generated during the shipment of 15 ppm diesel fuel by pipeline. After the
implementation of EPA's 15 ppm highway diesel requirement and the envisioned off-highway
diesel fuel controls, the pipelines that transport the majority of the nation's diesel fuel are
projected to continue to carry HSD fuel and/or 500 ppm diesel fuel. These pipelines would blend
their downgraded 15 ppm diesel into the 500 ppm and/or HSD 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
EPA's highway program is implemented. These pipelines currently carry only 500 ppm highway
diesel fuel. In EPA's highway diesel final rule, EPA projected that these pipelines would install
an additional storage tank to contain the relatively low volumes of downgraded 15 ppm diesel
fuel generated during pipeline transportation of the product. EPA projected that this downgraded
material would be sold into the off-highway diesel market. The implementation of the
envisioned nonroad diesel fuel controls would not change this practice. We expect that these
pipeline operators would continue to find 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
1-37
-------
Draft Regulatory Support Document
handle two grades of nonroad diesel fuel, pipeline operators would have to engage in a rate case
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 almost all of the overall increase in total product movements, but they
have 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. .
1-38
-------
Industry Characterization
References to Chapter 1
1. RTI. 2003. Industry Profile for Nonroad Diesel Tier 4 Rule. Prepared for the U. S.
Environmental Protection Agency. EPA Contract Number 68-D-99-024, April 2003. (Docket A-
2001-28).
2. Power Systems Research(PSR). 2002. OELink Sales Database.
3. See endnote 1.
Allegro Energy Group. 2001. How Pipelines Make the Oil Market Work—Their Networks,
Operations, and Regulations. New York: Allegro.
American Petroleum Institute (API). 1998. "All About Petroleum." As accessed on November
20, 2002. .
American Petroleum Institute (API). 2001. "Pipelines Need Operational Flexibility to Meet
America's Energy Needs." As accessed on November 20, 2002. .
American Petroleum Institute (API). 2002. "Marketing Basic Facts." As accessed on September
25, 2002. .
Association of Oil Pipe Lines (AOPL). 2000. "Fact Sheet: U.S. Oil Pipe Line Industry." As
accessed on November 20, 2002. .
Association of Oil Pipe Lines (AOPL). 2001. "Shifts in Petroleum Transportation." As
accessed on November 20, 2002. .
Association of Oil Pipe Lines (AOPL). 2002. "Why Pipelines?" As accessed on November 20,
2002. .
Business & Company Resource Center. .
Chevron. 2002. "Diesel Fuel Refining and Chemistry." As accessed on August 19, 2002.
.
Colonial. 2002. "Frequently Asked Questions." As accessed on September 24, 2002.
.
1-39
-------
Draft Regulatory Support Document
Considine, Timothy J. 2002. "Inventories and Market Power in the World Crude Oil Market."
As accessed on November 1, 2002. .
Dun & Bradstreet. Million Dollar Directory, .
Federal Energy Regulatory Commission (FERC). 2002. FERC Form No. 6, Annual Report of
Oil Pipelines. .
Flint Hills Resources. 2002. "Refining Overview." As accessed on September 10, 2002.
.
Federal Trade Commission (FTC). Midwest Gasoline Price Investigation, March 29, 2001, p.7.
As accessed September 25, 2002. .
Freedonia Group. 2001. "Diesel Engines and Parts in the United States to 2005—Industry
Structure." .
Hoover's Online, .
National Petroleum Council. 1989. "Petroleum Storage and Transportation." System
Dynamics. Volume n. Washington, DC: National Petroleum Council.
U.S. Department of Agriculture, National Agricultural Statistics Service (USDA-NASS). 2002.
Agricultural Statistics 2002. Washington, DC: U.S. Department of Agriculture.
U.S. Department of Energy, Energy Information Administration (EIA). 200 la. Annual Energy
Review, 2000. Washington, DC: Department of Energy.
U.S. Department of Energy, Energy Information Administration (EIA). 2001b. Fuel Oil and
Kerosene Sales, 2000, Tables 7-12. Washington, DC: Department of Energy.
U.S. Department of Energy, Energy Information Administration (EIA). 2002a. Petroleum
Supply Annual 2001. Washington, DC: Department of Energy.
U.S. Department of Energy, Energy Information Administration (EIA). 2002b. Refinery
Capacity Data Annual. As accessed on September 23, 2002.
. Washington, DC: Department of Energy.
1-40
-------
Industry Characterization
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.
1992 data from U.S. Census Bureau. 1992 Census of Manufactures, Industry Series
MC920I-29A. Table 1A.
1997 data from US Census Bureau, 1997 Economic Census - Manufacturing, Industry Series
EC97M-3241A, Table 1.
1998-2000 data from US Census Bureau, Annual Survey of Manufactures-2000, 2000, Statistics
for Industry Groups and Industries MOO(AS)-1, Table 2.
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 PM25 NAAQS: Current and Future Air Quality .... 2-11
2.1.2.1 Current PM Air Quality 2-11
2.1.2.2 Risk of Future Violations 2-22
2.1.3 Welfare Effects of Particulate Matter 2-33
2.1.3.1 Visibility Degradation 2-33
2.1.3.2 Other Effects 2-46
2.2 Air Toxics 2-50
2.2.1 Diesel Exhaust PM 2-50
2.2.1.1 Potential Cancer Effects of Diesel Exhaust 2-50
2.2.1.2 Other Health Effects of Diesel Exhaust 2-53
2.2.1.3 Diesel Exhaust PM Ambient Levels 2-55
2.2.1.4 Diesel Exhaust PM Exposures 2-65
2.2.2 Gaseous Air Toxics 2-69
2.2.2.1 Benzene 2-73
2.2.2.2 1,3-Butadiene 2-77
2.2.2.3 Formaldehyde 2-80
2.2.2.4 Acetaldehyde 2-83
2.2.2.5 Acrolem 2-85
2.2.2.6 Polycyclic Organic Matter 2-87
2.2.2.7 Dioxins 2-87
2.3 Ozone 2-87
2.3.1 Health Effects of Ozone 2-89
2.3.2 Attainment and Maintenance of the 1-Hour and 8-Hour Ozone NAAQS 2-91
2.3.2.1 1-Hour Ozone Nonattainment Areas and Concentrations 2-92
2.3.2.2 8-Hour Ozone Levels: Current and Future Concentrations 2-95
2.3.2.3 Potentially Counterproductive Impacts on Ozone Concentrations fromNOx Emissions
Reductions 2-107
2.3.3 Welfare Effects Associated with Ozone and its Precursors 2-112
2.4 Carbon Monoxide 2-115
2.4.1 General Background 2-115
2.4.2 Health Effects of CO 2-116
2.4.3 CO Nonattainment 2-117
-------
Air Quality, Health, and Welfare Effects
CHAPTER 2: Air Quality, Health, and Welfare Effects
With today's proposal, EPA is acting to extend highway types of emission controls to another
major source of diesel engine emissions: nonroad diesel engines. These emissions are significant
contributors to atmospheric pollution of particulate matter (PM), ozone and a variety of toxic air
pollutants among other pollutants. In our most recent nationwide inventory used for this
proposal (1996), the nonroad diesels affected by this proposal contribute over 43 percent of
diesel PM emissions from mobile sources, up to 18 percent of total PM25 emissions in urban
areas, and up to 14 percent of NOx emissions in urban areas.
Without further control beyond those standards we have already adopted, by the year 2020,
these engines will emit 61 percent of diesel PM from mobile sources, up to 19 percent of all
direct PM25 emissions in urban areas, and up to 20 percent of NOx.emissions in urban areas.
When fully implemented, today's proposal would reduce nonroad diesel PM25 and NOx
emissions by more than 90 percent. It will also virtually eliminate nonroad diesel SOx
emissions, which amounted to nearly 300,000 tons in 1996, and would otherwise grow to
approximately 380,000 tons by 2020.
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 other environmental effects such as visibility. Based on the most recent
monitoring data available for this rule (1999-2001), such problems are widespread in the United
States. There are over 70 million people living in counties with PM2.5 levels exceeding the
PM2.5 NAAQS, and 111 million people living in counties exceeding the 8-hour ozone NAAQS.
Figure 2.-1 illustrates the widespread nature of these problems. Shown in this figure are counties
exceeding either or both of the two NAAQS plus mandatory Federal Class I areas, which have
particular needs for reductions in haze.
2-1
-------
Draft Regulatory Impact Analysis
Figure 2-1
Nonroad Diesel-related Air Quality Problems are Widespread
Areas
fc&jivl Federal Class! Areas (Visibility)
^^ Counties Exceeding 8-hr Ozone NAAQS
| Counties Exceeding PM2.5NAAQS
^B Counties Exceeding Both NAAQS
Air quality data derived from AQS (1999-2001)
with data handling per Agency guidance except PM2.5 data
includes monitors with complete data in at least 10 quarters.
As we will describe later in Chapter 9, the air quality improvements expected from this
proposal would produce major benefits to human health and welfare, with a combined value in
excess of half a trillion dollars between 2007 and 2030. By the year 2030, this proposed rule
would be expected to prevent approximately 9,600 deaths per year from premature mortality, and
16,000 nonfatal heart attacks per year. By 2030, it would also prevent 14,000 annual acute
bronchitis attacks in children, 260,000 respiratory symptoms in children, nearly 1 million lost
work days among adults because of their own symptoms, and 6 million days where adults have to
restrict their activities due to symptoms in 2030.
In this chapter we will describe in more detail the air pollution problems associated with
emissions from nonroad diesel engines and air quality benefits we expect to realize from the fuel
and engine controls in this proposal. The emissions from nonroad diesel engines that are being
directly controlled by this rulemaking are NOx, PM and NMHC, and to a lesser extent, CO.
2-2
-------
Air Quality, Health, and Welfare Effects
Gaseous air toxics from nonroad diesel engines will also be reduced as a consequence of the
proposed standards. In addition, there will be a substantial reduction in SOx emissions resulting
from the proposed reduction in sulfur level in diesel fuel. SOx is transformed in the atmosphere
to form PM (sulfate).
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 and
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 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 (> 10 micrometers) tend to be removed by the respiratory clearance
mechanisms whereas smaller particles are deposited deeper in the lungs. Also, particulate
scatters light obstructing visibility.
The emission sources, formation processes, chemical composition, atmospheric residence
times, transport distances and other parameters of fine and coarse particles are distinct. Fine
AAmbient particulate matter from nonroad diesel engine is associated with the direct emission
of diesel particulate matter, and with particulate matter formed indirectly in the atmosphere by
NOx and SOx emissions (and to a lesser extent NMHC emissions). Both NOx and NMHC
participate in the atmospheric chemical reactions that produce ozone.
2-3
-------
Draft Regulatory Impact Analysis
particles are directly emitted from combustion sources and are formed secondarily from gaseous
precursors such as sulfur dioxide, oxides of nitrogen, or organic compounds. Fine particles are
generally composed of sulfate, nitrate, chloride, ammonium compounds, organic carbon,
elemental carbon, and metals. Nonroad diesels currently emit high levels of NOx which react in
the atmosphere to form secondary PM2 5 (namely ammonium nitrate). 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). Combustion of coal, oil, diesel, gasoline, and wood, as well as
high temperature process sources such as smelters and steel mills, produce emissions that
contribute to fine particle formation. In contrast, coarse particles typically result from
mechanical crushing or grinding in both natural and anthropogenic sources. They include
resuspended dusts, plant material, and crustal material from paved roads, unpaved roads,
construction, farming, and mining activities. Fine particles can remain in the atmosphere for days
to weeks and travel through the atmosphere hundreds to thousands of kilometers, while coarse
particles 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 (which is attributable to a number of sources including
diesel) contributes to a series of adverse health effects. These health effects are discussed in
detail in the EPA Air Quality Criteria Document for PM as well as the draft updates of this
document released in the past year.1 In addition, EPA recently released its final "Health
Assessment Document for Diesel Engine Exhaust," which also reviews health effects
information related to diesel exhaust as a whole including diesel PM, which is one component of
ambient PM.2
As detailed in these documents, health effects associated with short-term variation in ambient
particulate matter (PM) have been indicated by 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
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
2-4
-------
Air Quality, Health, and Welfare Effects
Two major cohort studies, the Harvard Six Cities and the ACS studies suggest an association
between exposure to ambient PM and premature mortality from cardiorespiratory causes.25'26
These are two prospective cohort studies that 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 5.27 This most recent finding is of special
interest in this proposal, because of the association of diesel exhaust and lung cancer in
occupational studies of varying design.
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.28'29'30'31'32'33 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.34'35'36'37
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 (white blood cell - WBC) 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.38'39'40'41 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.42'43'44'
45 These blood indices have been associated with coronary heart disease and cardiac events such
as heart attack.46'47 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.48'49
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.50'51
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.
Recently, 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.52 The 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
2-5
-------
Draft Regulatory Impact Analysis
studies employing generalized linear models or other parametric methods, as well as case-
crossover studies, are not affected. As discussed in HEI materials provided to EPA and to
CASAC, researchers working on the NMMAPS found problems in the default "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. These and other scientists have begun to reanalyze the results of several important time
series studies with alternative approaches that address these issues and have found a downward
revision of some results. For example, the mortality risk estimates for short-term exposure to
PM10 from NMMAPS were overestimated (this study was not used in this benefits analysis of
fine particle effects). However, both the relative magnitude and the direction of bias introduced
by the convergence issue is case-specific. In most cases, the concentration-response relationship
may be overestimated; in other cases, it may be underestimated. The preliminary reanalyses of
the mortality and morbidity components of NMMAPS suggest that analyses reporting the lowest
relative risks appear to be affected more greatly by this error than studies reporting higher
relative risks.53'54
During the compilation of the draft Air Quality Criteria Document, 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, reduced lower respiratory
symptoms, and reduced premature mortality due to short-term PM exposures. While resolution
of these issues is likely to take some time, the preliminary results from ongoing reanalyses of
some of the studies (Dominici et al, 2002; Schwartz and Zanobetti, 2002; Schwartz, personal
communication 2002) suggest a more modest effect of the S-plus error than reported for the
NMMAPS PM10 mortality study.55 In December 2002, a number of researchers submitted
reanalysis reports, and the HEI is currently coordinating review of these reports by a peer review
panel. The final report on these reanalyses is expected by the end of April 2003, and the results
will be incorporated in the fourth external review draft of the Criteria Document that will be
released in summer 2003. While we wait for further clarification from the scientific community,
we are not presenting the tables of short-term exposure effects from the draft Air Quality Criteria
Document. EPA will continue to monitor the progress of this concern, and make appropriate
adjustments as further information is made available.
The long-term exposure health effects of PM are summarized in Table 2.1.1-1 which is taken
directly from the draft Air Quality Criteria Document referenced earlier that was released in
2002. This document is continuing to undergo expert and public review.
2-6
-------
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
Range of City
PM Levels *
Indicator Increment in PM1 Means (ug/m3)
Change in Health Indicator per
Increment in PM1
Increased Total Mortality in Adults
Relative Risk (95% CI)
Six CityE
PM15/10 (20 ug/m3)
PM25 (10 ug/m3)
SOJ (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
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)
SOJ (15 ug/m3)
PM15/10 (20 ug/m3)
PM25 (10 ug/m3)
PM15/10 (20 ug/m3)
(SSI)
PM25 (10 ug/m3)
PM25 (10 ug/m3)
PM10 (50 ug/m3)
PM10 (cutoff =
30 days/year
>100 ug/m3)
PM10 (50 ug/m3)
PM10 (cutoff =
30 days/year
>100 ug/m3)
PM15/10 (50 ug/m3)
TSP (100 ug/m3)
H+ (100 nmol/m3)
SO: (15 ug/m3)
PM2 j (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)
1.02(0.99-1.04)
1.07(1.04-1.10)
1.04(1.01-1.08)
1.242 (0.955-1.616) (males)
1.082 (1.008-1. 162) (males)
0.879 (0.713-1.085) (females)
0.958 (0.899-1.021) (females)
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)
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
Table 2.1.1-1 (continued)
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
Increased Cough in Children
12 Southern California
communities'
(all children)
12 Southern California
communitiesK
(children with asthma)
PM10 (25 ug/m3)
Acid vapor (1. 7 ppb)
PM10 (19 ug/m3)
PM25 (15 ug/m3)
Acid vapor (1.8 ppb)
Change in Health Indicator per
Increment in PMa
Odds Ratio (95% CI)
1.06 (0.93, 1.21)
1.13 (0.92, 1.38)
1.1 (0.0.8, 1.7)
1.3 (0.7, 2.4)
1.4(0.9,2.1)
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
Increased Obstruction in Adults
Southern CaliforniaL
Decreased Lung Function in
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)
PM10 (cutoff of
42 days/year
>100 ug/m3)
Children
PM15/10 (50 ug/m3)
TSP (100 ug/m3)
H+ (52 nmoles/m3)
PM2 , (15 ug/m3)
SOI (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)
PM2 5 (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)
1.09 (0.92, 1.30)
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
NR
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
-------
Air Quality, Health, and Welfare Effects
Table 2.1.1-1 (continued)
Effect Estimates per Increments'1 in Long-term
Mean Levels of Fine and Inhalable Particle Indicators From U.S. and Canadian Studies
Range of City
PM Levels*
Means (ug/m3)
Type of Health
Effect and Location
Indicator
Change in Health Indicator per
Increment in PMa
Decreased Lung Function in Adults
Southern California1"
(% predicted FEVb
females)
Southern California1"
(% predicted FEVb males)
Southern California1"
(% predicted FEVb males
whose parents had asthma,
PM10 (cutoff of
54.2 days/year
>100 ug/m3)
PM10 (cutoff of
54.2 days/year
>100 ug/m3)
PM10 (cutoff of
54.2 days/year
>100 ug/m3)
+0.9 % (-0.8, 2.5) FEVj 52.7 (21.3, 80.6)
+0.3 % (-2.2, 2.8) FEVj 54.1 (20.0, 80.6)
-7.2 % (-11.5, -2.7) FEVj 54.1 (20.0, 80.6)
bronchitis, emphysema)
Southern California1"
(% predicted FEVb
females)
Southern California1"
(% predicted FEVb males)
I (1.6 ug/m3)
I (1.6 ug/m3)
Not reported
-1.5% (-2.9, -0.1)FEV!
7.4(2.7, 10.1)
7.3(2.0, 10.1)
*Range of mean PM levels given unless, as indicated, studies reported overall study mean (min, max), or mean
(±SD); NR=not reported.
AResults calculated using PM increment between the high and low levels in cities, or other PM increments given
in parentheses; NS Changes = No significant changes.
Schwartz, I; Dockery, D. W.; Neas, L. M. (1996) Is daily mortality associated specifically with fine particles?
J. Air Waste Manage. Assoc. 46: 927-939.
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.
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.
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.
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.
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.
Fairley, D. (1999) Daily mortality and air pollution in Santa Clara County, California: 1989-1996. Environ.
Health Perspect. 107: 637-641.
Burnett, R. T.; Brook, J.; 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.
2-9
-------
Draft Regulatory Impact Analysis
i Burnett, R. T.; Cakmak, S.; Brook, J. R.; Krewski, D. (1997) The role of paniculate size and chemistry in the
association between summertime ambient air pollution and hospitalization for cardiorespiratory diseases.
Environ. Health Perspect. 105: 614-620.
j Burnett, R. T.; Smith-Doiron, M; Stieb, D.; Cakmak, S.; Brook, J. R. (1999) Effects of paniculate 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.
n 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, PN2 5, SO42_, H+, and
O3. Am. J. Respir. Crit. Care Med. 160: 117-125.
O 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: paniculate matter and health.
Inhalation Toxicol. 12(suppl. 4): 75-90.
Most diesel PM is smaller than 2.5 microns based on extensive emissions characterization
studies and as reviewed in the recently release Diesel HAD (Health Assessment Document for
Diesel Exhaust).56'57 Since there are other sources of PM between the 2.5 to 10 micron range
(such as earth crustal material), diesel PM constitutes a smaller fraction of PM10 than it does of
PM25. EPA is also evaluating the health effects of PM between 2.5 and 10 microns in the draft
revised Air Quality Criteria Document.
In addition to the information in the draft revised Air Quality Criteria Document, further
conclusions about health effects associated with mobile source PM on-road diesel engine-
generated PM being relevant to nonroad application is supported by the observation in the Diesel
HAD that the particulate characteristics in the zone around nonroad diesel engines is likely to be
substantially the same as the characteristics of diesel particles in general (such as those found
along heavily traveled roadways).
Another body of studies have examined health effects associated with living near a major
roads. 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.58 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.59 Other studies have shown children living near roads with
2-10
-------
Air Quality, Health, and Welfare Effects
high truck traffic density have decreased lung function and greater prevalence of lower
respiratory symptoms compared to children living on other roads.60 Another recently published
study from Los Angeles found that maternal residence near heavy traffic during pregnancy is
associated with adverse birth outcomes, such as preterm birth and low birth weight.61 However,
these studies are not specifically related to PM, but to fresh emissions from mobile sources,
which includes other components as well.
Another recent cohort study examined the association between mortality and residential
proximity to major roads in the Netherlands. Examining a cohort of 55 to 69 year-olds from
1986 to!994, the study indicated that long-term residence near major roads, an index of exposure
to primary mobile source emissions (including diesel exhaust), was significantly associated with
increased cardiopulmonary mortality.62
Other studies have shown that living near major roads results in substantially higher
exposures to ultrafme 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
(WBCs) contained PM compared with children living on quiet streets.63 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 ultrafme
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.64'65'66
The particle characteristics in the zone around nonroad diesel engines is not likely to differ
substantially from published air quality measurements made along busy roadways. 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 of exposure misclassification in these studies. 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 PM25 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
2-11
-------
Draft Regulatory Impact Analysis
standards. Thus, since most PM emitted by diesel nonroad engines is fine PM, the emission
controls proposed today should contribute to attainment and maintenance of the existing PM
NAAQS. More broadly, the proposed standards will benefit public health and welfare through
reductions in direct diesel PM and reductions of NOx, SOx, and HCs which 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.
The reductions from today's proposed rules 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. L2. L J PM10 Levels
The current NAAQS for PM10 were first 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 ug/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 ug/m3 averaged over three years.
Currently, 29.5 million people live in PM10 nonattainment areas, including moderate and
serious areas. There are presently 58 moderate PM10 nonattainment areas with a total population
of 6.8 million. The attainment date for the initial moderate PM10 nonattainment areas, designated
by operation of 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. Four serious
PM10 nonattainment areas (Phoenix, Arizona; 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.B While all of these areas are
expected to be in attainment before the emission reductions from this proposed rule are expected
to occur, these reductions will be important to assist these areas in maintaining the standards.
Many PM10 nonattainment areas continue to experience exceedances. Of the 29.5 million
people living in designated PM10 nonattainment areas, approximately 25 million people are living
in nonattainment areas with measured air quality violating the PM10 NAAQS in 1999-2001.
BEPA has also proposed to grant Las Vegas, Nevada, an extension until December 31, 2006.
2-12
-------
Air Quality, Health, and Welfare Effects
Among these are the seven serious areas listed in Table 2.1.1-2 and 4 moderate areas: Nogales,
AZ, Imperial Valley, CA, Mono Basin, CA, and El Paso, TX.
Table 2.1.1-2
Serious PM10 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
Proposed
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
1999-2001 Measured
Violation
Yes
Yes
Yes
Yes
Yes
Yes
No
Yes
22.7 million
In addition to these designated nonattainment areas, there are 19 unclassified areas, where 8.7
million live, for which States have reported PM10 monitoring data for 1999-2001 period
indicating a PM10 NAAQS violation. Although we do not believe that we are limited to
considering only designated nonattainment areas a part of this rulemaking, we have focused on
the designated areas in the case of PM10. 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 1999-
2001 in these 19 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. The emission reductions from today's proposal would help States
improve their PM10 air quality levels and maintain the PM10 NAAQS.
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
2-13
-------
Draft Regulatory Impact Analysis
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 Fed. Reg., 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 to the standard is
referred to as the "design value.") The long-term standard specifies an expected annual
arithmetic mean not to exceed 15 ug/m3 averaged over three years.
Current PM2 5 monitored values for 1999-2001, which cover counties having about 75
percent of the country's population, indicate that at least 65 million people in 129 counties live in
areas where annual design values of ambient fine PM violate the PM2 5 NAAQS. There are an
additional 9 million people in 20 counties where levels above the NAAQS are being measured,
but there are insufficient data at this time to calculate a design value in accordance with the
standard, and thus determine whether these areas are violating the PM2 5 NAAQS. In total, this
represents 37 percent of the counties and 64 percent of the population in the areas with monitors
with levels above the NAAQS. Furthermore, an additional 14 million people live in 40 counties
that have air quality measurements within 10 percent of the level of the standard. These areas,
although not currently violating the standard, will also benefit from the additional reductions
from this rule in order to ensure long term maintenance.
Figure 2.1.1-4 is a map of currently available PM2 5 monitoring data, highlighting monitor
locations near or above the annual PM2.5 NAAQS. As can be seen from this figure, high
ambient levels are widespread throughout the East and California.
Figure 2.1.1-5 graphically presents the numbers of people currently exposed to various
unhealthy levels of PM25.67 As shown in Table 2.1.1-3 of the 74 million people currently living
in counties with measurements above the NAAQS, 22 million live in counties above 20 ug/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.
2-14
-------
Air Quality, Health, and Welfare Effects
Table 2.1.1-3
1999/2001 Monitored Populationa Living in Counties with Annual Average13 PM25
Concentrations Shown (70 Percent of Total U.S. Population)
Measured 1999/2000 Annual
Average PM2 5
Concentration
(Hg/m3)
(A)
>25
>20 <=25
>15 <=20
<=15
Number of Counties
Within The Concentration
Range
3
10
136
402
2000 Population Living
in Monitored Counties
Within The
Concentration Range
(Millions, 2000 Census
Data)
(B)
12.8
9.2
52.3
115.6
Cumulative Percent of
2000 Monitored
Population Living in
Counties Within The
Concentration Range0
(C)
7
5
27
61
a Monitored population estimates represent populations living in monitored counties (with community based monitors)
based on monitors with at least 10 quarter with at least 11 samples per quarter between 1999 and 2001.
b Annual average represents the monitor reading with the highest average in each monitored county.
0 The monitored population is 189.2 million (as reflected in column C, where C=B/Monitored Population). Total
monitored population is 191 million; the Census total county-based 2000 population is 272.7 million.
2-15
-------
Draft Regulatory Impact Analysis
Figure 2.1.1-4
Current Fine PM Monitoring Data
PM2.5: Status of 1999-2001 Monitoring
Data from AQS 7/8/32. Counties with s/fes thatoperated anytime 1999-2001 (1202 sites in 706 counties)
Counties with at least 1 complete site w/ d.v. > 15.0 [129]
Counties with at least 1 complete site w/ d.v. < 15.0 (and none above) [182]
Counties without a complete site [395]
2-16
-------
Air Quality, Health, and Welfare Effects
Figure 2.1.1-5
Populations Exposed to PM2.5 Levels
Above the NAAQS
15 17 19 21 23 25
PM2.5 design value
27
29
The relative contribution of various chemical components to PM2 5 varies by region of the
country. Data on PM2 5 composition are available from the EPA Speciation Trends Network in
2001 and the IMPROVE Network in 1999 covering both urban and rural areas in numerous
regions of the U. S. These data show that carbonaceous PM25 makes up the major component for
PM25 in both urban and rural areas in the Western U.S. Carbonaceous PM2 5 includes both
elemental and organic carbon. Nitrates formed from NOx also plays a major role in the western
U.S., especially in the California area where it is responsible for about a quarter of the ambient
PM2 5 concentrations. Sulfate plays a lesser role in these regions by mass, but it remains
important to visibility impairment discussed below. For the Eastern and mid U.S., these data
show that both sulfates and carbonaceous PM25 are major contributors to ambient PM2 5both
urban and rural areas. In some eastern areas, carbonaceous PM25 is responsible for up to half of
ambient PM2 5 concentrations. Sulfate is also a major contributor to ambient PM2 5 in the Eastern
U.S. and in some areas make greater contributions than carbonaceous PM25.
Nonroad engines, especially nonroad diesel engines, contribute significantly to ambient PM2 5
levels, largely through emissions of carbonaceous PM25. Carbonaceous PM25 is a major portion
of ambient PM25, especially in populous urban areas. Much of the total carbon PM excess is
organic carbon. Nonroad diesels also emit high levels of NOx which react in the atmosphere to
form secondary PM2 5 (namely ammonium nitrate). 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). Figure 2.1.1-1 shows the levels and composition of ambient PM2 5 in some
2-17
-------
Draft Regulatory Impact Analysis
urban areas.
Figure 2.1.1-2 shows the levels and composition of PM25 in rural areas where the total PM25
levels are generally lower. From Figures 2.1.1-1 and 2.1.1-2, one can compare the levels and
composition of PM25 in various urban areas and a corresponding rural area. This comparison, in
Figure 2.1.1-3, shows that much of the excess PM25 in urban areas (annual average concentration
at urban monitor minus annual average concentration at corresponding rural monitor) is indeed
from carbonaceous PM.68'69 See the AQ TSD for details.
2-18
-------
Figure 2.1.1-1
Annual Average PM2.5 Concentrations (pg/m3)
and Particle Type in Urban Areas, 2001
O 10ug/m3
I I Sulfate
I I Ammonium
•I Nitrate
^HTotal Carbon
•• Crustal Material
Source: EPA Sped til! on Arff/wA'h 2001
-------
Figure 2.1.1-2
Annual Average PM2.5 Concentrations
and Particle Type in Rural Areas,1999
ISulfete
Ammonium
Nitrate O 10 ug
Total Carbon O 15 ug/m3
Crustal Material fj 20 ug/m3
\^s
-------
Figure 2.1.1-3
Composition of Urban Excess PM25 at Selected Sites, 1999
(Source: U.S. EPA (2003) AQ TSD;Roa and Frank 2003)
Sulfate:
_ a D
0.0 0.4 0.9
Ammonium:
_ D D
0.0 0.9 1.9
Nitrate:
D
0.4 3.5 6.5
TCM(k=1.8):
D
li
2.9 8.1 13.2
Crustal:
_ a D
0.0 0.4 0.8
ichmond
-------
Draft Regulatory Impact Analysis
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, the proposed standards
would 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 proposed standards will also reduce SOx and VOC.
Nonroad diesel engines emissions also contribute to national SOx and VOC emissions
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.
2.1.2.2 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 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.
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 which will
require additional emission reductions to attain and maintain the PM2 5 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. This modeling supports the conclusion
that there is a broad set of areas with predicted PM25 concentrations at or above 15 ug/m3
between 1996 and 2030 in the baseline scenarios without additional emission reductions.
2-22
-------
Air Quality, Health, and Welfare Effects
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
U.S. for the various emissions scenarios. Version 7 of REMSAD was used for this proposed
rule. 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
proposed rule on U.S. PM concentrations. 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.
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 emissions 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).70 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
c Given the potential impact of the porposed 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.
2-23
-------
Draft Regulatory Impact Analysis
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 U.S. 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, there are few PM2 5 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 U.S.
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 U.S. mostly at
national parks, national wilderness areas, and other protected pristine areas.71 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 U.S., 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 to ambient observations (e.g.
1.23 is a 23 percent overprediction). Numbers less than 1 indicate underpredictions.
2-24
-------
Air Quality, Health, and Welfare Effects
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 U.S. 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 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
2-25
-------
Draft Regulatory Impact Analysis
that the results are limited by our current knowledge of PM science and chemistry, and by the
emissions 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 proposal also indicates that the present
widespread number of counties with annual averages above 15 ug/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 ug/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 within 10 percent
of the standard in 2020, which will increase to 64 million people in 2030. The AQ Modeling
TSD lists the specifics.
Our modeling also indicates that the reductions we are expecting from today's proposal will
make a substantial contribution to reducing these exposures.0 In 2020, the number of people
living in counties with PM25 levels above the NAAQS would be reduced from 66 million to 60
million living in 67 counties. That is a reduction of 9 percent in exposed population and 15
percent of the number of counties. In 2030, there would be a 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 exposed
population and a 21 percent of the number of counties. Furthermore, our modeling also shows
that the emission reductions would 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 ug/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 ambient monitoring data
sufficient to calculate current 3-year design values.
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
DThe results illustrate the type of PM changes for the preliminary control option, as discussed
in the Draft RIA in Section 3.6. The proposal differs from the modeled control case based on
updated information; however, we believe that the net results would approximate future
emissions, although we anticipate the PM reductions might be slightly smaller.
2-26
-------
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
GA
GA
IL
IL
IL
IL
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
Washington
Wilkinson
Cook
Du Page
Madison
St Clair
1999-2001
Design Value
(ug/m3)
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
16.5
18.1
18.8
15.4
17.3
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
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
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
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
21,176
10,220
5,376,741
904,161
258,941
256,082
-------
State
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
OH
OH
OH
OH
OH
OH
County
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
Mahoning
Montgomery
Scioto
Stark
Summit
Trumbull
1999-2001
Design Value
(ug/m3)
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
16.4
17.6
20
18.3
17.3
16.2
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
Control"
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
Population
in 2000
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
257,555
559,062
79,195
378,098
542,899
225,116
-------
State
PA
PA
PA
PA
SC
SC
TN
TN
TN
TN
TN
TX
TX
UT
VA
WV
WV
WV
WV
WV
WI
County
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)
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
Population of Violating Counties"
2020
Base
V
V
V
V
V
V
V
V
V
V
V
79
65,821,078
Control"
V
V
V
V
V
V
V
V
V
V
67
60,453,470
2030
Base
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
107
85,525,624
Control"
V
V
V
V
V
V
V
V
V
V
V
V
84
71,375,639
Population
in 2000
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 The proposal differs based on updated information; however, we believe that the net results would approximate future
emissions, although we anticipate the design value improvements would be slightly smaller.
b Populations are based on 2020 and 2030 estimates. See the AQ Modeling TSD for details.
Table 2.1.2-3 lists the counties with 2020 and 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!3.55 ug/m3, but less
than 15.05 ug/m3. Counties are marked with an "V" in the table if their projected design values
are greater than or equal to 15.05 ug/m3. The current 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 proposed rule will help assure that these counties continue to meet the standard.
-------
Table 2.1.3-3
Counties with 2020 and 2030 Projected Annual PM2.5 Design Values
within Ten Percent of the Annual PM2.5 Standard.a
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
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
Jefferson
Orleans
West Baton Rouge
1999-2001
Design Value
(ue/m3)
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
13.6
14.1
14.1
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
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
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
X
X
V
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
X
X
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
455,466
484,674
21,601
-------
State
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
PA
PA
PA
County
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
Lancaster
Washington
York
1999-2001
Design Value
rue/m3)
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
16.9
15.5
16.3
2020
Base
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
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
X
X
X
X
2030
Base
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
X
X
V
Control"
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
X
X
Population
in 2000
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
470,658
202,897
381,751
-------
State
SC
SC
SC
SC
TN
TN
TN
TN
TN
TX
UT
VA
VA
VA
VA
WV
WV
WV
WV
WI
WI
County
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
(ug/m3)
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%
Population of Counties within 10%b
2020
Base
X
X
X
X
X
X
X
X
X
X
X
X
X
X
V
X
70
23,836,367
Control"
X
X
X
X
X
X
X
X
X
X
X
X
X
62
24,151,782
2030
Base
X
V
X
X
V
X
V
V
X
V
V
X
V
X
X
X
X
X
V
V
X
64
16,870,324
Control"
X
X
X
V
X
X
X
X
X
X
X
X
X
X
X
V
X
70
24,839,565
Population
in 2000
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 The proposal differs based on updated information; however, we believe that the net results would approximate future
emissions, although we anticipate the design value improvements would be slightly smaller.
Populations are based on 2020 and 2030 estimates. See the AQ Modeling TSD for details.
We estimate that the reduction of this proposed rule would produce nationwide air quality
improvements in PM levels. On a population weighted basis, the average change in future year
annual averages would be a decrease of 0.33 ug/m3 in 2020, and 0.46 ug/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's current plans call for designating PM2 5 nonattainment areas in
late-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 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
-------
Air Quality, Health, and Welfare Effects
date of designation as nonattainment." Therefore, based on this information, we expect that most
or all areas will need to attain the PM2 5 NAAQS in the 2009 to 2014 time frame, and then be
required to maintain the NAAQS thereafter.
Since the emission reductions expected from today's proposal would begin in this same time
frame, the projected reductions in nonroad emissions would 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.72'73'74 Furthermore, this action would ensure 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 later.
2.1.3 Welfare 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.
2.1.3.1 Visibility Degradation
Visibility can be defined as the degree to which the atmosphere is transparent to visible
light.75 Visibility impairment has been considered the "best understood and most easily
measured effect of air pollution."76 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. Haze obscures the clarity, color, texture, and form of what
we see. Fine particles are the major cause of reduced visibility in parts of the U.S. 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.
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. Sulfates contribute to visibility
impairment especially on the haziest days across the U.S., accounting in the rural Eastern U.S.
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 Eastern U.S.
locations.77
2-33
-------
Draft Regulatory Impact Analysis
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.78 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. 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.79
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.
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.80
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.81
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.82 In the Eastern U.S., 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
particulate load than in the East. Because of significant differences related to visibility
conditions in the Eastern and Western U.S., 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
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.
2-34
-------
Air Quality, Health, and Welfare Effects
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
which 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 proposed 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 mandatory Federal Class I areas, we discuss the types of effects, current and
future visibility conditions absent the proposed reductions, and the changes we anticipate from
the proposed reductions in emissions from nonroad diesels. We conclude that the proposed
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
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.83 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 U.S., 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).
2-35
-------
Draft Regulatory Impact Analysis
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.84 Regional strategies 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.85
Without the effects of pollution, a natural visual range is approximately 140 miles (230
kilometers) in the West and 90 miles (150 kilometers) in the East. However, over the years, in
many parts of the U.S., fine particles have significantly reduced the range that people can see. In
the West, the current 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).86
Current PM2 5 monitored values for 1999-2001 indicate that at least 65 million people in 129
counties live in areas where design values of PM25 annual levels are at or above the PM25
NAAQS. There are an additional 9 million people in 20 counties where levels exceeding the
NAAQS are being measured, but there are insufficient data at this time to make a complete
comparison with the NAAQS. In total, this represents 37 percent of the counties and 64 percent
of the population in the areas with monitors with levels above the NAAQS. Taken together,
these data indicate that a total of 74 million people live in areas where long-term ambient fine
particulate matter levels are at or above 15 jig/m3.87 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 14 million people live in 41 counties that have air quality measurements for
1999-2001 within 10 percent of the level of the PM standard. These areas, although not currently
violating the standard, would also benefit from the additional reductions from this proposed rule
in order to ensure long term maintenance of the standard and to prevent deterioration in visibility
conditions.
2-36
-------
Air Quality, Health, and Welfare Effects
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 reductions from this
proposed 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 PM25 levels are
above 15 |ig/m3.88 In 2030, this number will rise to 107 counties with a population of 71 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 to 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 deciviews
(or visual range of 162 kilometers) in 2030, compared to 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.
2-37
-------
Draft Regulatory Impact Analysis
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
a 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 proposed rule contribute to these effects. 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 Proposed Reductions
For this proposal, we also modeled a preliminary control scenario which illustrates the likely
reductions from our proposal. 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 process
based on our best judgement at that time. As additional data regarding technical feasibility and
other factors became available, our judgement 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 below. It is important to note that these changes would not affect our estimates of the
baseline conditions without additional controls described above. For the final rule, considering
public comment, we plan to model the final control scenario. We anticipate that the proposed
nonroad diesel emissions reductions would improve to the projected visibility impairment, and
that there continues to be a need for reductions from those engines.
Based on our modeling, we predict that in 2020, there would be 12 counties with a population
of 6 million that would come into attainment with the annual PM25 because of the improvements
in air quality from the proposed emissions reductions. In 2030, a total of 24 counties (12
additional counties) with a population of 14 million (8 million additional people) would come
2-38
-------
Air Quality, Health, and Welfare Effects
into attainment with the annual PM2 5 because of the improvements in air quality from the
proposed emissions reductions. There would also be reductions in counties with levels close to
the standard that would improve visibility conditions and help them maintain the standards. All
of these areas and their populations would experience improvements in visibility as well as
health, described earlier.
We estimate that the reduction of this proposed rule would produce nationwide air quality
improvements in PM levels. On a population weighted basis, the average change in future year
annual averages would be a decrease of 0.33 ug/m3 in 2020, and 0.46 ug/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 proposed rule in 2030 would improve visibility by 0.33 deciview. Likewise,
we estimate visibility in the West to be about 8.83 deciviews (or visual range of 162 kilometers)
in 2030, and we estimate emission reductions from this proposed rule in 2030 would improve
visibility by 0.25 deciview. These improvements are needed in conjunction with other sulfur
strategies in the East and a combination of strategies in the West to make reasonable progress
toward visibility goals.89 Thus, in the future, a substantial percent of the population may
experience improvements visibility in areas where they live, work and recreate because of the
proposed nonroad emission reductions.
2-39
-------
Draft Regulatory Impact Analysis
Table 2.1.3-2
Summary of Future National Visibility Improvements
from Nonroad Diesel Emission Reductions (Annual Average Deciviews)
Regions*
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 proposal differs based on updated information; however, we believe that the net results would approximate future
PM emissions, although we anticipate the visibility improvements would be slightly 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.90 In the East, there are and will
continue to be sizable areas above 15 ug/m3 and where light extinction is significantly above
natural background. Thus, large areas of the Eastern U.S. 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 in these and other park and wilderness areas each year.
2-40
-------
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.91 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 emissions 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.92
The conclusions draw upon the Interagency Monitoring of Protected Visual Environments
(IMPROVE) network data.93
As described in the EPA Trends Report, most of the IMPROVE sites in the intermountain
West and Colorado Plateau have annual average impairment of 12 deciviews or less, with the
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.
2-41
-------
Draft Regulatory Impact Analysis
worst days ranging up to 17 deciviews (compared to 5.3 deciviews of natural background
visibility).94 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 Eastern U.S.
Class I areas trend sites, visibility on the haziest 20 percent of days remains significantly
impaired with a mean visual range of 23 kilometers for 1999 as compared to 84 kilometers for
the clearest days in 1999. In the 26 Western U.S. Class I areas trends sites, 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.95 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.
2-42
-------
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 Emissions 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
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).
2-43
-------
Figure 2.1.3-1
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.
-------
Air Quality, Health, and Welfare Effects
2.1.3.1.2.3 Future Improvements in Mandatory Federal Class I Visibility/Torn the
Proposed 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 emissions 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 which illustrates the
likely reductions from our proposal. 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 Regions*
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 proposal differs based on updated information; however, we believe that the net results would approximate future
PM emissions, although we anticipate the visibility improvements would be slightly smaller.
2-45
-------
Draft 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.96 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 U.S. 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 emissions trends and acidity of water bodies in the Eastern U.S. 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.97
The decrease in sulfates is consistent with emissions 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.98 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 U.S.
2-46
-------
Air Quality, Health, and Welfare Effects
The SOx and NOx reductions from today's action 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 today's action 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 US 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
2-47
-------
Draft Regulatory Impact Analysis
associated impairments of freshwater and estuarine resources for human uses." Since 1970,
eutrophic conditions worsened in 48 estuaries and improved in 14. In 26 systems, there was no
trend in overall eutrophication conditions since 1970. 10° 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 U.S., 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.101 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.102 A review of peer reviewed
literature in 1995 on the subject of air deposition suggests a typical contribution of 20 percent or
higher.103 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 1997.104
The National Exposure Research Laboratory, US EPA, estimated based on prior studies that 20
to 35 percent of the nitrogen loading to the Chesapeake Bay is attributable to atmospheric
deposition.105 The mobile source portion of atmospheric NOx contribution to the Chesapeake
Bay was modeled at about 30 percent of total air deposition.106
Deposition of nitrogen from nonroad diesel engines contributes to elevated nitrogen levels in
waterbodies. The proposed standards for nonroad diesel engines will reduce total NOx emissions
by 831,000 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.107 One of these
15 compounds, a group known as polycyclic 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 7 most toxic
POM is at least 62 tons/year108 and represents only those POM that are adhered to mobile source
parti culate emissions. 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.
2-48
-------
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 PM reductions from today's proposed action will help reduce not only the PM emissions
from land-based nonroad diesel engines but also the deposition of the POM adhering to the
particles, thereby helping to reduce health effects of POM in lakes and streams, accelerate the
recovery of affected lakes and streams, and revive the ecosystems adversely affected.
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
2-49
-------
Draft Regulatory Impact Analysis
degradation of its appearance. When such accumulation produces sufficient changes in 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 which 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).109 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 emissions controls is appropriate.
EPA recently 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.110 In accordance with earlier EPA guidelines, diesel
exhaust would be similarly classified as a probable human carcinogen (Group Bl).111'112 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 US Department of Health and Human Services) have made similar
classifications.113'114'115'116'117 The Health Effects Institute has also made numerous studies and
2-50
-------
Air Quality, Health, and Welfare Effects
report on the potential carcinogen!city of diesel exhaust.118'119' 12° Numerous animal and
bioassay/genotoxic tests have been done on diesel exhaust.121'122 Also, case-control and cohort
studies have been conducted on railroad engine exposures123'124'125 in addition to studies on truck
workers. 126'127'128 Also, there are numerous other epidemiologic studies including some studying
mine workers and fire fighters.129' 13°
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 cleaner 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 onroad and nonroad engines. The EPA Diesel HAD
acknowledges that the studies were done on engines with older technologies generally for onroad
and that "there have been changes in the physical and chemical composition of some DE [diesel
exhaust] emissions (onroad vehicle emissions) over time, 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 onroad engines which have been subject to PM standards since 1988.
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.131 Another cohort study of operators of heavy construction equipment also
showed increased lung cancer incidence for these workers.132
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.133'134
That is, these two studies show an overall increase in lung cancer for the exposed groups of 35
percent and 47 percent compared to 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 ug/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
and the absence of quantitative exposure characterization in retrospective studies.
2-51
-------
Draft Regulatory Impact Analysis
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 to 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 ug./m3 for diesel PM from on-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 ug/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 judgement. 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
occupational settings where increased lung cancer risk is reported and ambient environmental
2-52
-------
Air Quality, Health, and Welfare Effects
exposures. The potential for small exposure differences underscores the concerns about the
appropriateness of extrapolation from occupational risk to ambient environmental exposure
levels should be more confidently judged to be appropriate.
EPA also recently assessed air toxic emissions and their associated risk (the National-Scale
Air Toxics Assessment or NATA for 1996), and we concluded that diesel exhaust ranks with
other substances that the national-scale assessment suggests pose the greatest relative risk.135 This
national assessment estimates average population inhalation exposures to diesel PM in 1996 for
nonroad as well as onroad 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 |j.g/m3 used to develop the risk perspective in
the Diesel HAD. The average nationwide NATA mobile exposure levels are 1.44 |j.g/m3 total
with an onroad source contribution of 0.46 |j.g/m3 and a nonroad source contribution of 0.98
Hg/m3. The average urban exposure was 1.64 |j.g/m3 and the average rural exposure was 0.55
Hg/m3. In five percent of urban census tracts across the United States, average exposures were
above 4.33 |j.g/m3. The EPA Diesel HAD states that use of the NATA exposure estimates instead
of the 0.8 |j.g/m3 estimate results in a similar risk perspective.
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 has been designed a likely human
carcinogen due to inhalation at environmental exposure. Other organizations have made
similar determinations.
2. The entire 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 US 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
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
2-53
-------
Draft Regulatory Impact Analysis
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.136,137,138,139 The diesel RfC is based on a "no observable adverse effect" level of 144
ug/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
ug/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 is presently
lacking to derive an RfC.
While there have been relatively few human 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 is relevant to diesel PM is 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.140'141'142'143 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 ug/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 ug/m3
provides adequate public health protection for ambient PM2 5. In fact, there may be benefits to
reducing diesel PM below 5 ug/m3 since diesel PM is a major contributor to ambient PM2 5 .G
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 diesels in
GIt should again be noted that recent epidemiologic studies (such as by Schwartz, Laden, and
Zanobetti) of ambient PM2 5 do not indicate a threshold of effects at low concentrations.
2-54
-------
Air Quality, Health, and Welfare Effects
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 and others)
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. In a review discussed earlier
of studies (by R. Delfmo) of the respiratory health of people living near roadways, another
publication indicated that the risk of asthma and related respiratory disease appeared elevated in
people living near heavy traffic. These studies offer evidence that people exposed most directly
to emissions from mobile sources including those from diesels 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.
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 which are 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 on-highway diesel PM.
2.2.1.3.1 Toxics Modeling and Methods
In addition to the general ambient PM modeling conducted for this proposal, diesel PM
concentrations for 1996 were recently 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 which 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.
2-55
-------
Draft Regulatory Impact Analysis
From the NATA 1996 modeling, overall mean annual national ambient diesel PM levels of
2.06 |j.g/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 diesels. A map of
county median concentrations is provided in Figure 2.2.1-1. While the high median
concentrations are clustered in the Northeast, Great Lake States and California, areas of high
median concentrations are distributed throughout the U.S.
Table 2.2.1-1
Distribution of Average Ambient Concentrations of
Diesel PM at the National Scale in the 1996 NATA Assessment.
5th Percentile
25thPercentile
Average
75th Percentile
95th Percentile
Onroad Contribution
to Average
Nonroad Contribution
to Average
Nationwide (ng/m3)
0.33
0.85
2.06
2.45
5.37
0.63
1.43
Urban (jig/m3)
0.51
1.17
2.41
2.7
6.06
0.72
1.69
Rural (ng/m3)
0.15
0.42
0.74
0.97
1.56
0.27
0.47
2-56
-------
Figure 2.2.1-1
Estimated County Median Concentrations of Diesel Particulate Matter
1996 Estimated County Median Ambient Concentrations
Diesel Partbulate Matter — United States Counties
Distribution of U.S. Ambient Concentrations
HlghaetlnU.S. ^^^_ 15
95 | | 1.90
90
Percentile 75
so
£5
Lowaet In U.S.
1-47 County Median Ambient Pollutant Concentration
( micraqrams / cubic meter )
* "
0.38
0.014-
Source; U.S. EPA / QijQPS
National— Scale Afr Toxfcs Assessment
Source: EPA National-Scale Air Toxics Assessment for 1996.
-------
Draft 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.144
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.145
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.146 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 |j.g/m3 at were
estimated at a receptor versus a 3 |j.g/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.
2-58
-------
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 AvenmgQ Concentrations QigAn3]. ISC Roads. Houston. TX. 1996
DiaselPm, All Sources- no Background
220
240 260 2EO
UTM Zone 15 West-East Distance (km)
3OO
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
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
2-59
-------
Draft Regulatory Impact Analysis
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 particulate 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.147
Results from several source apportionment studies were used to develop these factors.148'149'150'
i5i, 152,153,154 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
• East or West US
• 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.155 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 |j.g/m3. EPA has
compared these estimates based on elemental carbon measurements to 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 inventory
model. The corresponding fractions of modeled values within 100 % of the monitored value are
2-60
-------
Air Quality, Health, and Welfare Effects
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 NAT A 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 onroad and
nonroad but, instead, gives diesel PM as a whole. One can allocate the diesel PM numbers based
on the inventory split between onroad 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 |J.g/m3 annual
average. UNMIX/PMF models show similar results.
2-61
-------
Draft 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 s concentration)
Ambient
Measurement
Technique: TOT
or TOR
TOT
East or
Wptf
East
East
East
East
West
Sipnsnn
Fall (Q4)
Spring (Q2)
Summer
(03)
Winter (Ql)
Unknown
Location
Type
rrpnprnl
Mixed
Mixed
Mixed
Mixed
Urban
TOT Total
TOR
Winter
Winter
Rural
Urban
Winter Total
TOR Total
Grand Total
MTNP
2.3
2.4
2.1
2.2
1.2
1.2
0.6
0.5
0.5
0.5
0.5
MAYa
2.3
2.4
2.1
2.2
2.4
2.4
1.0
1.0
1.0
1.0
2.4
AVFRArTFa
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.
a 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.
2-62
-------
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
Variah1pb
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
%
DiffnrRnp.R
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.
2-63
-------
Draft 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 on-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 1982.156 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 |j.g/m3. It should be noted 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).157 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 |j.g/m3 in an urban area,
and 1.2 |j.g/m3 in a rural area. Source profiles in this study were based on onroad vehicles.
The CMB model was applied in California's San Joaquin Valley during winter 1995-1996.158
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 |j.g/m3. In the Kern Wildlife Refuge, diesel PM source contribution
estimates were 1.32 and 1.75 |j.g/m3 during the two periods.
The CMB model was applied in the southeastern U.S. 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 |j.g/m3 in N.
Birmingham, AL, 1.02-2.43 |j.g/m3 in Gulfport, MS, 3.29-5.56 |j.g/m3 in Atlanta, GA, and
Pensacola, FL 1.91-3.07 |j.g/m3 which represented the urban sites in the study. Suburban sites in
the study were located outside Pensacola, FL (1.08-1.73 ng/m3). Rural sites were located in
Centreville, AL (0.79-1.67 |ag/m3), Oak Grove, MS (1.05-1.59 |ag/m3), and Yorkville, GA (1.07-
2.02 |ag/m3).
The CMB model was applied to ambient PM2 5 data collected during a severe photochemical
2-64
-------
Air Quality, Health, and Welfare Effects
smog event during 1993 in Los Angeles using organic tracers.159 Modeled concentrations of
diesel contributions to PM2 5 during this episode were conducted for Long Beach (8.33 ng/m3),
downtown Los Angeles (17.9 |j.g/m3), Azusa (14.9 |j.g/m3), and Claremont, CA (7.63 |j.g/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 onroad 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 |j.g/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 |j.g/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.160'161'162'163 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 |j.g/m3
level. Estimates for exposures to diesel PM for diesel fork lift operators have been made that
range from 7 to 403 |j.g/m3 as reported in the Diesel HAD. In addition, the Northeast States for
Coordinated Air Use Management (NESCAUM) is presently measuring occupational exposures
2-65
-------
Draft Regulatory Impact Analysis
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 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. These data, tentatively scheduled to be available in about a year, will be useful in
quantifying high localized exposure levels in the vicinity of nonroad equipment.164 Some initial
results are expected in late 2003.
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 on-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
carbon measurements.165 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 |j.g/m3 with levels of 1.64 |j.g/m3 for urban
counties and 0.55 |j.g/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.
2-66
-------
Air Quality, Health, and Welfare Effects
Table 2.2.1-4
Distribution of Average Exposure Concentrations to
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
(Hg/m3)
0.16
0.58
1.44
1.73
3.68
0.46
0.98
Urban (ng/m3)
0.29
0.81
1.64
1.91
4.33
0.52
1.12
Rural (ng/m3)
0.07
0.29
0.55
0.67
1.08
0.21
0.34
2-67
-------
Figure 2.2.1-3
Estimated County Median Exposure Concentrations of Diesel Particulate Matter
1996 Estimated County Median Exposure Concentration
Diesel Partbulate Matter — United States Counties
Saeram*
Distribution of U.S. hi halation Exposure Concentration
Hlghset In U.S. ^^^_ 10.2
95 | | US
90
75
50
25
Lewset In U.S.
Percent! le
Q
^•°unt/ Median Exposure Concentration
( micrograms / cubic meter)
Source: U.S. EPA / QAQPS
National— Scale Ai'r Toxfcs Assessment
Source: EPA National-Scale Air Toxics Assessment for 1996.
-------
Air Quality, Health, and Welfare Effects
As explained earlier, the fact that these levels are below the 5 ug/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 PM25 regardless of its composition. In other words, all of the health
effects cited in the implementation of the PM25 NAAQS apply to diesel PM.
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 onroad rather than nonroad
diesels, there are many similarities between on-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 (CARB) has
measured elemental carbon near the Long Beach Freeway in 1993.166 Levels measured ranged
from 0.4 to 4.0 |j.g/m3 (with one value as high as 7.5 ug/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 diesels.
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). For some of these pollutants, nonroad diesel engine emissions are believed to account
for a significant proportion of total nationwide emissions. 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. Mobile sources contribute significantly to total emissions of these air toxics. As
discussed later in this section, this proposed rulemaking will result in significant reductions of
these emissions.
2-69
-------
Draft Regulatory Impact Analysis
Nonroad engines are major contributors to nationwide cancer risk from air toxic pollutants, as
indicated by the NATA 1996.167 In fact, this study and the National Toxics Inventory (NTI) for
1996 are used throughout this section for toxics inventory information for nonroad sources.168
Also, a supplemental paper provides more detail on nonroad diesel.169 In addition, a paper
published by the Society of Automotive Engineers gives future projections to 2007 for these air
toxics.170 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.171 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.172 The
EPA proposed rulemaking will result in reductions of 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.
2-70
-------
Air Quality, Health, and Welfare Effects
Figure 2.2.2-1
1996 Risk Characterization
at lifeline cancer risk lor Irte US population, based on 199<6" expomre
In 23 Laftmaqerik an polluLanls from vaikxis MKJICC sectors
0.001
C01
Upcer-Bound Lffepn^ Career Hl&li par Million
fl 1 1 10
MX.
2-71
-------
Figure 2.2.2-2
Contribution of Source Sectors to Average Annual Nationwide Inhalation Exposure to Air Toxics in 1996
1
.a
o
O
100%
90%
80%
70%
60%
50%
40%
30%
20%
10%
0%
• Background
QArea
QMajor
• Nonroad Mobile
QOnroad Mobile
Benzene
1,3-Butadiene
Form aldehyde
Pollutant
Acetaldehyde
Acrolein
Source: National Scale Air Toxics Assessment.
-------
Air Quality, Health, and Welfare Effects
2.2.2.1 Benzene
Benzene is an aromatic hydrocarbon which 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.173'174 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.175
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).
2-73
-------
Draft Regulatory Impact Analysis
Figure 2.2.2-3
Contribution of Source Sectors to Total Average
Nationwide Mobile Source Ambient Concentrations in 1996
100%
90%
Locomotives9
D Commercial Marines
Aircraft?
D Diesel Nonroad EnginesG
• 4-Stroke Gasoline NonroadS
Q2-Stroke Gasoline Nonroad4
D Heavy Duty Diesel VehiclesS
• Light Duty DieseG
D Gasoline highway vehiclesl
Pollutant
The EPA's IRIS database lists benzene as a known human carcinogen (causing leukemia at
high, prolonged air exposures) by all routes of exposure.176 It is associated with additional
health effects including genetic changes in humans and animals and increased proliferation of
bone marrow cells in mice.177'178 EPA states in its IRIS database that the data indicate a causal
relationship between benzene exposure and acute lymphocyte leukemia and suggest a
relationship between benzene exposure and chronic non-lymphocytic leukemia and chronic
lymphocytic leukemia. Respiration is the major source of human exposure and at least half of
this exposure is attributable to gasoline vapors and automotive emissions. A number of adverse
noncancer health effects including blood disorders, such as preleukemia and aplastic anemia,
have also been associated with low-dose, long-term exposure to benzene.
Respiration is the major source of human exposure to benzene. Long-term respiratory
exposure to high levels of ambient benzene concentrations has been shown to cause cancer of the
2-74
-------
Air Quality, Health, and Welfare Effects
tissues that form white blood cells. Among these are acute nonlymphocytic leukemia,11 chronic
lymphocytic leukemia and possibly multiple myeloma (primary malignant tumors in the bone
marrow), although the evidence for the latter has decreased with more recent studies.179'180
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 genetic changes in humans and animals181 and increased
proliferation of mouse bone marrow cells.182 The occurrence of certain chromosomal changes in
individuals with known exposure to benzene may serve as a marker for those at risk for
contracting leukemia.183
The latest assessment by EPA places the excess risk of developing acute nonlymphocytic
leukemia at 2.2 x 10"6 to 7.7 x 10"6/|J.g/m3. In other words, there is a risk of about two to eight
excess acute nonlymphocytic leukemia cases in one million people exposed to 1 |j.g/m3 over a
lifetime (70 years).184 This range of unit risk represents the maximum likelihood estimate of risk.
Figure 2.2.2-4 depicts the distribution of upper bound lifetime cancer risk from inhalation of
benzene from ambient sources, based on average population exposure, from the 1996 NAT A
Assessment. Upper bound cancer risk is above 10 in a million across the entire U.S. EPA
projects a median nationwide reduction in ambient concentrations of benzene from mobile
sources of about 46percent 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.
HLeukemia 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.
2-75
-------
Figure 2.2.2-4
Distribution of Upper Bound Lifetime Cancer Risk from Inhalation of
Benzene from Ambient Sources, Based on Average Population Exposure
1996 Estimated County Median Cancer Risk
Benzene — United States Counties
Saerams
Upper—Bound Lifetime Cancer Risk
100 In a million
30 In a million
10 In a million
3 In a million
1 In a million
J In a million
0
Source: U.S. EPA / OAQPS
NATA National—Scale Afr Toxfcs Assessment
Source: 1996 NAT A Assessment.
-------
Air Quality, Health, and Welfare Effects
A number of adverse noncancer health effects, blood disorders such as preleukemia and
aplastic anemia, have also been associated with low-dose, long-term exposure to benzene.185'186
People with long-term exposure to benzene may experience harmful effects on the blood-forming
tissues, especially the bone marrow. These effects can disrupt normal blood production and
cause a decrease in important blood components, such as red 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 for blood to clot). Chronic inhalation exposure to benzene in humans
and animals results in pancytopenia,1 a condition characterized by decreased numbers of
circulating erythrocytes (red blood cells), leukocytes (white blood cells), and thrombocytes
(blood platelets).187'188 Individuals that develop pancytopenia and have continued exposure to
benzene may develop aplastic anemia/ whereas others exhibit both pancytopenia and bone
marrow hyperplasia (excessive cell formation), a condition that may indicate a preleukemic
state.18919° The most sensitive noncancer effect observed in humans is the depression of absolute
lymphocyte counts in the circulating blood.191
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 NAT A, 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 U.S. Of ambient butadiene levels due to mobile
Pancytopenia 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.
JAplastic 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).
2-77
-------
Draft Regulatory Impact Analysis
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.192
Recently EPA redesignated 1,3-butadiene as a known human carcinogen.193'194'195 The specific
mechanisms of 1,3-butadiene-induced carcinogenesis are unknown. However, it is virtually
certain 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 unit cancer risk estimate is 0.08/ppm or 3x10-5 per |J.g/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 |J.g/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.196 A twofold adjustment to the epidemiologic-based unit
cancer risk was 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. Figure 2.2.2-5 depicts the distribution of upper bound lifetime cancer
risk from inhalation of 1,3-butadiene from ambient sources, based on average population
exposure, from the 1996 NATA Assessment. Upper bound cancer risk is above 10 in a million
across the entire U.S. 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.
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.197 Based on this critical effect and the
benchmark concentration methodology, an RfC (i.e., a chronic exposure level presumed to be
"without appreciable risk" for noncancer effects) was calculated. This RfC for chronic health
effects was 0.9 ppb.
2-78
-------
Figure 2.2.2-5
Distribution of Upper Bound Lifetime Cancer Risk from
Inhalation of 1,3-Butadiene from Ambient Sources, Based on Average Population Exposure
1996 Estimated County Median Cancer Risk
1,3—Butadiene — United States Counties
Saoramsi
Upper—Bound Lifetime Cancer Risk
1QQ In a million
30 In o million
10 In a million
3 In a million
1 In a million
J3 In a mil lion
0
Source: US. EPA / QAQPS
NATA National-Scale AM- Toxfcs Assessirient
Source: 1996 NATA Assessment.
-------
Draft Regulatory Impact Analysis
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 which
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.198 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 NAT A, 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 has classified formaldehyde as a probable human carcinogen based on limited evidence
for carcinogen!city in humans and sufficient evidence of carcinogen!city in animal studies, rats,
mice, hamsters, and monkeys.199' 20° 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.201 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.202'203'204 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.205 Research has demonstrated that
formaldehyde produces mutagenic activity in cell cultures.206
The upper confidence limit estimate of a lifetime extra cancer risk from continuous
formaldehyde exposure is about 1.3 * 10"5/|J.g/m3. In other words, it is estimated that
approximately 10 persons in one million exposed to 1 |j.g/m3 formaldehyde continuously for their
lifetime (70 years) would develop cancer as a result of this exposure. The agency is currently
conducting a reassessment of risk from inhalation exposure to formaldehyde based on new
information including a study by the Chemistry Industry Institute of Toxicology.207'208 Figure
2.2.2-6 depicts the distribution of upper bound lifetime cancer risk from inhalation of
formaldehyde from ambient sources, based on the current unit risk and average population
2-80
-------
Air Quality, Health, and Welfare Effects
exposure from the 1996 NATA Assessment. Upper bound cancer risk is above 10 in a million
for more than one hundred million Americans. EPA projects a median nationwide reduction in
ambient concentrations of benzene from mobile sources of about 43 percent between 1996 and
2007, as a result of current and planned control programs (Cook et al., 2002).
Formaldehyde exposure also causes a range of noncancer health effects. At low
concentrations (e.g. 0.05-2.0 ppm), irritation of the eyes (tearing of the eyes and increased
blinking) and mucous membranes is the principal effect observed in humans. At exposure to 1-
11 ppm, 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.209 In persons with bronchial
asthma, the upper respiratory irritation caused by formaldehyde can precipitate an acute
asthmatic attack, sometimes at concentrations below 5 ppm.210 Formaldehyde exposure may also
cause bronchial asthma-like symptoms in non-asthmatics.211212
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.
2-81
-------
Figure 2.2.2-6
Distribution of Upper Bound Lifetime Cancer Risk from Inhalation
of 1,3-Butadiene from Ambient Sources, Based on Average Population Exposure
1996 Estimated County Median Cancer Risk
Formaldehyde — United States Counties
Upper—Bound Lifetime Cancer Risk
100 In a million
30 In a million
10 In a million
3 In a million
1 In o million
In a mil lion
Source: U.S. EPA / QAQPS
National-Scale Air Toxics Assessment
Source: 1996 NAT A Assessment.
-------
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 which
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 form 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.
Acetaldehyde is classified 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)'213'214'215'216'217 The upper confidence limit estimate of a lifetime extra cancer
risk from continuous acetaldehyde exposure is about 2.2 x 10"6 /|j.g/m3. In other words, it is
estimated that about 2 persons in one million exposed to 1 |j.g/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. Figure
2.2.2-7 depicts the distribution of upper bound lifetime cancer risk from inhalation of
formaldehyde from ambient sources, based on the current unit risk and average population
exposure from the 1996 NATA. Upper bound cancer risk is above one in a million for more
than one hundred million Americans. EPA projects a median nationwide reduction in ambient
concentrations of benzene 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.218 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
2-83
-------
Figure 2.2.2-7
Distribution of Upper Bound Lifetime Cancer Risk from Inhalation of
Acetaldehyde from Ambient Sources, Based on Average Population Exposure
1996 Estimated County Median Cancer Risk
Acetaldehyde — United States Counties
Saeramsi
Upper—Bound Lifetime Cancer Risk
1QQ In a million
jfl In a million
10 In a million
3 In a million
1 In a million
J In a million
Q
Source: US, EPA / 04QPS
MATA National— Scale Afr Toxfcs Assessment
Source: 1996 NAT A Assessment.
-------
Air Quality, Health, and Welfare Effects
acetaldehyde on reproductive and developmental effects. The in vitro and in vivo studies provide
evidence to suggest that acetaldehyde may be the causative factor in birth defects observed in
fetal alcohol syndrome, though evidence is very limited linking these effects to inhalation
exposure. Long-term exposures should be kept below the reference concentration of 9 |j.g/m3 to
avoid appreciable risk of these noncancer health effects (EPA, 1988).
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.219
2.2.2.5 Acrolein
In a recent test program which 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 acetaldehyde 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 extremely toxic to humans from the inhalation route of exposure, with acute
exposure resulting in upper respiratory tract irritation and congestion. The Agency developed a
reference concentration for inhalation (RfC) of acrolein of 0.02 |j.g/m3 in 1993. Figure 2.2.2-8
depicts the distribution of hazard quotients for acrolein across the U.S.K The hazard quotient is
greater than one for most of the U.S. population, indicating a potential for adverse noncancer
health effects.
Although no information is available on its carcinogenic effects in humans, based on
laboratory animal data, EPA considers acrolein a possible human carcinogen.220
KThe hazard quotient is the ratio of average ambient exposure over the reference
concentration (level below which adverse health effects are not expected to occur). A hazard
quotient above one indicates the potential for adverse health effects, but does not necessarily
mean adverse health effects will occur.
2-85
-------
Figure 2.2.2-8
Distribution of Noncancer Hazard Quotients for Inhalation
of Acrolein from Ambient Sources, Based on Average Population Exposure
1996 Estimated County Median Noncancer Hazard
Acrolein — United States Counties
Sacram*
Upper—Bound Lifetime Noncancer Hazard
Hazard
Quotient
Source: U.S. EPA / OAQPS
NATA National-Scale Air Toxics .Assessment
Source: 1996 NAT A Assessment.
-------
Air Quality, Health, and Welfare Effects
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
inventories are based on only 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.221
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.222
2.2.2.7 Dioxins
Recent studies have confirmed that dioxins are formed by and emitted from diesels (both
heavy-duty diesel trucks and non-road diesels although in very small amounts) and are estimated
to account for about 1 percent of total dioxin emissions in 1995. Recently EPA issued a draft
assessment designating one dioxin compound, 2,3,7,8-tetrachlorodibenzo-p-dioxin as a human
carcinogen and the complex mixtures of dioxin-like compounds as likely to be carcinogenic to
humans using the draft 1996 carcinogen risk assessment guidelines. EPA is working on its final
assessment for dioxin.223 An interagency review group is evaluating EPA's designation of dioxin
as a likely human carcinogen. These nonroad rules will have minimal impact on overall dioxin
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.
2-87
-------
Draft Regulatory Impact Analysis
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 on-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
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.
255
-OO
-------
Air Quality, Health, and Welfare Effects
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.224 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.
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.225'226'227'228'229' 23°
Repeated exposure to ozone can make people more susceptible to respiratory infection and lung
inflammation and can aggravate preexisting respiratory diseases, such as asthma.231'232'233'234'235 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 the lungs
and/or chronic respiratory illnesses, such as emphysema and chronic bronchitis.236'237'238'239
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.240 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.241'242 For example, summer camp studies in the Eastern U.S. and southeastern Canada
have reported significant reductions in lung function in children who are active outdoors.243'244'
245,246,247,248,249,250 purmer, 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.251'252'253'254
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.255 Attaining the 8-hour
2-89
-------
Draft Regulatory Impact Analysis
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.256 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 (CASAC).
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
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.257'258 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).259 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.260 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.261
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
LIn communities with high ozone (i.e., mean 8-hour 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.
2-90
-------
Air Quality, Health, and Welfare Effects
conducted in Nevada in kindergarten to 6th grade 262 and in Southern California in grades 4 to
6.263 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 which
may predispose elderly and other sensitive individuals to become more susceptible to the adverse
health effects of other air pollutants, such as PM.264'265 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.266 Similarly, Thurston and Ito (1999) have reported
associations between ozone and mortality.267 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).268
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 recent time-series studies of short-term exposures to air
pollution and health effects.269
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
level of the 8-hour standard which is requiste 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 U.S. and large areas of California.
Nonroad engines contribute a substantial fraction of ozone precursors in metropolitan areas.
In presenting these values, we examine concentrations in counties as well as calculating
design values. 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. Thus, we follow this convention in these analyses.
2-91
-------
Draft Regulatory Impact Analysis
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 defining the current design value of a
given area, the 1999-2001 design values were chosen to provide the most recent set of air quality
data for identifying areas likely to have an ozone problem in the future. The 1999-2001 design
values are listed in the AQ TSD, which is available in the docket to this rule.
2.3.2.1 1-Hour Ozone Nonattainment Areas and Concentrations
Currently, there are 116 million people living in 56 1-hour ozone nonattainment areas
covering 233 counties. Of these, there are 1 extreme and 10 severe 1-hour ozone nonattainment
areas with a total affected population of 86.5 million as shown in Table 2.3-1. We focus on these
designated areas because the timing of their attainment dates relates to the timing of the proposed
reductions. Five severe 1-hour ozone nonattainment areas have attainment dates of December
31, 2007. While all of these areas are expected to be in attainment before the emission
reductions from this proposed rule are expected to occur, these reductions will be important to
assist these areas in maintaining the standards. The Los Angeles South Coast Air Basin is
designated as an extreme nonattainment area and has a compliance date of December 31, 2010.
The reductions from this rule will be an important part of their overall strategy to attain and
maintain the standard.
2-92
-------
Air Quality, Health, and Welfare Effects
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
Baltimore, MD
Philadelphia- Wilmington-Trenton, PA-
NJ-DE-MD
Sacramento, CA
San Joaquin Valley, CA
Ventura County, CA
Total Population
Attainment
Date
December 31, 2010a
December 3 1,2007
December 3 1,2007
December 3 1,2007
December 3 1,2007
December 3 1,2007
2005
2005
2005
2005
2005
2000
Population
(millions)
14.6
8.9
4.5
1.7
20.2
0.5
0.8
6.0
1.2
7.8
0.1
1999-2001
Measured
Violation?
Yes
No
Yes
No
Yes
Yes
Yes
Yes
Yes
Yes
Yes
86.5 million
' Extreme 1-Hour nonattainment areas. All other areas are severe nonattainment areas.
The extreme nonattainment area will need additional reductions to attain the ozone standard
and will also be able to rely on additional reductions from today's proposed action in order to
maintain the standard. The severe areas will be able to rely on the reductions from today's
proposed action in order to maintain the standard.
The emission reductions from this proposed rule would also help these areas reach attainment
at lower overall cost, with less impact on small businesses, as discussed in other chapters of this
document. Following implementation of controls for regional NOx reductions, States will have
already adopted emission reduction requirements for most large sources of NOx for which cost-
effective control technologies are known and for which they have authority to control. Those
that must adopt measures to complete their attainment demonstrations and maintenance plans,
therefore, will have to consider their remaining alternatives. Many of the alternatives that areas
may consider could be more costly, and the NOx emissions impact from each additional
emissions source subjected to new emissions controls could be considerably smaller than the
emissions impact of the standards being proposed today. Therefore, the emission reductions
from the standards we are finalizing today will ease the need for States to find first-time
2-93
-------
Draft Regulatory Impact Analysis
reductions from the mostly smaller sources that have not yet been controlled, including area
sources that are closely connected with individual and small business activities. The emission
reductions from nonroad diesel engines also reduce the need for States to seek even deeper
reductions from large and small sources already subject to emission controls.
Each of the areas in Table 2.3-1 is adopting additional measures to address specific emission
reduction shortfalls in attainment SIPs submitted for New York, Houston, the South Coast Basin,
Philadelphia, and Baltimore based on the local ozone modeling and other evidence. The San
Joaquin Valley will need additional reductions to attain and maintain the standards. There is
some risk that New York will fail to attain the standard by 2007, and thus a transferred risk that
Connecticut will also fail. A similar situation exists in Southern California, where attainment of
the South Coast is a precondition of the ability of downwind to reach attainment by their
respective attainment dates. Additional reductions from this rule will assist New York and
Greater Connecticut, and the South Coast and its downwind nonattainment areas, in reaching the
standard by each areas' respective attainment dates and maintaining the standard in the future.
The Los Angeles (South Coast Air Basin) ozone attainment demonstration is fully approved,
but it is based in part on reductions from new technology measures that have yet to be identified
(as allowed under CAA Section 182(e)(5)). Thus, additional reductions would be helpful to this
area, as discussed in the draft plan.270 The 2007 attainment demonstration for the Southeast
Desert area is also approved. However, a transport situation exists between the Southeast Desert
areas and the South Coast Air Basin, such that attainment in the Southeast Desert depends on
progress in reducing ozone levels in the South Coast Air Basin.
Even if the SIPs were approved and all shortfalls were filled in an area, there would still be a
risk that ozone levels in such an area could exceed the NAAQS. EPA's approval of an
attainment demonstration generally indicates our belief that a nonattainment area is reasonably
likely to attain by the applicable attainment date with the emission controls in the SIP. However,
such approval does not indicate that attainment is certain. Moreover, no ozone forecasting is 100
percent certain, so attainment by these deadlines is not certain, even though we believe it is more
likely than not. There are significant uncertainties inherent in predicting future air quality, such
as unexpected economic growth, unexpected vehicle miles traveled (VMT) growth, the year-to-
year variability of meteorological conditions conducive to ozone formation, and modeling
approximations. There is at least some risk in each of these areas that even assuming all
shortfalls are filled, attainment will not be reached by the applicable dates without further
emission reductions. The Agency's mid-course review in the SIP process—as well as the Clean
Air Act's provisions for contingency measures—is part of our strategy for dealing with some of
these uncertainties, but does not ensure successful attainment.
Many 1-hour ozone nonattainment areas continue to experience exceedances. Approximately
51 million people are living in counties with measured air quality violating the 1-hour NAAQS
2-94
-------
Air Quality, Health, and Welfare Effects
in 1999-2001.M See the AQ TSD for more details about the counties and populations
experiencing various levels of measured 1-hour ozone concentrations.
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 proposed 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 today's proposed action will help areas to attain and maintain the 1-hour standards.
2.3.2.2 8-Hour Ozone Levels: Current and Future Concentrations
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.271 The 8-hour standard greatly limits ozone exposures of concern for the general
population and sensitive populations. This section describes the current measured 8-hour
concentrations and describes our modeling to predict future 8-hour ozone concentrations.
2.3.2.2.1 Current 8-Hour Ozone Concentrations
Based upon the measured data from years 1999 - 2001, there are 291 counties with measured
values that violate the 8-hour ozone NAAQS, with a population totaling 111 million, as shown in
Figure 2-1. Of these, 61 million people live in counties that meet the 1-hour standard but violate
the 8-hour standard. There may be additional areas above the level of the NAAQS for which no
monitoring data are available.
An additional 37 million people live in 155 counties that have air quality measurements
within 10 percent of the level of the standard. These areas, though currently not violating the
standard, will also benefit from the emission reductions from this proposed rule.
Approximately 48 million people lived in counties with at least a week (7 days) of 8-hour
ozone concentrations measurements at or above 0.085 ppm in 2000. Approximately 8 million
people lived in counties experiencing 20 days and 4 million experienced 40 days of 8-hour ozone
concentrations at or above 0.085 ppm in 2000. See the AQ TSD for more details about the
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 are experiencing recent exceedances, as indicated in the table. Further, the analysis is limited to areas
with monitors.
2-95
-------
Draft Regulatory Impact Analysis
counties and populations experiencing various levels of measured 8-hour ozone concentrations.
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 proposal, we
modeled a preliminary control scenario which illustrates the likely reductions from our proposal.
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 process based on our best judgement at that
time. As additional data regarding technical feasibility and other factors became available, our
judgement 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 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. For the final rule, considering public comment, we plan to
model the final control scenario. This proposed 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 U.S. using Comprehensive Air Quality Model with
Extension (CAMx). The model simulations were performed 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.
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 which will
require additional emission reductions to attain and maintain the ozone NAAQS. The impacts of
the proposed controls were determined by comparing the model results in the future year control
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 emissions changes relate to what
was modeled.
2-96
-------
Air Quality, Health, and Welfare Effects
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.
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 U.S. for the various emissions 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/non-attainment 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
on-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 U.S., 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 U.S. 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 U.S.: June 12-24, July
5-15, and August 7-21. Two multi-day meteorogical scenarios during the summer of 1996 were
used in the model simulations over the western U.S.: 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 emissions scenarios (1996 base year,
2020 base, 2020 control, 2030 baseline, 2030 control) were simulated for the selected episodes.
2-97
-------
Draft Regulatory Impact Analysis
The meteorological data required for input into CAMx (wind, temperature, vertical mixing,
etc.) were developed by separate meteorological models. For the eastern U.S., 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 U.S., 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 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 on-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 which
adequately replicates observed pollutant concentrations in the base year can be used to assess the
effects of future year emissions 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,
2-98
-------
Air Quality, Health, and Welfare Effects
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.
In general, the model tends to underestimate observed ozone, especially in the modeling over
the western U.S. as shown in Table 2.3-2. When all hourly observed ozone values greater than a
60 ppb threshold are compared to 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 2.3-2.
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 proposal 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/economic benefits due to the
proposed emissions reductions.
2.3.2.2.5 Results of Photochemical Ozone Modeling: Areas at Risk of Future 8-Hour
Violations
This next section summarizes the results of our modeling of ozone air quality impact 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
2-99
-------
Draft Regulatory Impact Analysis
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
attainment and maintenance. Those interested in greater detail should review the AQ TSD and
AQ Modeling TSD, which are both 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 emissions 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.272 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.0 273 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 proposal, 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 proposed nonroad diesel reductions), 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.274 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 without the
nonroad diesel emissions reductions proposed today 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
°For the one-hour NAAQS we used a cut-off of 80 ppb. Please see the On-highway
Passenger Vehicle Tier 2 Air Quality Modeling TSD for more details (EPA 1999b).
2-100
-------
Air Quality, Health, and Welfare Effects
attainment with the ozone 8-hour NAAQS. EPA's current plans call for designating ozone 8-
hour nonattainment areas in April 2004. EPA is planning to propose that States submit SIPs that
address how areas will attain the 8-hour ozone standard within three years after nonattainment
designation regardless of their classification. EPA is also planning to propose that certain SIP
components, such as those related to reasonably available control technology (RACT) and
reasonable further progress (RFP) be submitted within 2 years after designation. We therefore
anticipate that States will submit their attainment demonstration SIPs by April 2007. Section
172(a)(2) of the Clean Air Act requires that SIP revisions for areas that may be covered only
under subpart 1 of part D, Title I of the Act demonstrate that the nonattainment areas will attain
the ozone 8-hour standard as expeditiously as practicable but no later than five 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 these provisions, we expect that most or all areas covered under
subpart 1 will attain the ozone standard in the 2007 to 2014 time frame. For areas covered under
subpart 2, the maximum attainment dates will range from 3 to 20 years after designation,
depending on an area's classification. Thus, we anticipate that areas covered by subpart 2 will
attain in the 2007 to 2014 time period.
Furthermore, the inventories that underlie the ozone modeling conducted for this
rulemaking included reductions from all current or committed federal, State and local controls
and, for the control case, the proposed nonroad diesel program itself. It did not did not 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). Therefore,
Tables 2.2-3 and 2.2-4 below should be interpreted as indicating what areas 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-
3 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 today's proposal would begin in the same time
period in which areas will need reductions to attain by their attainment dates, the projected
reductions in nonroad emissions would be extremely important to States in meeting the new
NAAQS. It is our expectation that States will be relying on such nonroad reductions in order 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 average
2-101
-------
Draft Regulatory Impact Analysis
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 non-attaining 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 proposal, but
they could substantially change the count 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 today's proposal might change the actual attainment status of very few
counties. Bearing this limitation in mind, our modeling indicates that the nonroad diesel
emissions reductions would decrease the net number of nonattainment counties by 2 in 2020 and
by 4 in 2030, without consideration of new state programs.
Areas presented in Table 2.3-3 and 2.3-4 have monitored 1999-2001 air quality data
indicating violations of the 8-hour ozone NAAQS, or are within 10 percent of the standard, and
are predicted to have exceedances in 2020 or 2030 without the reductions from this rule. Table
2.3-3 lists those counties with predicted exceedances of the 8-hour ozone standard in 2020 or
2030 without emission reductions from this rule (i.e., base cases). These areas are listed in
columns with a "b" after the year (e.g., 2020b). Table 2.3-2 also lists those counties with
predicted exceedances of the 8-hour ozone standard in 2020 and 2030, with emission reductions
from this rule (i.e., control case). These areas are listed in columns with a "c" after the year (e.g.,
2020c). An area was considered likely to have future exceedances if exceedances were predicted
by the model, and the area is currently violating the 8-hour ozone standard, or is within 10
percent of violating the 8-hour ozone standard.
In Table 2.3-3 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 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 ambient monitoring data sufficient to calculate current design values.
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 the proposed rule is positive.
2-102
-------
Table 2.3-3: Counties with 2020 and 2030 Projected Ozone Design Values
in Violation of the 8-Hour Ozone Standard.a
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
(r>r>b)
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 proposed emission reductions differs 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.
-------
Draft Regulatory Impact Analysis
In Table 2.3-4 we present the counties with 2020 and 2030 projected 8-hour ozone 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 77
ppb, but less 85 ppb. Counties are marked with a "V" in the table if their projected design values
are greater than or equal to 85 ppb. The current 3-year average 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 proposed rule will help assure that these counties continue to meet the standard.
Table 2.3-4
Counties with 2020 and 2030 Projected Ozone Design Values
within Ten Percent of the 8-Hour Ozone Standard.a
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
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 Rou
Iberville
Jefferson
Livingston
St Charles
St James
1999-2001
Design Value
foot))
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
88
86
83
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
X
X
Control"
X
X
X
X
X
X
X
X
X
X
X
V
X
X
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
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
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
91,814
48,072
21,216
2-104
-------
Air Quality, Health, and Welfare Effects
State
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
PA
RI
County
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
Philadelphia
Kent
1999-2001
Design Value
(rob)
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
88
94
2020
Base
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
X
X
Control"
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
V
X
X
X
X
X
X
X
2030
Base
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
X
X
Control"
X
X
X
X
X
X
V
X
X
X
X
X
X
V
X
X
X
X
X
X
Population
in 2000
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
1,517,550
167,090
2-105
-------
Draft Regulatory Impact Analysis
State
RI
TN
TX
TX
TX
TX
TX
TX
TX
VA
VA
VA
WI
WI
WI
WI
WI
WI
WI
WI
County
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
footo
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
79
40.465.492
Control"
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
82
44.013.587
Control"
X
X
X
X
X
X
X
X
X
X
X
X
X
X
54
35.631.215
Population
in 2000
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
The proposed emission reductions differs 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.
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 0.085 to 0.12 ppm 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.
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 0.085 ppm. We then grouped the counties which had days with
ozone in this range according to the number of days this was predicted to happen, and summed
their projected populations.
In the base case (i.e., before the application of emission reductions resulting from this rule),
2-106
-------
Air Quality, Health, and Welfare Effects
we estimated that in 2020 53 million people are predicted to live in counties with at least 2 days
with 8-hour average concentrations of 0.085 ppm 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 0.085 ppm or higher. About 30 million people live in counties with at least 7
days of 8-hour ozone concentrations at or above 0.085 ppm 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 0.085 ppm in 2020 and 2030 without additional
controls. 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
Emissions Reductions
While the proposed rule would 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 emissions 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.275
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.276'277 There are other hypotheses for the cause of the "weekend effect."278 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 emissions 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.
EPA maintains that the best available approach for determining the value of a particular
emissions reduction strategy is the net air quality change projected to result from the rule,
2-107
-------
Draft Regulatory Impact Analysis
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 models279.
Model scenarios of 2020 and 2030 with and without the proposed controls are compared to
determine the expected changes in future pollutant levels resulting from the proposed rule. There
are several factors related to the air quality modeling and inputs which should be considered
regarding the disbenefit issue. First, our future year modeling conducted does not contain any
local governmental actions beyond the controls proposed 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 proposal (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 CARB 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 the assessment of the proposed emissions
reductions. Three of the most important assessments are: 1) the effect of the proposed rule on
projected future-year ozone violations, 2) the effect of the proposed 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 today's 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
the proposed controls is projected to be strongly positive due to the expected reductions in fine
PM.
The net impact of the proposed rule on projected 8-hour ozone violations in 2020 is that three
counties would no longer violate the NAAQS280. Conversely, one county in the New York City
CMS A (Bronx County) which is currently not in violation of the NAAQS is projected to violate
the 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. 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
2-108
-------
Air Quality, Health, and Welfare Effects
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.
2-109
-------
Draft 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 Rule"
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
a The proposal 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 which are 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/national ozone
levels1. Consistent with that fact, the photochemical modeling completed for this rule indicates
that the emissions reductions proposed today 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 emissions reductions that States, local governments, and Tribes need to reach their clean
air goals. It is expected that future 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 support the rule because they believe the
resulting NOx reductions would reduce both ozone and PM281. EPA believes that a balanced air
2-111
-------
Draft Regulatory Impact Analysis
quality management approach that includes NOx emissions reductions from nonroad engines is
needed as part of the Nation's progress toward clean air.
Another category of potential effects that may change in response to ozone reduction
strategies results from the shielding provided by ozone against the harmful effects of ultraviolet
radiation (UV-B) derived from the sun. The great majority of this shielding results from
naturally occurring ozone in the stratosphere, but the 10 percent of total "column"ozone present
in the troposphere also contributes.282 A variable portion of this tropospheric fraction of UV-B
shielding is derived from ground level ozone related to anthropogenic air pollution. Therefore,
strategies that reduce ground level ozone could, in some small measure, increase exposure to
UV-B from the sun.
While it is possible to provide quantitative estimates of benefits associated with globally
based strategies to restore the far larger and more spatially uniform stratospheric ozone layer, the
changes in UV-B exposures associated with ground level ozone reduction strategies are much
more complicated and uncertain. Comparatively smaller changes in ground-level ozone
(compared to the total ozone in the troposphere) and UV-B are not likely to measurably change
long-term risks of adverse effects.
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.283 Because the proposed rule would reduce ground-level ozone and
nitrogen deposition, benefits are expected to accrue to the welfare effects categories described in
the paragraphs (subsections) below.
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."284
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.285 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.286 287 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.288 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
2-112
-------
Air Quality, Health, and Welfare Effects
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.289
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).290 291
292 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.293 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
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.294 295 In terms of forest productivity and ecosystem diversity, ozone may be
the pollutant with the greatest potential for regional-scale forest impacts.296 Studies have
demonstrated repeatedly that ozone concentrations commonly observed in polluted areas can
have substantial impacts on plant function.297 298 2"
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.300 In most instances, responses to chronic or
2-113
-------
Draft Regulatory Impact Analysis
recurrent exposure are subtle and not observable for many years. These injuries can cause stand-
level forest decline in sensitive ecosystems.301302 303 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,304 welfare benefits are also 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.305 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 U.S."306 In addition, economic studies have shown a relationship
between observed ozone levels and crop yields.307 308 309
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 economic benefits 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.310 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.
2.3.3.2 Nitrogen (NOx)-related welfare effects.
Agriculture. The proposed rule, 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' 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.
2-114
-------
Air Quality, Health, and Welfare Effects
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 deposition of nitrogen could have negative
effects on forest and vegetation growth in ecosystems where nitrogen is a limiting factor.311
On the other hand, there is evidence that forest ecosystems in some areas of the United States
are already or are becoming nitrogen saturated.312 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.313
The reductions in ground-level ozone and nitrogen deposition that would result from the
proposed rule would be expected to reduce the adverse impacts described above. In particular, it
is expected that economic impacts, such as those related to reduced crop yields and forest
productivity, would be reduced.
2.4 Carbon Monoxide
The standards being proposed today would also help 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 November 4, 2002, there were 24 areas designated as
non-attainment with the SO2 standard, and 14 designated CO non-attainment 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
2-115
-------
Draft Regulatory Impact Analysis
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 non-
attainment area (65 FR 76790, December 7, 2000). In that action EPA found that engines subject
to this proposed rule contribute to CO non-attainment in areas such as Los Angeles, Phoenix,
Spokane, Anchorage, and Las Vegas. Nonroad land-based diesel engines emitted 927,500 tons
of CO in 1996 (1 percent of mobile source CO). Thus, nonroad diesel engines contribute to CO
non-attainment 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 non-attainment areas. Thus, the emissions benefits
from this proposed rule will help areas to attain and maintain the CO NAAQS.
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.314'
315 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.316 The subsequent hypoxia in brain tissue then produces behavioral
effects, including decrements in continuous performance and reaction time.317
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.318 Persons
with heart disease are especially sensitive to carbon monoxide poisoning and may experience
chest pain if they breathe the gas while exercising.319 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.320
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 U.S. 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
2-116
-------
Air Quality, Health, and Welfare Effects
studies as part of the Criteria Document review process.321
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 22 million people currently live in the 13 non-attainment 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.322 Over the next decade, we would
expect there to be a minor decreasing trend from the highway segment due primarily to the more
stringent standards for certain light-duty trucks.323 CO standards for passenger cars and other
light-duty trucks and heavy-duty vehicles did not change as a result of other recent rulemakings.
As explained in Chapter 9, EPA currently does not have appropriate tools for modeling
changes in ambient concentrations of CO or air toxics for input into a national benefits analysis.
As noted above, CO has been linked to numerous health effects; however, we are unable to
quantify the CO-related health or welfare benefits of the Nonroad Diesel Engine rule at this time.
However, nonroad diesel engines do contribute to nonattainment in some areas. Thus, the
emissions benefits from this proposed rule would help areas to attain and maintain the CO
NAAQS.
2-117
-------
Draft Regulatory Impact Analysis
Chapter 2 References
1. U.S. EPA (1996) Air Quality Criteria for Particulate Matter - Volumes I, H, and HI,
EPA/600/P-95/001aF, EPA/600/P-95/001bF, EPA/600/P-95/001cF. Docket No. A-99-06.
Document Nos. II-A-18 to 20. and U.S. EPA (2002). Air Quality Criteria for Particulate Matter
- Volumes I and II (Third External Review Draft, 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.
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.
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.
2-118
-------
Air Quality, Health, and Welfare Effects
12. Schwartz, I; 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., HI. (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. HI. 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., HI; 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 U: morbidity and mortality from air pollution in the United States.
Res Rep Health Eff Inst Number 94, Part H. 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.
2-119
-------
Draft Regulatory Impact Analysis
25. Dockery, DW; Pope, CA, IE; 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.
26. Pope, CA, IE; 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.
27. 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.
28. 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.
29. 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
11:116-122.
30. 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.
31. Pope, C.A. HI; 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.
32. Pope, C.A. HI; Verrier, R.L.; Lovett, E.G.; et al. (1999) Heart rate variability associated with
particulate air pollution. Am Heart J 138: 890-899.
33. Gold, D.R.; Litonjua, A; Schwartz, J; et al. (2000) Ambient pollution and heart rate
variability. Circulation 101: 1267-1273.
34. 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.
35. 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.
36. 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.
37. 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.
2-120
-------
Air Quality, Health, and Welfare Effects
38. 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.
39. 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.
40. 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.
41. Ohio, 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
42. Seaton et al. (1999) Paniculate air pollution and the blood. Thorax 54: 1027-1032.
43. Peters et al. (2001a) Particulate air pollution is associated with an acute phase response in
men; results from the MONICA-Augsburg study. Eur Heart J 22(14): 1 198-1204.
44. 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.
45. Peters et al. (1997) Increased plasma viscosity during and air pollution episode: a link to
mortality? Lancet 349: 1582-87.
46. Zimmerman, M.A.; Selzman, C.H.; Cothren, C.; et al. (2003) Diagnostic implications of
C-reactive protein. Arch Surg 138(2):220-4.
47. 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.
48. 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.
49. 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.
50. Peters, A.; Liu, E.; Verrier, R.L.; et al. (2000) Air pollution and incidence of cardiac
arrhythmia. Epidemiology 1 1 : 11-17.
51. 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.
2-121
-------
Draft Regulatory Impact Analysis
52. 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
attached. Docket No. A-2000-01. Document No. IV-A-145.
53. Dominici, F.; McDermott, A.; Daniels, M.; et al. (2002) Report to the Health Effects
Institute: reanalyses of the National Morbidity, Mortality, and Air Pollution Study (NMMAPS)
Database. [Accessed at www.biostat.jhsph.edu/~fominic/HEI/nmmaps.html.]
54. 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.
55. Colburn, KA and PRS Johnson (2003). Air Pollution Concerns Not Changed by S-PLUS
Flaw [sic]. Science. 299:665-666. January 31, 2003.
56. 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.
57. U.S. EPA (1985). Size specific total particulate emission factor for mobile sources. EPA
460/3-85-005. Office of Mobile Sources, Ann Arbor, MI.
58. Delfmo 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.
59. 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.
60. 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.
61. 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.
62. 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.
63. Bunn, H.J.; Dinsdale, D.; Smith, T.; et al. (2001) Ultrafme particles in alveolar macrophages
from normal children. Thorax 56(12):932-4.
2-122
-------
Air Quality, Health, and Welfare Effects
64. Zhu, Y.; Hinds, W.C.; Kim, S.; et al. (2002) Concentration and size distribution of ultrafme
particles near a major highway. J Air Waste Manage Assoc 52: 1032-1042.
65. Zhu, Y.; Hinds, W.C.; Kim, S.; et al. (2002) Study of ultafine particles near a major highway
with heavy-duty diesel traffic. Atmos Environ 36:4323-4335.
66. 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.
67. 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. EPA-
452VR-96-013. Docket No. A-99-06. Document No. H-A-23.
68. 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.
69. EPA (2002) Latest Finds on National Air Quality, EPA 454/K-02-001.
70. Mansell (2000). User's Instructions for the Phase 2 REMSAD Preprocessors, Environ
International. Novato, CA. 2000.
71. IMPROVE (2000). Spatial and Seasonal Patterns and Temporal Variability of Haze and its
Constitutents in the United States. Report in. Cooperative Institute for Research in the
Atmosphere, ISSN: 0737-5352-47.
72. CARB and New York State Department of Environmental Conservation (April 9, 2002).
Letter to EPA Administrator Christine Todd Whitman.
73. 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.
74. Western Regional Air Partnership (WRAP) January 28, 2003), Letter to Governor Christine
Todd Whitman.
75.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 Particulate Matter (PM)" Vol I - HI.
EPA600-P-99-002a; Docket No. A-99-06. Document Nos. U-A-18 to 20.
US EPA (1996). Review of the National Ambient Air Quality Standards for Particulate Matter:
Policy Assessment of Scientific and Technical Information OAQPS Staff Paper. EPA-452/R-96-
2-123
-------
Draft Regulatory Impact Analysis
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, US EPA.
Review of the National Ambient Air Quality Standards for Parti culate Matter: Policy Assessment
of Scientific and Technical Information, OAQPS Staff Paper. Preliminary Draft. June 2001.
Docket A-2000-01, Document IV-A-199.
76. Council on Environmental Quality, 1978. Visibility Protection for Class I Areas, the
Technical Basis. Washington DC. Cited in US 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.
77.US EPA Trends Report 2001. This document is available on the internet at
http://www.epa. gov/airtrends/.
78. 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. U-B-21.
79.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 parti culate matter air quality criteria documents
are also available at http://www.epa.gov/ncea/partmatt.htm. Also, US 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.
80. National Research Council, 1993 (Ibid). This document is available on the internet at
http://www.nap.edu/books/0309048443/html/.
81.National Research Council, 1993 (Ibid). This document is available on the internet at
http://www.nap.edu/books/0309048443/html/.
82. 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 US 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, US 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.
2-124
-------
Air Quality, Health, and Welfare Effects
83.U.S. EPA. (2003). Air Quality Technical Support Document for the proposed Nonroad Diesel
rulemaking. OAQPS. April 2003.
84.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.
85. 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,
EPA452-R-96-013, July, 1996, at IV-7. This document is available from Docket A-99-06,
Document U-A-23.
86. See 64 FR 35722, July 1, 1999.
87.Memorandum to Docket A-99-06 from Eric O. Ginsburg, Senior Program Advisor,
"Summary of 1999 Ambient Concentrations of Fine Particulate Matter," November 15, 2000.
Air Docket A-2000-01, Document No. II-B-12.
88. 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 U-B-14.
89.Western Regional Air Partnership (WRAP) letter dated Jan 28, 2003 to Administrator
Christine Todd Whitman.
90. U.S. EPA. (1993). Effects of the 1990 Clean Air Act Amendments on Visibility in Class I
Areas: An EPA Report to Congress. EPA-452/R-93-014, Docket A-2000-01, Document IV-A-
220. And see also 64 FR 35722, July 1, 1999.
91 .This goal was recently upheld by the US 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.
92.U.S. EPA. (1993). Effects of the 1990 Clean Air Act Amendments on Visibility in Class I
Areas: An EPA Report to Congress. EPA-452/R-93-014, Docket A-2000-01, Document IV-A-
20.U.S. EPA Trends Report 2002.
93. For more information and the IMPROVE data, see
http://vista.cira.colostate.edu/improve/data/IMPROVE/improve_data.htm.
94.U.S. EPA Trends Report 2002.
2-125
-------
Draft Regulatory Impact Analysis
95. Chestnut, L.G., and R.D. Rowe. 1990a. Preservation for Visibility Protection at the National
Parks: Draft Final Report. Prepared for Office of Air Quality Planning and Standards, US
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.
96.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
20460, November 1995.
91 .Acid Rain: Emissions Trends and Effects in the Eastern United States, US General
Accounting Office, March, 2000 (GOA/RCED-00-47).
98. Acid Deposition Standard Feasibility Study: Report to Congress, EPA430-R-95-001a,
October, 1995.
99. Deposition of Air Pollutants to the Great Waters, Third Report to Congress, June, 2000.
WO.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.
101. 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.
102. Deposition of Air Pollutants to the Great Waters, Third Report to Congress, June, 2000.
103. Valigura, Richard, et al., Airsheds and Watersheds II: A Shared Resources Workshop, Air
Subcommittee of the Chesapeake Bay Program, March, 1997.
104. The Impact of Atmospheric Nitrogen Deposition on Long Island Sound, The Long Island
Sound Study, September, 1997.
105. 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.
106. 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.
2-126
-------
Air Quality, Health, and Welfare Effects
107. 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, EPA-453/R-97-011. You are referred to that document for a more detailed
discussion.
108. The 1996 National Toxics Inventory, Office of Air Quality Planning and Standards, October
1999.
109. U.S. EPA. Control of Emissions of Hazardous Air Pollutants from Mobile Sources; Final
Rule (66 FR 17230-17273, March 29, 2001).
110. 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.
111. U.S. EPA. (1986) .Guidelines for carcinogen risk assessment. Federal Register
51(185):33992-34003.
112. 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.
113. International Agency for Research on Cancer - IARC. (1997). Monographs on the
evaluation of carcinogenic risks to humans. Vol. 68. Silica, some silicates, coal dust and
para-aramid fibrils. Lyon, France: IARC, pp. 362-375.
114. 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.
115. 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.
116. California Environmental Protection Agency. (Cal EPA, OEHHA) (1998). Health risk
assessment for diesel exhaust. Public and Scientific Review Draft.
117. 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.
118. Health Effects Institute (HEI). (1995). Diesel exhaust: a critical analysis of emissions,
exposure, and health effects. Cambridge, MA.
2-127
-------
Draft Regulatory Impact Analysis
119. 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.
120. 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.
121. 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.
122. Lewtas, J. (1983). Evaluation of the mutagenicity and carcinogenicity of motor vehicle
emissions in short-term bioassays. Environ Health Perspect 47:141-152/.
123. 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.
124. 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.
125. 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.
126. 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.
127. 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.
128. 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.
129. Saverin, R. (1999). German potash miners: cancer mortality. Health Effects Institute
Number 7. March 7-9, Stone Mountain, GA, pp. 220-229.
130. 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.
131. 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.
132. 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
2-128
-------
Air Quality, Health, and Welfare Effects
Ind Med 42:435-448. U.S. Environmental Protection Agency.
133. Bhatia, R., Lopipero, P., Smith, A. (1998). Diesel exhaust exposure and lung cancer.
Epidemiology 9(1):84-91.
134. Lipsett, M: Campleman, S.; (1999). Occupational exposure to diesel exhaust and lung
cancer: a meta-analysis. Am J Public Health 80(7): 1009-1017.
135. U.S. EPA (2002), National-Scale Air Toxics Assessment for 1996. This material is
available electronically at http://www.epa.gov/ttn/atw/nata/.
136. 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.
137. 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.
138. 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.
139. 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.
140. Reger, R; Hancock, J; Hankinson, J; et al. (1982) Coal miners exposed to diesel exhaust
emissions. Ann Occup
Hyg 26:799-815.
141. 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.
142. El Batawi, MA; Noweir, MH. (1966) Health problems resulting from prolonged exposure
to air pollution in diesel bus garages. Ind Health 4:1-10.
143. Wade, JF, HI; Newman, LS. (1993) Diesel asthma: reactive airways disease following
overexposure to locomotive exhaust. J Occup Med 35:149-154
144. 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.
2-129
-------
Draft Regulatory Impact Analysis
EPA454-B-95-003b.
145. 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.
146. 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.
147. 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.
148. 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.
149. 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.
150. 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.
151. 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.
152. 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.
153. Air Improvement Resources. (1997). Contribution of Gasoline Powered Vehicles to
Ambient Levels of Fine Particulate Matter. CRC Project A-18.
154. 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.
155. 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.
2-130
-------
Air Quality, Health, and Welfare Effects
156. 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.
157. 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.
158. 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
159. 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
Technol 36: 3806-3814.
160. Cal-EPA. (1998) Measuring concentrations of selected air pollutants inside California
vehicles. Final report.
161. 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.
162. Groves, J; Cain, JR. (2000). A Survey of Exposure to Diesel Engine Exhaust Emissions in
the Workplace. Ann Occ Hyg 44(6):435-447.
163. 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.
164. Northeast States for Coordinated Land Use Management (2001). EPA Grant
X-82963001-1.
165. 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.
166. California EPA. (1998). Proposed Identification of Diesel Exhaust as a Toxic Air
Contaminant. Appendix in, 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.
167. U.S. EPA (2002). National-Scale Air Toxics Assessment. This material is available
electronically at http://www.epa.gov/ttn/atw/nata/.
168. U.S. EPA (2001). 1996 National Toxics Inventory. This material is available
electronically at http://www.epa.gov/ttn/chief/nti/.
2-131
-------
Draft Regulatory Impact Analysis
169. Cook R., 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.
170. 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.
171. U.S. EPA. (2002). Comparison of ASPEN Modeling System Results to Monitored
Concentrations, http://www.epa.gov/ttn/atw/nata/draft6.html#SecI.
172. 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.
173. 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/.
174. 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.
175. 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.
176. 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.
177. 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.
178. 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.
179. 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
2-132
-------
Air Quality, Health, and Welfare Effects
Planning and Standards, Washington, DC., 1985.
180. Clement Associates, Inc. (1991). Motor vehicle air toxics health information, for U.S. EPA
Office of Mobile Sources, Ann Arbor, MI, September 1991.
181. 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.
182. 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.
183. Lumley, M., H. Barker, and J.A. Murray (1990). Benzene in petrol, Lancet 336:1318-1319.
184. U.S. EPA (1998). Environmental Protection Agency, Carcinogenic Effects of Benzene: An
Update, National Center for Environmental Assessment, Washington, DC. 1998. EPA600-P-97-
001F. http://www.epa.gov/ncepihom/Catalog/EPA600P97001F.html.
185. Aksoy, M. (1989). Hematotoxi city and carcinogeni city of benzene. Environ. Health
Perspect. 82: 193-197.
186. Goldstein, B.D. (1988). Benzene toxi city. Occupational medicine. State of the Art
Reviews. 3: 541-554.
187. 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.
188. Goldstein, B.D. (1988). Benzene toxicity. Occupational medicine. State of the Art
Reviews. 3: 541-554.
189. Aksoy, M., S. Erdem, and G. Dincol. (1974). Leukemia in shoe-workers exposed
chronically to benzene. Blood 44:837.
190. 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.
191. 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.
2-133
-------
Draft Regulatory Impact Analysis
192. 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
193. 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-98-00IF.
194. U.S. EPA (2002). Health Assessment of Butadiene, This material is available electronically
at http://cfpub.epa.gov/ncea/cfm/recordisplay.cfm7dei d=54499.
195. U.S. EPA (1998). A Science Advisory Board Report: Review of the Health Risk
Assessment of 1,3-Butadiene. EPA-SAB-EHC-98.
196. 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.
197. 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.
198. 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.
199. 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.
200. 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.
201. 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.
202. 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.
203. 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.
204. 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.
2-134
-------
Air Quality, Health, and Welfare Effects
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. 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.
207. Chemical Industry Institute of Toxicology (1999). Formaldehyde: Hazard Characterization
and Dose-Response Assessment for Carcinogenicity by the Route of Inhalation.
208. 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.
209. Wilhelmsson, B. and M. Holmstrom. (1987). Positive formaldehyde PAST after prolonged
formaldehyde exposure by inhalation. The Lancet: 164.
210. 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.
211. 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.
212. Nordman, H., H. Keskinen, and M. Tuppurainen. (1985). Formaldehyde asthma - rare or
overlooked? J. Allergy din. Immunol. 75:91-99.
213. 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.
214. 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.
215. 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.
216. Woutersen, R.A. and L.M. Appelman. (1984). Lifespan inhalation carcinogenicity study of
acetaldehyde in rats. HI. Recovery after 52 weeks of exposure. Report No. V84.288/190172.
CIVO-Institutes TNO, The Netherlands.
217. 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.
218. California Air Resources Board (CARB) (1992). Preliminary Draft: Proposed identification
of acetaldehyde as a toxic air contaminant, Part B Health assessment, California Air Resources
2-135
-------
Draft Regulatory Impact Analysis
Board, Stationary Source Division, August, 1992.
219. 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.
220. U.S. EPA (1994). Integrated Risk Information System File of Acrolein. This material is
available electronically at http://www.epa.gov/iris/subst/0364.htm.
221. Dubowsky, S.D.; Wallace, L.A.; and Buckley, T.J. (1999) The contribution of traffic to
indoor concentrations of polycyclic aromatic hydrocarbons. J Expo Anal Environ Epidemiol
9(4):312-21.
222. Perera, F.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.
223. U.S. EPA (June 2000). Exposure and Human Health Reassessment of 2,3,7,8-
Tetrachlorodibenzo-p-Dioxin (TCDD) and Related Compounds, External Review Draft,
EPA600-P-00-001Ag. This material is available electronically at
http://www.epa.gov/ncea/dioxin.htm.
224.U.S. EPA (1996). Air Quality Criteria for Ozone and Related Photochemical Oxidants,
EPA/600/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
http:/www.epa.gov/ttn/naaqs/standards/ozone/s.03.index.html.
225.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.
226.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.
227.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.
228.Lipfert, F.W.; Hammerstrom, T. (1992) Temporal patterns in air pollution and hospital
admissions. Environ. Res. 59: 374-399.
229.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.
2-136
-------
Air Quality, Health, and Welfare Effects
230. U.S. EPA (1996). Air Quality Criteria for Ozone and Related Photochemical Oxidants,
EPA/600/P-93/004aF. Docket No. A-99-06. Document Nos. II-A-15 to 17. (See page 9-33)
231 .U.S. EPA (1996). Air Quality Criteria for Ozone and Related Photochemical Oxidants,
EPA/600/P-93/004aF. Docket No. A-99-06. Document Nos. II-A-15 to 17. (See page 7-167)
232.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.
233.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.
234.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.
235.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.
236.U.S. EPA (1996). Air Quality Criteria for Ozone and Related Photochemical Oxidants,
EPA/600/P-93/004aF. Docket No. A-99-06. Document Nos. II-A-15 to 17. (See page 7-171)
237.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.
238.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.
239.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.
240.U.S. EPA. (1996). Review of National Ambient Air Quality Standards for Ozone,
Assessment of Scientific and Technical Information, OAQPS Staff Paper, EPA-452/R-96-007.
Docket No. A-99-06. Document No. H-A-22.
241.U.S. EPA (1996). Air Quality Criteria for Ozone and Related Photochemical Oxidants,
EPA/600/P-93/004aF. Docket No. A-99-06. Document Nos. II-A-15 to 17.
2-137
-------
Draft Regulatory Impact Analysis
242.U.S. EPA. (1996). Review of National Ambient Air Quality Standards for Ozone,
Assessment of Scientific and Technical Information, OAQPS Staff Paper, EPA-452/R-96-007.
Docket No. A-99-06. Document No. U-A-22.
243.U.S. EPA (1996). Air Quality Criteria for Ozone and Related Photochemical Oxidants,
EPA/600/P-93/004aF. Docket No. A-99-06. Document Nos. II-A-15 to 17. (See page 7-170)
244.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.
245.Higgins, I. T. T.; D'Arcy, J. B.; Gibbons, D. I; 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.
246.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.
247.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.
248.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).
249.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.
250.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.
251.U.S. EPA (1996). Air Quality Criteria for Ozone and Related Photochemical Oxidants,
EPA/600/P-93/004aF. Docket No. A-99-06. Document Nos. II-A-15 to 17. (See pages 7-160 to
7-165)
2-138
-------
Air Quality, Health, and Welfare Effects
252.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.
253.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.
254.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.
255.U.S. EPA. (1996). Review of National Ambient Air Quality Standards for Ozone,
Assessment of Scientific and Technical Information, OAQPS Staff Paper, EPA-452/R-96-007.
Docket No. A-99-06. Document No. U-A-22.
256. 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, US Environmental Protection Agency, Research Triangle Park, NC 27711
(7/2002) Docket No. A-2001-11. Document No. IV-A-19.
257. 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)
258. 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.
259.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.
260.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.
261.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.
262. Chen, L.; Jennison, B. L.; Yang, W.; Omaye, S. T. (2000) Elementary school absenteeism
and air pollution. Inhalation Toxicol. 12: 997-1016.
263. 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
2-139
-------
Draft Regulatory Impact Analysis
school absenteeism due to respiratory illnesses Epidemiology 12:43-54.
264.Devlin, R. B.; Folinsbee, L. I; 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.
265.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.
266.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 H:
Morbidity, Mortality and Air Pollution in the United States. Research Report No. 94, Part n.
Health Effects Institute, Cambridge MA, June 2000. (Docket Number A-2000-01, Document
Nos. IV-A-208 and 209)
267.Thurston, G. D.; Ito, K. (2001) Epidemiological studies of acute ozone exposures and
mortality. J. Exposure Anal. Environ. Epidemiol. 11: 286-294.
268.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.
269. 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
attached. Docket No. A-2000-01. Document No. IV-A-145.
270. South Coast Air Management District draft plan. (See
http ://www. aqmd.gov/aqmp/03 aqmp.htm)
271.U.S. EPA (1996). Review of National Ambient Air Quality Standards for Ozone,
Assessment of Scientific and Technical Information, OAQPS Staff Paper, EPA-452/R-96-007.
Docket No. A-99-06. Document No. U-A-22.
272. 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
273.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. H-A-30.
2-140
-------
Air Quality, Health, and Welfare Effects
274.U.S. EPA (2003). Technical Support Document for Nonroad Diesel Proposed Rulemaking.
275. NARSTO Synthesis Team (2000). An Assessment of Tropospheric Ozone Pollution: A
North American Perspective.
276. 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.
111. 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. Techno!., 36, 4099-4106.
278. Larsen, L.C. (2003). The Ozone Weekend Effect in California: Evidence Supporting NOx
Emissions Reductions, Submitted to the J. Air & Waste Manage. Assoc.
279. U.S. EPA (2003). Air Quality Technical Support Document for the proposed Nonroad
Diesel rulemaking.
280. Two counties in the Atlanta CMSA and one in the Baltimore-Washington CMSA.
281. 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
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.)
282. (NAS, 1991)
283. 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.
284. U.S. EPA (1996). Air Quality Criteria for Ozone and Related Photochemical Oxidants,
EPA/600/P-93/004aF. Docket No. A-99-06. Document Nos. II-A-15 to 17.
285. 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.
286. U.S. EPA (1996). Air Quality Criteria for Ozone and Related Photochemical Oxidants,
EPA/600/P-93/004aF. Docket No. A-99-06. Document Nos. II-A-15 to 17.
2-141
-------
Draft Regulatory Impact Analysis
287. 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.
288. U.S. EPA (1996). Air Quality Criteria for Ozone and Related Photochemical Oxidants,
EPA/600/P-93/004aF. Docket No. A-99-06. Document Nos. II-A-15 to 17.
289. U.S. EPA (1996). Air Quality Criteria for Ozone and Related Photochemical Oxidants,
EPA/600/P-93/004aF. Docket No. A-99-06. Document Nos. II-A-15 to 17.
290. U.S. EPA (1996). Air Quality Criteria for Ozone and Related Photochemical Oxidants,
EPA/600/P-93/004aF. Docket No. A-99-06. Document Nos. II-A-15 to 17.
291. 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.
292. Winner, W.E., 1994. Mechanistic analysis of plant responses to air pollution. Ecological
Applications, 4(4):651-661.
293. U.S. EPA (1996). Air Quality Criteria for Ozone and Related Photochemical Oxidants,
EPA/600/P-93/004aF. Docket No. A-99-06. Document Nos. II-A-15 to 17.
294. U.S. EPA (1996). Air Quality Criteria for Ozone and Related Photochemical Oxidants,
EPA/600/P-93/004aF. Docket No. A-99-06. Document Nos. II-A-15 to 17.
295. 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.
296. 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.
297. U.S. EPA (1996). Air Quality Criteria for Ozone and Related Photochemical Oxidants,
EPA/600/P-93/004aF. Docket No. A-99-06. Document Nos. II-A-15 to 17.
298. De Steiguer, J., J. Pye, C. Love. 1990. Air pollution Damage to U.S. forests. Journal of
Forestry, Vol 88(8) pp. 17-22.
299. Pye, J.M. Impact of ozone on the growth and yield of trees: A review. Journal of
Environmental Quality 17 pp.347-360., 1988.
300. U.S. EPA (1996). Air Quality Criteria for Ozone and Related Photochemical Oxidants,
EPA/600/P-93/004aF. Docket No. A-99-06. Document Nos. II-A-15 to 17.
2-142
-------
Air Quality, Health, and Welfare Effects
301. U.S. EPA (1996). Air Quality Criteria for Ozone and Related Photochemical Oxidants,
EPA/600/P-93/004aF. Docket No. A-99-06. Document Nos. II-A-15 to 17.
302. 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.
303. 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; EPA report no. EPA-600/D-82-276.
304. U.S. EPA (1996). Air Quality Criteria for Ozone and Related Photochemical Oxidants,
EPA/600/P-93/004aF. Docket No. A-99-06. Document Nos. II-A-15 to 17.
305. 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, US Environmental Protection Agency.
306. U.S. EPA (1996). Air Quality Criteria for Ozone and Related Photochemical Oxidants,
EPA/600/P-93/004aF. Docket No. A-99-06. Document Nos. II-A-15 to 17.
307. 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.
308. 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.
309. 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.
310. Abt Associates, Inc. 1995. Urban ornamental plants: sensitivity to ozone and potential
economic losses. US 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.
311. U.S. EPA (1993). Air Quality Criteria for Oxides of Nitrogen, EPA/600/8-91/049aF.
Docket No. A-2000-01. Document Nos. II-A-89.
312. U.S. EPA (1993). Air Quality Criteria for Oxides of Nitrogen, EP A/600/8-9 l/049aF.
Docket No. A-2000-01. Document Nos. II-A-89.
313. 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
2-143
-------
Draft Regulatory Impact Analysis
Quality Planning and Standards, US Environmental Protection Agency.
314. U.S. EPA (2000). AIR QUALITY CRITERIA FOR CARBON MONOXIDE. USEPA
EPA600/P-99/001F. 01 Jun 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 A-II-29
315. Coburn, R.F. (1979) Mechanisms of carbon monoxide toxicity. Prev. Med. 8:310-322.
316. Helfaer, M.A., and Traystman, RJ. (1996) Cerebrovascular effects of carbon monoxide.
In: Carbon Monoxide (Penney, D.G., ed). Boca Raton, CRC Press, 69-86.
317. 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.
318. U.S. EPA (2000). AIR QUALITY CRITERIA FOR CARBON MONOXIDE. USEPA
EPA600/P-99/001F. 01 Jun 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 A-II-29
319. U.S. EPA (2000). AIR QUALITY CRITERIA FOR CARBON MONOXIDE. USEPA
EPA600/P-99/001F. 01 Jun 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 A-II-29
320. U.S. EPA (2000). AIR QUALITY CRITERIA FOR CARBON MONOXIDE. USEPA
EPA600/P-99/001F. 01 Jun 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 A-II-29
321. U.S. EPA (2000). AIR QUALITY CRITERIA FOR CARBON MONOXIDE. USEPA
EPA600/P-99/001F. 01 Jun 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 A-II-29
322. 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 (EPA-454/R-00-002), March, 2000. These documents are available at Docket No. A-2000-
01, Document No. U-A-72. See also Air Quality Criteria for Carbon Monoxide, U.S. EPA, EPA
600/P-99/001F, June 2000, at 3-10. Air Docket A-2001-11. This document is also available at
http://www.epa.gov/ncea/coab stract. htm.
323. The more stringent standards refer to light light-duty trucks greater than 3750 pounds
loaded vehicle weight, up through 6000 pounds gross vehicle weight rating (also known as
LDT2).
2-144
-------
Air Quality, Health, and Welfare Effects
2-145
-------
CHAPTER 3: Emissions Inventory
3.1 Nonroad Diesel Baseline Emissions Inventory Development 3-1
3.1.1 Land-Based Nonroad Diesel Engines—PM25, NOX, SO2, VOC, and CO Emissions 3-2
3.1.1.1 Overview 3-2
3.1.1.2 NONROAD's Major Inputs 3-3
3.1.1.3 Emissions Estimation Process 3-7
3.1.1.4 Estimation of VOC Emissions 3-9
3.1.1.5 Estimation of SO2 Emissions 3-9
3.1.1.6 Estimation of PM25 Emissions 3-10
3.1.1.7 Estimation of Fuel Consumption 3-10
3.1.1.8 Baseline Inventory 3-10
3.1.2 Land-Based Nonroad Diesel Engines—Air Toxics Emissions 3-12
3.1.3 Commercial Marine Vessels and Locomotives 3-14
3.1.4 Recreational Marine Engines 3-19
3.1.5 Fuel Consumption for Nonroad Diesel Engines 3-22
3.2 Contribution of Nonroad Diesel Engines to National Emission Inventories 3-24
3.2.1 Baseline Emissions Inventory Development 3-24
3.2.2 PM25 Emissions 3-25
3.2.3 NOX Emissions 3-26
3.2.4 SO2 Emissions 3-26
3.2.5 VOC Emissions 3-27
3.2.6 CO Emissions 3-27
3.3 Contribution of Nonroad Diesel Engines to Selected Local Emission Inventories 3-35
3.3.1 PM25 Emissions 3-35
3.3.2 NOX Emissions 3-38
3.4 Nonroad Diesel Controlled Emissions Inventory Development 3-41
3.4.1 Land-Based Diesel Engines—PM25, NOX, SO2, VOC, and CO Emissions 3-41
3.4.1.1 Standards and Zero-Hour Emission Factors 3-42
3.4.1.2 Transient Adjustment Factors 3-42
3.4.1.3 Deterioration Rates 3-45
3.4.1.4 In-Use Sulfur Levels, Certification Sulfur Levels, and Sulfur Conversion Factors 3-45
3.4.1.5 Modeling 50-75 hp and 75-100 hp Within the NONROAD 50-100 hp Bin 3-47
3.4.1.6 Controlled Inventory 3-47
3.4.2 Land-Based Diesel Engines—Air Toxics Emissions 3-50
3.4.3 Commercial Marine Vessels and Locomotives 3-51
3.4.4 Recreational Marine Engines 3-53
3.5 Anticipated Emission Reductions With the Proposed Rule 3-55
3.5.1 PM25 Reductions 3-56
3.5.2 NOX Reductions 3-64
3.5.3 SO2 Reductions 3-66
3.5.4 VOC and Air Toxics Reductions 3-73
3.5.5 CO Reductions 3-76
3.6 Emission Inventories Used for Air Quality Modeling 3-77
-------
CHAPTER 3: Emissions Inventory
This chapter presents our analysis of the emission impact of the proposed 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 proposed for the land-based diesel engine category. For the other three nonroad diesel
categories, no new engine controls are being proposed; however, the diesel fuel sulfur
requirements are expected to decrease particulate matter less than 2.5 microns (PM25) and sulfur
dioxide (SO2) emissions 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 proposed standards and fuel sulfur requirements. Section 3.5 follows with the
expected emission reductions associated with the proposed rule. Section 3.6 concludes the
chapter by describing the changes in the inputs and resulting emissions inventories between the
preliminary baseline and control scenarios used for the air quality modeling and the updated
baseline and control scenarios in this proposal.
The estimates of baseline emissions and emission reductions from the proposed rule 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: PM25, 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.
3.1 Nonroad Diesel Baseline Emissions Inventory Development
This section describes how the baseline emissions inventories were developed for the four
categories of nonroad diesel engines affected by this proposal: 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.
-------
Draft Regulatory Impact Analysis
3.1.1 Land-Based Nonroad Diesel Engines—PM25, NOX, SO2, VOC, and CO Emissions
The baseline emissions inventories for land-based diesel engines were generated using the
draft NONROAD2002 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 NONROAD2002 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 draft NONROAD2002 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.
3.1.1.1 Overview
The draft NONROAD2002 model estimates emissions 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 the years 1970 and 2050.
The draft NONROAD2002 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 emissions inventories.
This section describes how the draft NONROAD2002 model estimates emissions particularly
relevant to this analysis, including particulate 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 proposed 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
O O
3-2
-------
Emissions Inventory
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
subsection 3.1.1.4.
3.1.1.2 NONROAD's Major Inputs
The draft NONROAD2002 model uses three major sets of inputs in estimation of exhaust
emission inventories: (1) emissions calculation variables, (2) projection variables, and (3)
scenario option variables.
3.1.1.2.1 Emissions Calculation Variables
The draft NONROAD2002 model estimates exhaust emissions using the equation
/exh = £exh -A-L-P-N
where each term is defined as follows:
/exh = the exhaust emission inventory (gram/year, gram/day),
E1^ = exhaust emission factor (gram/hp-hr),
A = equipment activity (operating hours/year),
L = Load factor (average proportion of rated power used during operation (%)),
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.).
Exhaust Emission Factor. The emission factor in a given simulation year consists of three
components, a "zero-hour" emission level (ZFIL) , a transient adjustment factor (TAF) 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 emissions data has 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-
5-3
-------
Draft Regulatory Impact Analysis
state operation. The TAFs are estimated by collecting emissions measurements on specific
engines using both transient and steady-state cycles, and calculating the ratio
FF
rp . -p _ transient
-^ "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., TAP = 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 factors1.
During a model run, the model applies emissions 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 ZFIL-TAF-DF. Deterioration factors vary from year to year; we describe their
calculation in more detail in subsection 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 TAP 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
a given scenario, NONROAD adjusts the PM emission factor by the margin SPMadj (g/hp-hr)
calculated as
SPMadj = BSFC ->nso4,S •WpM.s '0.01 '(S^ - Sm_m)
where: BSFC = brake-specific fuel consumption (g fuel/hp-hr),
mso4 s = a constant, representing the sulfate fraction of total particulate sulfur, equal to
7.0gPMSO4/gPMS,
mpu,s = a constant, representing the fraction of fuel sulfur converted to parti culate 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),
= in-use diesel sulfur level as specified by user (wt%).
3-4
-------
Emissions Inventory
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 NONROAD2002, 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 NONROAD2002, we obtained estimates from the
Partslink database, maintained by Power Systems Research/Compass International, Inc. The
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 NONROAD2002 model projects future or past
emissions. As a reference point, the input file contains populations in the model's base year
1998. 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.
5-5
-------
Draft Regulatory Impact Analysis
(PSR). This database contains sales estimates for nonroad equipment for model years 1973
through 1999. 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 emissions
deterioration.
Annual Population Growth Rate (%/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 tiers4. 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% 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 (2/y); during the first
median life, 50% of the engines are scrapped, and over the second, the remaining 50% are
scrapped. For a more detailed description of median life, see the model documentation.2
Relative Deterioration Rate (% increase in emission factor/% 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
NONROAD2002 calculates the deterioration factor as
DF = 1 + d
pollutant,tier,year pollutant,tier I 7
where:
DFpollutantyear = 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 (% increase in emission factor /
% useful life expended) and regulatory tier,
age = the age of a specific model-year group of engines in the simulation year,
-------
Emissions Inventory
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.
Table 3.1-1
Modeled Baseline In-Use Diesel Fuel Sulfur Content
for Land-Based Nonroad Diesel Engines
Fuel Sulfur (ppm)
2318
2271
2237
2217
Calendar Year
through 2005
2006
2007-2009
2010+
3.1.1.3 Emissions Estimation Process
To project emissions in a given year, the draft NONROAD2002 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 (JV^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,
5-7
-------
Draft Regulatory Impact Analysis
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 2ly5
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.
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.
Emissions Calculation. For a given pollutant, the calculations described above are performed
and the resulting inputs multiplied in the exhaust emissions 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
sum over all equipment types
7
exh,poll
=1
sum over all rated-power classes
within an equipm ent type
A
sum over all model-year cohorts
within a rated-power class
-------
Emissions Inventory
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 non-
reactive 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 NONROAD2002 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 - 0 02 -FF
ijl crank,HC,year ~~ u-u^ -^ 1 exh,HC,year
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, VOCcmnk= 1.053-THCcmnk
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 NONROAD2002 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.
EFS02 =[BSFC.(l-mPM,)-EFTHC].Sm_use.ms02,
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,
3-9
-------
Draft Regulatory Impact Analysis
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
mso2 s = 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 (/HPMjS)
and for the sulfur remaining in the unburned fuel (EF^). 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 92% of PM10 emissions. This is based on an analysis of size distribution
data for diesel vehicles.7
3.1.1.7 Estimation of Fuel Consumption
The draft NONROAD2002 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 (%
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 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.
3-10
-------
Emissions Inventory
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
191,858
175,155
169,360
163,684
157,726
152,310
147,050
142,043
138,140
135,640
133,495
131,530
130,288
129,691
129,674
129,932
130,388
130,986
131,765
132,672
133,767
135,146
136,655
138,195
139,797
141,410
143,091
144,798
146,471
148,187
149,915
151,660
153,451
155,260
157,088
158,922
160,748
162,618
164,511
166,681
168,853
171,019
PM25
176,510
161,143
155,811
150,589
145,108
140,125
135,286
130,680
127,089
124,789
122,815
121,007
119,865
119,316
119,300
119,537
119,957
120,507
121,224
122,059
123,065
124,334
125,723
127,140
128,613
130,097
131,644
133,214
134,753
136,332
137,922
139,527
141,175
142,839
144,521
146,208
147,888
149,609
151,350
153,346
155,345
157,337
NOX
1,583,664
1,569,902
1,556,973
1,544,395
1,522,881
1,503,228
1,483,942
1,450,762
1,414,673
1,374,171
1,331,986
1,291,533
1,255,472
1,225,493
1,202,185
1,183,043
1,167,635
1,156,099
1,147,635
1,142,299
1,140,236
1,140,727
1,143,660
1,148,710
1,155,440
1,163,558
1,172,971
1,183,408
1,194,643
1,206,483
1,218,884
1,231,995
1,245,794
1,259,909
1,274,280
1,288,943
1,303,901
1,319,167
1,334,609
1,350,619
1,366,795
1,383,101
S02
147,926
167,094
171,957
176,819
181,677
186,532
191,385
192,228
194,003
198,657
203,311
206,104
210,737
215,366
219,992
224,615
229,235
233,809
238,381
242,952
247,521
252,089
256,656
261,222
265,786
270,350
274,913
279,446
283,978
288,510
293,042
297,573
302,104
306,635
311,165
315,695
320,226
324,755
329,285
333,814
338,344
342,873
voc
221,403
200,366
191,785
183,584
176,201
169,541
163,193
156,295
149,518
142,310
135,259
128,391
122,161
116,940
112,619
108,942
105,800
103,210
101,137
99,415
97,952
96,855
96,055
95,488
95,170
95,066
95,144
95,373
95,729
96,186
96,724
97,348
98,059
98,822
99,628
100,482
101,380
102,336
103,325
104,415
105,529
106,664
CO
1,010,518
923,886
886,722
850,751
817,858
790,468
764,918
742,184
724,213
709,119
695,970
684,552
675,805
671,268
670,147
670,842
672,944
676,412
681,217
686,723
692,845
700,017
707,986
716,295
724,914
733,953
743,434
753,165
763,023
773,136
783,449
793,923
804,566
815,321
826,151
837,047
847,953
858,992
870,072
881,159
892,281
903,406
3-11
-------
Draft Regulatory Impact Analysis
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
192,750
175,981
170,165
164,467
158,487
153,045
147,761
142,732
138,814
136,306
134,154
132,184
130,942
130,347
130,335
130,598
131,060
131,665
132,452
133,368
134,470
135,858
137,377
138,926
140,537
142,161
143,853
145,570
147,254
148,981
150,719
152,475
154,278
156,098
157,937
159,783
161,621
163,503
165,407
167,589
169,774
171,952
PM25
177,330
161,903
156,552
151,310
145,808
140,802
135,940
131,314
127,708
125,401
123,422
121,609
120,466
119,919
119,908
120,150
120,575
121,132
121,856
122,698
123,713
124,990
126,387
127,812
129,294
130,788
132,345
133,925
135,474
137,062
138,662
140,277
141,936
143,610
145,302
147,001
148,691
150,423
152,174
154,182
156,192
158,195
NOX
1,592,025
1,578,148
1,565,144
1,552,490
1,530,854
1,511,087
1,491,692
1,458,315
1,422,017
1,381,288
1,338,867
1,298,193
1,261,939
1,231,796
1,208,367
1,189,132
1,173,656
1,162,082
1,153,602
1,148,271
1,146,227
1,146,750
1,149,727
1,154,830
1,161,619
1,169,801
1,179,283
1,189,792
1,201,104
1,213,023
1,225,506
1,238,701
1,252,586
1,266,789
1,281,249
1,296,002
1,311,051
1,326,409
1,341,945
1,358,049
1,374,321
1,390,723
S02
148,729
167,999
172,889
177,777
182,662
187,544
192,424
193,272
195,057
199,736
204,416
207,225
211,884
216,538
221,189
225,837
230,483
235,083
239,680
244,276
248,870
253,464
258,056
262,647
267,237
271,826
276,414
280,972
285,529
290,086
294,643
299,199
303,755
308,311
312,867
317,422
321,977
326,532
331,087
335,642
340,196
344,750
voc
222,517
201,386
192,765
184,524
177,107
170,414
164,035
157,104
150,293
143,050
135,965
129,063
122,803
117,555
113,212
109,517
106,359
103,757
101,676
99,946
98,478
97,378
96,575
96,006
95,688
95,584
95,663
95,894
96,253
96,713
97,255
97,882
98,598
99,367
100,178
101,037
101,940
102,902
103,898
104,994
106,114
107,257
CO
1,015,773
928,674
891,304
855,132
822,062
794,522
768,838
745,994
727,946
712,797
699,604
688,153
679,389
674,849
673,736
674,446
676,570
680,068
684,911
690,458
696,625
703,845
711,863
720,223
728,896
737,990
747,528
757,318
767,234
777,407
787,780
798,316
809,021
819,838
830,731
841,690
852,660
863,763
874,908
886,060
897,248
908,439
3.1.2 Land-Based Nonroad Diesel Engines—Air Toxics Emissions
EPA focused on 5 major air toxics pollutants for the proposed rule: benzene, formaldehyde,
acetaldehyde, 1,3-butadiene, and acrolein. These pollutants are VOCs and are included in the
total land-based nonroad diesel VOC emissions estimate. EPA developed the baseline inventory
3-12
-------
Emissions Inventory
estimates for these pollutants by multiplying the baseline VOC emissions from the draft
NONROAD2002 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
Emissions Inventory.8
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
4,428
4,007
3,264
2,990
2,568
2,116
1,937
1,903
1,947
Formaldehyde
26,126
23,643
19,257
17,643
15,150
12,484
11,429
11,227
11,487
Acetaldehyde
11,734
10,619
8,649
7,924
6,805
5,607
5,133
5,043
5,159
1,3 -Butadiene
443
401
326
299
257
212
194
190
195
Acrolein
664
601
490
449
385
317
291
285
292
3-13
-------
Draft Regulatory Impact Analysis
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
4,450
4,028
3,281
3,006
2,581
2,127
1,948
1,913
1,958
Formaldehyde
26,257
23,764
19,356
17,735
15,229
12,550
11,491
11,288
11,550
Acetaldehyde
11,793
10,673
8,694
7,966
6,840
5,637
5,161
5,070
5,188
1,3 -Butadiene
445
403
328
301
258
213
195
191
196
Acrolein
668
604
492
451
387
319
292
287
294
3.1.3 Commercial Marine Vessels and Locomotives
Although no new engine controls are being proposed for diesel commercial marine and
locomotive engines, these engines do use diesel fuel and the effects of the proposed fuel changes
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 locomotive emissions
final rule, (b) the December 1999 final rule for commercial marine diesel engines, and (c) the
January 2001 heavy duty highway diesel fuel rule that takes effect in June 2006.
Since the draft NONROAD2002 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)
were taken from the existing JrtDDV inventory. These are presented in Table 3.1-5. Only 48-
state emission estimates are available for these pollutants. VOC emissions in this inventory were
calculated by multiplying THC emissions by a factor of 1.053, which is also the factor used for
land-based diesel engines.
3-14
-------
Emissions Inventory
Table 3.1-5
Baseline (48-State) NOX, VOC, and CO Emissions
for Locomotives and Commercial Marine Vessels (short tons)
Year
1996
2020
2030
NOX
Locomotives
921,556
612,722
534,520
CMV
959,704
819,201
814,827
VOC
Locomotives
48,381
36,546
31,644
CMV
31,545
37,290
41,354
CO
Locomotives
112,171
119,302
119,302
CMV
126,382
159,900
176,533
The baseline SO2 and PM emission inventory estimates were revised to reflect changes to the
base sulfur levels. Tables 3.1-6a and 3.1-6b provide the 48-state and 50-state baseline fuel sulfur
levels, PM2 5, and SO2 emissions. The fuel sulfur levels were calculated as weighted average in-
use levels of (a) uncontrolled nonroad diesel fuel at 3400 ppm sulfur, (b) "spillover" of low sulfur
highway diesel fuel into use by nonroad applications outside of California, and (c) full use of low
sulfur California fuel in all nonroad applications in California. 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 sulfur levels is discussed in more detail in Chapter 7.
Locomotive diesel fuel volumes were calculated by the following methodology. Calendar
year 2000 distillate fuel consumption for railroads was taken from the US Energy Information
Administration (EIA) Fuel Oil & Kerosene Supply (FOKS) 2000 report. The volume of diesel
fuel consumed by railroad was assumed to represent 95 percent of the reported distillate value.
Also, the fuel volumes reported in FOKS appear to represent fuel usage by locomotives as well
as by rail maintenance equipment, so the fuel consumption specific to locomotives was
calculated by subtracting one percent, which is our estimate of rail maintenance fuel
consumption, from the railroad diesel volume estimate. Calendar year 2001-2020 locomotive
fuel consumption values were computed by multiplying the year 2000 fuel volume by a growth
factor computed as the ratio of projected calendar year railroad sector energy consumption to
year 2000 energy consumption from the EIA Annual Energy Outlook (AEO) 2002, Table 7,
Transportation Sector, Key Indicators and Delivered Energy Consumption, Energy Use by Mode,
Railroad. Calendar year 2021-2040 locomotive gallons were computed by growing the year 2000
locomotive fuel volume using the EIA/AEO 2000-2020 average annual compound growth of
0.892% (e.g., 2030CY growth factor = 1.0089230 = 1.305). The methodology for determining
the locomotive diesel fuel consumption is further documented in Chapter 7. The locomotive
diesel fuel consumption for 1996 was estimated by multiplying the calendar year 2000
locomotive fuel volume by the ratio of the FOKS railroad distillate fuel volumes for 1996 and
2000.
3-15
-------
Draft Regulatory Impact Analysis
Table3.1-6a
Baseline (48-State) Fuel Sulfur Levels, SO2,
Sulfate PM, and PM2 s Emissions for Locomotives and Commercial Marine 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
Locomotive
Usage
(109gal/yr)
3.039
2.821
2.966
2.918
2.969
3.010
3.051
3.081
3.111
3.124
3.146
3.169
3.223
3.237
3.246
3.255
3.270
3.303
3.322
3.340
3.358
3.369
3.399
3.430
3.460
3.491
3.522
3.554
3.585
3.617
3.650
3.682
3.715
3.748
3.782
3.815
3.849
3.884
3.918
3.953
3.989
Commercia
1 Marine
Usage
(109gal/yr)
1.560
1.611
1.629
1.646
1.664
1.683
1.701
1.720
1.739
1.758
1.778
1.797
1.817
1.838
1.858
1.879
1.900
1.921
1.943
1.965
1.988
2.010
2.033
2.057
2.080
2.105
2.129
2.154
2.179
2.205
2.231
2.257
2.284
2.312
2.340
2.368
2.397
2.426
2.456
2.486
2.517
Base
Sulfur
Level
(ppm)
2396
2396
2396
2396
2396
2396
2396
2352
2321
2321
2321
2302
2302
2302
2302
2302
2302
2302
2302
2302
2302
2302
2302
2302
2302
2302
2302
2302
2302
2302
2302
2302
2302
2302
2302
2302
2302
2302
2302
2302
2302
Base
S02
Loco
50,534
46,911
49,328
48,521
49,366
50,047
50,738
50,288
50,109
50,331
50,679
50,625
51,487
51,720
51,861
52,006
52,240
52,776
53,079
53,355
53,655
53,832
54,312
54,796
55,285
55,778
56,276
56,778
57,285
57,796
58,311
58,832
59,356
59,886
60,420
60,959
61,503
62,052
62,605
63,164
63,727
CMV
25,948
26,791
27,083
27,379
27,679
27,983
28,291
28,078
28,015
28,326
28,638
28,718
29,036
29,359
29,686
30,019
30,356
30,698
31,045
31,397
31,754
32,117
32,485
32,859
33,239
33,624
34,016
34,413
34,817
35,227
35,644
36,068
36,498
36,936
37,380
37,832
38,292
38,759
39,235
39,718
40,210
Sulfate PM
Loco
4,066
3,774
3,969
3,904
3,972
4,026
4,082
4,046
4,031
4,049
4,077
4,073
4,142
4,161
4,172
4,184
4,203
4,246
4,270
4,293
4,317
4,331
4,370
4,409
4,448
4,488
4,528
4,568
4,609
4,650
4,691
4,733
4,775
4,818
4,861
4,904
4,948
4,992
5,037
5,082
5,127
CMV
2,088
2,155
2,179
2,203
2,227
2,251
2,276
2,259
2,254
2,279
2,304
2,310
2,336
2,362
2,388
2,415
2,442
2,470
2,498
2,526
2,555
2,584
2,614
2,644
2,674
2,705
2,737
2,769
2,801
2,834
2,868
2,902
2,936
2,972
3,007
3,044
3,081
3,118
3,157
3,195
3,235
PM10 EF
Loco
(g/gallon
6^8
6.8
6.8
6.8
6.8
6.8
6.6
6.4
6.2
6.0
5.9
5.7
5.7
5.6
5.5
5.4
5.3
5.2
5.1
5.1
5.0
4.9
4.8
4.7
4.7
4.6
4.5
4.4
4.4
4.3
4.2
4.2
4.1
4.0
4.0
3.9
3.9
3.9
3.8
3.8
3.7
Total PM2 5
Loco
20,937
19,436
20,438
20,103
20,453
20,735
20,403
19,977
19,541
18,994
18,807
18,300
18,612
18,368
18,089
17,809
17,559
17,404
17,167
17,257
17,013
16,728
16,533
16,333
16,478
16,272
16,060
15,843
15,985
15,761
15,531
15,670
15,433
15,191
15,327
15,077
15,211
15,347
15,087
15,222
14,953
CMV
36,366
37,186
37,397
37,608
37,819
38,031
38,243
38,416
38,601
38,813
39,026
39,221
39,434
39,647
39,860
40,074
40,288
40,503
40,718
40,933
41,149
41,365
41,767
42,171
42,574
42,978
43,382
43,787
44,192
44,598
45,004
45,411
45,818
46,226
46,634
47,043
47,453
47,863
48,273
48,684
49,096
3-16
-------
Emissions Inventory
2040 I 4.024 I 2.548 I 2302 I 64.296 I 40.710 I 5.173 I 3.275 I 3.7 I 15.087 I 49.509 II
Table3.1-6b
Baseline (50-State) Fuel Sulfur Levels, SO2,
Sulfate PM, and PM2 s Emissions for Locomotives and Commercial Marine 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
Locomotive
Usage
(109gal/yr)
3.043
2.825
2.970
2.922
2.973
3.014
3.055
3.085
3.115
3.129
3.150
3.173
3.227
3.242
3.251
3.260
3.274
3.308
3.327
3.344
3.363
3.374
3.404
3.434
3.465
3.496
3.527
3.559
3.590
3.622
3.655
3.687
3.720
3.753
3.787
3.821
3.855
3.889
3.924
Commercia
1 Marine
Usage
(109gal/yr)
1.642
1.695
1.713
1.732
1.751
1.770
1.790
1.809
1.830
1.850
1.870
1.891
1.912
1.933
1.955
1.977
1.999
2.021
2.044
2.067
2.091
2.115
2.139
2.164
2.189
2.214
2.240
2.266
2.292
2.320
2.347
2.375
2.403
2.432
2.461
2.491
2.521
2.552
2.583
Base
Sulfur
Level
(ppm)
2396
2396
2396
2396
2396
2396
2396
2352
2321
2321
2321
2302
2302
2302
2302
2302
2302
2302
2302
2302
2302
2302
2302
2302
2302
2302
2302
2302
2302
2302
2302
2302
2302
2302
2302
2302
2302
2302
2302
Base
S02
Loco
50,604
46,976
49,396
48,589
49,434
50,116
50,808
50,358
50,178
50,401
50,749
50,695
51,559
51,792
51,933
52,078
52,313
52,849
53,153
53,429
53,729
53,906
54,387
54,872
55,362
55,856
56,354
56,857
57,364
57,876
58,392
58,913
59,439
59,969
60,504
61,044
61,588
62,138
62,692
CMV
27,297
28,184
28,491
28,802
29,118
29,437
29,762
29,538
29,471
29,798
30,127
30,210
30,545
30,885
31,229
31,579
31,934
32,293
32,658
33,029
33,405
33,787
34,174
34,567
34,967
35,372
35,784
36,202
36,627
37,059
37,497
37,943
38,396
38,856
39,323
39,799
40,282
40,774
41,274
Sulfate PM
Loco
4,071
3,779
3,974
3,909
3,977
4,032
4,088
4,051
4,037
4,055
4,083
4,079
4,148
4,167
4,178
4,190
4,209
4,252
4,276
4,299
4,323
4,337
4,376
4,415
4,454
4,494
4,534
4,574
4,615
4,656
4,698
4,740
4,782
4,825
4,868
4,911
4,955
4,999
5,044
CMV
2,196
2,267
2,292
2,317
2,343
2,368
2,394
2,376
2,371
2,397
2,424
2,431
2,457
2,485
2,512
2,541
2,569
2,598
2,627
2,657
2,688
2,718
2,749
2,781
2,813
2,846
2,879
2,913
2,947
2,981
3,017
3,053
3,089
3,126
3,164
3,202
3,241
3,280
3,321
PM10 EF
Loco
(g/gallon
6^8
6.8
6.8
6.8
6.8
6.8
6.6
6.4
6.2
6.0
5.9
5.7
5.7
5.6
5.5
5.4
5.3
5.2
5.1
5.1
5.0
4.9
4.8
4.7
4.7
4.6
4.5
4.4
4.4
4.3
4.2
4.2
4.1
4.0
4.0
3.9
3.9
3.9
3.8
Total PM2 5
Loco
20,966
19,463
20,466
20,131
20,482
20,764
20,432
20,004
19,568
19,021
18,833
18,325
18,637
18,393
18,114
17,834
17,583
17,428
17,191
17,281
17,037
16,751
16,556
16,355
16,501
16,294
16,082
15,865
16,007
15,782
15,553
15,692
15,455
15,212
15,348
15,098
15,233
15,368
15,108
CMV
38,257
39,119
39,340
39,563
39,785
40,008
40,231
40,413
40,608
40,831
41,054
41,259
41,483
41,708
41,932
42,157
42,382
42,608
42,834
43,061
43,288
43,515
43,939
44,363
44,787
45,212
45,637
46,063
46,490
46,916
47,344
47,772
48,200
48,629
49,059
49,489
49,919
50,351
50,783
3-17
-------
Draft Regulatory Impact Analysis
2038
2039
2040
3.959
3.994
4.030
2.615
2.648
2.680
2302
2302
2302
63,252
63,816
64 385
41,783
42,300
42.826
5,089
5,134
5.180
3,362
3,403
3.445
3.8
3.7
3.7
15,243
14,974
15.108
51,215 1
51,648
52.082 I
Vessel bunkering (commercial and recreational marine) distillate fuel consumption for
calendar year 2000 was also taken from the EIA FOKS 2000 report. The volume of diesel fuel
consumed by vessel bunkering was assumed to represent 90 percent of the reported distillate
value. The fuel consumption specific to commercial marine was then calculated by subtracting
the recreational marine fuel consumption as generated by the draft NONROAD2002 model.
Calendar year 2001-2040 commercial marine diesel fuel consumption values were computed by
multiplying the year 2000 volume by the growth factor of CO emission projections for the
combination of Category 1 and 2 vessels in the 2002 diesel marine engine final rule. CO
emission projections were used due to availability and applicability as an appropriate surrogate
for fuel consumption. The commercial marine diesel fuel consumption for 1996 was estimated
by multiplying the calendar year 2000 commercial marine fuel volume by the ratio of the CO
emission projections for 1996 and 2000.
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% 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 NONROAD2002 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 not used for two reasons: 1) THC emission factors were not
available for years other than 1996, 2020, and 2030, and 2) this correction factor 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
3-18
-------
Emissions Inventory
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
PM10 is assumed to be equivalent to total PM, and PM2 5 is estimated by multiplying PM10
emissions by a factor of 0.92. This is the factor used for all nonroad diesel engines; the basis is
described in Section 3.1.1.6.
Annual PM10 emission estimates for commercial marine vessels in calendar years 1996
and 2000 were taken from the inventory done for the HD07 rule. For years 2001 - 2030, the year
2000 inventory was adjusted according to the commercial marine growth factor mentioned above
from the 2002 diesel marine engine final rule. The fuel sulfate portion was then adjusted to
account for the revised sulfur levels.
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 NONROAD2002 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 NONROAD2002 model. The
emission inventory numbers presented here assume that recreational marine applications would
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 these inventory values do not 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% - 37% 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% 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.
3-19
-------
Draft Regulatory Impact Analysis
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
529
594
611
627
643
660
676
688
700
716
732
745
760
776
791
806
821
836
851
865
880
895
909
924
938
953
967
982
996
1,011
1,026
1,042
1,057
1,072
1,088
1,103
1,119
1,134
1,150
1,166
1,182
1,198
PM25
487
547
562
577
592
607
622
633
644
659
673
686
700
714
728
741
755
769
783
796
810
823
837
850
863
877
890
903
917
930
944
958
973
987
1,001
1,015
1,029
1,044
1,058
1,073
1,087
1,102
NOX
19,440
21,899
22,548
23,196
23,844
24,492
25,139
25,790
26,439
27,088
27,736
28,384
29,028
29,671
30,314
30,957
31,600
32,244
32,888
33,531
34,174
34,817
35,460
36,103
36,746
37,388
38,031
38,673
39,316
39,959
40,604
41,250
41,896
42,543
43,189
43,836
44,483
45,131
45,779
46,428
47,076
47,725
SO2
2,251
2,537
2,613
2,689
2,765
2,841
2,917
2,939
2,974
3,049
3,123
3,171
3,244
3,317
3,390
3,463
3,536
3,610
3,683
3,756
3,830
3,903
3,976
4,050
4,123
4,196
4,270
4,343
4,416
4,489
4,563
4,636
4,709
4,782
4,856
4,929
5,002
5,075
5,149
5,222
5,295
5,368
VOC
803
900
923
947
970
992
1,015
1,037
1,059
1,081
1,102
1,124
1,145
1,166
1,186
1,207
1,227
1,247
1,268
1,288
1,308
1,328
1,347
1,367
1,387
1,406
1,426
1,446
1,465
1,486
1,507
1,528
1,550
1,571
1,592
1,614
1,636
1,658
1,680
1,703
1,725
1,748
CO
3,215
3,613
3,713
3,814
3,913
4,013
4,112
4,211
4,309
4,406
4,503
4,599
4,695
4,790
4,884
4,979
5,072
5,166
5,260
5,353
5,445
5,538
5,630
5,722
5,814
5,906
5,997
6,089
6,181
6,275
6,370
6,465
6,561
6,656
6,752
6,848
6,945
7,041
7,139
7,238
7,336
7,435
3-20
-------
Emissions Inventory
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
532
598
615
631
648
664
680
692
705
721
736
750
765
781
796
811
826
841
856
871
886
900
915
930
944
959
973
988
1,003
1,018
1,033
1,048
1,064
1,079
1,095
1,110
1,126
1,141
1,157
1,173
1,189
1,205
PM25
490
550
566
581
596
611
626
637
648
663
677
690
704
718
732
746
760
774
787
801
815
828
842
855
869
882
895
909
922
936
950
964
979
993
1,007
1,021
1,036
1,050
1,065
1,079
1,094
1,109
NOX
19,562
22,036
22,689
23,342
23,994
24,646
25,297
25,952
26,605
27,258
27,911
28,563
29,210
29,858
30,505
31,152
31,798
32,446
33,094
33,742
34,389
35,036
35,683
36,330
36,977
37,623
38,270
38,916
39,563
40,210
40,860
41,510
42,160
42,810
43,461
44,112
44,763
45,414
46,067
46,719
47,372
48,025
SO2
2,265
2,553
2,629
2,706
2,783
2,859
2,936
2,957
2,993
3,068
3,142
3,191
3,264
3,338
3,411
3,485
3,559
3,632
3,706
3,780
3,854
3,928
4,001
4,075
4,149
4,223
4,297
4,370
4,444
4,518
4,591
4,665
4,739
4,812
4,886
4,960
5,034
5,107
5,181
5,255
5,328
5,402
voc
808
906
929
953
976
999
1,021
1,044
1,066
1,088
1,109
1,131
1,152
1,173
1,194
1,214
1,235
1,255
1,276
1,296
1,316
1,336
1,356
1,376
1,395
1,415
1,435
1,455
1,475
1,495
1,516
1,538
1,559
1,581
1,602
1,624
1,646
1,668
1,691
1,713
1,736
1,759
CO
3,236
3,635
3,737
3,838
3,938
4,038
4,138
4,237
4,336
4,434
4,531
4,628
4,724
4,820
4,915
5,010
5,104
5,199
5,293
5,386
5,480
5,573
5,665
5,758
5,850
5,943
6,035
6,127
6,220
6,314
6,410
6,506
6,602
6,698
6,795
6,891
6,988
7,086
7,184
7,283
7,382
7,482
3-21
-------
Draft Regulatory Impact Analysis
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 NONROAD2002 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.
Some of the estimates in Table 3.1-8 are different than those presented in Chapter 7,
which are ultimately used in estimating the cost of the proposed fuel regulations. As described
above, the diesel fuel consumption volumes for land-based nonroad engines in this chapter were
obtained from the draft NONROAD2002 model. In Chapter 7, land-based diesel fuel
consumption demand was developed from an independent source of fuel consumption
information. Specifically, those estimates are based primarily on data contained in the EIA
FOKS 2000 report. That document broadly reports fuel sales for all uses, including stationary
sources. Therefore, a number of assumptions must be applied to the information contained in
FOKS 2000 to obtain an estimate of diesel fuel for nonroad engines.
When comparing the two methods of developing fuel consumption estimates, there is
some difference in the results. Rather than adopt one of the two for all uses, we have decided to
maintain the NONROAD2002 based estimates for inventory generation and the land-based diesel
fuel estimates of Chapter 7 for the cost analyses. Use of the nonroad estimates for cost or the
Chapter 7 estimates for emissions would introduce internal inconsistencies in the resulting cost
or inventory results. These two estimates differ by a relatively small amount, approximately 15
percent in 2030, so we decided that maintaining consistency within the emissions modeling and
within the cost estimation was preferable to enforcing consistency between these two areas. The
Agency will continue to investigate how to resolve the differences between the two approaches
for the final rule, if appropriate.
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 NONROAD2002
model for each calendar year, and subtracted from the respective combined marine end use
volume to produce the commercial marine estimate. Also, the combined marine volume
estimates in the two chapters differ by about one percentage for years prior to 2008 due to the use
of a slightly different computational methodology for that period.
3-22
-------
Emissions Inventory
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
9,254
10,440
10,740
11,040
11,340
11,640
11,940
12,238
12,535
12,833
13,131
13,429
13,728
14,028
14,327
14,627
14,926
15,223
15,519
15,816
16,112
16,409
16,706
17,002
17,299
17,595
17,892
18,186
18,481
18,776
19,070
19,365
19,660
19,954
20,249
20,544
20,838
21,133
21,428
21,722
22,017
22,312
50-State
9,304
10,496
10,798
11,100
11,401
11,703
12,005
12,304
12,603
12,903
13,202
13,502
13,803
14,104
14,405
14,706
15,007
15,306
15,604
15,902
16,200
16,498
16,797
17,095
17,393
17,691
17,989
18,286
18,582
18,878
19,175
19,471
19,767
20,063
20,358
20,656
20,952
21,249
21,545
21,841
22,137
22,434
Recreational Marine
48-State
136
153
157
162
167
171
176
180
185
190
194
199
203
208
212
217
222
226
231
235
240
245
249
254
258
263
268
272
277
281
286
291
295
300
304
309
314
318
323
327
332
336
50-State
136
154
158
163
168
172
177
181
186
191
195
200
205
209
214
218
223
228
232
237
242
246
251
255
260
265
269
274
279
283
288
292
297
302
306
311
315
320
325
329
334
339
Locomotives
48-State
3,039
2,821
2,966
2,918
2,969
3,010
3,051
3,081
3,111
3,124
3,146
3,169
3,223
3,237
3,246
3,255
3,270
3,303
3,322
3,340
3,358
3,369
3,399
3,430
3,460
3,491
3,522
3,554
3,585
3,617
3,650
3,682
3,715
3,748
3,782
3,815
3,849
3,884
3,918
3,953
3,989
4,024
50-State
3,043
2,825
2,970
2,922
2,973
3,014
3,055
3,085
3,115
3,129
3,150
3,173
3,227
3,242
3,251
3,260
3,274
3,308
3,327
3,344
3,363
3,374
3,404
3,434
3,465
3,496
3,527
3,559
3,590
3,622
3,655
3,687
3,720
3,753
3,787
3,821
3,855
3,889
3,924
3,959
3,994
4,030
Commercial Marine
48-State
1,560
1,611
1,629
1,646
1,664
1,683
1,701
1,720
1,739
1,758
1,778
1,797
1,817
1,838
1,858
1,879
1,900
1,921
1,943
1,965
1,988
2,010
2,033
2,057
2,080
2,105
2,129
2,154
2,179
2,205
2,231
2,257
2,284
2,312
2,340
2,368
2,397
2,426
2,456
2,486
2,517
2,548
50-State
1,642
1,695
1,713
1,732
1,751
1,770
1,790
1,809
1,830
1,850
1,870
1,891
1,912
1,933
1,955
1,977
1,999
2,021
2,044
2,067
2,091
2,115
2,139
2,164
2,189
2,214
2,240
2,266
2,292
2,320
2,347
2,375
2,403
2,432
2,461
2,491
2,521
2,552
2,583
2,615
2,648
2,680
3-23
-------
Draft Regulatory Impact Analysis
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 in order to be consistent with the air quality
modeling region. The baseline cases represent current and future emissions with only the
existing standards. For nonroad engines, the baseline inventories were developed prior to
promulgation of standards that cover large spark-ignition engines (>25 hp), recreational
equipment, and recreational marine diesel engines (>50 hp).9 Although the future inventories
presented here do not account for the impact of the standards for those nonroad categories,
qualitative impacts of those standards on the inventories will be discussed. We intend to account
for the impact of these standards in the final rule analysis.
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
proposed 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 PM25, contributions of nonroad diesel engines to both total diesel PM25 and total
manmade PM2 5 are also presented.
The development of the 1996, 2020, and 2030 baseline emissions inventories for the
nonroad sector and for the sectors not affected by this proposed rule will be briefly described,
followed by discussions for each pollutant of the contribution of nonroad diesel engines to
national baseline inventories.
3.2.1 Baseline Emissions 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 emissions estimates for
all nonroad sources, with the exception of commercial marine engines, locomotives, and aircraft.
The methodology has been documented elsewhere.10
The on-highway estimates are based on the MOBILESb model, but with some further
adjustments to reflect MOBILE6 emission factors. The on-highway inventories are similar to
those prepared for the Heavy-Duty Diesel (HDD) rulemaking n, with the exception of
adjustments to NOX and VOC for California counties, based on county-level estimates from the
California Air Resources Board.
3-24
-------
Emissions Inventory
The stationary point and area source estimates are also based on the HDD rulemaking,
with the exception of adjustments to NOX and VOC for California counties, based on county-
level estimates from the California Air Resources Board. There were also some stack parameter
corrections made to the point source estimates.
The model inputs for the diesel nonroad sources have been described in detail in Section
3.1. Although county-level-based inventories were developed by Pechan for the land-based
diesel and recreational marine diesel categories, these were not used in this section. Instead, the
emission estimates for these categories were based on national level runs. 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). It was not possible to develop revised county-level estimates for these
categories due to resource and time constraints. Second, county-level estimates were only
developed for 2020 and 2030. Estimates for interim years are also needed to fully evaluate the
anticipated emission benefits of the proposed rule. Interim year estimates are generated using
national level model runs. In order to be consistent with other sections of the 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 a more detailed comparison of national level and county level results, see
Section 3.6.
3.2.2 PM2 5 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 diesels are 43% of the total diesel PM25
emissions in 1996, and this percentage increases to 64% by 2030. Emissions from land-based
nonroad diesels actually decrease from 176,510 tons in 1996 to 124,334 tons in 2020 due to the
existing emission standards. From 2020 to 2030, however, emissions increase to 139,527 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 proposal due to the fuel sulfur
requirements. For all nonroad diesel sources affected by this proposal, the contribution to total
diesel PM25 emissions increases from 57% in 1996 to 92% 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 diesels
are 8% of the total manmade PM25 emissions in 1996, and this percentage drops slightly to 6% in
2020 and 2030. The contribution of land-based diesels to total mobile source PM25 emissions is
32% in 1996, rising to 37% by 2030. For all nonroad diesel sources, the contribution to total
3-25
-------
Draft Regulatory Impact Analysis
manmade PM2 5 emissions is 11% in 1996, and this percentage drops slightly to 9% in 2020 and
2030.
The recently promulgated standards for large spark-ignition engines, recreational
equipment, and recreational marine diesel engines (>50 hp) include PM standards for the
recreational equipment and recreational marine diesel categories. PM2 5 emissions from
recreational equipment would be reduced roughly 50% by 2030, whereas PM2 5 emissions from
recreational marine diesel engines over 50 hp would be reduced roughly 25% by 2030 with these
standards. Since PM2 5 emissions from recreational equipment and recreational marine diesel
engines constitute less than 1% of the total emissions, the impact of these PM standards will have
a negligible effect on the inventories provided in Tables 3.2-1 and 3.2-2.
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 diesels are 6% of the total emissions in 1996,
and this percentage increases to 8% by 2030. The contribution of land-based diesels to total
mobile source NOX emissions is 12% in 1996, rising to 24% by 2030. Emissions from land-
based nonroad diesels actually decrease from 1,583,664 tons in 1996 to 1,140,727 tons in 2020
due to the existing emission standards. From 2020 to 2030, however, emissions increase to
1,231,995 tons, as growth in this sector offsets the effect of the existing emission standards.
NOX emissions from recreational marine diesel engines, commercial marine diesel
engines, and locomotives will not be affected by this proposal. For these categories combined,
the contribution to total NOX emissions remains stable at 9% from 1996 to 2030.
The recently promulgated standards for large spark-ignition engines, recreational
equipment, and recreational marine diesel engines (>50 hp) include NOX standards for the
recreational marine diesel and large spark-ignition categories. NOX emissions from recreational
marine diesel engines over 50 hp would be reduced roughly 25% by 2030, whereas NOX
emissions from large spark-ignition engines would be reduced roughly 90% by 2030 with these
standards. Although the contribution from these categories will decrease due to the standards,
the contribution of land-based diesel engines to the total NOX inventory remains stable at 8% in
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 diesels are 1% of the total emissions in 1996,
and this percentage increases to 2% by 2030. The contribution of land-based diesels to total
3-26
-------
Emissions Inventory
mobile source SO2 emissions is 20% in 1996, rising to 44% 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 proposal due to the fuel sulfur
requirements. For all nonroad diesel sources affected by this proposal, the contribution to total
SO2 emissions increases from 1% in 1996 to 3% in 2030.
The recently promulgated standards for large spark-ignition engines, recreational
equipment, and recreational marine diesel engines (>50 hp) do not impact SO2 emissions;
therefore, the SO2 emissions inventories presented in Table 3.2-4 are not affected by these
standards.
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 diesels are 1% of the total emissions in 1996,
and this percentage remains stable at 1% by 2030. The contribution of land-based diesels to total
mobile source VOC emissions is 3% in 1996, decreasing slightly to 2% by 2030. Emissions
from land-based nonroad diesels actually decrease from 221,403 tons in 1996 to 96,855 tons in
2020 due to the existing emission standards. From 2020 to 2030, however, emissions increase to
97,348 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 proposal. For these categories combined,
the contribution to total VOC emissions increases slightly from 1% from 1996 to 2% in 2030.
The recently promulgated standards for large spark-ignition engines, recreational
equipment, and recreational marine diesel engines (>50 hp) include VOC standards for each
category. VOC emissions from large spark-ignition engines would be reduced roughly 65% by
2030 with these standards. VOC emissions from recreational equipment would be reduced
roughly 70%, whereas VOC emissions from recreational marine diesel engines over 50 hp would
be reduced roughly 35% by 2030. Although the contribution from these categories will decrease
due to the standards, the contribution of land-based diesel engines to the total VOC inventory
remains stable at 1% in 2030.
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.
3-27
-------
Draft Regulatory Impact Analysis
CO emissions from land-based nonroad diesels are 1% of the total emissions in 1996, and
this percentage remains stable at 1% by 2030. The contribution of land-based diesels to total
mobile source CO emissions is also 1% in 1996, remaining at 1% by 2030. Emissions from
land-based nonroad diesels actually decrease from 1,010,518 tons in 1996 to 700,017 tons in
2020 due to the existing emission standards. From 2020 to 2030, however, emissions increase to
793,923 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 proposal. For these categories combined, the
contribution to total CO emissions is less than 1% in 1996 and 2030.
The recently promulgated standards for large spark-ignition engines, recreational
equipment, and recreational marine diesel engines (>50 hp) include CO standards for the large
spark-ignition and recreational equipment categories. CO emissions from large spark-ignition
engines would be reduced roughly 90% by 2030 with these standards, whereas CO emissions
from recreational equipment would be reduced roughly 20% by 2030. Although the contribution
from these categories will decrease due to the standards, the contribution of land-based diesel
engines to the total CO inventory remains stable at 1% in 2030.
3-28
-------
Emissions Inventory
Table 3.2-1
Annual Diesel PM2S Baseline Emission Levels for Mobile and Other Source Categories"
Category
Land-Based Nonroad
Diesel
Recreational Marine
Diesel < 50 hp
Recreational Marine
Diesel >50 hp b
Commercial Marine
Diesel
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
176,510
62
425
36,367
20,937
234,301
167,384
401,685
12,199
413,884
97%
%of
mobile
source
43.9%
0.0%
0.1%
9.1%
5.2%
58%
42%
100%
—
—
—
%of
total
42.6%
0.0%
0.1%
8.8%
5.1%
57%
40%
97%
3%
2020
short
tons
124,334
70
753
41,365
16,727
183,249
18,426
201,675
4,010
205,685
98%
%of
mobile
source
61.7%
0.0%
0.4%
20.5%
8.3%
91%
9%
100%
—
—
—
%of
total
60.4%
0.0%
0.4%
20.1%
8.1%
89%
9%
98%
2%
2030
short
tons
139,527
64
894
45,411
15,670
201,566
13,948
215,514
4,231
219,745
98%
%of
mobile
source
64.7%
0.0%
0.4%
21.1%
7.3%
94%
6%
100%
—
—
—
%of
total
63.5%
0.0%
0.4%
20.7%
7.1%
92%
6%
98%
2%
a These are 48-state inventories. They do not include Alaska and Hawaii.
b These inventories do not account for the final rule to control emissions from nonroad large spark-ignition engines,
recreational marine diesel engines >50 hp, and recreational vehicles, published November 8, 2002.
0 This category includes point sources burning either diesel, distillate oil (diesel), or diesel/kerosene fuel.
5-29
-------
Draft Regulatory Impact Analysis
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 c
Recreational
Marine SI
Nonroad SI <25 hp
Nonroad SI >25hp c
Recreational SI °
Commercial Marine
Diesel
Commercial
Marine SI
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
176,510
62
425
35,147
24,130
1,370
4,632
36,367
1,370
20,937
27,891
328,841
230,684
559,525
1,653,392
2,212,917
25%
%of
mobile
source
31.5%
0.0%
0.1%
6.3%
4.3%
0.2%
0.8%
6.5%
0.2%
3.7%
5.0%
59%
41%
100%
—
—
—
%of
total
8.0%
0.0%
0.0%
1.6%
1.1%
0.1%
0.2%
1.6%
0.1%
1.0%
1.3%
15%
10%
25%
75%
2020
short tons
124,334
70
753
26,110
29,998
2,297
5,557
41,365
1,326
16,727
30,024
278,561
72,377
350,938
1,712,004
2,062,942
17%
%of
mobile
source
35.4%
0.0%
0.2%
7.4%
8.5%
0.6%
1.6%
11.8%
0.4%
4.8%
8.6%
79%
21%
100%
—
—
—
%of
total
6.0%
0.0%
0.0%
1.3%
1.5%
0.1%
0.3%
2.0%
0.1%
0.8%
1.5%
14%
4%
17%
83%
2030
short tons
139,527
64
894
27,223
34,435
2,687
5,912
45,411
1,427
15,670
30,606
303,856
75,825
379,681
1,824,609
2,204,290
17%
%of
mobile
source
36.7%
0.0%
0.2%
7.2%
9.1%
0.7%
1.6%
12.0%
0.4%
4.1%
8.1%
80%
20%
100%
—
—
—
%of
total
6.3%
0.0%
0.0%
1.2%
1.6%
0.1%
0.3%
2.1%
0.1%
0.7%
1.4%
14%
3%
17%
83%
25%
a These are 48-state inventories. They do not include Alaska and Hawaii.
b Excludes natural and miscellaneous sources.
0 These inventories do not account for the final rule to control emissions from nonroad large spark-ignition engines,
recreational marine diesel engines >50 hp, and recreational vehicles, published November 8, 2002.
-------
Emissions Inventory
Table 3.2-3
Annual NOX 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 b
Recreational
Marine SI
Nonroad SI <25 hp
Nonroad SI >25hp b
Recreational SI b
Commercial Marine
Diesel
Commercial
Marine SI
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
1,583,664
523
18,917
33,304
63,584
281,068
8,606
959,704
6,428
921,556
165,018
4,042,371
9,066,489
13,108,860
11,449,752
24,558,612
53%
%of
mobile
source
12.1%
0.0%
0.1%
0.3%
0.5%
2.1%
0.1%
7.3%
0.0%
7.0%
1.3%
31%
69%
100%
—
—
—
%of
total
6.4%
0.0%
0.1%
0.1%
0.3%
1.1%
0.0%
3.9%
0.0%
3.8%
0.7%
17%
37%
53%
47%
2020
short tons
1,140,727
682
34,136
61,749
100,119
484,504
13,065
819,201
4,551
612,722
228,851
3,500,307
1,984,611
5,484,917
10,050,213
15,535,130
35%
%of
mobile
source
20.8%
0.0%
0.6%
1.1%
1.8%
8.8%
0.2%
14.9%
0.1%
11.2%
4.2%
64%
36%
100%
—
—
—
%of
total
7.3%
0.0%
0.2%
0.4%
0.6%
3.1%
0.1%
5.3%
0.0%
3.9%
1.5%
22%
13%
35%
65%
2030
short tons
1,231,995
706
40,544
67,893
116,514
567,696
13,539
814,827
4,355
534,520
258,102
3,650,691
1,577,788
5,228,479
10,320,361
15,548,840
34%
%of
mobile
source
23.6%
0.0%
0.8%
1.3%
2.2%
10.9%
0.3%
15.6%
0.1%
10.2%
4.9%
70%
30%
100%
—
—
—
%of
total
7.9%
0.0%
0.3%
0.4%
0.7%
3.7%
0.1%
5.2%
0.0%
3.4%
1.7%
24%
10%
34%
66%
a These are 48-state inventories. They do not include Alaska and Hawaii.
b These inventories do not account for the final rule to control emissions from nonroad large spark-ignition engines,
recreational marine diesel engines >50 hp, and recreational vehicles, published November 8, 2002.
-------
Draft Regulatory Impact Analysis
Table 3.2-4
Annual SO2 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
Commercial
Marine SI
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
147,926
57
2,194
2,170
6,530
882
1,673
25,948
191,813
50,534
11,305
441,032
302,938
743,970
17,636,602
18,380,572
4%
%of
mobile
source
19.9%
0.0%
0.3%
0.3%
0.9%
0.1%
0.2%
3.5%
25.8%
6.8%
1.5%
59%
41%
100%
—
—
—
%of
total
0.8%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.1%
1.0%
0.3%
0.1%
2%
2%
4%
96%
2020
short tons
252,089
100
3,803
2,522
8,347
1,060
2,679
32,117
196,918
53,832
15,267
568,734
35,311
604,045
14,510,426
15,114,471
4%
%of
mobile
source
41.7%
0.0%
0.6%
0.4%
1.4%
0.2%
0.4%
5.3%
32.6%
8.9%
2.5%
94%
6%
100%
—
—
—
%of
total
1.7%
0.0%
0.0%
0.0%
0.1%
0.0%
0.0%
0.2%
1.3%
0.4%
0.1%
4%
0%
4%
96%
2030
short tons
297,573
119
4,517
2,698
9,714
1,211
2,774
36,068
210,060
58,832
16,813
640,379
40,788
681,167
14,782,220
15,463,387
4%
%of
mobile
source
43.7%
0.0%
0.7%
0.4%
1.4%
0.2%
0.4%
5.3%
30.8%
8.6%
2.5%
94%
6%
100%
—
—
—
%of
total
1.9%
0.0%
0.0%
0.0%
0.1%
0.0%
0.0%
0.2%
1.4%
0.4%
0.1%
4%
0%
4%
96%
' These are 48-state inventories. They do not include Alaska and Hawaii.
-------
Emissions Inventory
Table 3.2-5
Annual VOC Baseline Emission Levels for Mobile and Other Source Categories
Category
Land-Based Nonroad
Diesel
Recreational Marine
Diesel <50 hp
Recreational Marine
Diesel >50 hp b
Recreational
Marine SI
Nonroad SI <25 hp
Nonroad SI >25hp b
Recreational SI b
Commercial Marine
Diesel
Commercial
Marine SI
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
221,403
128
676
804,488
1,330,229
44,926
403,984
31,545
960
48,381
176,394
3,063,114
5,286,948
8,350,062
10,249,136
18,599,198
45%
%of
mobile
source
2.7%
0.0%
0.0%
9.6%
15.9%
0.5%
4.8%
0.4%
0.0%
0.6%
2.1%
37%
63%
100%
—
—
—
%of
total
1.2%
0.0%
0.0%
4.3%
7.2%
0.2%
2.2%
0.2%
0.0%
0.3%
0.9%
17%
28%
45%
55%
2020
short tons
96,855
108
1,219
380,891
650,158
42,504
719,031
37,290
998
36,546
239,654
2,205,255
2,055,843
4,261,098
9,648,376
13,909,474
31%
%of
mobile
source
2.3%
0.0%
0.0%
8.9%
15.3%
1.0%
16.9%
0.9%
0.0%
0.9%
5.6%
52%
48%
100%
—
—
—
%of
total
0.7%
0.0%
0.0%
2.7%
4.7%
0.3%
5.2%
0.3%
0.0%
0.3%
1.7%
16%
15%
31%
69%
2030
short tons
97,348
80
1,448
372,970
751,883
47,411
749,134
41,354
1,079
31,644
265,561
2,359,912
2,296,972
4,656,884
10,751,134
15,408,018
30%
%of
mobile
source
2.1%
0.0%
0.0%
8.0%
16.1%
1.0%
16.1%
0.9%
0.0%
0.7%
5.7%
51%
49%
100%
—
—
—
%of
total
0.6%
0.0%
0.0%
2.4%
4.9%
0.3%
4.9%
0.3%
0.0%
0.2%
1.7%
15%
15%
30%
70%
a These are 48-state inventories. They do not include Alaska and Hawaii.
b These inventories do not account for the final rule to control emissions from nonroad large spark-ignition engines,
recreational marine diesel engines >50 hp, and recreational vehicles, published November 8, 2002.
-------
Draft Regulatory Impact Analysis
Table 3.2-6
Annual CO Baseline Emission Levels for Mobile and Other Source Categories
Category
Land-Based Nonroad
Diesel
Recreational Marine
Diesel <50 hp
Recreational Marine
Diesel >50 hp b
Recreational
Marine SI
Nonroad SI <25 hp
Nonroad SI >25hp b
Recreational SI b
Commercial Marine
Diesel
Commercial
Marine SI
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
1,010,518
365
2,850
1,995,907
16,735,812
2,144,654
1,824,753
126,382
6,010
112,171
949,313
24,908,737
53,585,364
78,494,101
16,318,451
94,812,552
83%
%of
mobile
source
1.3%
0.0%
0.0%
2.5%
21.3%
2.7%
2.3%
0.2%
0.0%
0.1%
1.2%
32%
68%
100%
—
—
—
%of
total
1.1%
0.0%
0.0%
2.1%
17.7
%
2.3%
1.9%
0.1%
0.0%
0.1%
1.0%
26%
56%
83%
17%
2020
short tons
700,017
395
5,143
1,977,403
24,675,763
2,785,383
2,765,874
159,900
6,702
119,302
1,387,178
34,583,061
48,333,986
82,917,047
15,648,555
98,565,602
84%
%of
mobile
sources
0.8%
0.0%
0.0%
2.4%
29.8%
3.4%
3.3%
0.2%
0.0%
0.1%
1.7%
42%
58%
100%
—
—
—
%of
total
0.7%
0.0%
0.0%
2.0%
25.0%
2.8%
2.8%
0.2%
0.0%
0.1%
1.4%
35%
49%
84%
16%
2030
short tons
793,923
356
6,109
2,075,666
28,728,492
3,198,141
2,891,759
176,533
7,233
119,302
1,502,265
39,499,779
55,609,767
95,109,546
16,325,306
111,434,852
85%
%of
mobile
source
0.8%
0.0%
0.0%
2.2%
30.2%
3.4%
3.0%
0.2%
0.0%
0.1%
1.6%
42%
58%
100%
—
—
—
%of
total
0.7%
0.0%
0.0%
1.9%
25.8
%
2.9%
2.6%
0.2%
0.0%
0.1%
1.3%
35%
50%
85%
15%
a These are 48-state inventories. They do not include Alaska and Hawaii.
b These inventories do not account for the final rule to control emissions from nonroad large spark-ignition engines,
recreational marine diesel engines >50 hp, and recreational vehicles, published November 8, 2002.
3-34
-------
Emissions Inventory
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. The EPA selected a collection of typical cities
spread across the United States in order 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 PM25 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.
A
Construction, industrial, and commercial nonroad diesel equipment comprise most of the land-based nonroad
emissions inventory. These types of equipment are more concentrated in urban areas where construction projects,
manufacturing, and commercial operations are prevalent.
3-35
-------
Draft Regulatory Impact Analysis
Table 3.3-1
Land-Based Nonroad Percent Contribution
to PM9, 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)
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%
' Includes only direct exhaust emissions; see Chapter 2 for a discussion of secondary fine PM levels.
3-36
-------
Emissions Inventory
Table 3.3-2
Annual Land-Based Nonroad Diesel Contributions
to PM9, 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)
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%
' Includes only direct exhaust emissions; see Chapter 2 for a discussion of secondary fine PM levels.
3-37
-------
Draft Regulatory Impact Analysis
Table 3.3-3
Land-Based Nonroad Percent Contribution
to PM9, Inventories in Selected Urban Areas in 2030a
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
777
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%
' Includes only direct exhaust emissions; see Chapter 2 for a discussion of secondary fine PM levels.
3.3.2 NOX 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.
3-38
-------
Emissions Inventory
Table 3.3-4
Land-Based Nonroad Percent Contribution
to NOV Inventories in Selected Urban Areas in 1996
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%
3-39
-------
Draft Regulatory Impact Analysis
Table 3.3-5
Annual Land-Based Nonroad Diesel Contributions
to NOV Inventories in Selected Urban Areas in 2020
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%
3-40
-------
Emissions Inventory
Table 3.3-6
Land-Based Nonroad Percent Contribution
to NOV Inventories in Selected Urban Areas in 2030
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%
3.4 Nonroad Diesel Controlled Emissions Inventory Development
This section describes how the controlled emissions inventories were developed for the
four categories of nonroad diesel engines affected by this proposal: 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 emissions development.
3.4.1 Land-Based Diesel Engines—PM25, NOX, SO2, VOC, and CO Emissions
The emission inventory estimates used in this proposed rule were generated using the
draft NONROAD2002 model with certain input modifications to account for the in-use diesel
fuel sulfur reductions and the additional controls being proposed for the Tier 4 engines. This
section will only describe these modifications to the model inputs, since the other aspects of the
3-41
-------
model, including inputs for earlier engines, are covered in detail in the technical reports that
document the draft NONROAD2002 model.
3.4.1.1 Standards and Zero-Hour Emission Factors
The proposed standards that are presented in Section 3 of the preamble are shown in
Table 3.4-1. The modeled emission factors corresponding to the proposed standards are shown
in Table 3.4-2. These emission factors are derived from the standards by applying an assumed
8% compliance margin to the standard. This compliance margin was derived from data for
highway diesel vehicles and used in the HD07 rule. Additionally, a transient adjustment factor is
applied, as described below, if the engine power and model year place it in a category subject to a
steady-state certification test cycle instead of a transient test.
Besides exhaust emissions, the proposed rule includes changes in crankcase hydrocarbon
emissions. Crankcase losses prior to 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
As shown in Table 3.4-2, the proposed new standards for engines over 75 hp beginning in
2011 or 2012, and for those under 75 hp beginning in 2008, call for use of a transient
certification test cycle. Thus, there was no Transient Adjustment Factor (TAP) applied to the
emission factors for these engines (i.e., the model applies a TAP of 1.0); the zero-hour emission
factor was modeled simply as the value of the standard minus an assumed 8% compliance
margin.
3-42
-------
Table 3.4-1
Proposed Tier 4 Exhaust Emissions Certification Standards
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)
Emissions Standard
(g/bhp-hr)a
transitional
or final
final
transitional4
final
transitional
final
transitional
final
transitional
final
PM
0.30
0.22
0.02
0.01
(100%)
0.01
(100%)
0.01
(100%)
0.01
(100%)
0.01
(50%)
0.01
(100%)
NOX
NMHC
5.6b'c
5.6/3. 5b'c
3.5b
0.30
(50%)
0.30
(100%)
0.30
(50%)
0.30
(100%)
0.30
(50%)
0.30
(100%)
0.14
(50%)
0.14
(100%)
0.14
(50%)
0.14
(100%)
0.14
(50%)
0.14
(100%)
CO
4.9
3.7
3.7 c
3.7 c
3.7 c
2.6 c
2.6 c
2.6 c
2.6 c
Model
Year
2008
2008
2013
2012-2013
2014
2011-2013
2014
2011-2013
2014
a Percentages are model year sales fractions required to comply with the indicated standard.
b This is a combined NMHC + NOX standard.
0 This emissions 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/bhp-hr for engines below 50 hp and 3.5 g/bhp-hr for
engines at or above 50 hp.
d Manufacturers may optionally skip the transitional standards for 25-75 hp engines; the final standards would then take
effect for these engines in the 2012 model year.
3-43
-------
Draft Regulatory Impact Analysis
Table 3.4-2
NONROAD Model EF Inputs for Proposed Tier 4 Exhaust Emissions Standards
Engine
Power
hp< 11
1 1< hp < 25
25 < hp < 50
50 < hp < 75
75 750
Emission Factor Modeling Inputs
g/bhp-hr
Type of standard
final e
final e
transitional f
final
transitional f
final
transitional
final
transitional
final
transitional
final
transitional
final
transitional
final
trans- 50%
itional8
50%
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
0.13
0.01
0.01
NOxab
4.
4.
4.
3
3
3
3.0 (50%)
0.
2.5 (50%)
0.
2.5 (50%)
0.
2.5 (50%)
0.
2.5 (50%)
0.
4
0.
0.
30
44
73
.0
.0
.0
0.28 (50%)
28
0.28 (50%)
28
0.28 (50%)
28
0.28 (50%)
28
0.28 (50%)
28
.1
28
30
THC b'c
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
0.17
0.13
0.13
cod
4.11
2.16
1.53
0.15
2.4
0.24
0.24
0.24
0.87
0.087
7.5
0.075
8.4
0.084
1.3
0.13
0.76
0.076
0.076
Model
Year
2008
2008
2008
2013
2008
2013
2012-2013
2014
2012-2013
2014
2011-2013
2014
2011-2013
2014
2011-2013
2014
2011-2013
2014
a Percentages are model-year sales fractions required to comply with the indicated standard.
b 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.
0 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.
d Tier 4 CO is assumed to decrease by 90% from its prior levels in any cases where particulate traps are expected for PM control.
" Final standards and emission factor inputs for engines under 25 hp take effect in 2008, starting in 2008 the modeling of these inputs changes to
reflect the start of a transient certification test requirement at which time Transient Adjustment Factors are no longer applied to the
emission factors.
f Transitional standards and emission factor inputs for 25-75 hp engines are based on transient use, so Transient Adjustment Factors will not be
applied to the emission factors shown here.
s The transitional standards for engines >750 hp consist of 50% engines meeting Tier 2 standards with the 8-mode test and 50% meeting the final
Tier 4 standards with transient test. TAFs will only get applied to the emissions of the engines meeting the Tier 2 standards. Application
of TAF's is described in technical report NR-009b.'
3-44
-------
Emissions Inventory
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-OOQb.1
Table 3.4-3
Deterioration Rates for Nonroad Diesel Engines
Pollutant
HC
CO
NOX
PM
Relative Deterioration Rate (% increase per %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 was assumed for modeling the engines that would be 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 proposed 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/bhp-hr (30% conversion instead of
2.247% that is used for all earlier non-trap equipped engines).
3-45
-------
Draft Regulatory Impact Analysis
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
(175 560
(hp > 750)
Standards
Tier 2
transitional
final
Tier 3 transitional a
final
Tier 3
final
Tier 3
final
Tier 2
transitional b
final
Modeled Certification Fuel
Sulfur Content, PPM
2000
500
15
500
15
2000
15
2000
15
2000
50% 2000
50% 15
15
Model
Year
through 2007
2008
2013
2008-2011
2012
2007-2011
2012
2006-2010
2011
2006-2010
2011-2013
2014
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 would allow manufacturers to take advantage of
the availability of 500 ppm fuel that year.
b The engines remaining at the Tier 2 level would be allowed to continue certifying on the same fuel as earlier Tier 2
engines, but those meeting the Tier 4 0.01 PM standard are assumed to certify on 15 ppm fuel.
3-46
-------
Emissions Inventory
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
Standards
Baseline
June intro of 500 ppm
500 ppm standard
June intro of 1 5 ppm
Final 1 5 ppm standard
Baseline
June intro of 500 ppm
Final 500 ppm standard
Modeled In-Use Fuel
Sulfur Content, ppm
2318
2271
1075
245
100
11
2396
2352
1114
252
233
Calendar
Year
through 2005
2006
2007
2008-2009
2010
2011+
through 2005
2006
2007
2008-2009
2010+
3.4.1.5 Modeling 50-75 hp and 75-100 hp Within the NONROAD 50-100 hp Bin
The proposed standards call for different treatment of diesel engines above and below 75
hp (56 kW), but the NONROAD model is not currently designed to handle a 75 hp cutpoint
within its 50-100 hp bin. Thus, a modeling method was used in which the NONROAD model
was run twice for each scenario - one time applying the 50-75 hp standards to the 50-100 hp bin,
and one time applying the 75-100 hp standards to that bin. Then a weighted average of the two
sets of emission inventory outputs was calculated, with the weighting based on overall diesel
population and horsepower within the 50-100 hp range. The population weighting was essentially
50/50 (half 50-75 hp and half 75-100 hp), but when the average hp of these two power sub-
ranges is taken into account, the resulting inventory weighting was 57% for the 75-100 hp
outputs and 43% for the 50-75 hp outputs.
The engine population and power data that was used to calculate this weighting was
based on detailed sales data from PSR12 as described in technical report NR-006b, "Nonroad
Engine Population Estimates."
3.4.1.6 Controlled Inventory
Tables 3.4-6a and 3.4-6b present the PM10, PM2 5, 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.
3-47
-------
Draft Regulatory Impact Analysis
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
191,858
175,155
169,360
163,684
157,726
152,310
147,050
142,043
130,006
120,783
117,672
113,732
108,633
101,846
94,540
87,051
79,578
72,412
65,636
59,412
53,834
48,976
44,686
40,803
37,312
34,073
31,098
28,431
26,095
24,198
22,486
20,912
19,617
18,444
17,347
16,334
15,427
14,653
13,942
13,548
13,236
12.957
FM,<
176,510
161,143
155,811
150,589
145,108
140,125
135,286
130,680
119,606
111,120
108,258
104,633
99,943
93,698
86,976
80,086
73,211
66,619
60,385
54,659
49,527
45,057
41,111
37,539
34,327
31,347
28,610
26,156
24,008
22,262
20,687
19,239
18,048
16,968
15,959
15,027
14,193
13,481
12,827
12,464
12,177
11.921
NOT
1,583,664
1,569,903
1,556,973
1,544,395
1,522,881
1,503,228
1,483,942
1,450,762
1,414,673
1,373,870
1,331,368
1,290,526
1,234,897
1,173,275
1,114,569
1,030,363
950,060
874,829
804,895
742,607
686,592
637,025
595,511
560,026
529,072
502,436
479,114
459,646
443,460
429,909
418,492
410,084
403,630
398,093
393,543
389,845
387,098
385,165
383,784
383,422
383,769
384.422
SO,
147,926
167,094
171,957
176,819
181,677
186,532
191,385
192,228
93,229
21,757
22,267
9,297
1,032
1,032
1,027
1,021
1,014
1,009
1,005
1,003
1,003
1,005
1,009
1,015
1,022
1,029
1,038
1,047
1,058
1,070
1,083
1,096
1,109
1,123
1,137
1,151
1,166
1,180
1,195
1,210
1,226
1.241
VOC
221,403
200,366
191,785
183,584
176,201
169,541
163,193
156,295
149,518
142,282
135,201
128,301
121,298
114,587
108,169
102,315
97,013
92,325
88,273
84,706
81,526
78,822
76,555
74,606
72,984
71,635
70,520
69,626
68,961
68,470
68,103
67,861
67,795
67,826
67,919
68,080
68,309
68,617
68,967
69,432
69,944
70.495
CO
1,010,518
923,887
886,723
850,751
817,858
790,468
764,918
742,184
724,213
709,158
696,085
684,775
661,317
625,141
580,076
534,590
490,424
449,037
411,280
377,046
345,972
318,530
294,889
273,610
254,823
237,880
222,570
209,119
197,634
188,366
180,519
173,579
168,098
163,618
159,577
156,020
153,038
150,653
148,547
146,684
145,351
144.359
3-48
-------
Emissions Inventory
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 PM7, NOT SO, VOC CO
192,750
175,981
170,165
164,467
158,487
153,045
147,761
142,732
130,636
121,368
118,245
114,290
109,168
102,347
95,004
87,475
79,963
72,758
65,944
59,687
54,078
49,194
44,881
40,979
37,472
34,220
31,232
28,555
26,210
24,304
22,585
21,004
19,704
18,526
17,426
16,409
15,500
14,722
14,009
13,613
13,300
13.020
177,330
161,903
156,552
151,310
145,808
140,802
135,940
131,314
120,185
111,658
108,786
105,147
100,435
94,159
87,403
80,477
73,566
66,937
60,669
54,912
49,752
45,258
41,290
37,701
34,474
31,482
28,734
26,270
24,113
22,360
20,778
19,323
18,128
17,044
16,032
15,096
14,260
13,544
12,888
12,524
12,236
11.979
1,592,025
1,578,148
1,565,144
1,552,490
1,530,854
1,511,087
1,491,692
1,458,315
1,422,017
1,380,984
1,338,243
1,297,178
1,241,223
1,179,237
1,120,193
1,035,489
954,717
879,052
808,721
746,089
689,768
639,935
598,209
562,551
531,454
504,702
481,283
461,742
445,499
431,900
420,447
412,011
405,538
399,987
395,428
391,724
388,976
387,045
385,668
385,312
385,667
386.330
148,729
167,999
172,889
177,777
182,662
187,544
192,424
193,272
93,735
21,875
22,388
9,347
1,038
1,038
1,033
1,026
1,020
1,014
1,010
1,009
1,009
1,011
1,015
1,020
1,027
1,035
1,044
1,053
1,064
1,076
1,089
1,102
1,115
1,129
1,143
1,157
1,172
1,187
1,202
1,217
1,232
1.248
222,517
201,386
192,765
184,524
177,107
170,414
164,035
157,104
150,293
143,022
135,906
128,973
121,935
115,188
108,735
102,849
97,519
92,806
88,733
85,149
81,954
79,238
76,959
75,001
73,371
72,015
70,894
69,996
69,328
68,834
68,466
68,223
68,157
68,189
68,282
68,444
68,676
68,985
69,338
69,805
70,321
70.875
1,015,773
928,674
891,304
855,132
822,062
794,522
768,838
745,994
727,946
712,836
699,719
688,377
664,805
628,436
583,127
537,388
492,974
451,351
413,376
378,949
347,705
320,115
296,347
274,958
256,081
239,060
223,684
210,176
198,641
189,329
181,441
174,465
168,958
164,459
160,400
156,827
153,833
151,436
149,320
147,450
146,113
145.119
3-49
-------
Draft Regulatory Impact Analysis
3.4.2 Land-Based Diesel Engines—Air Toxics Emissions
Since air toxics emissions are part of the VOC emissions inventory, NMHC standards
being proposed in this rule would 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 proposed rule. The EPA uses the
same fractions used to calculate the base air toxic emissions without the proposed rule (see
section 3.1.2), along with the estimated VOC emissions resulting from the proposed rule, to
calculate the air toxics emissions resulting from the proposed 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
4,007
3,264
2,990
2,556
1,940
1,576
1,410
1,357
Formaldehyde
23,643
19,257
17,643
15,140
11,448
9,301
8,321
8,008
Acetaldehyde
10,619
8,649
7,924
6,800
5,142
4,178
3,738
3,597
1,3 -Butadiene
401
326
299
257
194
158
141
136
Acrolein
601
490
449
385
291
236
212
204
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
4,028
3,281
3,006
2,579
1,950
1,585
1,418
1,364
Formaldehyde
23,764
19,356
17,735
15,219
11,507
9,350
8,366
8,050
Acetaldehyde
10,673
8,694
7,966
6,836
5,168
4,200
3,757
3,616
1,3 -Butadiene
403
328
301
258
195
158
142
136
Acrolein
604
492
451
387
293
238
213
205
3-50
-------
Emissions Inventory
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 are being proposed for these
engines. However, due to the diesel fuel sulfur changes that are being proposed, 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 essentially
the same as 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 would use
diesel fuel meeting a 500 ppm sulfur standard beginning in June 2007. This was modeled as 340
ppm sulfur outside of California and 120 ppm in California, based on available fuel survey data
for in-use highway fuel relative to the existing 500 ppm highway diesel fuel sulfur standards.
Additional sulfur adjustments were made to account for the "spillover" of low sulfur highway
fuel meeting a 15 ppm standard in the applicable years prior to the start of the proposed 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 prior to their use of aftertreatment (2.247%). The fuel sulfur levels were
calculated as weighted average in-use levels of (a) uncontrolled nonroad diesel fuel at 3400 ppm
sulfur, (b) controlled locomotive and marine diesel fuel at 340 ppm, (c) "spillover" of low sulfur
highway diesel fuel into use by nonroad applications outside of California, and (d) full use of low
sulfur California fuel in all nonroad applications in California. 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 5 months at the higher sulfur level of the prior year
plus 7 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.
3-51
-------
Draft Regulatory Impact Analysis
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)
1114
252
252
233
233
233
233
233
233
233
233
233
233
233
233
233
233
233
233
233
233
233
233
233
233
233
233
233
233
233
233
233
233
233
Control
SO2
Loco
(tons/yr)
24,051
5,457
5,495
5,119
5,205
5,230
5,245
5,260
5,285
5,339
5,370
5,399
5,430
5,449
5,498
5,547
5,597
5,647
5,698
5,749
5,801
5,853
5,905
5,959
6,012
6,065
6,119
6,174
6,229
6,285
6,341
6,397
6,454
6.512
CMV (tons/yr)
13,446
3,071
3,105
2,904
2,936
2,969
3,002
3,036
3,071
3,106
3,141
3,177
3,214
3,251
3,288
3,326
3,365
3,404
3,444
3,485
3,526
3,567
3,610
3,653
3,697
3,741
3,786
3,832
3,878
3,926
3,974
4,023
4,072
4.123
Sulfate PM
Loco
(tons/yr)
1,935
439
442
412
419
421
422
423
425
430
432
434
437
438
442
446
450
454
458
463
467
471
475
479
484
488
492
497
501
506
510
515
519
524
CMV
(tons/yr)
1,082
247
250
234
236
239
242
244
247
250
253
256
259
262
265
268
271
274
277
280
284
287
290
294
297
301
305
308
312
316
320
324
328
332
Total PM2 5
Loco (tons/yr)
17,612
15,673
15,462
14,932
15,186
14,927
14,639
14,350
14,083
13,893
13,636
13,707
13,444
13,147
12,920
12,687
12,801
12,561
12,316
12,066
12,174
11,916
11,652
11,756
11,485
11,208
11,308
11,022
11,120
11,219
10,923
11,020
10,714
10.810
CMV (tons/yr)
37,523
36,944
37,136
37,310
37,502
37,693
37,885
38,077
38,269
38,460
38,652
38,844
39,036
39,228
39,606
39,985
40,363
40,741
41,119
41,498
41,876
42,255
42,633
43,012
43,390
43,769
44,148
44,527
44,905
45,284
45,663
46,042
46,421
46.801
3-52
-------
Emissions Inventory
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)
1114
252
252
233
233
233
233
233
233
233
233
233
233
233
233
233
233
233
233
233
233
233
233
233
233
233
233
233
233
233
233
233
233
233
Control
SO2
Loco
(tons/yr)
24,084
5,465
5,503
5,126
5,213
5,237
5,252
5,268
5,292
5,347
5,378
5,407
5,438
5,456
5,505
5,555
5,605
5,655
5,706
5,757
5,809
5,861
5,914
5,967
6,020
6,074
6,128
6,183
6,238
6,293
6,350
6,406
6,463
6.521
CMV (tons/yr)
14,145
3,231
3,267
3,055
3,088
3,123
3,158
3,194
3,231
3,267
3,304
3,342
3,381
3,420
3,459
3,499
3,540
3,581
3,623
3,666
3,709
3,753
3,798
3,843
3,889
3,935
3,983
4,031
4,080
4,130
4,180
4,232
4,284
4.337
Sulfate PM
Loco
(tons/yr)
1,938
440
443
412
419
421
423
424
426
430
433
435
437
439
443
447
451
455
459
463
467
472
476
480
484
489
493
497
502
506
511
515
520
525
CMV
(tons/yr)
1,138
260
263
246
248
251
254
257
260
263
266
269
272
275
278
282
285
288
291
295
298
302
306
309
313
317
320
324
328
332
336
340
345
349
Total PM2 5
Loco (tons/yr)
17,637
15,695
15,484
14,952
15,207
14,947
14,659
14,369
14,103
13,912
13,655
13,726
13,462
13,165
12,938
12,705
12,818
12,579
12,333
12,083
12,191
11,932
11,669
11,773
11,501
11,223
11,323
11,037
11,136
11,235
10,938
11,035
10,729
10.825
CMV (tons/yr)
39,473
38,864
39,066
39,249
39,451
39,653
39,854
40,056
40,258
40,460
40,662
40,863
41,065
41,267
41,665
42,063
42,461
42,859
43,257
43,655
44,053
44,451
44,849
45,248
45,646
46,044
46,443
46,841
47,240
47,638
48,037
48,436
48,834
49.233
3.4.4 Recreational Marine Engines
Even though this proposed rule does not include any emission standards for marine
engines, there are PM and SO2 benefits associated with these engines due to the proposed fuel
sulfur standards. The emission inventory estimates presented in Tables 3.4-9a and 3.4-9b assume
3-53
-------
Draft Regulatory Impact Analysis
that recreational marine applications would use diesel fuel meeting the same standards as
locomotive and commercial marine diesel fuel, which means an in-use sulfur content of 1114
ppm in the 2007 transition year and 232 ppm in 2010 and later as shown in Table 3.4-5.
Consistent with the baseline inventory described above, these inventory values do not include the
benefits associated with the standards promulgated in September 2002 for diesel recreational
marine engines.
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
PM,n
529
594
611
627
643
660
676
688
576
497
507
516
526
535
545
555
565
574
584
593
603
612
621
631
640
649
658
667
677
686
696
706
716
726
736
746
757
767
111
788
798
809
PM,,
487
547
562
577
592
607
622
633
530
457
467
474
484
493
502
511
520
528
537
546
555
563
572
580
589
597
605
614
622
631
640
650
659
668
677
687
696
705
715
725
734
744
NOT
19,440
21,899
22,548
23,196
23,844
24,492
25,139
25,790
26,439
27,088
27,736
28,384
29,028
29,671
30,314
30,957
31,600
32,244
32,888
33,531
34,174
34,817
35,460
36,103
36,746
37,388
38,031
38,673
39,316
39,959
40,604
41,250
41,896
42,543
43,189
43,836
44,483
45,131
45,779
46,428
47,076
47.725
SO,
2,251
2,537
2,613
2,689
2,765
2,841
2,917
2,939
1,428
331
339
321
328
336
343
351
358
365
373
380
388
395
402
410
417
425
432
440
447
454
462
469
477
484
491
499
506
514
521
529
536
543
voc
803
900
923
947
970
992
1,015
1,037
1,059
1,081
1,102
1,124
1,145
1,166
1,186
1,207
1,227
1,247
1,268
1,288
1,308
1,328
1,347
1,367
1,387
1,406
1,426
1,446
1,465
1,486
1,507
1,528
1,550
1,571
1,592
1,614
1,636
1,658
1,680
1,703
1,725
1.748
CO
3,215
3,613
3,713
3,814
3,913
4,013
4,112
4,211
4,309
4,406
4,503
4,599
4,695
4,790
4,884
4,979
5,072
5,166
5,260
5,353
5,445
5,538
5,630
5,722
5,814
5,906
5,997
6,089
6,181
6,275
6,370
6,465
6,561
6,656
6,752
6,848
6,945
7,041
7,139
7,238
7,336
7.435
3-54
-------
Emissions Inventory
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
532
598
615
631
648
664
680
692
579
500
511
519
529
539
549
559
568
578
588
597
607
616
625
635
644
653
662
671
681
690
700
710
721
731
741
751
761
772
782
793
803
814
PM,,
490
550
566
581
596
611
626
637
533
460
470
477
487
496
505
514
523
532
541
549
558
567
575
584
592
601
609
618
626
635
644
654
663
672
682
691
700
710
720
729
739
749
NOT
19,562
22,036
22,689
23,342
23,994
24,646
25,297
25,952
26,605
27,258
27,911
28,563
29,210
29,858
30,505
31,152
31,798
32,446
33,094
33,742
34,389
35,036
35,683
36,330
36,977
37,623
38,270
38,916
39,563
40,210
40,860
41,510
42,160
42,810
43,461
44,112
44,763
45,414
46,067
46,719
47,372
48,025
SO,
2,265
2,553
2,629
2,706
2,783
2,859
2,936
2,957
1,437
333
341
323
330
338
345
353
360
368
375
383
390
398
405
412
420
427
435
442
450
457
465
472
480
487
495
502
509
517
524
532
539
547
voc
808
906
929
953
976
999
1,021
1,044
1,066
1,088
1,109
1,131
1,152
1,173
1,194
1,214
1,235
1,255
1,276
1,296
1,316
1,336
1,356
1,376
1,395
1,415
1,435
1,455
1,475
1,495
1,516
1,538
1,559
1,581
1,602
1,624
1,646
1,668
1,691
1,713
1,736
1,759
CO
3,236
3,635
3,737
3,838
3,938
4,038
4,138
4,237
4,336
4,434
4,531
4,628
4,724
4,820
4,915
5,010
5,104
5,199
5,293
5,386
5,480
5,573
5,665
5,758
5,850
5,943
6,035
6,127
6,220
6,314
6,410
6,506
6,602
6,698
6,795
6,891
6,988
7,086
7,184
7,283
7,382
7,482
3.5 Anticipated Emission Reductions With the Proposed Rule
Emissions from nonroad diesel engines will continue to be a significant part of the
emissions 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
3-55
-------
Draft Regulatory Impact Analysis
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 proposal will affect fine paniculate 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 proposed fuel standards will affect PM2 5 and SO2.
This section discusses the expected emission reductions associated with this proposal.
The baseline case represents future emissions with current standards. The controlled case
estimates the future emissions of these engines based on the proposed standards and fuel
requirements in this notice. 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 PM25 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
proposal. PM2 5 will be reduced due to the proposed PM exhaust 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 2014. Nonroad
diesel fuel sulfur is reduced to a 500 ppm standard in June of 2007, and further reduced for land-
based nonroad diesel engines 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 in order to reflect close to complete turnover of the
fleet to engines meeting the proposed standards. For comparison purposes, emissions 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.
3-56
-------
Table3.5-la
Total Emission Reductions (48-State) from Proposed Rule
Year
2000
2005
2007
2008
2009
2010
2011
2015
2020
2025
2030
PM25
0
0
10,605
19,061
19,998
21,864
25,496
52,476
85,254
109,325
126,910
NOX
0
0
0
301
619
1,007
20,574
217,575
503,701
693,857
821,911
SO2
0
0
142,948
249,746
254,544
270,977
285,003
305,639
331,840
358,863
385,932
VOC
0
0
0
29
59
90
862
8,788
18,033
24,624
29,487
CO
0
0
0
0
0
0
14,487
182,520
381,487
520,864
620,345
Benzene
0
0
0
1
1
2
17
176
361
492
590
Formaldehyde
0
0
0
3
7
11
102
1,037
2,128
2,906
3,479
Acetaldehyde
0
0
0
2
3
5
46
466
956
1,305
1,563
1,3 -Butadiene
0
0
0
0
0
0
2
18
36
49
59
Acrolein
0
0
0
0
0
0
3
26
54
74
88
Table3.5-lb
Total Emission Reductions (50-State) from Proposed Rule
Year
2000
2005
2007
2008
2009
2010
2011
2015
2020
2025
2030
PM2,
0
0
10,705
19,238
20,181
22,058
25,712
52,851
85,827
110,026
127,708
NOX
0
0
0
304
624
1,015
20,717
218,939
506,815
698,000
826,690
SO2
0
0
144,298
252,100
256,935
273,470
287,583
308,386
334,799
362,041
389,337
VOC
0
0
0
29
59
91
868
8,841
18,141
24,769
29,660
CO
0
0
0
0
0
0
14,585
183,596
383,730
523,844
623,851
Benzene
0
0
0
1
1
2
17
177
363
495
593
Formaldehyde
0
0
0
3
7
11
102
1,043
2,141
2,923
3,500
Acetaldehyde
0
0
0
2
3
5
46
469
961
1,313
1,572
1,3 -Butadiene
0
0
0
0
0
0
2
18
36
50
59
Acrolein
0
0
0
0
0
0
3
27
54
74
89
Table3.5-2a
Estimated National (48-State) PM2 5
Emissions and Reductions From Nonroad Land-Based Diesel Engines"
-------
Draft Regulatory Impact Analysis
Year
2000
2005
2007
2008
2009
2010
2011
2015
2020
2025
2030
PM2 5 Emissions [short tons]
Without
Rule
161,143
135,286
127,089
124,789
122,815
121,007
119,865
119,957
124,344
131,644
139,527
With fuel sulfur
reduced to 500 ppm in
2007;
No Tier 4 standards
161,143
135,286
119,606
111,657
109,378
107,265
105,816
104,682
107,543
113,335
119,710
With Rule
(Fuel sulfur reduced
to 15 ppm in 2010;
Tier 4 standards)
161,143
135,286
119,606
111,120
108,258
104,633
99,943
73,211
45,057
28,610
19,239
PM2 5 Reductions [short tons]
With fuel sulfur
reduced to 500 ppm in
2007;
No Tier 4 standards
0
0
7,483
13,132
13,437
13,742
14,049
15,275
16,801
18,309
19,817
With
Rule
0
0
7,483
13,669
14,557
16,374
19,922
46,745
79,277
103,034
120,288
' PM25 represents 92% of PM10 emissions.
3-58
-------
Emissions Inventory
Table3.5-2b
Estimated National (50-State) PM2 5
Emissions and Reductions From Nonroad Land-Based Diesel Engines'
Year
2000
2005
2007
2008
2009
2010
2011
2015
2020
2025
2030
PM2 5 Emissions [short tons]
Without
Rule
161,903
135,940
127,708
125,401
123,422
121,609
120,466
120,575
124,990
132,345
140,277
With fuel sulfur
reduced to 500 ppm in
2007;
No Tier 4 standards
161,903
135,940
120,185
112,197
109,911
107,792
106,341
105,218
108,106
113,936
120,352
With Rule
(Fuel sulfur reduced
to 15 ppm in 2010;
Tier 4 standards)
161,903
135,940
120,185
111,658
108,786
105,147
100,435
73,566
45,258
28,734
19,323
PM2 5 Reductions [short tons]
With fuel sulfur
reduced to 500 ppm in
2007;
No Tier 4 standards
0
0
7,524
13,204
13,510
13,817
14,125
15,357
16,883
18,409
19,925
With
Rule
0
0
7,524
13,743
14,636
16,462
20,032
47,010
79,732
103,611
120,954
' PM25 represents 92% 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 proposed rule requirements. PM2 5 emissions
are reduced roughly 120,000 tons with the proposed rule by 2030.
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 proposed PM2 5 rule. By 2030, we estimate that
PM2 5 emissions from this source would be reduced by 86 percent in that year.
3-59
-------
Draft Regulatory Impact Analysis
Figure 3.5-1: Estimated Reductions inPM2.5 Emissions
From Land-Based Nonroad Engines (tons/year)
• Base 50-State
•Control 50-State
2000 2005 2010 2015 2020 2025 2030
Nonroad diesel engines used in locomotives, commercial marine vessels, and recreational
marine vessels are not affected by the emission standards of this proposal. PM25 emissions from
these engines would be reduced by the reductions in diesel fuel sulfur for these types of engines
from an in-use average of 2400 ppm today to an in-use average of about 240 ppm in 2010. The
estimated 48-state and 50-state reductions in PM2 5 emissions from these engines based on the
proposed change in diesel fuel sulfur are given in Tables 3.5-3a and 3.5-3b, respectively. Total
PM2 5 reductions reach roughly 6,700 tons in 2030 for these diesel nonroad 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
85,000 tons in 2020, increasing to 127,000 tons in 2030. Simply reducing the fuel sulfur level to
500 ppm in 2007 would result in PM2 5 reductions of 23,000 tons in 2020 and 26,000 tons in
2030.
3-60
-------
Emissions Inventory
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
1,929
3,321
3,345
3,368
3,426
3,476
3,581
3,744
3,914
Commerical
Marine Diesel
0
0
1,078
1,869
1,890
1,911
1,932
2,019
2,137
2,263
2,399
Recreational
Marine Diesel
0
0
114
202
206
212
216
235
260
285
308
Total PM2 5
Reductions
0
0
3,121
5,392
5,441
5,491
5,574
5,730
5,978
6,292
6,621
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
1,931
3,326
3,349
3,373
3,430
3,480
3,586
3,749
3,919
Commerical
Marine Diesel
0
0
1,134
1,966
1,988
2,010
2,032
2,124
2,248
2,380
2,524
Recreational
Marine Diesel
0
0
115
203
208
213
217
237
262
286
311
Total PM2 5
Reductions
0
0
3,181
5,495
5,545
5,595
5,680
5,842
6,095
6,415
6,754
3-61
-------
Draft Regulatory Impact Analysis
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
2015
2020
2025
2030
PM2 5 Emissions [short tons]
Without
Rule
218,311
194,554
185,875
183,256
181,321
179,213
178,610
178,559
183,250
191,976
201,567
With fuel sulfur
reduced to 500
ppm in 2007
218,311
194,554
175,270
164,731
162,443
159,981
158,987
157,554
160,481
167,376
175,128
With fuel sulfur
further reduced to 1 5
ppm in 2010 for land-
based diesels
218,311
194,554
175,270
164,195
161,323
157,349
153,114
126,083
97,996
82,651
74,657
PM25 Reductions [short tons]
With fuel sulfur
reduced to 500
ppm in 2007
0
0
10,605
18,525
18,878
19,232
19,622
21,005
22,769
24,600
26,438
With fuel sulfur
further reduced to 1 5
ppm in 2010 for land-
based diesels
0
0
10,605
19,061
19,998
21,864
25,496
52,476
85,254
109,325
126,910
3-62
-------
Emissions Inventory
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
2015
2020
2025
2030
PM2 5 Emissions [short tons]
Without
Rule
221,035
197,228
188,532
185,916
183,986
181,883
181,291
181,301
186,084
194,960
204,705
With fuel sulfur
reduced to 500
ppm in 2007
221,035
197,228
177,828
167,217
164,931
162,471
161,486
160,101
163,105
170,135
178,026
With fuel sulfur
further reduced to 1 5
ppm in 2010 for land-
based diesels
221,035
197,228
177,828
166,678
163,805
159,826
155,579
128,449
100,257
84,933
76,997
PM25 Reductions [short tons]
With fuel sulfur
reduced to 500
ppm in 2007
0
0
10,705
18,699
19,055
19,412
19,805
21,199
22,979
24,825
26,679
With fuel sulfur
further reduced to 1 5
ppm in 2010 for land-
based diesels
0
0
10,705
19,238
20,181
22,058
25,712
52,851
85,827
110,026
127,708
3-63
-------
Draft Regulatory Impact Analysis
Figure 3.5-2: Estimated Reductions in
Emissions From Land-Based Nonroad Engines,
CMVs, RMVs, and Locomotives (tons/year)
250,000 -i
• Base 50-State
•Control 50-State
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 the proposed rule and the estimated
emissions reductions. The 50-state results are shown graphically in Figure 3.5-3. By 2030, we
estimate that NOX emissions from these engines will be reduced by 67 percent in that year.
NOX emissions from locomotives, commercial marine diesel vessels, and recreational
marine diesel vessels are not affected by this proposal.
3-64
-------
Emissions Inventory
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]
1,569,903
1,483,942
1,291,533
1,167,635
1,140,727
1,231,995
NOX Emissions With
Rule
1,569,903
1,483,942
1,290,526
950,060
637,025
410,084
NOX Reductions With
Rule
0
0
1,007
217,575
503,701
821,911
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]
1,578,148
1,491,692
1,298,193
1,173,656
1,146,750
1,238,701
NOX Emissions With
Rule
1,578,148
1,491,692
1,297,178
954,717
639,935
412,011
NOX Reductions With
Rule
0
0
1,015
218,939
506,815
826,690
3-65
-------
Draft Regulatory Impact Analysis
Figure 3.5-3: Estimated Reductions in NOx Emissions
From Land-Based Nonroad Engines (tons/year)
1,800,000
1,600,000
1,400,000
1,200,000 -
1,000,000 -
800,000
600,000
400,000
200,000
•BaseSO-State
• Control 50-State
2000 2005 2010 2015 2020 2025 2030
3.5.3 SO2 Reductions
As part of this proposal, sulfur levels in fuel would 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 would be reduced from the current average in-use level of roughly 2300
ppm to an average in-use level of about 1100 ppm. By 2010, the sulfur in diesel fuel used by
land-based nonroad engines would be reduced to an average in-use level of 11 ppm with a
maximum level of 15 ppm. The sulfur in diesel fuel used by locomotives, commercial marine,
and recreational marine engines would remain at an average in-use level of about 230 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 reductions from this
proposal. 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, emissions 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.
3-66
-------
Emissions Inventory
Table3.5-6a
Estimated National (48-State) SO2
Emissions and Reductions From Nonroad Land-Based Diesel Engines
Year
2000
2005
2007
2008
2009
2010
2011
2015
2020
2025
2030
SO2 Emissions [short tons]
Without
Rule
167,094
191,385
194,003
198,657
203,311
206,104
210,737
229,235
252,089
274,913
297,573
With fuel sulfur
reduced to 500 ppm in
2007
167,094
191,385
93,229
21,757
22,267
20,917
21,387
23,265
25,584
27,901
30,200
With Rule
(Fuel sulfur reduced
to 15 ppm in 2010)
167,094
191,385
93,229
21,757
22,267
9,297
1,032
1,014
1,005
1,038
1,096
SO2 Reductions [short tons]
With fuel sulfur
reduced to 500 ppm in
2007
0
0
100,774
176,900
181,044
185,187
189,350
205,970
226,505
247,012
267,373
With
Rule
0
0
100,774
176,900
181,044
196,807
209,705
228,221
251,084
273,875
296,477
3-67
-------
Draft Regulatory Impact Analysis
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
167,999
192,424
195,057
199,736
204,416
207,225
211,884
230,483
253,464
276,414
299,199
With fuel sulfur
reduced to 500 ppm in
2007
167,999
192,424
93,735
21,875
22,388
21,031
21,504
23,391
25,724
28,053
30,365
With Rule
(Fuel sulfur reduced
to 15 ppm in 2010)
167,999
192,424
93,735
21,875
22,388
9,347
1,038
1,020
1,011
1,044
1,102
SO2 Reductions [short tons]
With fuel sulfur
reduced to 500 ppm in
2007
0
0
101,321
177,861
182,028
186,194
190,380
207,092
227,740
248,361
268,834
With
Rule
0
0
101,321
177,861
182,028
197,878
210,846
229,464
252,453
275,370
298,098
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 reduced 298,000 tons with
the proposed rule by 2030. 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 not affected by the emission standards of this proposal. SO2 emissions from
these engines would be reduced by the reductions in diesel fuel sulfur for these types of engines
from an in-use average of 2400 ppm today to an in-use average of about 230 ppm in 2010. The
estimated 48-state and 50-state reductions in SO2 emissions from these engines based on the
proposed change in diesel fuel sulfur are given in Tables 3.5-7a and 3.5-7b, respectively. Total
50-state SO2 reductions reach 91,000 tons in 2030 for these diesel nonroad 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 with the
proposed rule are 334,000 tons in 2020, increasing to 389,000 tons in 2030. Simply reducing the
3-68
-------
Emissions Inventory
fuel sulfur level to 500 ppm in 2007 would result in SO2 reductions of 310,000 tons in 2020 and
360,000 tons in 2030.
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
26,058
44,874
45,184
45,506
46,282
46,955
48,383
50,578
52,873
Commerical Marine
Diesel Vessels
0
0
14,569
25,255
25,533
25,814
26,100
27,285
28,866
30,572
32,415
Recreational
Marine Diesel
Vessels
0
0
1,546
2,718
2,784
2,850
2,916
3,178
3,508
3,838
4,167
Total SO2
Reductions
0
0
42,173
72,847
73,501
74,170
75,298
77,418
80,757
84,988
89,455
3-69
-------
Draft Regulatory Impact Analysis
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
26,094
44,936
45,246
45,569
46,346
47,021
48,450
50,648
52,946
Commerical Marine
Diesel Vessels
0
0
15,326
26,568
26,860
27,156
27,457
28,703
30,367
32,161
34,100
Recreational
Marine Diesel
Vessels
0
0
1,556
2,735
2,801
2,868
2,934
3,198
3,530
3,862
4,193
Total SO2
Reductions
0
0
42,977
74,239
74,907
75,592
76,737
78,922
82,346
86,671
91,239
3-70
-------
Emissions Inventory
Table3.5-8a
Estimated National (48-State) SO2 Emissions
Land-Based Nonroad, Locomotive, Commercial Marine,
and Reductions from
and Recreational Marine Vessels
Year
2000
2005
2007
2008
2009
2010
2011
2015
2020
2025
2030
SO2 Emissions [short tons]
Without
Rule
243,333
273,331
275,101
280,363
285,750
288,617
294,504
315,367
341,941
369,475
397,109
With fuel sulfur
reduced to 500
ppm in 2007
243,333
273,331
132,153
30,617
31,206
29,261
29,857
31,978
34,679
37,475
40,281
With fuel sulfur
further reduced to 1 5
ppm in 2010 for land-
based diesels
243,333
273,331
132,153
30,617
31,206
17,640
9,502
9,728
10,100
10,612
11,176
SO2 Reductions [short tons]
With fuel sulfur
reduced to 500
ppm in 2007
0
0
142,948
249,746
254,543
259,356
264,648
283,389
307,262
332,000
356,828
With fuel sulfur
further reduced to 1 5
ppm in 2010 for land-
based diesels
0
0
142,948
249,746
254,543
270,977
285,003
305,639
331,840
358,863
385,932
3-71
-------
Draft Regulatory Impact Analysis
Table3.5-8b
Estimated National (50-State) SO2 Emissions
Land-Based Nonroad, Locomotive, Commercial Marine,
and Reductions from
and Recreational Marine Vessels
Year
2000
2005
2007
2008
2009
2010
2011
2015
2020
2025
2030
SO2 Emissions [short tons]
Without
Rule
245,712
275,929
277,699
283,004
288,434
291,320
297,252
318,288
345,084
372,849
400,720
With fuel sulfur
reduced to 500
ppm in 2007
245,712
275,929
133,401
30,904
31,499
29,534
30,135
32,274
34,998
37,817
40,647
With fuel sulfur
further reduced to 1 5
ppm in 2010 for land-
based diesels
245,712
275,929
133,401
30,904
31,499
17,851
9,669
9,903
10,285
10,807
11,383
SO2 Reductions [short tons]
With fuel sulfur
reduced to 500
ppm in 2007
0
0
144,298
252,100
256,935
261,786
267,117
286,014
310,086
335,032
360,073
With fuel sulfur
further reduced to 1 5
ppm in 2010 for land-
based diesels
0
0
144,298
252,100
256,935
273,470
287,583
308,386
334,799
362,041
389,337
3-72
-------
Emissions Inventory
Figure 3.5-4: Estimated SOx Benefits from Reducing
Sulfur for Land-Based Nonroad Engines,
CMVs, RMVs, and Locomotives
450,000 -,
400,000 -
350,000 -
300,000 -
250,000
200,000 -
150,000 -
100,000 -
50,000 -
0
•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 that EPA expects from implementing the proposed NMHC standards.
Although this proposal does not include specific standards for air toxics, these pollutants
would be reduced through the implementation of the proposed NMHC standards. Tables 3.5-10a
and 3.5-10b show our estimate of the proposed rule's beneficial impact on the key air toxics
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.
3-73
-------
Draft Regulatory Impact Analysis
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 tonsl
200,366
163,193
128,391
105,800
96,855
95,144
97,348
VOC
With Rule
[short tonsl
200,366
163,193
128,301
97,013
78,822
70,520
67,861
VOC Reductions
With Rule
[short tonsl
0
0
90
8,788
18,033
24,624
29,487
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]
201,386
164,035
129,063
106,359
97,378
95,663
97,882
VOC
With Rule
[short tons]
201,386
164,035
128,973
97,519
79,238
70,894
68,223
VOC Reductions
With Rule
[short tons]
0
0
91
8,841
18,141
24,769
29,660
3-74
-------
Emissions Inventory
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
4,007
4,007
0
3,264
3,264
0
2,990
2,990
0
2,568
2,566
2
2,116
1,940
176
1,937
1,576
361
1,903
1,410
492
1,947
1,357
590
Formaldehyde
23,643
23,643
0
19,257
19,257
0
17,643
17,643
0
15,150
15,140
11
12,484
11,448
1,037
11,429
9,301
2,128
11,227
8,321
2,906
11,487
8,008
3,479
Acetaldehyde
10,619
10,619
0
8,649
8,649
0
7,924
7,924
0
6,805
6,800
5
5,607
5,142
466
5,133
4,178
956
5,043
3,738
1,305
5,159
3,597
1,563
1,3 -Butadiene
401
401
0
326
326
0
299
299
0
257
257
0
212
194
18
194
158
36
190
141
49
195
136
59
Acrolein
601
601
0
490
490
0
449
449
0
385
385
0
317
291
26
291
236
55
285
212
73
292
204
88
3-75
-------
Draft Regulatory Impact Analysis
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
4,028
4,028
0
3,281
3,281
0
3,006
3,006
0
2,581
2,579
2
2,127
1,950
177
1,948
1,585
363
1,913
1,418
495
1,958
1,364
593
Formaldehyde
23,764
23,764
0
19,356
19,356
0
17,735
17,735
0
15,229
15,219
11
12,550
11,507
1,043
11,491
9,350
2,141
11,288
8,366
2,923
11,550
8,050
3,500
Acetaldehyde
10,673
10,673
0
8,694
8,694
0
7,966
7,966
0
6,840
6,836
5
5,637
5,168
469
5,161
4,200
961
5,070
3,757
1,313
5,188
3,616
1,572
1,3 -Butadiene
403
403
0
328
328
0
301
301
0
258
258
0
213
195
18
195
158
36
191
142
50
196
136
59
Acrolein
604
604
0
492
492
0
451
451
0
387
387
0
319
293
27
292
238
54
287
213
74
294
205
89
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
proposed rule and the estimated emissions reductions. Although for most engines, Tier 4 does
not revise the existing CO standards, CO is estimated to be reduced 90% with the advent of trap-
equipped engines (corresponding to the start of 0.02 or 0.01 g/bhp-hr PM standards). By 2030,
we estimate that 50-state CO emissions from these engines will be reduced 623,000 tons in that
year.
CO emissions from locomotives, commercial marine diesel vessels, and recreational
marine diesel vessels are not affected by this proposal.
3-76
-------
Emissions Inventory
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]
923,886
764,918
684,552
672,944
700,017
793,923
CO Emissions
With Rule
[short tons]
923,886
764,918
684,552
490,424
318,530
173,579
CO Reductions
With Rule
[short tons]
0
0
0
182,520
381,487
620,345
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]
923,674
768,838
688,153
676,570
703,845
798,316
CO Emissions
With Rule
[short tons]
928,674
768,838
688,377
492,974
320,115
174,465
CO Reductions
With Rule
[short tons]
0
0
0
183,596
383,730
623,851
3.6 Emission Inventories Used for Air Quality Modeling
The emissions inputs for the air quality modeling are required early in the analytical
process, in order to be able to conduct the air quality modeling and present the results in this
proposal. 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 3 of the preamble for this proposal and Chapter
4 of this document). As a result, we have revised the control scenario. We have also made
minor changes to the baseline fuel sulfur levels. This section describes the changes in the inputs
and resulting emissions inventories between the preliminary baseline and control scenarios used
3-77
-------
Draft Regulatory Impact Analysis
for the air quality modeling and the updated baseline and control scenarios in this proposal. This
section will focus on the four nonroad diesel categories that are affected by the proposed
standards and/or the fuel sulfur requirements: land-based diesel engines, recreational marine
diesel engines, commercial marine diesel engines, and locomotives. There have been no changes
to any other source categories (e.g., highway, stationary point and area sources). While these
other source categories are not affected by the proposed rule, revisions to the inventories for
these other source categories can impact air quality results.
The methodology used to develop the emissions inventories for the air quality modeling
is first briefly described, followed by comparisons of the preliminary and proposed 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 emissions 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 diesel nonroad categories affected by the proposed 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.
3-78
-------
Emissions Inventory
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
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
959,704
819,201
814,827
921,556
612,722
534,520
PM2,
[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
31,545
37,290
41,354
48,381
36,546
31,644
CO
[short tons]
1,010,501
702,145
793,899
3,215
5,537
6,464
126,382
159,900
176,533
112,171
119,302
119,302
For the proposed baseline inventories, we have made minor changes to the diesel fuel
sulfur levels. The diesel fuel sulfur inputs used for the preliminary and proposed baseline
inventories are provided in Table 3.6-2. The diesel fuel sulfur level is now reduced from
2500ppm to roughly 2300ppm, beginning in 2006. Both the preliminary and proposed 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.
There have also been reductions to the fuel volumes assigned to locomotives and
commercial marine vessels. For the preliminary inventory development, railroad distillate and
vessel bunkering 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. Similarly, 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 proposed inventory, the EIA railroad distillate and vessel bunkering distillate
estimates were first adjusted to estimate the fraction of distillate that is diesel fuel. The diesel
fractions used are 0.95 for railroad distillate and 0.90 for vessel bunkering distillate. Fuel
consumption estimates from rail maintenance and recreational marine engines were then
3-79
-------
Draft Regulatory Impact Analysis
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 estimate of recreational marine fuel
consumption continues to be that generated from the draft NONROAD2002 model. The
derivation of diesel fractions and the revised estimate of rail maintenance fuel consumption is
documented in Chapter 7.
As a result, the corrections to fuel sulfur levels and locomotive and commercial marine
fuel volumes will reduce the PM and SO2 baseline inventories in 2020 and 2030.
Table 3.6-2
Modeled Baseline In-Use Diesel Fuel Sulfur Content
Proposed Baseline vs Preliminary Baseline Used for Air Quality Modeling
Applications
Land-Based Diesel Engines
Commercial and Recreational Marine
Engines and Locomotives
Proposed Baseline
Fuel Sulfur
ppm
2318
2271
2237
2217
2396
2352
2321
2302
Calendar Year
through 2005
2006
2007-2009
2010+
through 2005
2006
2007-2009
2010+
Preliminary Baseline
Fuel Sulfur
ppm
2500a
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 commercial marine diesel and locomotive categories, revised PM and SO2
inventories were generated for the proposed baseline scenarios. For the land-based diesel and
recreational marine diesel categories, it was not possible to generate revised county-level baseline
inventories due to time and resource constraints. Instead, for the land-based diesel and
recreational marine diesel categories, national level NONROAD model runs were used as the
basis for comparison of the preliminary and proposed baseline scenarios. National level model
runs were done using the 48-state average fuel sulfur levels for both the preliminary and
proposed baseline scenarios in 1996, 2020 and 2030.
3-80
-------
Emissions Inventory
To examine the feasibility of using national level model results, Table 3.6-3 first provides
a comparison of the 48-state emissions derived from national level model runs to those derived
from a sum of county level runs for the same preliminary baseline scenario. The county-level
results were taken from Table 3.6-1. The national level and sum of county level results are quite
similar for NOX, VOC, and CO. Use of the national level model runs lowers the emissions
baseline slightly for SO2, and less so for PM. This is expected, since diesel NOX, VOC, and CO
emissions are insensitive to county-level differences in fuel characteristics and temperature. PM
and SO2 are sensitive to fuel sulfur levels, with SO2 exhibiting the most sensitivity.
Table 3.6-4 compares the proposed and preliminary 48-state baseline scenario inventories
for land-based diesel engines, recreational marine diesel engines, commercial marine diesel
engines, and locomotives. The national level model run results are used as the basis for
comparison for the land-based diesel and recreational marine diesel categories. Results are only
presented for PM2 5 and SO2 emissions, since these are the only pollutants affected by the
changes.
PM2 5 emissions are reduced roughly 2% with the proposed baseline scenario in 2020 and
2030, while SO2 is reduced 13%. These reductions to the baseline will serve to decrease the
emission reductions of the rule for PM25 and SO2.
3-81
-------
Table 3.6-3
Modeled 48-State Emissions for Preliminary Baseline Scenario Used for Air Quality Modeling
Comparison of Results Derived from National Level Model Runs vs. Sum of County Level Model Runs
Applications
Land-Based Diesel Engines
Recreational Marine Diesel
Engines
Year
1996
2020
2030
1996
2020
2030
NOX [short tons]
National
Level
1,583,664
1,140,727
1,231,995
19,440
34,817
41,250
County
Level
1,583,641
1,144,686
1,231,981
19,438
34,814
41,246
%
Difference
0.0%
-0.3%
0.0%
0.0%
0.0%
0.0%
PM2 5 [short tons]
National
Level
177,375
126,720
142,342
494
848
988
County
Level
178,500
127,755
143,185
511
876
1,021
%
Difference
-0.6%
-0.8%
-0.6%
-3.3%
-3.2%
-3.2%
SO2 [short tons]
National
Level
159,540
284,268
335,558
2,349
4,239
5,035
County
Level
172,175
308,075
360,933
2,535
4,562
5,418
%
Difference
-7.3%
-7.7%
-7.0%
-7.4%
-7.1%
-7.1%
Table 3.6-3, continued
Applications
Land-Based Diesel Engines
Recreational Marine Diesel
Engines
Year
1996
2020
2030
1996
2020
2030
VOC [short tons]
National
Level
221,403
96,855
97,348
803
1,328
1,528
County
Level
221,398
97,113
97,345
803
1,327
1,528
%
Difference
0.0%
-0.3%
0.0%
0.0%
0.0%
0.0%
CO [short tons]
National
Level
1,010,501
700,017
793,923
3,215
5,538
6,465
County
Level
1,010,501
702,145
793,899
3,215
5,537
6,464
%
Difference
0.0%
-0.3%
0.0%
0.0%
0.0%
0.0%
-------
Emissions Inventory
Table 3.6-4
Modeled 48-State Baseline PM2 s and SO2 Emission Reductions Due to Changes in Baseline
Applications
Land-Based
Diesel Engines
Recreational
Marine Diesel
Engines
Commercial
Marine Diesel
Engines
Locomotives
Total
Year
1996
2020
2030
1996
2020
2030
1996
2020
2030
1996
2020
2030
1996
2020
2030
PM2 5 Emissions [short tons]
Proposed
176,510
124,334
139,527
487
823
958
36,367
41,365
45,411
20,937
16,727
15,670
234,301
183,249
201,566
Preliminary
177,375
126,720
142,342
494
848
988
37,203
42,054
46,185
22,396
17,683
16,988
237,468
187,305
206,503
Reduction
865 (0.5%)
2,386(1.9%)
2,815 (2.0%)
7(1.4%)
25 (2.9%)
30 (3.0%)
836 (2.2%)
689(1.6%)
774(1.7%)
1,459 (6.5%)
956 (5.4%)
1,318(7.8%)
3,167(1.3%)
4,056 (2.2%)
4,937 (2.4%)
SO2 [short tons]
Proposed
147,926
252,089
297,573
2,251
3,903
4,636
25,948
32,117
36,068
50,534
53,832
58,832
226,659
341,941
397,109
Preliminary
159,540
284,268
335,558
2,349
4,239
5,035
37,252
43,028
48,308
57,979
62,843
70,436
257,120
394,378
459,337
Reduction
11,614
(7.3%)
32,179
(11.3%)
37,985
(11.3%)
98 (4.2%)
336 (7.9%)
399 (7.9%)
11,304
(30.3%)
10,911
(25.4%)
12,240
(25.3%)
7,445
(12.8%)
9,011
(14.3%)
11,604
(16.5%)
30,461
(11.8%)
52,437
(13.3%)
62,228
(13.5%)
a Based on 48-state national runs for land-based and recreational marine categories. Based on 48-state sum of county level
runs for commercial marine and locomotive engines.
3-83
-------
Draft Regulatory Impact Analysis
3.6.3 Control Inventories
Table 3.6-5 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
only affect future year emission estimates.
Table 3.6-5
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
819,201
814,827
612,722
534,520
PM2,
[short tons]
36,477
14,112
552
636
38,882
42,625
13,051
11,798
S02
[short tons]
3,340
1,159
20
24
184
206
272
305
voc
[short tons]
73,941
63,285
1,327
1,528
37,290
41,354
36,546
31,644
CO
[short tons]
249,734
133,604
5,537
6,464
159,900
176,533
119,302
119,302
The certification standards used for the preliminary and proposed control scenarios are
provided in Tables 3.6-6 and 3.6-7, respectively. In general, the preliminary control scenario is
more stringent in terms of levels and effective model years for PM and NOX than the proposed
control scenario for all horsepower categories. The NMHC standard is 0.14 g/bhp-hr with both
scenarios, although the phase-in of this standard is later in the proposed control scenario. The
CO standards are unchanged in both control scenarios, although CO is assumed to be reduced
90% in both scenarios with the advent of trap-equipped engines (corresponding to the start of
0.02 or 0.01 g/bhp-hr PM standards). As a result, the proposed standards will increase the
emissions of PM, NOX, NMHC, and CO in 2020 and 2030 relative to the preliminary standards.
3-84
-------
Table 3.6-6
Preliminary Tier 4 Exhaust Emissions Certification 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/bhp-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
N(\
5.6
0.30
NMHC
0.14
5.6 ^
0.30
0.14
3.5^
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
4.1b
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/bhp-hr for engines
below 11 hp and 4.9 g/bhp-hr for engines at or above 11 hp.
-------
Table 3.6-7
Proposed Tier 4 Exhaust Emissions Certification Standards
Engine Power
hp<25
25 < hp < 75
75 < hp < 175
175 < hp < 750
hp > 750
Emissions Standard
(g/bhp-hr)a
transitional or final
final
transitional4
final
transitional
final
transitional
final
transitional
final
PM
0.30
0.22
0.02
0.01
(100%)
0.01
(100%)
0.01
(100%)
0.01
(100%)
0.01
(50%)
0.01
(100%)
NO,
NMHC
5.6 b-c
5.6/3. 5 b'c
3.5 b
0.30
(50%)
0.30
(100%)
0.30
(50%)
0.30
(100%)
0.30
(50%)
0.30
(100%)
0.14
(50%)
0.14
(100%)
0.14
(50%)
0.14
(100%)
0.14
(50%)
0.14
(100%)
CO
4.9
3.7
3.7°
3.7 c
3.7 c
2.6 c
2.6 c
2.6 c
2.6 c
Model Year
2008
2008
2013
2012-2013
2014
2011-2013
2014
2011-2013
2014
a Percentages are model year sales fractions required to comply with the indicated standard.
b This is a combined NMHC + NOX standard.
0 This emissions 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/bhp-hr for engines below 50 hp and 3.5 g/bhp-hr for engines at or above 50 hp.
d Manufacturers may optionally skip the transitional standards for 25-75 hp engines; the final standards would then take effect for these engines in the 2012 model year.
-------
Emissions Inventory
The diesel fuel sulfur inputs used for the preliminary and proposed control scenarios are
provided in Tables 3.6-8 and 3.6-9, respectively. For land-based diesel engines, the modeled in-
use diesel fuel sulfur content is 1 Ippm 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 1 Ippm in 2020 and 2030 for the preliminary control scenario, but 233 ppm in
2020 and 2030 for the proposed control scenario. As a result, the proposed fuel sulfur levels 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 3.6-8
Modeled 48-State & 50-State In-Use
Diesel Fuel Sulfur Content Used for Air Quality Modeling
Applications
All Diesel Categories
Standards
Baseline + hwy 500 ppm
"spillover"
Baseline + hwy 1 5 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
3-87
-------
Draft Regulatory Impact Analysis
Table 3.6-9
Modeled 48-State & 50-State In-Use Diesel Fuel Sulfur Content for Proposed Standards
Applications
Land-based,
all power ranges
Recreational and
Commercial Marine Diesel
Engines and Locomotives
Standards
Baseline
June intro of 500 ppm
500 ppm standard
June intro of 1 5 ppm
Final 1 5 ppm standard
Baseline
June intro of 500 ppm
Final 500 ppm standard
Modeled In-Use
Fuel Sulfur
Content, ppm
2318
2271
1075
245
100
11
2396
2352
1114
252
233
Calendar Year
through 2005
2006
2007
2008-2009
2010
2011+
through 2005
2006
2007
2008-2009
2010+
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 proposed control scenarios are provided in Tables 3.6-
10 and 3.6-11, respectively. For 2020 and 2030, the certification diesel fuel sulfur levels are the
same for both the preliminary and proposed control scenarios.
For the commercial marine diesel and locomotive categories, inventories were generated
for the proposed control scenarios, using the fuel volume and fuel sulfur level estimates. For the
land-based diesel and recreational marine diesel categories, it was not possible to generate
revised county-level control inventories. Instead, for the land-based diesel and recreational
marine diesel categories, national level NONROAD model runs were used as the basis for
comparison of the preliminary and proposed control scenarios. National level model runs were
done using the 48-state average fuel sulfur levels for both the preliminary and proposed control
scenarios in 2020 and 2030.
To examine the feasibility of using national level model results, Table 3.6-12 first
provides a comparison of the 48-state emissions derived from national level model runs to those
derived from a sum of county level runs for the same preliminary control scenario. The county-
level results were taken from Table 3.6-5. The national level and sum of county level results are
quite similar. This is expected, since diesel NOX, VOC, and CO emissions are insensitive to
-------
Emissions Inventory
county-level differences in fuel characteristics and temperature. PM and SO2 are sensitive to fuel
sulfur levels, with SO2 exhibiting the most sensitivity.
Table 3.6-13 compares the proposed and preliminary 48-state control scenario inventories
for land-based diesel engines, recreational marine diesel engines, commercial marine diesel
engines, and locomotives. The national level model run results are used as the basis for
comparison for the land-based diesel and recreational marine diesel categories. 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 proposed 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 proposed and
preliminary scenarios is attributed to differences in aggregation of county-level runs compared to
using one national level run.
The recreational marine, commercial marine, and locomotive categories are not
controlled in either scenario; however, the in-use fuel sulfur level is 1 Ippm for the preliminary
control scenario and 233 ppm for the proposed control scenario. This affects the PM and SO2
emissions. Accordingly, the PM and SO2 emissions for these categories are higher for the
proposed control scenario.
Table 3.6-10
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
' Tier 4 refers to both transitional and final standards.
3-89
-------
Draft Regulatory Impact Analysis
Table 3.6-11
Modeled Certification Diesel Fuel Sulfur Content for Proposed Standards
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
transitionalb
final
Modeled Certification Fuel
Sulfur Content, PPM
2000
500
15
500
15
2000
15
2000
15
2000
50% 2000
50% 15
15
Model
Year
through 2007
2008
2013
2008-2011
2012
2007-2011
2012
2006-2010
2011
2006-2010
2011-2013
2014
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 would allow manufacturers to take advantage of
the availability of 500 ppm fuel that year.
b The engines remaining at the Tier 2 level would be allowed to continue certifying on the same fuel as earlier Tier 2
engines, but those meeting the Tier 4 0.01 PM standard are assumed to certify on 15 ppm fuel.
3-90
-------
Table 3.6-12
Modeled 48-State Emissions for Preliminary Control Scenario Used for Air Quality Modeling
Comparison of Results Derived from National Level Model Runs vs. Sum of County Level Model Runs
Applications
Land-Based Diesel Engines
Recreational Marine Diesel
Engines
Year
2020
2030
2020
2030
NOX [short tons]
National
Level
477,100
222,238
34,817
41,250
County
Level
481,068
222,237
34,814
41,246
%
Difference
-0.8%
0.0%
0.0%
0.0%
PM25 [short tons]
National
Level
35,991
14,031
535
616
County
Level
36,477
14,112
552
636
%
Difference
-1.3%
-0.6%
-3.1%
-3.1%
SO2 [short tons]
National
Level
968
1,078
19
22
County
Level
1,040
1,159
20
24
%
Difference
-6.9%
-7.0%
-5.0%
-8.3%
Applications
Land-Based Diesel Engines
Recreational Marine Diesel
Engines
Year
2020
2030
2020
2030
VOC [short tons]
National
Level
74,423
64,329
1,328
1,528
County
Level
73,941
63,285
1,327
1,528
%
Difference
0.7%
1.6%
0.1%
0.0%
CO [short tons]
National
Level
247,593
133,606
5,538
6,465
County
Level
249,734
133,604
5,537
6,464
%
Difference
-0.9%
0.0%
0.0%
0.0%
-------
Table 3.6-13
Modeled 48-State Controlled Emissions
Emissions Impact Due to Changes in Control Scenario
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]
Proposed
637,025
410,084
34,817
41,250
819,201
814,827
612,722
534,520
Preliminary
477,100
222,238
34,817
41,250
819,201
814,827
612,722
534,520
Difference
159,925
(+33.5%)
188,264
(+84.7%)
0
(0.0%)
0
(0.0%)
0
(0.0%)
0
(0.0%)
0
(0.0%)
0
(0.0%)
PM25 [short tons]
Proposed
45,057
19,239
563
650
39,228
43,012
13,147
11,756
Preliminary
35,991
14,031
535
616
38,882
42,625
13,051
11,798
Difference
9,066
(+25.2%)
5,208
(+37.1%)
28
(+5.2%)
34
(+5.5%)
346
(+0.9%)
387
(+0.9%)
96
(+0.7%)
42
(-0.4%)
SO2 [short tons]
Proposed
1,005
1,096
395
469
3,251
3,653
5,449
5,959
Preliminary
968
1,078
19
22
184
206
272
305
Difference
37
(+3.8%)
18
(+1.7%)
376
(+1979%)
447
(+2032%)
3,067
(+1667%)
3,447
(+1673%)
5,177
(+1903%)
5,654
(+1854%)
-------
Table 3.6-13 (cont.)
Modeled 48-State Controlled Emissions Impact Due to Changes in Control Scenario
Applications
Land-Based Diesel
Engines
Recreational Marine
Diesel Engines
Commercial Marine
Diesel Engines
Locomotives
Year
2020
2030
2020
2030
2020
2030
2020
2030
VOC [short tons]
Proposed
78,822
67,861
1,328
1,528
37,290
41,354
36,546
31,644
Preliminary
74,423
64,329
1,328
1,528
37,290
41,354
36,546
31,644
Difference
4,399
(+5.9%)
3,532
(+5.5%)
0
(0.0%)
0
(0.0%)
0
(0.0%)
0
(0.0%)
0
(0.0%)
0
(0.0%)
CO [short tons]
Proposed
318,530
173,579
5,538
6,465
159,900
176,533
119,302
119,312
Preliminary
247,593
133,606
5,538
6,465
159,900
176,533
119,302
119,312
Difference
70,937
(+28.7%)
39,973
(+29.9%)
0
(0.0%)
0
(0.0%)
0
(0.0%)
0
(0.0%)
0
(0.0%)
0
(0.0%)
-------
Draft 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, ML 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 U-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 U-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. Size Specific Total P articulate Emission Factors for
Mobile Sources. EPA 460/3-85-005. August, 1995. (Docket A-2001-28, Document II-A-35)
8. 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/EmisInventory/draftnei99ver3/haps/documentation/nonroad/)
9. 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)
10. 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,
3-94
-------
Emissions Inventory
2003.
11. 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)
12. Power Systems Research, OELink (Sales) Database, October 2001 updated version.
3-95
-------
CHAPTER 4: Technologies and Test Procedures for Low-Emission Engines
4.1 Feasibility of Emission Standards 4-1
4.1.1 PM Control Technologies 4-1
4.1.1.1 In-Cylinder PM Control 4-2
4.1.1.2 Diesel Oxidation Catalysts (DOCs) 4-3
4.1.1.3 Catalyzed Diesel Particulate Filters (CDPFs) 4-4
4.1.2 NOx Control Technologies 4-15
4.1.2.1 In-Cylinder NOx Control Technologies 4-15
4.1.2.2 LeanNOx Catalyst Technology 4-16
4.1.2.3 NOx Adsorber Technology 4-16
4.1.2.4 Selective Catalytic Reduction (SCR) Technology 4-61
4.1.3 Can These Technologies Be Applied to Nonroad Engines and Equipment? 4-62
4.1.3.1 Nonroad Operating Conditions and Exhaust Temperatures 4-63
4.1.3.2 Durability and Design 4-72
4.1.4 Are the Standards Proposed for Engines >25 hp and <75 hp Feasible? 4-74
4.1.4.1 What makes the 25 - 75 hp category unique? 4-75
4.1.4.2 What engine technology is used today, and will be used for Tier 2 and Tier 3? 4-76
4.1.4.3 Are the proposed standards for 25 -75 hp engines technologically feasible? 4-76
4.1.5 Are the Standards Proposed for Engines <25 hp Feasible? 4-81
4.1.5.1 What makes the < 25 hp category unique? 4-82
4.1.5.2 What engine technology is currently used in the <25 hp category? 4-82
4.1.5.3 What data indicates the proposed standards are feasible? 4-82
4.1.6 Meeting the Crankcase Emissions Requirements 4-85
4.1.7 Why Do We Need 15ppm Sulfur Diesel Fuel? 4-86
4.1.7.1 Catalyzed Diesel Particulate Filters and the Need for Low Sulfur Fuel 4-87
4.1.7.2 Diesel NOx Catalysts and the Need for Low Sulfur Fuel 4-91
4.2. Supplemental Transient Emission Testing 4-94
4.2.1. Background and Justification 4-94
4.2.1.1 Microtrip-Based Duty Cycles 4-96
4.2.1.2 "Day in the Life"-BasedDuty Cycles 4-96
4.2.2. Data Collection and Cycle Generation 4-97
4.2.2.1. Test Site Descriptions 4-97
4.2.2.2 Engine and Equipment Description 4-99
4.2.2.3 Data Collection Process 4-102
4.2.2.4 Cycle Creation Process 4-102
4.2.3 Composite Cycle Construction 4-109
4.2.4 Cycle Characterization Statistics 4-111
4.2.5 Cycle Normalization / Denormalization Procedure 4-112
4.2.6 Cycle Performance Regression Statistics 4-113
4.2.7 Constant-Speed Variable-Load Transient Test Procedure 4-113
4.2.7.1 Background on Cycles Considered 4-114
4.2.7.2 Justification of Selections 4-115
4.2.8 Cycle Harmonization 4-116
4.2.8.1 Technical Review 4-116
4.2.8.2 Global Harmonization Strategy 4-118
4.2.9 Supplemental Cold Start Transient Test Procedure 4-126
4.2.10 Applicability of Component Cycles to Nonroad Diesel Market 4-128
4.2.10.1 Market Representation of Component Cycles 4-129
4.2.10.2 Inventory Impact of Equipment Component Cycles 4-129
4.2.10.3 HP and Sales Analysis 4-130
4.2.10.4 Broad Application Control 4-130
4.2.11 Final Certification Cycle Selection Process 4-131
4.3 Feasibility of Not-to-Exceed Standards 4-132
4.3.1 What EPA concerns do all NTE standards address? 4-132
4.3.2 How does EPA characterize the highway NTE test procedures? 4-133
4.3.3 How does EPA characterize the alternate NTE test procedures mentioned above? 4-133
4.3.4 What limits might be placed on NTE compliance under the alternate test procedures? 4-133
4.3.5 How does the "constant-work" moving average work, and what does it do? 4-136
4.3.6 What data would need to be collected in order to calculate emissions results using the alternate NTE? . . . 4-138
4.3.7 Could data from a vehicle's on-board electronics be used to calculate emissions? 4-139
4.3.8 How would anyone test engines in the field? 4-140
4.3.9 How might in-use crankcase emissions be evaluated? 4-140
4.3.10 How might the agency characterize the technological feasability for manufacturers to comply with NTE
standards? 4-140
-------
Technologies and Test Procedures for Low-Emission Engines
CHAPTER 4: Technologies and Test Procedures for Low-
Emission Engines
4.1 Feasibility of Emission Standards
This section of Chapter 4 documents the technical feasibility analysis we conducted in developing
the proposed Tier 4 emissions standards for nonroad diesel engine. The proposed standards and a
summary of this analysis can be found in Section in of the preamble. This analysis incorporates
recent Agency analyses of diesel emission control technologies for on-highway vehicles and expands
those analyses with more recent data and additional analysis specific to the application of technology
to nonroad diesel engines.1'2
The 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 horsepower categories (<25 hp, 25 - 75 hp, and >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 nonroad and on-highway diesels has concentrated on
improvements to the engine itself to limit the emissions leaving the combustion chamber.
However, during the past 15 years, 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 on-highway diesel vehicles based on the emission reduction potential of
these devices. Through use of these devices, we believe emissions control similar to that attained by
gasoline three-way-catalyst applications will be possible for diesel powered on-highway vehicles and
nonroad equipment. However, without low sulfur diesel fuel, these technologies cannot be
implemented.
4.1.1 PM Control Technologies
Particulate matter from diesel engines is made of three components;
- solid carbon soot,
- volatile and semi-volatile organic matter, 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
4-1
-------
Draft Regulatory Impact Analysis
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 form 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 which passes through the engine with no or only partial oxidation
and which 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 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.3 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 of the need for low sulfur fuel below.
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. A number of technologies exist that 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 on-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 in order to meet the new emission standards proposed here offers
an opportunity for significant reductions in engine-out PM emissions and possibly for reductions in
fuel consumption.
Another means to reduce the soot portion of diesel PM engine-out is to operate the diesel
(compression-ignited) engine with a homogenous method of operation rather than the typical
heterogenous 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.
4-2
-------
Technologies and Test Procedures for Low-Emission Engines
Homogenous diesel combustion has been under development for more than twenty years, yet it is
still unable to overcome a number of developmental issues.4'5 Fundamental among these issues is the
ability to control the start of combustion.6 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 pressures which
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.7'8 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 possible, while still providing the full range of engine operation. Other
approaches that are similarly limited to low load engine operation have been proposed in order to
produce a dual combustion mode engine.9'1011
4.1.1.2 Diesel Oxidation Catalysts (DOCs)
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 on-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.12
Today, highly active DOC formulations cannot be used for NMHC and CO control because of sulfur
in current diesel fuel which will lead to unacceptable sulfate PM emissions as discussed later in this
section.
4-3
-------
Draft Regulatory Impact Analysis
DOCs are also very effective at reducing the air toxic emissions from diesel engines. Test data
shows that emissions of toxics such as polycyclic aromatic hydrocarbons (PAHs) can be reduced by
more than 80 percent with a DOC.13
DOCs are ineffective at controlling the solid carbon soot portion of PM. The solid (soot)
typically constitutes 60 to 90 percent of the total diesel PM. Therefore, even with 15 ppm sulfur fuel
DOCs would not be able to achieve the level of PM control needed to meet the standard proposed
today.
4.1.1.3 Catalyzed Diesel Particulate Filters (CDPFs)
4.1.1.3.1 CDPF PM and HC Control Effectiveness
Emission levels from CDPFs are determined by a number of factors. Filtering efficiencies for
solid particle emissions like soot are determined by the characteristics of the PM filter, including wall
thickness and pore size. Filtering efficiencies for diesel soot can be as high as 99 percent with the
appropriate filter design.14 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/bhp-hr). This level of soot
emission control is not 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.
The SOF portion of diesel PM consists of primarily gas phase hydrocarbons in engine exhaust due to
the high temperatures and only forms particulate in the environment when it condenses. Catalytic
materials applied to CDPFs can oxidize a substantial fraction of the SOF in diesel PM just as the SOF
portion would be 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/bhp-hr),
compliance with the 0.01 g/bhp-hr emission standard proposed today may require additional
technology beyond the application of a CDPF system alone.A
Modern on-highway diesel engines have controlled SOF emission rates in order to comply with
the existing 0.1 g/bhp-hr emission standards. Typically the SOF portion of PM from a modern on-
highway diesel engine contributes less than 0.02 g/bhp-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.).15 Nonroad diesel
engines may similarly need to control engine-out SOF emissions in order to comply with the standard
proposed today. 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 engine-out SOF
control (i.e., engine-out SOF < 0.02 g/bhp-hr) and the application of catalytic material to the DPF,
A 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/bhp-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/bhp-hr standard once the other constituents of
diesel PM (soot and sulfate) are added. In this case, SOF emissions will need to be reduced engine-out or SOF
control greater than 90 percent will need to be realized by the CDPF.
4-4
-------
Technologies and Test Procedures for Low-Emission Engines
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/bhp-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.16 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.17
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
1 ndeno(1 23-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 have proposed today to limit
nonroad diesel fuel sulfur levels to be 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/bhp-hr over some
representative operating cycles using conventional diesel engine oils. This level of emissions will
allow for compliance with our proposed PM emissions standard of 0.01 g/bhp-hr, and we further
believe that there is room for reductions from this level in order to provide engine manufacturers with
additional compliance margin. During our 2002 Highway Progress Review, we learned that a number
of engine lubricating oil companies are working to reduce the sulfur content in engine lubricating oils.
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 in order to reduce SOF
emissions and secondarily to reduce sulfate PM emissions. While we believe that sulfate PM
4-5
-------
Draft Regulatory Impact Analysis
emissions will be the single largest source of the total PM from diesel engines, we believe with the
combination of technology, and the appropriate control of engine out PM, that sulfate and total PM
emissions will be low enough to allow compliance with a 0.01 g/bhp-hr standard, except in the case
of small engines with higher fuel consumption rates as described later in this section.
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.18 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 proposed NTE zone, as well as to the test modes which
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 proposed 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 PM emissions performance with the application of a CDPF over a number of
transient and steady-state test cycles.19 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 below. 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 would expect engine-out PM emissions to be lower for
production Tier 3 compliant diesel engines that will form the technology baseline for Tier 4 engines
meeting today's proposed standard. The engine demonstrated PM emissions of 0.009 g/bhp-hr on the
proposed Nonroad Transient Cycle (NRTC) from an engine-out level of 0.256 g/bhp-hr, a reduction
of 0.247 g/bhp-hr. The engine also demonstrated excellent PM performance on the existing steady-
state ISO Cl cycle with PM emissions of 0.010 g/bhp-hr from an engine-out level of 0.127, a
reduction of 0.107 g/bhp-hr. Thus, this engine would be compliant with the emission standard
proposed today for >75 hp variable speed nonroad engines.
4-6
-------
Technologies and Test Procedures for Low-Emission Engines
When tested on the proposed optional constant speed variable load cycle (CSVL) (to which this
engine would not be subject, under this proposal), the engine-out PM emission levels were 0.407
g/bhp-hr and were reduced to 0.016 g/bhp-hr (a reduction of 0.391 g/bhp-hr) with the addition of the
PM filter. As tested, this engine would not be compliant with the proposed optional CSVL standard,
but this is not surprising given that this Tier 1 engine was designed for variable speed engine
operation and not for single speed operation. We have great confidence given the substantial PM
reduction realized in this testing over the proposed CSVL cycle with a CDPF that a properly designed
nonroad diesel engine will be able to meet the standard of 0.01 g/bhp-hr.
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 Chapter 4.2 of this draft RIA 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/bhp-hr is
demonstrated on the AW2 cycle).
The results summarized in the two tables are indicative of the feasibility of the proposed NTE
provisions of 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 proposed NTE. 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/bh-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
proposed NTE standard. There are two other test points in Table 4.1-3 which are above the proposed
NTE emission level, both at 10 percent engine load. However, both are outside the proposed NTE
zone which excludes emissions for engine loads below 30 percent. It is important to note that
4-7
-------
Draft Regulatory Impact Analysis
although the engine would not be constrained to meet the proposed NTE under these conditions, the
resulting reductions at both points are still substantial in excess of 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.20 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 cycle would be compliant with the proposed
PM standard of 0.01 g/bhp-hr once the appropriate rounding convention is applied.8 It is also
interesting to note that the on-highway FTP transient emissions are higher than for either of the
proposed nonroad transient tests. This suggests that developing PM compliant engines on the
proposed nonroad transient cycles may not be substantially different from developing compliant
technologies for on-highway engines.
The rounding procedures in ASTM E29-90 are applied to the emission standard, therefore, the emission
results are rounded to the same number of significant digits as the specified standard, i.e., 0.014 g/bhp-hr is rounded
to 0.01 g/bhp-hr, while 0.015 g/bhp-hr would be rounded to 0.02 g/bhp-hr.
4-8
-------
Technologies and Test Procedures for Low-Emission Engines
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
emissions points (see table 4.1-5) have some of the highest brake specific emission rates. These rates
are not high enough however to preclude compliance with the steady-state emission cycle, are not
within the proposed NTE zone, and still show substantial PM reduction levels.
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%
The proposed NTE requirement, unlike the proposed nonroad transient cycle (NRTC) or the
existing ISO Cl cycle, is not a composite test. In fact, a number of the individual modes within the
Cl cycle test fall within the proposed NTE engine control zone. As discussed above, CDPFs are very
efficient at capturing elemental carbon PM (up to 99 percent), but sulfate-make under certain
4-9
-------
Draft Regulatory Impact Analysis
operating conditions may exceed the proposed NRTC or Cl standard of 0.01 g/bhp-hr, which is part
of the reason the proposed PM NTE standard is greater than the NRTC and Cl PM standards.
The proposed 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 proposed 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 proposed NTE requirements apply up to ambient
temperatures which are only 14°F greater than standard test cell conditions (100°F under the proposed
NTE, versus 86°F for Cl laboratory conditions). At an altitude of 5,500 feet above sea-level, the
proposed NTE applies 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 comply with the PM NTE 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 proposed NTE, we conclude that the proposed PM NTE standard for engines of 75
hp or higher is technologically feasible, provided low sulfur diesel fuel (15 ppm or lower) is
available. A discussion of the technical feasibility for engines with rated power lower than 75 hp is
given later in this chapter.
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).C 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 which is infrequently realized in normal diesel engine operation (for many engines'
exhaust temperatures may 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.21
However, the catalytic materials that most effectively promote soot and SOF oxidation 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
c 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).
4-10
-------
Technologies and Test Procedures for Low-Emission Engines
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 temperatures
compared to 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: are the exhaust temperatures high
enough on aggregate to oxidize the engine out PM 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 on-highway vehicles.22'23'24 However, it is possible that for some nonroad applications the
engine out PM 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
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 rates will decrease.
Similarly, techniques to raise exhaust temperatures periodically in order to initiate soot oxidation in a
PM filter have been developed for on-highway diesel vehicles as typified by the PSA system used on
more than 400,000 vehicles in Europe.25
During our 2002 Highway Diesel Progress Review, we investigated the plans of on-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.26 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 on-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).27 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 in order to comply with
the standards we are proposing today. We have estimated costs for such systems in our cost analysis.
D 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 rate?" the answer would be no, for all but a
very few nonroad or on-highway diesel engines.
4-11
-------
Draft Regulatory Impact Analysis
4.1.1.3.3 Current Status ofCDPF 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.28 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 US Heavy-Duty Federal Test Procedure (HD FTP), 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.E 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 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.29 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 to 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).30
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 assured. 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 soot oxidation effectiveness will also provide additional assurance that
excessive soot loading which could lead to PM filter failure will not occur.
At a recent conference of the Society of Automotive Engineers (SAE) a paper was presented that
documented improvements in catalyzed diesel particulate filter system design in order to improve
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.31
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.
Although 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
E 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.
4-12
-------
Technologies and Test Procedures for Low-Emission Engines
applications, there remains the possibility that for some conditions active regeneration systems
(backup systems) may be desirable. This is perhaps most likely for vehicles which are operated
primarily as passenger vehicles (light duty cars and trucks, and some light heavy-duty trucks). For
this reason a number of vehicle manufacturers have 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 a number of 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.32 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 by EPA in the FID2007 RIA but at that time little was known as to the extent of these
synergistic benefits.33 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.34 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 in order 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. However, both the industry and EPA would like to see
ash-related PM filter maintenance reduced as much as possible. EPA has specific guidelines for
4-13
-------
Draft Regulatory Impact Analysis
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 a number of 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. One simple approach to address ash is to increase
PM filter size in order to tolerate higher levels of ash accumulation. This approach, although
effective, is undesirable due to the added cost and size of the resulting PM filter. A number of
companies are investigating means to develop PM filter mechanisms which are more ash tolerant.
These approaches include concepts to increase storage area within the filter itself and concepts which
promote self-cleaning of the filter perhaps driven by engine and vehicle vibrations during normal
vehicle operation. It was not clear during our review that these technologies would be able to fully
address ash accumulation, but they were indicative of the potential to increase the interval between
necessary ash removal maintenance activities.
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 acid neutralizing additives
in engine oil should also decrease. The concept of an engine oil with less ash content is often referred
to as "low-ash oil." A number of 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 on-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 on-highway HD2007 emission technology engines would also be appropriate for use on Tier 4
emission technology 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 in
order to reduce PM emissions and to reduce operating costs for engine owners. Further
improvements in oil consumption may be possible in order to reduce ash accumulation rates in PM
filters. If oil accumulation rates could be halved and engine oil ash content similarly decreased, the
PM filter maintenance interval would be increased fourfold. Current retrofit PM filter ash
maintenance intervals can range from 50k miles to more than 200k miles.35
4-14
-------
Technologies and Test Procedures for Low-Emission Engines
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
A number of 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.36
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/bhp-hr NOx emission standard that we have proposed today.37 Some engine manufacturers
are today producing engines which utilize this technology over a narrow range of engine operation.38
Unfortunately, it is not possible today to apply this technology over the full range of diesel engine
operation. We do 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 which allows 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
4-15
-------
Draft Regulatory Impact Analysis
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.39 NOx reductions over the
HD transient FTP are only on the order of 12 percent due to excursions outside the optimum NOx
reduction efficiency temperature range for these devices.40 Consequently, the active lean NOx
catalyst does not appear to be capable of enabling the significantly lower NOx emissions required by
the proposed NOx standard.
The "passive" lean NOx catalyst uses no reductant injection. Therefore, the passive lean NOx
catalyst is 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 NOx catalysis listed here can
provide the significant NOx reductions necessary for compliance with the proposed 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.F 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 on-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). In today's action, we are proposing
similarly stringent NOx emission standards for nonroad engines greater than 75 hp, again 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. A number of 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.41'42'43 This method for NOx control has been shown
to be highly effective when applied to diesel engines but has a number of technical challenges
associated with it. Primary among these is sulfur poisoning of the catalyst as described in Section
4.1.2.3.x below.
F NOx adsorber catalysts are also called, NOx storage catalysts (NSCs), NOx storage and reduction catalysts
(NSRs), and NOx traps.
4-16
-------
Technologies and Test Procedures for Low-Emission Engines
4.1.2.3.1 How do NOx Adsorbers Work?
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 which
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.
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 some
oxygen (O2) which is not consumed during combustion. The reducing potential in the exhaust comes
from HC and CO emissions, which represent 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 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
4-17
-------
Draft Regulatory Impact Analysis
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. In order 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 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 in order 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
4-18
-------
Technologies and Test Procedures for Low-Emission Engines
control in stationary power systems and for lean burn gasoline engines. As outlined below we believe
that there are several approaches to accomplish the required periodic operation on a diesel engine.
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 in order 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 which 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 in order 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 in order
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 in
order 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 in order to
achieve a net fuel rich operating condition. That is, the primary plus secondary injection events
introduce an excess of fuel when compared to the amount of oxygen in the combustion chamber. The
secondary injection event occurs very late in the combustion cycle so that no torque is derived from
its introduction. This is necessary so that 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 could be done in a variety of ways, though they fall
into two general categories: predictive and reactive. The predictive method would estimate or
measure 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, a NOx regeneration would be performed. 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 would determine how accurately the predictive algorithm performed,
and adjust it accordingly. The reactive method is envisioned to monitor NOx downstream of the NOx
adsorber and, if NOx slippage is detected, a regeneration event would be triggered. This method is
dependent on good NOx sensor technology. This method would also depend on the ability to
4-19
-------
Draft Regulatory Impact Analysis
regenerate under any given engine operating condition, since the algorithm would be reacting 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
comply with the Tier 3 emission standards and will be virtually identical to the systems used by on-
highway 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 driver or equipment 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 in
order 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 that the
exhaust flow can be partitioned and routed through two or more catalyst "beds" which operate 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 in
order 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 would
have the following characteristics:
• Half of the system would operate with a major flow in an "adsorption mode", where most of
the exhaust is well lean of stoichiometric (X > 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.0
The other half of the system would have its exhaust flow restricted to just a small fraction (~5
percent) of the total flow and would operate 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.
- At these conditions, NOx can also be very efficiently reduced to N2 and O2 over a
precious metal catalyst.
G A condition of A = 1 means that there are precisely the needed quantity of reactants for complete reaction at
equilibrium. A < 1 means that there is insufficient oxygen, A > 1 means that there is excess oxygen.
4-20
-------
Technologies and Test Procedures for Low-Emission Engines
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
NOx
Adsorber
NOx
Adsorber
^
•^
Flnw-
Partially Closed
Exhaust-flow
Control Valve
Fully Open
Exhaust-flow
Control Valve
Secondary
Fuel Injector
(on)
Exhaust Flow
/ I Secondary
/vrcr Fuel Injector
(off)
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 would provide 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.44 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 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 could 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.45
4-21
-------
Draft Regulatory Impact Analysis
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 4.1.3.1.2
below to predict the effectiveness of the NOx adsorber technology on the proposed nonroad test
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 effective 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 which 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 which 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.
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.46 Thermal stability is the primary factor determining the NOx control efficiency of the NOx
adsorber at temperatures higher than 400-500°C. NOx adsorber catalyst developers are continuing to
4-22
-------
Technologies and Test Procedures for Low-Emission Engines
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. Therefore, even though NOx storage and release can occur at cooler
temperatures, NOx control is 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 would be as short as
30 seconds, at the very 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 would be expected to increase in inverse proportion to
the NOx emission rate from the engine. Therefore, the period of NOx storage under light load
conditions could likewise be expected to increase dramatically as well.
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 NVFEL,
especially with regards to NOx control under idle conditions. It was observed that when idle
4-23
-------
Draft Regulatory Impact Analysis
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). Additionally it was discovered that
the stored NOx could 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.
Additionally, 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 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 pictural
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 emission (grams)
over the on-highway heavy-duty FTP test for both a light heavy-duty (LFtD) and a heavy heavy-duty
(F£HD) 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/bhp-hr (the 1998 on-highway FID emission standard) 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 atNVFEL (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 could 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
4-24
-------
Technologies and Test Procedures for Low-Emission Engines
lowest in this region. This is due to the light-off temperature threshold for the NOx three-way
reduction function.
Figure 4.1-2
NOx Adsorber Efficiency Characteristics versus Exhaust Temperature
LHD Diesel Estimated NOx Adsorber Effectiveness over HD FTP
.100
- MECA "B" NOx Adsorber (%)
^| Engine Out NOx FTP (4g NOx Engine)
I I Projected FTP Tailpipe NOx
250
300
350
400
450
500
E
ro
5
"o
O
D.
LL
Q
I
(D
O
-------
Draft Regulatory Impact Analysis
Since the conversion efficiencies are based upon steady-state operation, it is likely that the low
temperature performance could be better than estimated here due to 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 to the FTP NOx efficiency for the LHD
engine shown in Figure 4.1-2 above 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 FTP. 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 heavy-duty FTP driving cycle. The discussion in 4.1.3.1.2 below shows
similarly that the proposed nonroad transient test 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 operation.47
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.
The steady-state analysis discussed above is based on steady-state emission results (i.e., after
exhaust temperatures have stabilized), but the proposed NRTC also includes a cold-start test where
the catalyst initial temperature will be at ambient conditions. The complete proposed NRTC test
sequence will include both a cold start emission test and a hot start emission test as described more
fully in 4.2. The NRTC emission level for the engine is determined by weighting the cold-start
emissions by 1/10 (10 percent), and weighting the hot-start emission results by 9/10 (90 percent).
4-26
-------
Technologies and Test Procedures for Low-Emission Engines
Historically, for on-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 exhaust emission control
devices, such as a NOx adsorber, the cold-start test will become a design challenge for on-highway
diesel engine manufacturers and, with this proposal, for nonroad diesel engine manufacturers, just as
it has been a design challenge for light-duty gasoline vehicle manufacturers for more than 20 years.
As discussed above, NOx adsorbers do have optimal temperature operating windows, and thus will
represent a design challenge.
Manufacturers have a number of tools available to them 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 a 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 light-duty gasoline catalysts, and it would allow the NOx
adsorber to store NOx for some period of time prior to 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, manufactures have a
number of active tools which can be used to enhance the cold-start performance of the system all
based on technologies which may be used to comply with the Tier 3 emission standards (i.e.,
technologies which will form the baseline for engines meeting the proposed Tier 4 standards). These
include the use of engine start-up routines which have a primary purpose of adding heat to the exhaust
to enhance NOx adsorber light-off For example:
• retarded injection timing;
• 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 of these tools in order to choose the best
combination necessary to minimize light-off time and improve the cold-start NRTC performance.
4-27
-------
Draft Regulatory Impact Analysis
On-highway manufacturers must overcome this same challenge in order to comply with the HD2007
emissions standards some number of years before these proposed nonroad emission standards go into
effect. Additionally, on-highway manufacturers must do this with a higher cold start weighting of 1/7
rather than 1/10 as proposed today for nonroad engines. This means that on-highway engine
manufacturers must have lower cold start emissions relative to their hot start emissions than will
nonroad engine manufacturers having to meet our proposed Tier 4 standards. Therefore, we believe
that the technologies we expect on-highway engine manufacturers to use for compliance (i.e., the
technologies delineated above) for the 2007 standards will be more than capable of being applied to
nonroad diesel engines in order to comply with the proposed Tier 4 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 NOx emissions engine out.48
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.
How effective are NOx adsorbers over the proposed NTE?
We are proposing an NTE standard for nonroad Tier 4 engines that replicates the provisions for
on-highway diesel trucks. A complete discussion of that proposal can be found in chapter 4.3. In
short, the proposal would set an NTE emissions limit that is 1.5 times the NRTC emissions limit over
a broad range of engine operating conditions. As discussed below, a 90 percent NOx reduction is
technologically feasible across the range of engine operating conditions and ambient conditions
subject to the proposed NTE standards. Also, as discussed below, some modifications to the
proposed NTE provisions to address technical issues which arise from the application of advanced
NOx catalyst systems were included in the HD2007 standards and are carried over into this proposal.
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 is test data on four different NOx adsorbers for
which extensive steady-state mapping was performed in order 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 proposed NTE NOx control zone, and would not be subject to the proposed
NTE standard. The following modes listed in these four figures are within the proposed NTE NOx
control zone, 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 proposed NTE zone. The region of the proposed
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, therefore it is not unexpected that the NOx
4-28
-------
Technologies and Test Procedures for Low-Emission Engines
reductions along the torque curve for the test engine are not as high as in other regions of the
proposed NTE zone. We would expect manufacturers to choose a NOx adsorber formulation which
matches the exhaust gas temperature operating range of 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 would likely not see periods of sustained operation along
the torque curve, and therefore the likelihood the NOx adsorber bed itself would achieve temperatures
in excess of 500°C would be diminished. 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.49'50'51 As discussed above, only small
improvements in the current characteristics are necessary in order to achieve 90 percent NOx
reductions or greater across the proposed NTE control zone.
As discussed above, the use of advanced NOx adsorber based catalyst systems will present cold-
start challenges for on-highway heavy-duty diesel engines, and for nonroad diesel engines, under our
proposed 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 in order to achieve the NOx standard.
The majority of engine operation which occurs within the proposed NTE control 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 which are within the proposed NTE control 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, though the proposed NTE does not include
engine start-up conditions, it is conceivable that a diesel which has not been warmed up could be
started and very quickly be operated under conditions which are subject to the proposed NTE
standard; for example, within a minute or less of vehicle operation after the vehicle has left an idle
state. The proposed NTE regulations specify a minimum emissions sampling period of 30 seconds.
Conceivably the vehicle emissions could be measured against the proposed NTE provisions during
that first minute of operation, and in all likelihood it would not meet the proposed 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,
operation with the proposed NTE 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.
Therefore, we have proposed to set a low temperature exhaust gas threshold of 250°C, below which
the specified NTE requirements do not apply, a provision we also adopted (for the same reason) for
on-highway engines in our HD2007 program, and we are proposing a similar provision for Tier 4.
The proposed 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 proposed NTE NOx standard. In general, it
can be said that the performance of the NOx adsorbers are only effected by the exhaust gas stream to
which the adsorbers are exposed. Therefore, the impact of ambient humidity, temperature, and
altitude will only effect the performance of the adsorber to the extent these ambient conditions change
the exhaust gas conditions (i.e., exhaust gas temperature and gas constituents). The ambient humidity
4-29
-------
Draft Regulatory Impact Analysis
conditions subject to the proposed 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, if any. The proposed NTE test procedure regulations specify an
upper bound on NTE testing for altitude at 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 on-highway Phasel (2004) final rule, Phase 1 technology
HDDEs (and thus similar Tier 3 nonroad diesel engines) can be designed to target air-fuel ratios at
altitude which will maintain appropriate exhaust gas temperatures, as well as maintain engine-out PM
levels near the 0.1 g/bhp-hr level, within the ambient conditions specified by the on-highway NTE
test procedure and thus the similar NTE procedure proposed today for Tier 4 nonroad engines.
Finally, the proposed NTE regulations specify ambient temperatures which are broader than the
NRTC temperature range of 68-86T. The proposed NTE test procedure specifies no lower ambient
temperature bounds. However, as discussed above, we have proposed to 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. Therefore, low ambient temperatures
will not present any difficulties for NTE NOx compliance. The proposed NTE also applies under
ambient temperatures which are higher than the laboratory conditions. The proposed NTE applies up
to a temperature of 100°F at sea-level, and up to 86°F at 5,500 feet above sea-level. At altitudes in
between, the upper proposed NTE ambient temperature requirement is a linear fit between these two
conditions. At 5,500 feet, the proposed NTE ambient temperature requirement is the same as the
upper end of the FTP temperature range (86°F), and therefore will have no impact on the
performance of the NOx adsorbers, considering that majority of the test data described throughout
this chapter was collected under laboratory conditions. The proposed 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 manufacturer to choose an adsorber
formulation which is matched to a particular engine design, and we would expect the small increase
in exhaust gas temperature which can occur from the expanded ambient temperature requirements for
the proposed NTE will be taken into account by the manufacturer 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 proposed 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 below (NRTC cycle) and in section 4.2 (CSVL cycle).
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 proposed NOx standard.
4-30
-------
Technologies and Test Procedures for Low-Emission Engines
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 understood well today. 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 on-highway technologies to nonroad engines as anticipated in this rulemaking, all of the
issues highlighted in this section while not addressed today, are expected to have already been
addressed for on-highway engines well before the start of this nonroad program.
In this section we will describe the major technical hurdles to address in order to ensure that the
significant emission reductions enabled through the application of NOx adsorbers is realized
throughout the life of nonroad diesel engines. The section is organized into separate durability
discussions for the system components (hardware) and various near and long term durability issues
for the NOx adsorber catalyst itself.
4.1.2.3.4.1 NOx Adsorber Regeneration Hardware Durability
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. Although 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 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 would normally be encountered on a diesel
4-31
-------
Draft Regulatory Impact Analysis
engine.52'53 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
parti culate 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 is expected to 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 would expect to cycle far fewer times over its life when compared to the
current long life of gasoline injectors. However, these gasoline fuel injectors 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. Thus, while benefitting from the
extensive experience with gasoline-based injectors a designer can, 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 in order to allow for NOx adsorber regeneration and desulfation.54
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 in order 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 in order 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 could be applied in order to
accomplish the flow diversion required for diesel NOx adsorber regeneration and desulfation. Since
temperatures will be typically cooler at the NOx adsorber compared to the inlet to the exhaust turbine
on a turbocharger, the control valve would be expected to be equally reliable when applied 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 in order 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.
4-32
-------
Technologies and Test Procedures for Low-Emission Engines
For example, contaminants from engine oil, like phosphorous or zinc, could attach to catalysts
sites partially blocking the site from the exhaust constituents and slowing reactions. Similarly, lead
added to gasoline in order 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. In order 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 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
4-33
-------
Draft Regulatory Impact Analysis
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.55
Figure 4.1-3
Comparison of NOx Conversion Efficiency before and after Desulfation
100
80
o 60
St
LU
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)
•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
4-34
-------
Technologies and Test Procedures for Low-Emission Engines
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 which was preceded by the sulfur poisoning
and degradation shown in the solid lines.
The increase in sulfur poisoning rate is important to understand in order to look at the means to
overcome the dramatic sulfur poisoning shown here. 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. Therefore, for a doubling in fuel sulfur levels a
corresponding doubling in the SOx poisoning rate would be predicted.
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).56
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. Therefore a separate sulfur release and reduction cycle
(desulfation cycle) will always be needed in order 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.57'58'59'60'61'62 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.
4-35
-------
Draft Regulatory Impact Analysis
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 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 n program controlled the air to fuel ratio and
catalyst inlet temperatures to achieve the high temperatures required to release the sulfur from the
device. Air to 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 could 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 n 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 to fuel ratio), and to the NOx
adsorber catalyst formulation.63'64 Lower air to 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.65 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 to fuel ratio for diesels of
30:1).66 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 would be approximately equivalent to 34 hours
of operation on seven 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.
4-36
-------
Technologies and Test Procedures for Low-Emission Engines
Figure 4.1-4
Flow Reactor Testing of a NOx Adsorber with Periodic Desulfations
95%
60%
9 10 11 12 13 14
# 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.67'68 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 to 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 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
4-37
-------
Draft Regulatory Impact Analysis
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.69 High temperatures, promote
sintering of platinum catalysts especially under oxidizing conditions.70 Therefore, it is 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 below, which shows the NOx conversion
efficiency of three NOx adsorber catalysts evaluated after extended periods of sulfur poisoning
followed by sulfur regeneration periods.71 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 an optimum desulfation
temperature exists which balances the tradeoffs between rapid sulfur regeneration and thermal
degradation (thermal sintering) at high temperatures.
4-38
-------
Technologies and Test Procedures for Low-Emission Engines
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
c
o
e
0)
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 below 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 would suggest as discussed
above, that the desulfation cycle can itself be damaging to the catalyst. In actual use we would expect
that an engine operating on 3 ppm sulfur fuel would not desulfate until well beyond a ten hour
interval and would 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. In order to ensure continued high performance when operating
on 78 ppm sulfur fuel the catalyst would require frequent desulfations. From the figure it can be
inferred that the desulfation events would need to be spaced at intervals as short as one to two hours
in order to maintain acceptable performance.
4-39
-------
Draft Regulatory Impact Analysis
Figure 4.1-6
Integrated NOx Conversion Efficiency following Aging and Desulfation
100%
80%
70%
Aging 10 hrs
: 3 ppm fuel
O
"•5
O
&
X
O
60%
50% -
40%
30% -
20%
10%
0%
3
I/)
3
I/)
in
0)
Q
• DECSE Catalyst S7
- Aged on 78 ppm S
• DECSE Catalyst S5
- Aged on 3 ppm S
0 10 20 30 40
Time (hours) cycle of 10 hrs Sulfur Aging / 6 min Desulfation
50
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 could have been higher.72
Based on the work in DECSE Phase n, the researchers concluded that:
• The desulfurization procedure developed has the potential to meet in-service engine operating
conditions and to provide acceptable driveability conditions.
• 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.
4-40
-------
Technologies and Test Procedures for Low-Emission Engines
• 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 ^
(0 =
0) O
O +3
90% -
80% -
70% -
60% -
50% -
O X
C O 40% -
JE ° 30% -
1C
LU
X 20% -I
O
10% -
0%
DECSE Catalyst S7
DECSE Catalyst S8
10 20 30 40 50 60 70
Desulfation Events (# of desulfation cycles)
The data available today on current NOx adsorber formulations shows 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. In order to ensure a gradual
and controllable decline in performance fuel sulfur levels must be minimized. 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
4-41
-------
Draft Regulatory Impact Analysis
California LEV and future Federal Tier 2 standards using platinum+rhodium (Pt+Rh) catalysts.
Initial designs were marked by unacceptable levels of platinum sintering which 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.73'74 Stabilization components can function in a number of
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 to the stabilization of NOx adsorber catalyst
components that is 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.75 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. Therefore the NOx adsorber when applied to diesel engines is expected to see both
lower average temperatures and lower peak temperatures when compared to 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 in order to limit the exposure to high temperatures.
Also, improvements in how the regeneration process occurs may provide avenues for improvement.
Low air to 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.
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 4.1-8.76 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
4-42
-------
Technologies and Test Procedures for Low-Emission Engines
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
c
O
o
to
X
o
50
10
20
30
40 50 60
DeSOx event
70
80
90
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 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. Therefore, the technology benefits substantially from the considerable experience
gained over the past 30 years with the highly reliable and durable three-way catalyst systems of today.
4-43
-------
Draft Regulatory Impact Analysis
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 in order 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 MY2007 and
later engines will be durable over the life of heavy-duty diesel vehicles, provided fuel with a 15 ppm
sulfur cap is used and that the technology will prove to be similarly durable when applied some years
later to nonroad diesel engines to comply with the proposed 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 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.77 Average NOx control for these power generation facilities is in excess of 92
percent.78 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.79 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
4-44
-------
Technologies and Test Procedures for Low-Emission 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.11 The
2000 model year saw the first U.S. application of this technology with the introduction of the Honda
Insight, certified to the California LEV-IULEV category standard. Table 4.1-6 below lists some of
the 2002 European lean-burn direct-injection gasoline vehicles which uses NOx adsorber catalyst
technology.80 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.
Table 4.1-6 2002 European Lean Burn Gasoline Direct-Injection Engines
Model
Audi A2 FSI
Audi A4 FSI
BMW 760 il_
Citroen C5 HPI
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 Galant 2.4 GDI
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(KW/PS)
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 (NVFEL)
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 the National
Vehicle and Fuel Emission Laboratory (NVFEL) revealed that formulations were capable of reducing
NOx emissions by more than 90 percent over the broad range of operation in the on-highway steady-
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.
4-45
-------
Draft Regulatory Impact Analysis
state SET procedure (sometimes called the EURO 4 test). At operating conditions representative of
"road-load" operation for a heavy duty on-highway truck, the catalysts showed NOx reductions as
high as 99 percent resulting in NOx emissions well below 0.1 g/bhp-hr from an engine out level of
nearly 5 g/bhp-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. Although not included in the test results shown in figures 4.1-10 through 4.1-12, the
ISO-C1 modes are closely approximated by a number of other test modes as can be seen in figure 4.1-
9. Therefore, we would expect similarly good performance on the ISO-C1 test modes. Testing on the
on-highway transient test procedure has shown similarly good results, with hot start FTP NOx
emissions 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 FTP 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 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
only be able to run 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 on only 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 set up,
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.
4-46
-------
Technologies and Test Procedures for Low-Emission Engines
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 emissions 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.81
Test Results — Single Leg NOx Adsorber System
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 prediction.1 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 which have duplicate mode numbers. The on-highway NTE (which, of course, is the same as
the proposed nonroad NTE) zone is also shown in Figure 4.1-9 to show that these two sets of modes
give comprehensive coverage of the proposed 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 proposed 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.
1 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 the transient FTP.
4-47
-------
Draft Regulatory Impact Analysis
Figure 4.1-9 Steady-State Test Modes from NVFEL Testing and ISO C-l Modes
700
600 --
500 -
400 -
Torque
NOx NTE
PM NTE
» AVL Modes
A SET modes
• ISOC-1 Modes
300
200
100
1200 1400
1600 1800 2000
Speed (rpm)
2200
2400
2600
2800
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 FE impact. 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. It should be noted 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 proposed 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 HC and FE impacts. But, like EPA mode 1, EPA mode 2 would adsorb for extended periods of
time without need for 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 in order to raise exhaust temperatures at these
operating modes.
4-48
-------
Technologies and Test Procedures for Low-Emission Engines
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 FE 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. It should be noted
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 engine-out emission prediction of the HDDE
transient 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 on-highway HDDE transient cycle. With exhaust
emission control devices, it loses some of its accuracy because of the inability of the emission control
devices to be regenerated at the low temperature modes (EPA modes 1, 2, 5). In real world
conditions, the HDDE does not come to steady-state temperatures at any of these 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 HDDE transient cycle performance is expected
to be much better than the composites would suggest (See 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.
4-49
-------
Draft 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%
C^ 80%
o
C
-------
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%
g.
u
C
01
'o 60%
E
UJ
c
o
1!
OL
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
4-51
-------
Draft 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
E
UJ
c
o
OL
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
4-52
-------
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
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
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
NOx Red
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%*
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% -i
u
C
0)
'o
E
UJ
C
o
'u
I
OL
80%
60%
40%
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
4-53
-------
Draft 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 RIA.82 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
HD2004 levels and the expected Tier 3 levels) 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 HD FTP transient test cycle.83 As discussed further in
4.1.3.1.2 below, performance over the NRTC is projected to be better than for the on-highway HD
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),
therefore emissions reductions would be 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 (X < 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.84
The transient HDDE FTP regeneration control was accomplished using a time-based regeneration
schedule. This control regenerated on a prescribed schedule of time and fuel quantities so that
regenerations occurred at predetermined engine conditions during the transient cycle.
The transient HDDE FTP results presented here are for hot-start cycles only. The adsorber system
was not optimized for cold start performance and would not provide a meaningful assessment of
adsorber warmup performance. In order to better simulate the "cold-soak-hot" procedure called for in
the HDDE FTP, a preconditioning mode was chosen to provide adsorber temperatures at the start of
the "hot" cycle that would be 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.
Test Results — Dual Leg NOx Adsorber System
The on-highway SET is made up of the 13 Euro in modes. Several modes were run twice by
different engineers, and the best calibration was chosen for the SET composite. Table 4.1-7 shows
4-54
-------
Technologies and Test Procedures for Low-Emission Engines
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.
Figure 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 FE 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 FE impact varied from virtually 0 to
2.3 percent depending on the mode with a composite FE 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.
HOPE Transient FTP Test Results
As with the steady-state test results, the hot-start FTP test results showed NOx and PM emissions
in excess of 90 percent. The baseline (without the catalyst system) NOx emissions of 2.7 g/bhp-hr
were reduced to 0.1 g/bhp-hr with the addition of the catalyst system, a better than 95 percent
reduction in NOx emissions. Similarly, the PM emissions were reduced to below 0.003 g/bhp-hr
from a baseline level of approximately 0.1 g/bhp-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 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).85
Durability Baseline NOx Adsorber Catalyst Testing
4-55
-------
Draft Regulatory Impact Analysis
Additional testing was conducted at NVFEL to provide baseline performance data to gauge
improvements in NOx adsorber durability performance in support of the 2007 Highway technology
reviews. The data provides 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.86 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 2 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.87 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
below. Efforts have also been made to reduce the cost of the system by using simpler injectors and
valve actuators.
Figure 4.1 -14 Prototype 4-leg System Compared to 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
4-56
-------
Technologies and Test Procedures for Low-Emission Engines
underway to confirm these steady-state results and to demonstrate the performance over transient
cycles.
Figure 4.1-15 Preliminary Results for Prototype 4 Leg System
100% , - . "°de.FE Penalty 2* ^8 3.3 3J
93%
10 11 12 13 CR
SET Mode
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 is itself one 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.88'89'90 91'92
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.93
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, emissions performance and
durability.94
4-57
-------
Draft Regulatory Impact Analysis
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 HD2007 emission standards including the NOx adsorber technology. As
noted in the 2002 Highway Diesel Progress Review, which documents in more detail progress by the
on-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 2003.95 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
4.1.2.3.5.3 by EPA, a dual leg system demonstrated 90 percent or higher NOx emissions control over
a wide range of operation.
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 three prototype diesel passenger cars with these technologies over the last
year. The results shown in figure 4.1-16 below demonstrate the potential for substantial reductions
with NOx adsorber and PM filter technologies when tested with low sulfur diesel fuel. All three
vehicles demonstrated substantial reductions in NOx and PM emissions when compared to a current
day relatively clean (compared to only a few years ago) diesel passenger car as represented by the
large black diamond on figure 4.1-16.
4-58
-------
Technologies and Test Procedures for Low-Emission Engines
Figure 4.1-16 Tier 2 Passenger Car Prototypes Tested at NVFEL on the FTP75 Cycle
0.8 -
0.7 -
0.6 -
§0.5-
o
'ft
X
O
.
Toyota Avensis DPNR Station Wagon 2.0 L
turbocharged Dl diesel, HPCR, cooled EGR, LTC,
DPNR, DOC
• Vehicle "B" 1.9 L turbocharged Dl diesel, HP inj.,
cooled EGR, NOx ads., PM-trap
Vehicle "D" 3.0 L turbocharged Dl diesel, HP inj.,
cooled EGR, NOx ads., PM-trap (unaged catalyst)
• Vehicle "D" 3.0 L turbocharged Dl diesel, HP inj.,
cooled EGR, NOx ads., PM-trap (100,000 km catalyst
aging)
* Vehicle "E" cooled EGR, DOC
•
•
•
0.01
0.02
0.03
PM Emissions (g/mi)
0.04
0.05
0.06
One vehicle in the test program, vehicle "D," 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 would appear 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 emissions levels shown in the figure, it only takes a very small breakthrough in
NOx emissions to significantly increase the emissions over the lowest control levels. At the current
time manufacturers are working to keep the number of regeneration episodes to the minimum number
in order to minimize stress on catalyst materials (i.e., limit thermal degradation as discussed in
section 4.2 above). We believe that 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 that 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 to the new system.
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
4-59
-------
Draft Regulatory Impact Analysis
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.96 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 today's proposed NOx
standards.
We have significant concerns regarding a technology that requires extensive user intervention in
order 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. Therefore,
unless the urea tank is prohibitively large, the urea must 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. Today, there is not a system in place 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.97
We are not aware of a proven mechanism that ensures that the user will replenish the urea supply
as necessary to maintain emissions performance. Further, we believe given the additional cost for
urea, that there will be significant disincentives for the end-user to replenish the urea because the cost
of urea could 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 proposed standards. Therefore, we have not based the feasibility or cost
analysis of this emission control program on the use or availability of the urea SCR technology.
However, we would not preclude its use for compliance with the emission standards provided that a
manufacturer could demonstrate satisfactorily to the Agency that urea would be used under all
conditions. We believe that only a few unique applications will be able to be controlled in a manner
such that urea use can be assured, and therefore believe it is inappropriate to base a national emission
control program on a technology which can serve effectively only in a few niche applications.
This section has described a number of 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 HI 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
standards that we are proposing today for engines of 75 hp, or greater, will begin to go into effect four
years later. This time lag between equivalent on-highway and nonroad diesel engine standards is
necessary in order to allow time for engine and equipment manufacturers to further develop these on-
highway technologies for nonroad engines and to align this program with nonroad Tier 3 emission
standards that begin to go into effect in 2006.
4-60
-------
Technologies and Test Procedures for Low-Emission Engines
The test procedures and regulations for the HD2007 on-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 2007 highway emission standards would comply with the equivalent nonroad
emission standards proposed today if it were to be tested over the transient and steady-state nonroad
emission test procedures proposed today, which cover the same regions and types engine operation.
Said in another way, an on-highway diesel engine produced in 2007 could be certified in compliance
with the transient and steady-state standards proposed today for nonroad diesel engines several years
in advance of the date when these standards would go into effect. However, that engine, while
compliant with certain of the nonroad emission standards proposed today, 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
on-highway diesel engines or produce nonroad engines that are developed from on-highway products.
Therefore, there is 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 which 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 of these issues in setting the emission standards and the timing of those
standards as proposed today.
This section describes some of the challenges to applying advanced emission control technologies
to nonroad engines and equipment, and why we believe that technologies developed for on-highway
diesel engines can be further refined to address these issues in a timely manner for nonroad engines
consistent with the emission standards proposed today.
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 only important 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,
4-61
-------
Draft Regulatory Impact Analysis
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.98 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. Therefore, there is 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 on-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 on-highway engine manufacturers and has been a focus of our 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. On-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 on-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 in order to comply with 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
4-62
-------
Technologies and Test Procedures for Low-Emission Engines
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 below. In the case of the Cummins engine, the testing was done at three different engine
ratings (250hp, 169hp, and 124hp) in order 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 earlier in this chapter, passive filter regeneration occurs when the exhaust
temperatures are high enough that on aggregate the PM accumulation rate on the filter is less 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." Another CDPF manufacturer has stated that catalyzed diesel paniculate 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.100
EPA used the more restrictive of these guidelines to evaluate the likelihood that passive
regeneration would be realized on a number of 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
4-63
-------
Draft Regulatory Impact Analysis
60 percentile temperature exceeded 250°C for most of the engine tests on all three engines. The runs
which did not result in at least 250°C for 40% of the duty cycle were from the on-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.
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)
169 hp (test #1529)
124 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
> 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 > 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 > 275°C
26%
48%
24%
15%
97%
4-64
-------
Technologies and Test Procedures for Low-Emission Engines
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 may be that passive
filter regeneration is not realized. In the case of those operating conditions, we believe that active
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 is contained in a recent report from the Engine Manufacturers Association (EMA) and the
European Association of Internal Combustion Engine Manufacturers (Euromot).101 This report
contains data from a range of applications and power categories. The similar to the data presented
above, the EMA/Euromot indicates that while a number of nonroad applications do generate
temperatures high enough to passively regenerate a filter, there are also a number of applications
which would require active regeneration.
We have assumed in our cost analysis that all nonroad engines complying with a PM standard of
0.02 g/bhp-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 regeneration system, although 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 on-highway diesel engines. We do not believe that there are any nonroad engine applications
above 25 horsepower 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
in order to match the engine performance characteristics to the full range of nonroad equipment.
We believe that given the timing of the emissions standards proposed today, and the availability
and continuing development of technologies to address temperature management for on-highway
engines which technologies are transferrable to all nonroad engines with greater than 25 hp power
rating, that nonroad engines can be designed to meet the proposed standards in a timely manner.
Matching the operating temperature window of the broad range of nonroad equipment may be
somewhat more challenging for nonroad engines than for many on-highway diesel engines simply
because of the diversity in equipment design and equipment use. Nonetheless, the problem has been
successfully solved in on-highway applications facing low temperature performance situations as
difficult to address as any encountered faced by nonroad applications. The most challenging
temperature regime for on-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
emissions control 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.102 Similarly, PSA has
demonstrated effective CDPF regeneration under demanding light-load taxi cab conditions with
4-65
-------
Draft Regulatory Impact Analysis
current production technologies.103 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 nonroad equipment in-use operating cycles which would be considered to
be more demanding of low temperature performance than on-highway urban driving. Both the
Toyota and PSA systems are designed to function even with extended idle operation as would be
typified by a taxi waiting to pick up a fare/ 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 as would be typified by the broad range of in-use nonroad duty cycles
and the new nonroad transient test proposed today.
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 emission
standards proposed today. 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
PM standard proposed today, especially in the case of engines which may be able to comply with the
proposed NOx standards with engine-out emission control technologies (i.e., engines rated between
25 and 75 horsepower).
4.1.3.1.2 NOx Adsorbers and Nonroad 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
J 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 400,000 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.
4-66
-------
Technologies and Test Procedures for Low-Emission Engines
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 on-highway FTP test cycle. We have extensive
experience testing NOx adsorber catalyst systems on the on-highway FTP procedure (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 on-highway FTP cycle. To accomplish that, we tested a
Cummins ISB (see Table 4.1-8 above) 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 below.
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
t The estimates are based on the absorber B curve shown in Figure 4.1-11.
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
t The estimates are based on the absorber B curve shown in Figure 4.1-11.
4-67
-------
Draft Regulatory Impact Analysis
Table 4.1-14
Estimated NOx Adsorber Efficiency on Cummins ISB - FTP Cycle
Engine Power (hp)
124
169
250
6" from turbo outlet (%)
60.3
72.4
83.0
A 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 to that achieved from the on-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 to the on-highway
FTP cycle, thus indicating that the transfer of on-highway 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 proposed 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 in order to
understand what levels of power density to expect in the future for nonroad diesel engines. For this
analysis, data from Power Systems Research 2002 database (PSR) was examined. 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 a number of power categories
from 1985 to 2000. The PSR database reports both rated power and engine displacement, from which
power was calculatedK. The data was 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.
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/liter.
4-68
-------
Technologies and Test Procedures for Low-Emission Engines
Table 4.1-15
Sales Weighted Power Density, 1985 - 2000
Year
1985
1986
1987
1988
1989
1990
1991
1992
1993
1994
1995
1996
1997
1998
1999
2000
% Change
1 985 - 2000
Sales Weighted Fower Density by Fower Category (hp/literj
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
1 5%
600hn+
27.;
27.(
27.<
28.
31. <
32.'
33.'
35.(
33. <
34.'
35.:
35.;
35.'
35.:
34.<
34.<
2 I'M
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.
4-69
-------
Draft Regulatory Impact Analysis
Figure 4.1-17 Power Density Trends for Nonroad Diesel Engines
1985 - 2000, >50 horsepower engines
1985
1987
1989
1991
1993
1995
1997
1999
Year
-50-100hp -A-100-175hp —O— 175-300hp -X-300-600hp -O-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
on-highway diesel vehicles. If no consideration is given to differences in operating conditions in
engine and equipment design eventual failure of the equipment would be expected.
Specific efforts to design for the nonroad operating conditions will be required in order 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 emissions
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.
Deutz, a nonroad engine manufacturer, sold approximately 2,000 diesel particulate filter systems
for nonroad equipment in the period from 1994 through 2000. Many of these systems were sold for
use in mining equipment. No other applications are likely to be more demanding than this. Mining
4-70
-------
Technologies and Test Procedures for Low-Emission Engines
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.104
The Deutz system utilized a conventional cordierite PM filter substrate as is commonly used for
heavy-duty on-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 in order 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.105 We believe that 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 which 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 do not believe that any new or fundamentally different solutions will need to be invented
in order to address the vibration and impact constraints for nonroad equipment. Our cost analysis
includes the hardware costs for mounting and shielding the aftertreatment equipment as well as the
engineering cost for equipment redesign.
Certain nonroad applications, including some forms of harvesting equipment and mining
equipment, may have specific limits on maximum surface temperature for equipment components in
order 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 that 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 already commonly experienced by diesel engines (i.e,
catalyst temperatures are expected to be below 800°C).L CDPF temperatures are not expected to
exceed approximately 700°C in normal use and are expected to only reach the 650°C temperature
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.F 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 emission system function then supplemental heating as described earlier may be necessary
but would not be expected to produce temperatures higher than the maximum levels normally
L 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.
4-71
-------
Draft Regulatory Impact Analysis
encountered in diesel exhaust. Furthermore, even if it were necessary to raise exhaust temperatures to
a higher level in order to promote effective emission control, there are technologies available to
isolate the higher exhaust temperatures from flammable materials such as dust. One approach would
be 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 in order to maintain higher exhaust temperatures inside the catalyst in order to promote
better catalyst function. Another technology to control surface temperature already used by some
nonroad equipment manufacturers is water cooled exhaust systems.106 This approach is similar to the
air-gapped system but uses engine coolant water to actively cool the exhaust system. We do not
believe that flammable dust concerns will 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. 107
We agree that nonroad equipment must be designed to address durable performance for a wide
range of operating conditions and applications that would not commonly be experienced by on-
highway vehicles. We believe further as demonstrated by retrofit experiences around the world that
technical solutions exist which allow catalyst based emission control technologies to be applied to
nonroad equipment.
4.1.4 Are the Standards Proposed for Engines >25 hp and <75 hp Feasible?
As discussed in Section in of the preamble, our proposal for standards for engines between 25
and 75 hp consists of a 2008 transitional standard and long-term 2013 standards. The proposed
transitional standard is a 0.22 g/bhp-hr PM standard. The 2013 standards consist of a 0.02 g/bhp-hr
PM standard and a 3.5 g/bhp-hr NMHC+NOx standard. The transitional standard is optional for 50-
75 hp engines, as the proposed 2008 implementation date is the same as the effective date of the Tier
3 standards. Manufactures may decided, at their option, not to undertake the 2008 transitional PM
standard, in which case their implementation date for the 0.02 g/bhp-hr PM standard begins in 2012.
In addition, we have proposed a minor revision to the CO standard for the 25-50 hp engines
beginning in 2008 to align these engines with the 50-75 hp engines. This proposed CO standard is
3.7 g/bhp-hr.
The remainder of this section discusses:
- what makes the 25-75 hp category unique;
- what engine technology is used today, and will be used for applicable Tier 2 and Tier 3
standards; and,
- why the proposed 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 a number of common traits with highway diesel engines. These include
4-72
-------
Technologies and Test Procedures for Low-Emission Engines
similarities in displacement, aspiration, fuel systems, and electronic controls. Table 4.1-16 contains a
summary of a number of key engine parameters from the 2001 engines certified for sale in the U.S.M
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
(estimated)
Percent of 2001 U.S. Production"
0-25 hp
83%
17%
0%
47%
not available
today
25-75 hp
47%
53%
7%
3%
limited availability
today
75-100 hp
4%
96%
62%
0%
available today
>100 hp
<0.1%
>99%
91%
0%
commonly available
today
1 Based on sales weighting of 2001 engine certification data
As can be seen in Table 4.1-16, the engines in the 25-75 hp category have a number of technology
differences from the larger engines. These include a higher percentage of indirect-injection fuel
systems, and a low fraction of turbocharged engines. (The distinction in the <25 hp category is quite
different, with no turbocharged engines, nearly one-half 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, 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). The research and development work being performed today for the heavy-duty highway
market is targeted at engines which are 4-cylinders or more, direct-injection, electronically controlled,
turbocharged, and with per-cylinder displacements greater than 0.5 liters. As discussed in more detail
below, 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 for the 25-75 hp
category (as well as for the <25 hp category).
4.1.4.2 What engine technology is used today, and will be used for Tier 2 and Tier 3?
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. Therefore, all engines in the
M
Data in Table 4.1-16 is 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.
4-73
-------
Draft Regulatory Impact Analysis
25-75 hp range have been meeting Tier 1 standards for the past several years, and the data presented
in Table 4.1-17 represent performance of Tier 1 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.108 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.109
We expect a significant shift in the engine technology which 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.110 In addition,
we have heard from some engine manufactures 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
which are built on the same engine platform. For the Tier 2 standards for the 25-50 hp products, a
large number of engines meet these standards today, and therefore we expect to see only moderate
changes in these engines, including the potential additional use of turbocharging on some models.111
4.1.4.3 Are the proposed standards for 25 -75 hp engines technologically feasible?
This section will discuss the feasibility of both the proposed 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 meet Tier 1
NMHC+NOx and PM standards today. We have examined the model year 2002 engine certification
data for engines in the 25-50 hp category.112 A summary of this data is presented in Table 4.1-17.
These data indicate that over 10 percent of the engine families meet the proposed 2008 0.22 g/bhp-hr
PM standard and 5.6 g/bhp-hr NMHC+NOx standard (unchanged from Tier 2 in 2008) today. 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.N Five engine families are more than
20 percent below the proposed 0.22 g/bhp-hr PM standard, and an additional 24 engine families
which already meet the 2008 NMHC+NOx standards would require no more than a 30 percent PM
reduction to meet the proposed 2008 PM standards. Unfortunately, similar data do not exist for
engines between 50 and 75 hp. There is no Tier 1 PM standard for engines in this power range, and
therefore engine manufacturers are not required to report PM emission levels until Tier 2 starts in
2004. However, in general, the 50-75 hp engines are more technologically advanced than the smaller
horsepower engines and would be expected to perform as well as, if not better than, the engines in the
25 - 50 hp range.
N 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.
4-74
-------
Technologies and Test Procedures for Low-Emission Engines
Table 4.1-17
2002 Model Year Certification Data for 25-50 hp Nonroad Diesel Engines
JM Emissions Relative to
3roposed 0.22 g/bhp-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
> 2008 NMHC+NOx std.
Total # ot Engine Families
1D1 Engines
(test cycle/aspiration)
5 -mode/
NA
0
1
2
3
2
8
8-mode/
NA
0
5
1
15
17
38
5 -mode/
TC
0
1
0
0
1
2
8-mode/
TC
0
2
1
4
3
10
Dl Engines
(test cycle/aspiration)
5 -mode/
NA
0
0
0
0
8
8
8-mode/
NA
1
0
1
2
40
44
8-mode/
TC
0
0
0
0
8
8
Totals
1
c
*;
24
IS
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 a number of engines in this power range can already achieve the proposed 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. Despite being certified to the same
emission standards with similar engine technology, the emission levels from these engines vary
widely. Figure 4.1-18 is a graph of the model year 2002 HC+NOx and PM data for engines in the 25-
50 hp range. As can be seen in Figure 4.1-18, the emission levels cover a wide range. The figure
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/bhp-hr, the PM rates vary from approximately 0.2 to more than 0.5 g/bhp-
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 proposed 2008 PM standard on the combination of
diesel oxidation catalysts and the lowest engine-out emissions being achieved today, because it is
uncertain whether or not additional engine-out improvements would lower all engines to the proposed
2008 PM standard. Instead, we believe there are two likely means by which companies can comply
with the proposed 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. Therefore, we believe some engine companies will choose to use a
combination of engine-out techniques and diesel oxidation catalysts, as discussed below.
4-75
-------
Draft Regulatory Impact Analysis
Figure 4.1-18 Emission Certification Data for 25-50 HP Model Year 2002 Engines
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
For those engines which do not already meet the proposed 2008 Tier 4 PM standard, a number of
engine-out technologies are available to achieve the standards by 2008. In our recent Staff Technical
Paper on the feasibility of the Tier 2 and Tier 3 standards, we projected that in order to comply with
the Tier 3 standards, engines greater than 50 hp would rely on some combination of a number of
technologies, including electronic fuel systems such as electronic rotary pumps or common-rail fuel
systems.113 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.114 Even for mechanical fuel systems, increased
injection pressures can reduce PM emissions substantially.115 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 proposed 2008 PM standard. As demonstrated
by the 2002 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 proposed
standards for many of the engines in the 25-75 hp category without the need to add turbocharging or
electronic fuel systems.
For those engines which are not able to achieve the proposed standards with known engine-out
techniques, we project that diesel oxidation catalysts can be used to achieve the proposed standards.
DOCs are passive flow-through emission control devices which 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
4-76
-------
Technologies and Test Procedures for Low-Emission Engines
heavy-duty diesel applications. In addition, DOCs have already been used to control PM or carbon
monoxide on some nonroad applications.116
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 paniculate sulfates.
However, even with today's 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/bhp-hr standard).117 However, given the high
level of sulfur in nonroad fuel today, 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 shows
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.118 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 emissions necessary to meet
the proposed standards for most engine manufacturers. With the availability of 500 ppm sulfur fuel,
DOC's can be designed to provide PM reductions on the order of 20 to 50%, while suppressing
particulate sulfate reduction.119 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 proposed 2008 PM standard and are at or below the 2008
NMHC+NOx standard for this power range, indicating that use of DOCs should readily achieve the
incremental improvement necessary to meet the proposed 2008 PM standard.
As discussed in Section in of the preamble, we have also proposed a minor change in the CO
standard for the 25-50 hp engines, in order to align it with the standard for the 50-75 hp engines.
This small change in the CO standard is intended to simplify EPA's regulations as part of our
decision to propose a reduction in the number of engine power categories for Tier 4. The current CO
standard for this category is 4.1 g/bhp-hr, and the proposed standard is 3.7 g/bhp-hr (i.e., the current
standard for engines in the 50-75 hp range). The model year 2002 certification data shows that more
than 95 percent of the engine families in the 25-50 hp engine range meet the proposed CO standard
today. In addition, a recent EPA test program run by a contractor on two nonroad diesel engines in
this power range showed that CO emissions were well below the proposed standards not only when
tested on the existing steady-state 8-mode test procedure, but also when tested on the nonroad
transient duty cycle we are proposing in today's action.120 Finally, DOCs typically reduce CO
emissions on the order of 50 percent or more, on both transient and steady-state duty cycles.121 Given
that more than 95 percent of the engines in this category meet the proposed standard today, and the
ready availability of technology which can easily achieve the proposed standard, we project this CO
standard will be achievable by model year 2008.
4.1.4.3.2 2013 Standards
For engines in the 25-50 range, we are proposing standards commencing in 2013 of 3.5 g/bhp-hr
for NMHC+NOx and 0.02 g/bhp-hr for PM. For the 50-75 hp engines, we are proposing a 0.02
g/bhp-hr PM standard which 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).
4-77
-------
Draft Regulatory Impact Analysis
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 proposed PM standard for engines >75
hp. As also discussed in Section 4.1.3, PM filters may require active back-up regeneration systems
for many nonroad applications. A number of secondary technologies are likely required to enable
proper regeneration, including possibly electronic fuel systems such as common rail systems which
are capable of multiple post-injections which can be used to raise 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 CDPF's 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 higher for the smaller engines,
however, we have addressed this issue in our proposal. The PM filter regeneration systems described
in 4.1.1 and 4.1.3 are also applicable to engines in this size range and are therefore likewise feasible.
There are specific trap regeneration technologies which we believe engine manufacturers in the 25-75
hp category may prefer over others. Specifically, an electronically-controlled secondary fuel injection
system (i.e., a system which 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.122
We are, however, proposing a slightly higher PM standard (0.02 g/bhp-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 results in higher fuel
consumption. Because of the higher fuel consumption rate, we expect a higher particulate sulfate
level, and therefore we have proposed a 0.02 g/bhp-hr standard.
Test data confirm that this proposed standard, as well as the proposed NTE of 1.5 times the
standard, are achievable. In 2001, EPA completed a test program run by a contractor on two small
nonroad diesel engines (a 25 hp IDI engine and a 50 hp IDI engine) which demonstrated the proposed
0.02 g/bhp-hr standard can be achieved with the use of a CDPF.123 This test program included testing
on the existing 8-mode steady-state test cycle as well as the nonroad transient cycle proposed in
today's action. The 0.02g/bhp-hr level was achieved on each engine over both test cycles. In
addition, the 0.02 g/bhp-hr level was achieved on a variety of nonroad test cycles which are intended
to represent several specific applications, such as skid-steer loaders, arc-welders, and agricultural
tractors. We believe these data are indicative of the robust emission reduction capability of
particulate filters and demonstrates the proposed NTE standard of 1.5 x 0.02 g/bhp-hr standard (0.03
g/bhp-hr) can be achieved using the proposed not-to-exceed test requirements. This test program also
4-78
-------
Technologies and Test Procedures for Low-Emission Engines
demonstrates why EPA has proposed a slightly higher PM standard for the 25 - 75 hp category (0.02
g/bhp-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.
4.1.4.3.2.2 NMHC+NOx Standard
We have proposed a 3.5 g/bhp-hr NMHC+NOx standard for engines in the 25 - 50 hp range for
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 are implemented in 2008. EPA's recent Staff
Technical paper which reviewed the technological feasibility of the Tier 3 standards contains a
detailed discussion of a number of technologies which are capable of achieving a 3.5 g/bhp-hr
standard. These include cooled EGR, uncooled EGR, as well as advanced in-cylinder technologies
relying on electronic fuel systems and turbocharging.124 These technologies are capable of reducing
NOx emission by as much as 50 percent. Given the Tier 2 NMHC+NOx standard of 5.6 g/bhp-hr, a
50 percent reduction would allow a Tier 2 engine to comply with the 3.5 g/bhp-hr NMHC+NOx
standard proposed in this action. In addition, because this NMHC+NOx standard is concurrent with
the 0.02 g/bhp-hr PM standards which we project will be achievable with the use of 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.
4.1.5 Are the Standards Proposed for Engines <25 hp Feasible?
As discussed in Section in of the preamble, our proposal for standards for engines less than 25 hp
is a new PM standard of 0.30 g/bhp-hr beginning in 2008. As discussed below, we are not proposing
to set a new standard more stringent than the existing Tier 2 NMHC+NOx standard for this power
category at this time. This section describes:
- what makes the <25 hp category unique;
- engine technology currently used in the <25 hp category; and,
- what data shows the proposed 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 4.1-16). Nearly 50 percent of the products have two-cylinders
or less, and 14 percent of the engines sold in this category are single-cylinder products, a number of
these have no batteries and are crank-start machines, much like today's simple walk behind
lawnmower engines. In addition, given what we know today and taking into account the Tier 2
standards which 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 5 to 10 years.
4-79
-------
Draft Regulatory Impact Analysis
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 will be
met by basic engine-out emission optimization strategies.125 We are not predicting that Tier 2 will
require electronic fuel systems, EGR, or turbocharging. As discussed in the Staff Technical Paper, a
large number of engines in this power category already meet the Tier 2 standards by a wide margin.126
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 indicates the proposed standards are feasible?
We project the proposed Tier 4 PM standard can be met by 2008 based on:
—the existence of a large number of engine families which meet the proposed standards today;
—the use of engine-out reduction techniques; and
—the use of diesel oxidation catalysts.
We have examined the recent model year (2002) engine certification data for nonroad diesel
engines less than 25 hp category.127 A summary of this data is presented in Table 4.1-18 These data
indicate that a number of engine families meet the proposed Tier 4 PM standard (and the 2008
NMHC+NOx standard, unchanged from Tier 2) today. The current data indicates approximately
28% of the engine families are at or below the proposed PM standard today, while meeting the 2008
NMHC+NOx standard. These include both IDI and DI engines, as well as a range of certification test
cycles.0 Many of the engine families are certified well below the proposed Tier 4 standard while
meeting the 2008 NMHC+NOx level. Specifically, 15 percent of the engine families are more than
20 percent below the proposed Tier 4 PM standard. An additional 15 percent of the engine families
which already meet the 2008 NMHC+NOx standards would require no more than a 30 percent PM
reduction to meet the proposed 2008 PM standards. The public certification data indicate that these
engines do not use turbocharging, electronic fuel systems, exhaust gas recirculation, or aftertreatment
technologies.
Table 4.1-18
2002 Model Year Certification Data for <25 hp Nonroad Diesel Engines
3M Emissions Relative to Proposed 0.30
z/bhp-hr Standard
IDI Engines
(test cycle)
5 -mode
6-mode
8-mode
DI Engines
(test cycle)
5 -mode
6-mode
8-mode
Totals
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.
4-80
-------
Technologies and Test Procedures for Low-Emission Engines
0-5% below T4a
5-20% below T4a
>20% below T4a
require <30%PM
reduction to meet T4a
requires >30%PM reduction and/or
> 2008 NMHC+NOx std.
Total # ot Engine Families
1
4
1
5
7
18
0
6
9
4
8
27
1
1
5
4
4
15
0
0
0
0
18
18
0
0
1
2
18
21
0
0
0
0
3
3
2
11
It
15
58
102
Engine also meets 2008 NMHC+NOx
These model year 2002 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. As with 25-75 hp engines, these data have a two-fold significance. First, they
indicate that a number of engines in this power category can already achieve the proposed 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. Despite being certified to the
same emission standards with similar engine technology, the emission levels from these engines vary
widely. Figure 4.1-19 is a graph of the model year 2002 HC+NOx and PM data. As can be seen in
the figure, the emission levels cover a wide range. Figure 4.1-19 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/bhp-
hr, the PM rates vary from approximately 0.15 to 0.5 g/bhp-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 proposed 2008 PM standard on the combination of diesel oxidation catalysts and the
lowest engine-out emissions being achieved today, because it is uncertain whether or not additional
engine-out improvements would lower all engines to the proposed 2008 PM standard. Instead, we
believe there are two likely means by which companies can comply with the proposed 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. Therefore, we
believe some engine companies will choose to use a combination of engine-out techniques and diesel
oxidation catalysts, as discussed below.
4-81
-------
Draft Regulatory Impact Analysis
Figure 4.1-19 Emission Certification Data for 25-50 HP Model Year 2002 Engines
x
x
lf>
O
'en 4
1
LU
X -5
O 6
%2
X
XX
X x Xy
XX A. S\
X
X
x
X
x_
-tr-x
-x*-
X
B
x
n
x
x
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 & Dl DIDI 6-mode data
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 are a number of technologies which exist 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.128 The wide-range of emission characteristics present in the existing engine
certification data is likely a result of differences in fuel systems and combustion chamber designs.
For many of the engines which have higher emission levels, further optimization of the fuel system
and combustion chamber can provide additional PM reductions.
Diesel oxidation catalysts (DOC) also offer the opportunity to reduce PM emissions from the
engines in this power category. DOCs are passive flow through emission control devices which 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.129 However, as discussed in Section 4.1.1, certain DOC formulations can be
sensitive to diesel fuel sulfur level. Specifically, precious-metal based oxidation catalysts (which
have the greatest potential for reducing PM) can oxidize the sulfur in the fuel and form paniculate
sulfates. Given the high level of sulfur in nonroad fuel today, 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 shows 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.130 Without the availability of 500 ppm sulfur fuel in 2008,
4-82
-------
Technologies and Test Procedures for Low-Emission Engines
DOCs would be of limited use for nonroad engine manufacturers and would not provide the
emissions necessary to meet the proposed standards for most engine manufacturers. With the
availability of 500 ppm sulfur fuel, DOC's can be designed to provide PM reductions on the order of
20 to 50%, while suppressing particulate sulfate reduction.131 These levels of reductions have been
seen on transient duty cycles as well as highway and nonroad steady-state duty cycles.
As discussed in Section in of the preamble, we have also proposed a minor change in the CO
standard for the <11 hp engines, in order to align those standards with the standards for the 11-25 hp
engines. The small change in the CO standard is intended to simplify EPA's regulations as part of
our decision to propose a reduction in the number of engine power categories for Tier 4. The current
CO standard for this category is 6.0 g/bhp-hr, and the proposed standard is 4.9 g/bhp-hr (i.e., the
current standard for engines in the 11-25 hp range). The model year 2002 certification data shows
that more than 90 percent of the engine families in this power category meet the proposed standards
today. In addition, DOCs typically reduce CO emissions on the order of 50 percent or more.132 Given
that more than 90 percent of the engines in this category meet the proposed standard today, and the
ready availability of technology which can easily achieve the proposed standard, we project this CO
standard will be achievable by model year 2008.
4.1.6 Meeting the Crankcase Emissions Requirements
The most common way to eliminate crankcase emissions has been to vent the blow-by gases into
the engine air intake system, so that 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 on-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 eliminated this exception for
highway engines in 2007 and are proposing to eliminate the exception for nonroad diesel engines as
well. 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 proposed
provision has been written such that if adequate control can be had without "closing" the crankcase
then the crankcase can remain "open." Compliance would be ensured by adding the emission from
the crankcase ventilation system to the emissions from the engine control system downstream of any
emission control equipment.
We expect that in order to meet the stringent tailpipe emission standards set here, that
manufacturers will have to utilize closed crankcase approaches as described here. 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.133 By
eliminating 90 percent of the oil that would normally 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
4-83
-------
Draft Regulatory Impact Analysis
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 15ppm 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. In order to evaluate the effect of sulfur on diesel exhaust
control technologies, we used three key factors to categorize the impact of sulfur in fuel on emission
control function. These factors were efficiency, reliability, and fuel economy. Taken together these
three factors lead us to believe that diesel fuel sulfur levels of 15 ppm will be required for the nonroad
emission standards proposed here to be feasible. Brief summaries of these factors are provided
below.
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 paniculate 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 paniculate 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 paniculate filter and increased work to pump exhaust
4-84
-------
Technologies and Test Procedures for Low-Emission Engines
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. Thus, for
both NOx and PM technologies the lower the fuel sulfur level the lower 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 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 particulate 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 could
be assured to occur at exhaust temperatures lower than 300°C, then diesel particulate filters would be
expected to 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 which 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.134 Some
platinum group metals are known to be good catalysts to promote the oxidation of NO to NO2.
Therefore in order to promote more effective passive regeneration of the diesel particulate filters,
significant amounts of platinum group metals (primarily platinum) are 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 in order 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.
4-85
-------
Draft Regulatory Impact Analysis
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.135 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 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 low NO2 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. In order 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 below 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 which included intercity trains, airport buses, mail trucks, city
buses and garbage trucks, and the extended time periods of operation (some vehicles have been
operating with traps for more than 5 years and in excess of 300,000 miles), there is a strong indication
of the robustness of this technology on 10 ppm low sulfur diesel fuel.136 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, although 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.137
Other possible reasons for the high failure rate in Finland when contrasted with the Swedish
P 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.
4-86
-------
Technologies and Test Procedures for Low-Emission Engines
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.138 Given that the fleets in Sweden and Finland
were substantially similar, 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.Q
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.139 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 would be 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. Therefore, we believe that, in order to ensure reliable and
economical operation over a wide range of expected operating conditions, nonroad diesel fuel sulfur
levels should be at or below 15 ppm.
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 paniculate 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 parti culate 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 which might be used in
Q The average temperature in Helsinki, Finland, for the month of January is 2IT. The average temperature in
Stockholm, Sweden, for the month of January is 26T. The average temperature at the University of Michigan in
Ann Arbor, Michigan, for the month of January is 24T. 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).
4-87
-------
Draft Regulatory Impact Analysis
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 on-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 US 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 below showing the expected
total direct PM emissions from a diesel engine on the supplemental emission test cycle.R The SET
test cycle, a 13 mode steady-state cycle, that this data was 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 would be 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 Emissions Performance
Tailpipe PMb
[g/bhp-hr]
0.003
0.006
0.009
0.017
0.071
PM Increase
Relative to 3 ppm Sulfur
--
100%
200%
470%
2300%
a 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.
b Total exhaust PM (soot, SOF, sulfate).
Table 4.1-19 makes it clear that there are significant PM emission reductions possible with the
application of catalyzed diesel particulate filters and low sulfur diesel fuel. At the observed sulfate
PM conversion rates, the DECSE program results show that the 0.01 g/bhp-hr total PM standard is
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.
4-88
-------
Technologies and Test Procedures for Low-Emission Engines
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, in order to realize the PM emissions
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 Particulate 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 which 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).
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.140 Under
these conditions, the stored sulfate is released and reduced in the catalyst.141 While research to date
on this procedure has been very favorable with regards to 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)
4-89
-------
Draft Regulatory Impact Analysis
reducing the effectiveness of the catalyst.142 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 is
expected to 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 is
expected to 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 could meet the HD2007 NOx standard using diesel fuel with a 30 ppm
average sulfur level. This would imply that the NOx adsorber could tolerate as much as a four fold
increase in desulfation frequency (when compared to 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 in
order 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 on-highway NOx adsorber catalyst technology that this
proposal is based upon and will be equally true for nonroad engines applying the NOx adsorber
technology to comply with our proposed Tier 4 standards.
Nonroad and on-highway diesel engines are similarly durable and thus over their lifetimes
consume a similar amount of diesel fuel. This means that both nonroad and on-highway diesel
engines will have the same exposure to sulfur in diesel fuel and thus will 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 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 in order to meet the NOx emission standards proposed
today. 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 in order to remove the stored sulfur.
Therefore, we believe that fuel with a sulfur level at or below 15 ppm sulfur will be necessary in
4-90
-------
Technologies and Test Procedures for Low-Emission Engines
order 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 in order to ensure
high NOx control efficiencies. As discussed more fully in Section 4.F. 1, 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 only trap that small portion of NOx emissions from a diesel engine which 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. Therefore, future Compact-SCR systems
would need to rely on a platinum oxidation catalyst in order to provide the required NOx emission
control. This use of an oxidation catalyst in order 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.
Without the oxidation catalyst promoted conversion of NO to NO2, neither of these NOx control
technologies can meet the proposed NOx standard. Therefore, each of these technologies will 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. Thus, we believe that diesel fuel sulfur
levels will need to be at or below 15 ppm in order to apply the NOx control technology.
4-91
-------
Draft Regulatory Impact Analysis
4.2. Supplemental 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 emissions control, a cycle that appropriately
characterizes the activity of the subject equipment would afford a greater level of control. The basics
of any nonroad transient duty cycle should include fulfillment of the following goals:
• represents nonroad activity broadly, with a basis in real world activities through diverse data
segments;
• exercises the engine over its operating range. Cycle would not be limited to a specific speed
or load, but traverses the operating range over the engine's full power range;
• measures particulate matter (PM) on a transient basis;
• captures the basic characteristics of PM, as currently defined, including:
- organic and inorganic carbon fractions;
- volatile fraction;
- sulfate fraction;
- ash, etc.
• ensures that control measures developed to control emissions over the cycle encourage and
afford greater assurance that adequate control measures 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 which provides a
representation of a broad range of nonroad diesel equipment activity. 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:
- 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, the combining of these two strategies has ensured that the strengths of
both approaches would be 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 event today as they appear
to match fairly well with EPA modeling data for the impacts of those applications.
4-92
-------
Technologies and Test Procedures for Low-Emission Engines
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. Please see Figure 4.2-1. It
should be clear that no single transient cycle, of reasonable length, could capture the full body of
nonroad diesel activity in the real world. 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 could not replace an in-use
characterization, however it does drive development of engine control strategies to focus emissions
and performance parameters on a broader set of activity that is much more likely to be seen in-use.
4-93
-------
Draft Regulatory Impact Analysis
Figure 4.2-1
Backhoe Loader and Crawler Tractor Cycle Data versus the ISO 8178-4 Cl Cycle
Backhoe\Loader Cycle Data (with ISOC-1 \Afeightings)
Crawler Tractor Cycle Data(with ISOC-1 V\feightings)
-20
40
Speed (%)
BO 100
20 40 60 80 100
Speed (%)
A much broader set of data from the nonroad duty cycle generation may be found in
Memorandum from Cleophas Jackson to the 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 would likely
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 was
compared to 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 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
4-94
-------
Technologies and Test Procedures for Low-Emission Engines
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"-based application 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.
4.2.2. 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 was actual work and not artificial
scenarios, so that the data accuracy was ensured.
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 if 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 was not included in the eventual Agricultural Tractor cycle.
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 was 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.
4-95
-------
Draft Regulatory Impact Analysis
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 was 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 cars143. The material being moved was typical quarry
material which included aggregate of various material densities such as crushed stone, gravel, and
sand. Twenty-six hours of data was 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 was 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 was collected over a single work day. The equipment was used
to perform repairs on a large, mobile steel crusher tower by a private contract firm, Holt. Eight hours
of data was 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 was 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 was 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 was 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 was gathered for this excavator.
4.2.2.2 Engine and Equipment Description
In generating the microtrip-based and the day-in-the-life-based 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 engines144 may be seen in Table 4.2-1 through Table 4.2-3.
4-96
-------
Technologies and Test Procedures for Low-Emission Engines
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
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 that were used for data generation for the "day in the life" -based approach were
based on a skid steer loader, an arc welder, and a wheel loader. The engine parameters of the
Caterpillar 988F Series n rubber tire loader, the Lincoln arc welder and the Daewoo skidsteer loader
are listed in Table 4.2-4 through Table 4.2-6.
4-97
-------
Draft Regulatory Impact Analysis
Table 4.2-4
Rubber Tired Loader—1997 Caterpillar 988F Series
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 Shi eld-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
in-line 4 cyl, naturally
L Displacement,
aspirated
Two pieces of equipment were selected for generating the excavator duty cycle based on estimates
of equipment population and horsepower distribution among excavators in the U.S. nonroad
equipment inventory at that time145. 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 kilowatts (208 horsepower) and a Mitsubishi/CAT 320 BL excavator at
4-98
-------
Technologies and Test Procedures for Low-Emission Engines
95 kilowatts (128 horsepower). 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 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 could be determined based on rack position and engine speed as measured
by the tachometer.
4-99
-------
Draft Regulatory Impact Analysis
Data loggers were used to record field data during operation and the data loggers were equipped
with flash memory media. The data loggers only recorded engine on operation, therefore data was
not gathered 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 was 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 was 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-squares 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)
• 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.
£(Oi -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.
4-100
-------
Technologies and Test Procedures for Low-Emission Engines
2. Load first activity file.
3. Calculate power. Add to raw data file.
4. Normalize speed using 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.1 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 deltau 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 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.
TSteady State is defined as any instantaneous change in normalized speed or normalized torque with a
magnitude less than 2%.
uDelta is used to describe the instantaneous rate of change of the specified quantity.
4-101
-------
Draft Regulatory Impact Analysis
Figure 4.2-2
Agricultural Tractor
Turnaround
Planter
5% 4% ldle 6%
Bedder D/0 4/0
5% /^^ ^^Oxboard Ripper
18%
Plow
18%
•^^^r
/
_ 29%
Field Cultivator ^
Tandem Offset Disc
15%
Figure 4.2-3
Backhoe Loader
Grade and Level
, A- *-^
Loading xT^X Idle
1 1 %
36%
12^*-^
x^T^X
/ \
/ \
\^^^^^/ Reading
Trenching 6%
37%
4-102
-------
Technologies and Test Procedures for Low-Emission Engines
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 to 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.
4-103
-------
Draft Regulatory Impact Analysis
Reading
RFO Differences from Activity to Microtrip
Normalized Torque Range
-20 to 20
20 to 40
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 would be used for creation of the nonroad transient composite duty cycle.
4.2.2.4.2 "Day in the Life " 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 versus 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.
1. Define "bins" sized at 100 rpm for speed by 200 ft-lb for torque.
4-104
-------
Technologies and Test Procedures for Low-Emission Engines
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 speedv duty cycle, and a high transient torque duty cycle for each
application.
9. For each window of activity, the data used was 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. Therefore, by choosing a segment that closely matched
the characteristic RFO compilation, it is 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 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:
vHigh 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.
4-105
-------
Draft Regulatory Impact Analysis
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 was
normalized using torque map information obtained for the Isuzu A-6SD1T engine on the steady-state
test bench.
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
4-106
-------
Technologies and Test Procedures for Low-Emission Engines
analyses for that population). This resulted in a composite cycle which 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 operationw was not included in the transient cycle). 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 was 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 normalization or GCSX 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.
wSteady State Operation is defined as an instantaneous speed or torque change less than 2% of the maximum
magnitude.
XGCS 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.
4-107
-------
Draft Regulatory Impact Analysis
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
Ag Tractor
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
4-108
-------
Technologies and Test Procedures for Low-Emission Engines
4.2.4 Cycle Characterization Statistics
The characterization of the operational data was 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.146 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
OU
fin
ACi
'tU
zU
n
F
F
F
[•
Speed accels/min Speed decels/min Torque accels/min Torque decels/min
. Speed
Torque
D NRC Mean+CL Mean-CL • Mean
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 to normalized values before any application cycle could be used on an engine other than
4-109
-------
Draft Regulatory Impact Analysis
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 duty cycle schedule of speed and load values which, because the
schedule points are in percent of an engine's maximum measured rpm (speed) and Ibs-ft (load), the
cycle can be converted over to the true speed and load values, in rpm and Ibs-ft, required to run that
cycle on any other engine, if one has the new engine's maximum achievable power (MAP) validation.
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)Y. 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 speadsheet calculation 147 and is discussed in a report
entitled "Summary and Analysis of Comments: Control of Emissions from Marine Diesel Engines",
document # EPA420-R-99-028 in Chapter 8, "Test Procedures"148. As detailed in Figure 4.2-8 below,
the maximum speed can be found below the point on the engine power curve that is the farthest
distance from the point of 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.
Y 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.
4-110
-------
Technologies and Test Procedures for Low-Emission Engines
Figure 4.2-8
Maximum Test Speed Determination
4-111
-------
Draft Regulatory Impact Analysis
140
120
100
20
0 250 500 750 1000 1250 1500 1750 2000 2250 2500
Max Test
Speed
rpm
4.
2.
6
C
y
cl
e
P
e
rf
o
r
m
a
n
c
e
R
egression Statistics
In assessing the nonroad transient duty cycles, ten nonroad diesel engines were exercised over the
nonregulatory149 nonroad duty cycles to assess emissions 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 the 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 Transient Test Procedure
Some nonnroad diesel engines are required by the equipment which they have been designed to
power, and the equipment's specific application, to operate in a constant-speed manner. While the
operating speed in many cases is not truly constant, it is generally true that the unit's speed will vary
little during operation. This equipment is more tolerant of changes in operating load than other more
closely governed constant-speed nonroad engines. 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. For these engines which are designed to operate
under restricted transient conditions, the Agency is proposing an alternative transient duty cycle over
which a nonroad engine manufacturer may choose to operate their engine(s) to comply with EPA's
new transient nonroad testing requirement. This cycle, the Constant-Speed Variable-Load (CSVL)
application duty cycle, is derived from EPA's Arc Welder Highly Transient Torque application duty
cycle. The cycle schedule, its commanded speed and load values, may be found at proposed
4-112
-------
Technologies and Test Procedures for Low-Emission Engines
regulations in 40 CFR Part 1039, Appendix "I". A manufacturer certifying their equipment to this
cycle would be constrained to allow their engines to operate equipment in-use only in a constant-
speed, variable-load manner.
EPA recognizes that some constant speed equipment is designed to operate near or at its rated
engine rpm during much or most of that equipment's useful life. The Agency's CSVL cycle, while
designed for a broad range of constant speed nonroad engines, has an average speed which may be
lower than the speed at which a manufacturer has designed their engine to operate at maximum
efficiency. While the CSVL cycle would still test that engine in a manner which might be
encountered by these types of engines under real world operation, EPA has given the nonroad engine
manufacturer who believes that their engine to be certified will be sensitive to the speed fluctuations
found in the CSVL cycle the option, within the transient test regulations, to operate their engine at a
set speed, determined by the operating specifications of the engine, for the course of the CSVL cycle.
This variation is described in proposed regulations at 40 CFR Part 1048, section 510. Set speed
operation will allow the manufacturer to run the CSVL cycle at the rated speed of the equipment to be
certified for the entire length of the cycle or to specify some constant percent of rated speed at which
to operate the equipment or engine to be certified. The load portion of the CSVL duty schedule
would remain unchanged under this modification. As in the previous case, a manufacturer certifying
their equipment on this cycle would be constrained to specify this restricted set speed, variable load
manner of operation in-use for their engines.
4.2.7.1 Background on Cycles Considered
As has been described earlier in Part 4 of the Draft RIA, the Arc Welder application cycles were
developed on an arc welder/electric generator, running a constant speed application at a variable
load, with a direct-injection, naturally-aspirated, 30kW (40 hp) engine. The Arc Welder Highly
Transient Torque cycle, one of the four cycles developed on this engine, is based on a single twenty-
minute segment of all the real time operating data collected on this engine.
Some manufacturers of constant-speed application engines have raised issues with EPA's
proposed CSVL cycle. One issue arises from manufacturers of high brake-mean effective pressure
(BMEP), i.e., high rated power, constant-speed engines. They point 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 feel that their engines
may be penalized by the number and magnitude of torque changes in the CSVL cycle. We do not
believe that this concern will significantly affect emissions performance for Tier 4 engines because
PM control is realized through mechanical filtration of the PM and as such is largely independent of
operating conditions (see the PM emissions performance noted in tables 4.1-2 through 4.1-5).
Similarly, NOx and NMHC control is expected to be realized with catalyst systems, that although
temperature dependent, are expected to be fully functional over the range of operation for the CSVL
cycle as evidenced by the steady-state emissions results shown in figure 4.1-11 (EPA modes 5,6,7,8
and 17,18,19,20). Further, manufacturers can improve catalyst performance at low torque operating
modes by increasing the use of EGR to both lower engine-out NOx emissions while simultaneously
raising exhaust temperatures to promote more effective catalyst function (see discussion in 4.1.3
above). Therefore, we can conclude with confidence that compliance with the proposed Tier 4
standards are feasible over the CSVL cycle.
4-113
-------
Draft Regulatory Impact Analysis
A second issue involves the average load experienced by an engine running on the proposed
CSVL cycle. The average load factor of the normalized application cycle is approximately 25% of
engine capacity. Manufacturers of constant speed engines which have a significantly higher load
factor on their engines during operation, closer to the 90% and greater range of normalized engine
load at constant speed, have argued that their engines will not be able to pass cycle regression
statistics for certification without significant re-tuning of the engines to operate over the CSVL cycle.
EPA will follow developments with specific manufacturers where the CSVL cycle is anticipated to
require substantial re-tuning, or even redesign, of engine controls in order to pass this cycle for engine
compliance with transient testing requirements.
4.2.7.2 Justification of Selections
The CSVL cycle should assure manufacturers that their constant-speed engines are able to meet
in-use emission and NTE standards. While the CSVL cycle may not be able to accommodate the
particular operating parameters required to run every constant speed engine, it is a fairly robust cycle
for many types of constant-speed equipment and applications. With a manufacturer's option to use a
set operating speed over the course of the test cycle, even fewer concerns should arise as to operating
an engine over this transient cycle. EPA has shown further that one can pass cycle regression
statistics on a typical nonroad engine with fairly mild cycle control procedures in place150. The
Agency feels that "tailoring" the constant speed cycle to multiple engines and applications will further
fragment the certification process for constant-speed engine manufacturers and in the end, will afford
less control over in-use engine emissions than maintaining the CSVL cycle alone as the single
certification option for the class of constant-speed engines. Given future engine technology and
trends in emission control, this cycle will represent a boundary for operating emissions from these
engines. By certifying engines to this testing procedure, manufacturers can be assured that their
engines will be as clean as, and may be even cleaner, operating in-use than operating over the CSVL
cycle for certification.
Manufacturers may choose to run the CSVL cycle for certification purposes at a set speed instead
of following the intended speed trace from the CSVL cycle, where their constant speed engine(s) is
governed very closely during actual engine operation. EPA is proposing this set speed option to the
CSVL cycle as a form of relief for these engines which normally are used to power applications like
electricity-generating sets and some refrigeration units. Details of this option may be found at
proposed regulations in 40 CFR Part 1068.
To pass cycle regression statistics for cycle performance on the CSVL cycle, an engine must have
achieved the performance parameters for speed, load and power found in table 4.2-14. These values
are the same regression statistics used to determine pass or fail on EPA's NRTC cycle and they do not
impose any greater burden on constant-speed engine manufacturers who choose engine compliance
with the CSVL transient test cycle than might be seen with operating their engines over the NRTC
cycle for compliance.
The Arc Welder cycles were corrected in the course of developing EPA's composite nonroad duty
cycle to accommodate for the fact that idle, as listed in early cycle versions, was actually an
intermediate speed for the engine (due to its limited high-speed range during transient operation) and
that the actual engine idle speed was closer to 800 rpm. As any idle point listed in the original cycles
was also idle for the application and not for the engine, this change was seen as appropriate. Speed
4-114
-------
Technologies and Test Procedures for Low-Emission Engines
changes that were based on low idle to high rated excursions were changed to peak torque speed to
rated excursions.
Contractor testing of EPA's Arc Welder cycles has yielded mixed results in passing regression
statistics upon later analysis, but passing regression statistics on these cycles was not the principal
goal of many of the individual cycle development and testing programs151. Many times, the
contractor was only required to optimize the test engine to the case of running a transient FTP for
passing statistics on the test engine152 and, subsequently, many different application cycles were then
run on the same engine in conjunction with the transient FTP cycle.
Most recently, though, EPA has run both its Arc Welder Typical 1 and Highly Transient Torque
application cycles in-house on an electric dynamometer with a turbocharged 93 kW (124 hp) test bed
engine153. The realtime control strategies employed on the dynamometer cycle runs included simple
PID (proportional, integral, and derivative) control algorithm-derived time constants and a half
second earlier time-shift of commanded throttle (torque) values in the denormalized cycle. These
control measures correct for anticipated differences in feedback from the actual speed and torque
values for the engine during operation and the command values from the dynamometer during testing.
The commanded throttle (torque) values were time-shifted, again, in a post-analysis of cycle data in
half-second increments. Out of 20 half-second values, cycle regression statistics were passed on all
parameters, for speed, torque and power, for three consecutive time values for the Arc Welder Typical
cycle and on six consecutive time shift values for the Arc Welder Torque cycle. This form of post-
processing is fairly typical for duty cycle testing conducted in a dynamometer test cell and would not
create a significant added burden on the manufacturer.
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. There are two aspects of this which 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 real world based duty cycles.
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 provide 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
would tend to demand a desired torque the engine's response to that demand, although 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 could be
addressed through the performance statistics.
Engine manufacturers sought to address the first concern through a playback analysis which
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
4-115
-------
Draft Regulatory Impact Analysis
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
Yanmar
Engine Model
3204-D17
3114-D17
3408E - TA
A-6SD1 TQA
6081
3066T
'97D3.152
'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
Ta
bl
e
4.
2-
10
En
gine 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 was 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.2
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.
Please see Draft RIA Section 4.2.3. of this rulemaking
4-116
-------
Technologies and Test Procedures for Low-Emission Engines
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 helps
reduce the testing burden a given manufacturer would have to face when selling and distributing their
product 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 emissions
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 arena 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 heavy
duty, highway diesel applications, it was seen as an adequate basis for which European industry and
government staff could assess the proposed EPA 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.
4-117
-------
Draft Regulatory Impact Analysis
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
30,10 Nm
5,80 Nm
+/- 20 Nm
out of tolerance
near to tolerance limit
5,67%
max 8 %
0,968
0,976
0,89 to 1,03
0,973
0,981
min 0,91
3,62 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 which subsequent development and testing work would be conducted. 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 then
"candidate" cycle would be used by the Joint Research center 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 4.2-9 and
4.2-10. The average load and average speed of each cycle are shown in Table 4.2.6.2.2.-1. 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.AA
AAA
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
4-118
-------
Technologies and Test Procedures for Low-Emission Engines
Figure 4.2-9
EPA Nonroad Transient Test Cycle as of March 2001
Draft Nonroad Transient Duty Cycle
Time (seconds)
Figure 4.2-10
JRC Nonroad Transient Test Cycle after Bessel Filtering
Joint EPA-EU Nonroad Transient Cycle, March, 2002
CO CM CD
Time (seconds)
"Speed
4-119
-------
Draft Regulatory Impact Analysis
Table 4.2-12
Comparison of Cycle Averages
Duty Cycles
EPANRTC
JRC Modified NRTC
Average Normalized Speed
63%
68%
Average Normalized Torque
47%
39%
The following figures 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
March EPA Cycle
JRC NRTC
Average Speed Average Load
4-120
-------
100
Technologies and Test Procedures for Low-Emission Engines
Figure 4.2-12
Average Speed Changes of the EPA NRTC
Speed Changes EPA March NRC
Time (seconds)
Figure 4.2-13
Average Speed Changes of JRC Modified NRTC
JRC_NRC Speed Changes
Tim e (second s)
Figure 4.2-14
Average Load Changes of JRC Modified NRTC
4-121
-------
Draft Regulatory Impact Analysis
JRC_NRC Load Changes
Tim e (seconds)
Figure 4.2-15
Average Load Changes of the EPA Generated NRTC
March EPA_NRC Load Changes
Time (seconds)
Gi
ve
n
th
e
m
od
ifi
ca
ti
on
s
in
th
e
du
ty
cy
cl
it was critical to assess the impact on the emissions signature of the cycle. The table below (Table
4-122
-------
Technologies and Test Procedures for Low-Emission Engines
4.2-13) shows that the emissions signature, based on tests at the National Vehicle and Fuel Emissions
Laboratory and at Southwest Research Institute as of May 2001, were relatively unchanged.
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 3 508
Heavy Duty
850 hp
Sep-OC
Mar-01
JRC
NOx
Mean Standard Dev.
PM Speed
Mean Standard Dev. SE
Mean Stddev. Mean Stddev. Mean Stddev. Mean Stddev.
10.30
10.14
11.198
0.02 0.20
0.03 0.20
0.03 0.20
0.004
0.002
0.004
79
90
68
1.41
2.12
0.71
1.03
1.01
1.03
0
0.01
0.00
0.949
0.939
0.962
0.001
0.002
0.001
-35
-9
-33
2.83
3.54
1.41
Mean Stddev. Mean Stddev. Mean Stddev. Mean Stddev. Mean Stddev. Mean Stddev. Mean Stddev. Mean Stddev.
15 0 0.8 0 0.734 0.004 184 0 14 0 0.88 0 0.801 0.283 29.6
15 0 0.83 0.007 0.734 0.001 188.5 3.54 14 0 0.9 0 0.804 0.002 29.5 1.273
12 0 0.91 0.007 0.765 0.001 56 1.41 11 0 0.95 0 0.823 0 6.1 0.1
Cummins ISB
Medium Dutv
Sep-00
Mar-01
JRC-Max Spd
JRC-ETC Pk Spd
Mean Standard Dev.
PM Speed
Mean Standard Dev. SE
3.76
3.79
4.06
4.09
0.01 0.08
0.03 0.08
0.03 0.08
0.01 0.08
Mean Stddev. Mean Stddev. Mean Stddev. Mean Stddev.
0.001 54.7 24.62 0.987 0.011 0.987 0.010 30.0 3.11
0.003 68 18.67 0.98 0.01 0.982 0.008 32 14.48
0.002 66 6.22 0.98 0.00 0.978 0.005 34 5.23
0.009 50 8.15 0.98 0.00 0.991 0.003 37 6.68
Mean Std dev. Mean Std dev. Mean Std dev. Mean Std dev. Mean Std dev. Mean Std dev. Mean Std dev. Mean Std dev.
69.7
67.5
43.5
48.4
2.06 0.955
3.12 0.96
0.14 0.981
2.63 0.985
0.011 0.930
0.008 0.933
0.002 0.960
0.00306 0.946
0.005 30.0
0.007 26.7
0.001 12.0
0.005 11.6
3.11
2.64
0.354
1.386
14.8
14.9
9.9
10.0
0.35 0.979
0.61 0.981
0.21 0.994
0.68 0.999
0.009
0.007
0.002
0.002
0.943
0.943
0.961
0.958
0.003
0.005
0.002
0.005
4.5
4.2
1.6
1.6
0.361
0.404
0.141
0.265
As has been noted earlier in this chapter, the cycle was modified by EPA 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 emissions 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 emissions signature™. 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 stringent to ensure an accurate and repeatable emissions
Memorandum from Cleophas Jackson to EPA Air Docket A-2001-28, JRC December 5, 2001, Report on
Cycle Performance
4-123
-------
Draft Regulatory Impact Analysis
signature was achieved00. With the conclusion of the international workgroup's efforts, the cycle
was considered complete by EPA. In 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 below in Table 4.2-14.
Table 4.2-14
NRTC Cycle Regression Statistics154
Speed rom
lorque IJSI-m
Power IkWI
Regression Line Tolerances
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
100 rpm
0.95 to 1.03
min 0.970
±50 rpm
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
8% of power map
maximum
0.89-1.03 (hot)
0.87 -1.03 (cold)*
min 0.9 100 (hot)
min 0.8500 (cold)
± 4 kW or ± 2.0% of max
power, whichever is
greater
: Under consideration by ISO workgroup.
4.2.9 Supplemental Cold Start Transient Test Procedure
We are proposing to include a requirement for a cold-start transient test to be run in conjunction
with the hot-start run of the proposed transient test procedure. The proposed cold-start measurement
is meant to recognize and quantify nonroad diesel engine missions generated for short periods at
engine start-up. We further propose to weight these cold start emission results as 1/10 of total
emissions, with the hot-start transient emissions making up the remainder. Cold start most often
refers to nonroad engine emissions created during a short period after the first key-on event of the
day, the first "cold start" for that piece of equipment in its workday. Given that the equipment has sat
at ambient temperature for a minimum of six hours and in most cases overnight, engine startup will
entail warming up the unit's operating and emission control equipment to normal operating
temperatures. Likewise, a short period of engine operation, after a longer period of engine inactivity,
may be characterized as having emissions similar to the earlier cold start operation as the unit must
warm up its operating systems once more before running at peak efficiency. With this as background,
EPA targeted the second-by-second operation of a population of some forty pieces of nonroad
equipment for analysis to characterize the "average" workday of each unit and to determine if some
portions of that workday were spent at a significantly higher rate of engine emissions than others.
Generally, times when an engine is operating at cold start or less than stable operating temperature,
frequently characterized by lower exhaust temperatures, engine emission rates can be seen to be
higher than those during "warmer" engine operation.
CCA
"Memorandum from Cleophas Jackson to EPA Air Docket A-2001-28, Nonroad Transient Duty Cycle
Development Report, Cornetti, G., Hummel, R., and Jackson, C.
4-124
-------
Technologies and Test Procedures for Low-Emission Engines
In one such analysis, EPA examined over 435 hours of second-by-second operating (actual "key-
on" operation time) and NOx emission data from 13 pieces of construction equipment in the field155
and is summarized in table 4.2-15 below.
Table 4.2-15
Portable Emissions Testing in a Construction Equipment Population
iqurpment lype
brawler Dozer
brawler Dozer
brawler Dozer
brawler Dozer
brawler Dozer
brawler Dozer
brawler Dozer
brawler Dozer
ixcavator
Dff -Highway Truck
Dff -Highway Truck
Dff -Highway Truck
A/hpp T.nnHpr
Model Year
1999
1985
1987
1988
1990
1995
1998
2001
1989
1999
2001
2001
1QS«
1 otal Cold-Start
Key-on Periods
11
4
11
10
6
7
8
12
8
8
7
8
7
1 otal Operating
Time (hr)
25.85
30.53
31.72
41.87
20.86
49.29
49.16
63.09
20.48
48.32
26.20
23.82
4 41
Cold-start JNUx as
Weighted Proportion
of Total
0.038
0.021
0.065
0.025
0.042
0.018
0.037
0.032
0.062
0.016
0.053
0.062
0 OQ7
All data was recorded in real time with EPA's Simple Portable On-board Testing emission
monitoring equipment. The test population consisted of eight bulldozers, three sediment haulers, one
excavator and one wheel loader. Each piece of equipment recorded one or two cold starts per
workday. Cold starts included in its definition the first ten minutes of operation after the first "key-
on" in a day and the first ten minutes of operation of that equipment after a period of an hour or more
of inactivity within that same workday, as defined by the Julian date. Most of the crawler/dozers and
the excavator exhibited three cold starts in a day and one dozer recorded four cold start periods in the
same day. Grams of NOx were summed for all cold start periods for each piece of equipment per day
and then divided by the total number of grams of NOx emitted from that piece of equipment for the
entire "key-on" period of operation for that same day, giving the proportion of cold start statistic
below.
P
coldstarts
coldstart,day
day
where NOx coldstarts is the amount of NOx (g) emitted during the 10-minute cold-start periods
during the day, and NOx day is the total amount of NOx (g) emitted during the day.
The proportion of NOx emissions from cold start operation for each day per piece of equipment,
P above, was multiplied by the total time that that unit spent running for each day that the equipment
had at least one "key-on". This statistic of the cold-start emissions proportion for a single workday
for a single piece of equipment was then summed, and divided by the total time that piece of
equipment spent in a "key-on" mode, i.e., running, over the course of the time that it was
4-125
-------
Draft Regulatory Impact Analysis
instrumented for study. This gives the time-weighted average of the proportion of NOx emitted
during cold start periods over a work day.
I
^coldstart,day op,day
Ptwa = —
L^ op,day
where/>twa is the time-weighted average of the cold-start proportion over all workdays and where
top,Aay is the total operating time for a given day (hours).
This proportion ranged from 1.8% to 9.7% over all the sampled equipment, with an average value
of 4.4%. The margin of error at a 95% confidence level (t-statistic) was +/- 1.4%.
Unresolved to date, is exactly what the rate of emissions from all pollutants might be in this
population of construction units with respect to both cold starts and at-temperature operations. EPA
has concerns that not all pollutants may be emitted at similar rates to those seen for NOx in this study.
Likewise, given the breadth of nonroad equipment and application types that exist, EPA will not
apply these results broadly to the nonroad equipment population but will use them as an indicator
that, in fact, cold start operation may account for a significant amount of nonroad operating emissions
over the course of a "typical" equipment workday. The time period over which these units were
tested is significant, as well. The model years of the sampled equipment ranged from 1985 to 2001,
and no piece of equipment was outfitted with either a diesel oxidation catalyst, PM filter trap or other
such emission control equipment. In future years, such emission control equipment may be more
common and the operating condition of the engine more a factor in when and how much pollutants
are emitted during various periods of engine operation, e.g., cold start.
At cold start "key-on", some units in this study were seen to operate at a slightly higher level of
engine idle than the unit's specified low-idle operation (approximately 10 percent higher engine
speed)156. After a short period, usually five to ten minutes, these engines dropped back to low-idle
speed operation. This type of operation, while the engine is still "cold", may be a contributor to
higher emission rates at start up for these engines, especially if higher exhaust temperatures will be
needed in the first few minutes of operation for on-board emission control systems in these types of
nonroad engines. In some of the equipment under EPA analysis, engine start up after periods of
inactivity during a typical workday lasting sixty minutes or longer exhibited exhaust temperatures
starting out below 100°C. Exhaust temperature remained under 150°C if the engine continued to
operate at low idle, sometimes falling back below 100°C. Total "warm" engine operating emissions
over the equipment's workday hours will surpass the day's "cold start" emissions by a large factor.
However, a nonroad diesel engine which is designed to emit less at cold start will have lower
emissions at other points in its operation as well and will on balance be a significantly cleaner engine
in complying with cold start regulations.
4.2.10 Applicability of Component Cycles to Nonroad Diesel Market
We started to pursue application-specific operating duty cycles which could be normalized for
laboratory testing of nonroad diesel engines in the 1997-1998 time frame. With a standardized set of
operating duty cycles, we would have a basis upon which to compare the brake-specific emission
4-126
-------
Technologies and Test Procedures for Low-Emission Engines
rates of nonroad engines both within and across horsepower 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, off-highway trucks
and 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, 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 would best represent the US nonroad equipment population
was aided by an analysis of the our nonroad equipment population database.157 Our source of data
placed the total 1995 nonroad equipment population figure at 7,100,113 units in the U.S. 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 engines which
operated at a constant speed with varying load requirements 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 emissions 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 below).
Table 4.2-16
Emissions Attributable to Base Nonroad Applications
Application
Ag tractor
Welder
Emission Distribution by
Application
34%
1%
4-127
-------
Draft Regulatory Impact Analysis
Application
Backhoe/loader
Crawler
Excavator
R/T Loader
Skid/steer
Total
Emission Distribution
Application
by
6%
7%
3%
6%
2%
59%
4.2.10.3 HP 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 (by horsepower).
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 all nonroad
horsepower in use. With similar equipment applications included, the equipment with an agricultural
tractor-like horsepower number or displacement approaches 30 percent. Backhoe loaders, crawler
dozers and rubber-tire loaders together accounted for 12 percent of the horsepower in the nonroad
population and, with similar applications included, accounted for approximately 35 percent of total
nonroad horsepower. The last three cycle component applications—excavators, skidsteer loaders and
arc welders, with arc welders and like equipment generally falling into the 50 horsepower and under
engine power band—constitute only 8 percent of total nonroad horsepower. 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 horsepower. Taking the sum of horsepower 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 below).
4-128
-------
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 Blow/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 4.2-16 below outlines the process by which a manufacturer of a particular nonroad
diesel engine might approach certification and compliance of that engine with EPA's proposed
nonroad transient (and steady-state) test requirements.
4-129
-------
Draft Regulatory Impact Analysis
Figure 4.2-16
Nonroad Diesel Engine Emission Testing Requirements
4-130
-------
Technologies and Test Procedures for Low-Emission Engines
Manf Seeks
Engine Duty
Cycle
Certification
CSVL Cycle
with set speed
optiofl
Cold Start
CSVL Cycle
with set speed
option
CSVL eifuipment
(welding or spray rig,
gen set, etc.)
1
Fire (5)
Mode Cycle
ISO 8178
"D-Z"
Weighting
Greater
T'tar, 19
-------
Draft Regulatory Impact Analysis
and may include a reasonable range of ambient and engine conditions. Second the contemplated Tier
4 nonroad standards anticipate the use of several emissions control technologies. NTE standards are
one of the most realistic and cost effective means to measure emissions from field-aged engines. This
helps ensure that emissions control systems are reliable for the useful life of an engine.
4.3.2 How does EPA characterize the highway NTE test procedures?
Refer to Chapter 4 of the Regulatory Impact Analysis of the highway HDDE rulemaking,
published January 19, 2001 for details on the highway NTE test procedures. Briefly, the highway
NTE provisions specify that averaging periods may be as short as 30 seconds in time, but under these
provisions testing is also restricted to a very limited region of engine operation, namely when all of
the following conditions are simultaneously met for at least 30 seconds (unless an aftertreatment
system regenerates, then the minimum time would be longer):
1. Speed greater than 15% above idle speed
2. Torque greater than or equal to 30% of maximum torque
3. Power greater than or equal to 30% of maximum power
4. Altitude less than or equal to 5500ft (~>82kPa)
5. Ambient temperature less than or equal to 100F at sea level to 86F at 5500ft (30-38C)
6. BSFC less than or equal to 105% minimum BSFC if not coupled to a multi-speed manual or
automatic transmissions
7. Outside of any manufacturer petitioned exclusion zone
8. Outside of any NTE region in which a manufacturer states that less than 5% of in-use time will be
spent
9. Intake manifold temperature greater than or equal to 86-100F (20-30C), depending upon intake
manifold pressure
10. Engine coolant temperature greater than or equal to 125-140F (52-60C), depending on intake
manifold pressure.
11. Exhaust aftertreatment temperature greater than or equal to 250C.
4.3.3 How does EPA characterize the alternate NTE test procedures mentioned above?
The alternate NTE test procedure would apply to all normal engine operation regardless of
speed-load combinations or a test period's frequency of idle, steady-state, or transient operation. This
all-inclusive range of engine operation is consistent with EPA-collected data from nonroad vehicles
and equipment. The alternate test procedure also requires only a few measured parameters, and this
facilitates simple on-vehicle measurements. The data reduction procedure utilizes a "constant work"
moving average that returns values weighted and calculated the same way that emissions data are
reduced from a CVS test cell or from a weighted steady-state test. This provides an engineering
target for manufacturers that is consistent with the FTP test procedures.
4.3.4 What limits might be placed on NTE compliance under the alternate test procedures?
The alternate NTE test procedures would apply to all normal operation. This may include
steady-state and/or transient engine operation. Given such NTE standards, the goal for the design
engineer is to ensure that engines are properly calibrated for controlling emissions under any
reasonably expected mode of engine operation. However, it may not be technologically feasible to
meet specific NTE standards under all ambient and engine conditions so we will adopt some
4-132
-------
Technologies and Test Procedures for Low-Emission Engines
restrictions to narrow the range of ambient conditions and engine operation under which an engine is
subject to the alternate NTE standards.
Engines are often designed to operate under extreme environmental conditions. To narrow
the NTE range of compliance for the design engineer, we are limiting emission measurements for
NTE testing to ambient temperatures up to +40° C, and to ambient pressures greater than or equal to
80.0 kPa. This allows testing over a wide range of conditions in addition to helping ensure that
engines are able to control emissions under the range of conditions under which they are likely to
operate. Because engine manufacturers already design entire engines to be reliable over an even
wider range of ambient conditions, it is reasonable to expect that similar design information is already
available to design low emissions engines that possess similar reliability with respect to emissions
performance. Information on these extreme conditions are already required for proper design and
construction of air intake systems, turbo-chargers, cooling systems, and lubricating systems.
And because of the catalysts expected to be utilized for NOx control, prolonged engine
operation that is insufficient for catalyst heating will also be considered in the alternative NTE. The
data reduction techniques for this is specified in the draft regulations in Appendix A. In summary, if
any 10% work interval (described below) has a flow-weighted average exhaust temperature of less
than 250C, then that interval's NOx emissions must be no greater than 4.00 g/kw-hr.
Other important limits that the Agency has considered in order to define NTE standards are
the averaging intervals and normalization procedures for data reduction. A longer averaging period
allows for greater data stability, due mainly to the smoothing effect of measuring over several events.
On the other hand, an overly long averaging period will likely mask areas of engine operation with
poor emission-control characteristics. Even if poor emissions occurred over a relatively short period
of a test, such high-emitting events may be indicative of a more serious deficiency once other engines
have been determined to possess similar deficiencies. This is especially true if additional testing
reveals that different in-use duty cycles cause the deficiency to result in poor emissions more
frequently.
In order to maintain consistent NTE multipliers between the on-highway and alternate NTEs,
similar on-highway and non-road data sets were analyzed using both procedures. The averaging
scheme of the alternate NTE was adjusted so that both NTE concepts "passed" and "failed" similarly.
Below is an example of such an analysis.
4-133
-------
Draft Regulatory Impact Analysis
On-Highway NTE Results
i>U/0
y|CO/
43/0
4Uyo
oco/
oO/o
IS ino/
— oU/o
>,
O
C oco/
0.09 0.15 0.21 0.26 0.32 0.38 0.43
NOx(g/Hp-hr)
40% T
35%
30%
J- 25% -
>.
1 20%
=
o-
J!
it 15%
10%
5%-\
0%
0.12
NOx Histogram of 10% Work Intervals
0.14
ShiftAverage=0.17 g/Hp-hr
0.17 0.20 0.22
NOx (g/Hp-hr)
0.25
Max
0.28
Analysis: The first chart depicts a histogram of the results by analyzing a real-world data set
with the alternative NTE. Note that at an arbitrary FTP standard of 0.2 g/hp-hr, the 1.5x multiplier
would result in a limit of 0.30 g/hp-hr. The alternative NTE reveals that the engine barely passes the
4-134
-------
Technologies and Test Procedures for Low-Emission Engines
0.30 NTE standard with a maximum value of 0.28. Note too that the average is just below the FTP
standard at 0.17 g/hp-hr. Now compare this to the same "barely passable" results returned by
analyzing the data according to the on-highway NTE. Since the multiplier is the same between the
two, one does observe that there is a small fraction of values greater than 0.30 in the highway NTE
analysis. This fraction is just less than 10% however, so the analysis indicates just greater than 90%
pass, which is part of the "pass" criteria outlined in the recent NTE settlement between EPA and
some engine manufacturers. The highway NTE's analysis also shows that none of the data points is
greater than 2x the NTE standard, which is also part of the settlement's "pass" criteria.
The alternate test procedure applies a "constant-work" moving average. In contrast to the
highway NTE, the alternate NTE comprises of engine operation within the same ambient bounds-but
at all speeds, loads, and BSFCs; there are neither manufacturer exclusion zones, nor exclusions based
on intake or coolant conditions. For NOx emissions only there is one minimum exhaust temperature
condition that needs to be met for a minimum duration. This results in a much more broad range of
operation covered by the alternate NTE. This is justified because the large number of applications in
the nonroad category (>6,000 different vehicle or equipment models) leads to a very broad range of
probable engine operation. This is evidenced in EPA's and other global organization's composite
non-road transient test cycles, and it is also obvious in the -6,000 hours of non-road on-vehicle data
that EPA collected during the years 2000 through 2002. Therefore, the alternate NTE procedure
requires that all collected data, over all speeds and loads, will be included as part of the data reduction
scheme.
4.3.5 How does the "constant-work" moving average work, and what does it do?
Since most engine operation will fall under the NTE standards, a sufficient emissions
averaging period is required to determine if higher emissions during a specific speed-load
combination is of significance. And because of the wide range of probable nonroad engine
operation-including extended periods of idle operation-the alternate NTE procedures employ a
"constant-work" moving average so that brake-specific emissions are not evaluated over short work
intervals.
Specifically, the procedure specifies that an individual "test period" consists of a minium of
six (6) hours engine-running operation. This test period is intended to be roughly consistent with a
typical operator work-shift using nonroad equipment. Engine-off time would essentially be skipped
until additional engine-running data can be merged with the previous engine-running data to continue
data averaging. The work over this test period is totaled and it is normalized to a 6-hr shift if more
than 6 hours of engine-running data was recorded. This means that if 8 hours were recorded, the total
work would be divided by 8-hours and multiplied by 6-hours. The averaging interval is 10% of the
normalized total work over the test period, and this 10% moving average will increment (or "move")
at increments of 1% of the total work. This moving average will cause averages to overlap in time.
This is desirable because it dampens the results' sensitivity to random combinations of speed and load
being averaged together because of their proximity in time. This moving average scheme always
returns at least 90 data points for comparison against the NTE standards. This reasonable number of
data points, each consisting of a significant portion of a test period's work (10%), allows for the use
of their maximum value to be compared to the NTE standards.
4-135
-------
Draft Regulatory Impact Analysis
EPA analyzed several statistical methods to compare these results to an NTE standard. It was
soon discovered that although the 90+ data points formed a convenient and continuous distribution of
emissions values, the distribution's characteristics changed shape from one real-world data set to
another. Some distributions seemed similar to normal or log-normal distributions, while others were
greatly skewed with long tails on one end or another. This meant that parametric statistics based on a
distributional assumption might be valid for one data set, but not on another. Therefore, statistics
such as arithmetic means, geometric means, standard deviations, and confidence intervals were useful
in some cases but not in all cases. Even powerful transformation algorithms were used in an attempt
to de-skew some of the data sets, but even these failed in certain cases. One of the few alternatives
for analyzing such data was to use non-parametric statistics, namely "tolerance intervals." In this case
a significant number of data points are collected (90+), and based only on the fact that this number of
points is collected, a level of confidence may be assigned as to whether the maximum emissions of
the data set reflects the maximum possible emissions from the engine. For the case of these 90+ data
points, non-parametric statistics suggest with 95% (+-4%) confidence that 95% of the engine's
emissions are less than the maximum of the 90+ data points. This statistic is consistent with the
intent of the NTE-it is a cap on the maximum allowable emissions. Note too that a 10% work
interval, based on a 6-hr shift was selected not only to return this number of data points but also it
was selected so that the NTE multipliers would be consistent with the highway multipliers.
There are several advantages to using the "constant-work" moving average approach. First,
all moving average data points from a complete test period represent brake-specific emissions over
the same work interval. This normalizes all of the data points so that they can be compared to each
other in a meaningful way. NTE standards set for the maximum value of this data set is effective
because the maximum of a set of significant (10% work) data points.
Note that this work-based normalization is similar to how an engine family is certified. A
representative engine from a family is "mapped" to determine its maximum output at each engine
speed, and this data set is used to "de-normalize" the certification test cycle so that the engine family
is always run over the same test interval of work, no matter whether it is tested by a manufacturer, by
an independent laboratory, or by EPA. This FTP constant work interval de-normalization also allows
high power engines to be certified to the same brake-specific standard over the same time interval,
but over a proportionally larger work interval. This element of certification testing is essential in
order to normalize engine emissions so that engines of different rated power can be compared to the
same brake-specific (e.g. work-specific) emissions standards.
In the alternate NTE the total in-use duty cycle of a 6 to!2-hr test period is used to determine
an appropriate work interval for the moving average of constant work. This use of a test period's
total work essentially utilizes characteristic "real-world" total work to determine the appropriate (and
constant) interval of work for data reduction. The total work of a test period will change from day-to-
day, so individual data points from different days may not be directly comparable, even within a
given engine family because the duty cycle on a given day will likely be different from another day.
However, it is precisely this variable-duty-cycle element of in-use testing that EPA intends to
preserve by using this data reduction scheme. By preserving this element of in-use testing, the engine
is compared against its real-world duty cycle, not a fixed certification test cycle or fixed test work
interval. This means that if an engine is tested under a qualitatively "high-load" test period, the work
intervals will be of a higher work value but still varying time intervals. Nevertheless, the engine will
be evaluated over a moving constant 10% work of that test period's total work. The opposite will
4-136
-------
Technologies and Test Procedures for Low-Emission Engines
also be true in that an engine evaluated over a "light-load" test period will be evaluated against lower
work values of constant work that will still occur over varying time intervals.
Another advantage to this data reduction scheme is that within a given test period, random
time intervals of relatively low power operation (i.e., idle) cause the work summing to take a longer
time period to complete. By using 10% work intervals, idle operation is consistently averaged with
other higher power output operation. This will always be true unless an in-use test period contains
more than 10% of continuous work at idle. Because a diesel engine uses about seventy-five times as
much fuel at rated power versus idle, this scenario is very unlikely. Before this could happen, the
data reduction scheme would require that over half of the test period's time would have to be spent
continuously, not intermittently, at idle in the case where the remainder of the test period averaged a
50% power duty-cycle. And if this were to be a real case, EPA would likely want to have such a
significant idle operation evaluated against the NTE standards.
Any low power operation less continuous than the extreme case described above will be
scaled relative to higher power operation because the field testing data reduction scheme "flow-
weights" each constant work interval's emissions. This is exactly the same way constant-volume
sampling (CVS) proportionally weights emissions during certification tests. Low-flow operation (i.e.,
idle) emissions are weighted proportionally less than other high-power emissions over a given work
interval.
EPA has evaluated this data reduction scheme on several highway and nonroad in-use data
sets collected from vehicles as they performed normal work in-use. We have determined that this
data reduction scheme reduces the data into a reasonable number of meaningful data points that can
be compared to the NTE standards in a consistent way. This data reduction scheme allows for the
maximum 10%-work emissions to be compared to NTE standards, and it allows the entire test
period's emissions to be compared to the FTP standard in a meaningful way.
4.3.6 What data would need to be collected in order to calculate emissions results using the
alternate NTE?
Emissions volume concentrations (i.e., ppm or %) from the raw exhaust would need to be
measured. These include total oxides of nitrogen (NO + NO2), total hydrocarbons (THC), carbon
monoxide (CO), and particulate matter mass (PM). THC needs to be converted to NMHC based on
proposed regulation §1039.240(e). Particulate matter mass may be measured over other varying work
intervals if a proportional integrating PM mass measurement technique is used rather than a PM mass
concentration measurement. In order to flow-weight concentrations similar to CVS sampling, a
signal linearly proportional to exhaust flow rate at standard conditions of 0 C, 101.325 kPa would
need to be measured. This value does not need to be an absolute value in engineering units because
the work calculation allows such unspecified units to be cancelled. Work may be calculated by first
calculating fuel consumed via carbon balance. Since complete combustion is an appropriate
assumption for diesel engines, fuel consumption may be determined by using carbon dioxide (CO2)
exhaust concentration multiplied by the signal proportional to exhaust flow. The fuel's atomic
hydrogen-to-carbon ratio also needs to be factored into this calculation. This fuel consumption is
then multiplied by the engine family's characteristic brake-specific fuel consumption to arrive at total
work. For the alternate NTE, the characteristic brake-specific fuel consumption will be the arithmetic
mean of the engine family's certification test cycles' brake-specific fuel consumptions. If an engine
4-137
-------
Draft Regulatory Impact Analysis
family is only certified to one test cycle, then its brake-specific fuel consumption over that test cycle
will be the characteristic brake-specific fuel consumption for testing.
Using this characteristic brake-specific fuel consumption has several advantages. First it
causes an engine family's characteristic certification efficiency to affect subsequent field-testing
results. This causes an engine with poor certification fuel economy (high brake-specific fuel
consumption), but favorable real-world fuel economy, to have proportionally higher emissions results
during field testing. This effect is small (i.e., about 5%), but this characteristic should help to
discourage manufacturers from designing low fuel economy solutions for meeting EPA certification
tests; especially with the knowledge that the engine is not likely to see certification-type operation in
use. According to recent manufacturer consent decrees this has resulted in low emissions and low
fuel economy at certification, but high emissions and improved fuel economy during "off-cycle"
operation. This off-cycle operation happened to occur frequently in-use, so that improved fuel
economy was realized with the consequence of significantly higher in-use emissions. Second, by
utilizing the characteristic certification efficiency, there is no requirement to measure engine output
torque in-use. In fact the Agency believes such a requirement might be cumbersome because torque
measurement may require attachment of a measurement device to the rotating output shaft, which
may be close-coupled and sealed in a transmission housing. Additionally, nonroad engines
sometimes have multiple output shafts, which would require multiple torque instrument installations;
further complicating testing. Another advantage of using an average BSFC is that when measuring
emissions at idle, a small amount of work is summed in the denominator of the emissions calculation.
Because of the likelihood that nonroad engines are actually performing some work at idle by
powering hydraulic, electric, or pneumatic accessories, this work should be included, and by using an
average BSFC, it is included.
Other than emissions concentrations and exhaust flow, the only other required measurements
would be those of ambient temperature and pressure for the purposes of determining if the data is
collected within the range of applicable ambient conditions.
In conclusion, the measurement requirements will likely be minimal: emissions volume
concentrations including CO2, a signal linearly proportional to standard exhaust flow, ambient
temperature, and ambient pressure.
4.3.7 Could data from a vehicle's on-board electronics be used to calculate emissions?
EPA will likely allow any data from a vehicle's on-board electronics to be used in the data
reduction scheme, provided that it meets the data accuracy and precision requirements specified in the
alternate NTE regulations. Additionally, the manufacturer would likely have to attest that such data
meets these requirements at the time of NTE testing.
4.3.8 How would anyone test engines in the field?
To test engines without removing them from equipment, analyzers would be connected to the
engine's exhaust to detect emission volume concentrations during normal operation. A signal linearly
proportional to standard exhaust volumetric flow rate should also be measured to convert the analyzer
responses to units of g/kW-hr for comparing to NTE standards. Ambient temperature and pressure
would also have to be measured to determine if the NTE standards were applicable.
4-138
-------
Technologies and Test Procedures for Low-Emission Engines
Available small analyzers and other equipment will likely be adapted for measuring
emissions. A portable heated flame ionization detector (HFID) will likely be used to measure total
hydrocarbon concentrations. A portable NDUV or Zirconia-based analyzer will likely be used to
measure total NOx emissions. A nondispersive infrared (NDIR) analyzer will likely be used to
measure CO and CO2. Technologies such as a Tapered Element Oscillating Microbalance or a Quartz
Crystal Microbalance will likely be used to inertially measure PM mass emissions.
Emission samples can best be drawn from the exhaust flow directly downstream from an
aftertreatment system to avoid diluting effects from the end of the tailpipe. Installing a sufficiently
long tailpipe extension will also likely be an acceptable way to avoid dilution.
4.3.9 How might in-use crankcase emissions be evaluated?
The anticipated crankcase emission-control technologies are best evaluated by visually
checking if they continue to function as designed. A visual inspection of in-use engine crankcase
emission-controls is appropriate to verify that these systems continue to function properly throughout
useful life. Furthermore, as stated in the preamble to this proposed rulemaking in section ni.B.2,
manufacturers that choose not to utilize the closed crankcase approach for addressing crankcase
emissions control would be responsible for ensuring that crankcase emissions would be readily
measurable in use.
4.3.10 How might the agency characterize the technological feasability for manufacturers to
comply with NTE standards?
The Agency acknowledges that compliance with NTE standards will require design engineers
to better understand their engines' emission behavior over a wide range of possible engine operation.
Though claims have been made that NTE standards might be interpreted to cover a theoretically
infinite degree of variability, we have determined that by evaluating a range of in-use duty cycles, a
consistent level of control for any additional operation may be predicted. Making careful
measurements over a statistically sound sampling plan provides reasonable certainty that any future
emissions from an engine is likely to be within certain bounds. Such statistics are frequently used to
ensure reliability of engine parts and engine performance, and we expect similar care to be taken
when designing engines to meet NTE standards. We do not believe manufacturers will need to test
an "infinite" or inappropriately large number of steady state and transient combinations. Rather,
manufacturers will be able to quickly narrow their test programs to focus in on those areas of
operation where the emissions are higher and come closer exceeding the NTE standards. Engineering
experience and logic dictates that manufacturers will not expend resources testing areas where
emissions are well understood and well below the NTE standards.
The same is true with respect to ambient conditions within the specified field-testing bounds.
The effects of temperature and pressure on emissions are well known, so manufacturers may limit
their testing to those ambient conditions that cause the highest emissions. Alternatively,
manufacturers might choose to not test under conditions representing the endpoints of the established
ranges, but rather they might test under "mid-range" conditions and rely on established extrapolation
methods to ensure that their engines will meet emission standards when tested throughout the range
of specified test conditions. If a manufacturer shows that engines meet emission standards under the
most challenging conditions, then engines will meet the standards under less challenging conditions.
4-139
-------
Draft Regulatory Impact Analysis
Because manufacturers have already demonstrated that they can build complex engines to be
reliable for very many years beyond EPA's regulatory useful life, even when these engines have been
operated throughout a wide range of extreme ambient conditions, the Agency believes that
manufacturers can utilize already known design parameters and engineering and testing techniques to
ensure that low emitting engines and aftertreatment systems are similarly reliable for emissions
reductions at least throughout regulatory useful life and under similar conditions.
We also expect the manufacturers' statements at certification to state that they meet the NTE
standards. These statements should be based on reasonable evidence of compliance, engineering
analysis and good engineering judgment. We do not expect manufacturers to have tested every
possible combination of points to be able to make their certifying statement.
In addition, we will put limits on the range of ambient conditions under which NTE standards
might be evaluated. For example, during emission tests ambient air temperature must be between -0°
C and 40° C and barometric pressure must be at least 80 kPa.
By restricting the NTE standards to "normal operation", we will likely allow manufacturers to
include in engine designs any limitations applicable to normal operation. For example, if a
manufacturer includes in the emission-related installation instructions a warning that the engine must
not be installed to power a pump greater than some specific pumping rate, and takes steps to enforce
that restriction, we would not consider such engine operation to be "normal operation" under the NTE
standards. In some cases, manufacturers may also program their engines with a governor or other
device to prevent engines from operating at certain speeds or loads.
Without NTE standards, we anticipate that some manufacturers might design their emission-
control systems to function effectively only over the narrow range of engine operation and ambient
conditions represented by the certification duty cycles. We feel that in these cases the NTE standards
might be interpreted as increasing the overall stringency of the regulation. However, the basis for
such a conclusion entirely depends upon a manufacturer's intended approach to meet emissions
regulations. EPA has always intended for manufacturers to design solutions that ensure emissions
control over a broad range of conditions and throughout useful life of an engine. Therefore, EPA
believes that NTE standards do not increase the stringency of the overall regulation, but rather NTE
standards ensure that engines are designed to meet the intention of the FTP standards.
In any case NTE standards evaluated via in-use testing will correspond directly with a more
effective control of emissions from in-use engines as they undergo normal operation in nonroad
applications. We also believe manufacturers will have available emission-control hardware (and
software) that allows for more robust control over a wide range of operation and conditions. With
some additional engineering, manufacturers can ensure that engines operate properly over the whole
range of normal operation.
We already have equipment available to measure emissions using NTE procedures.
Moreover, NTE standards take into account measurement tolerances and the variation in emissions
due to varying engine operation and ambient conditions. Given the very active interest in portable
measurement equipment in the rest of the industry, and given the lead time of this NPRM, we believe
that measurement equipment will be widely available well ahead of time so that the NTE standards
will likely apply to nonroad compression-ignition engines in 2011. We also believe that the
4-140
-------
Technologies and Test Procedures for Low-Emission Engines
measurement technology to meet the NTE standards is sufficiently known so that a technology review
is not likely to be necessary.
In the early years of any such program, manufacturers are more likely to devote more of their
effort to meet the NTE standards as they learn better how their engines behave under different types
of operation. However, as they gain experience in designing robust emission-control systems by
interpreting NTE test results, we would expect manufacturers to focus more on meeting the duty-
cycle standards, knowing that emission variability has been controlled enough that the NTE standards
no longer pose a significant additional constraint in their efforts to comply with all of the standards.
We have already set NTE standards for heavy-duty highway compression-ignition engines,
large spark-ignition engines, and marine engines, and we believe that any proposed nonroad NTE
standards take into account the unique aspects of operation and technology for nonroad compression-
ignition engines. We believe that the information available today is ample to support our conclusions
to propose NTE standards and field testing procedures for diesel nonroad engines.
We believe manufacturers will clearly do well by relying on these procedures to meet
emission-testing requirements at a substantially lower cost than would be involved with laboratory
testing.
The steady-state and transient test requirements clearly provide substantial assurance that
engines will be controlling emissions under the kinds of operation seen when installed in the various
types of nonroad equipment. We believe the NTE standards are an appropriate supplement to the
duty-cycle standards for two reasons. First, any duty cycle, even one with transient engine operation
cannot capture the whole range of "normal operation" from the multitude of different types of
nonroad equipment. This may be especially important, since some of these engines might be
operating in confined spaces where high emission levels pose a concern for individual exposures in
addition to the more general issue of pollution in urban areas. The certification duty cycles will
include many different combinations of speed, load, acceleration, and deceleration, but they cannot
include or substantially weight the whole range of operation that engines may experience. This is
underscored by in-use emission data generated to support the NTE standards. Second, without field-
testing procedures, manufacturers would only be able to meet in-use testing requirements by
removing engines from service and testing them in the laboratory.
4-141
-------
Draft Regulatory Impact Analysis
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, Air Docket A-2001-28.
3. Exhaust and Crankcase Emission Factors for Nonroad Engine Modeling - Compression-
Ignition, EPA420-P-02-016, NR-009B, Air Docket A-2001-28.
4. Onishi, S. et al, "Active Thermo-Atmosphere Combustion (ATAC) - A New Combustion
Process for Internal Combustion Engines," SAE 790840.
5. Najt, P. and Foster, D. "Compression-Ignited Homogeneous Charge Combustion," March
1983, SAE 830264.
6. Dickey, D. et al, "NOx Control in Heavy-Duty Diesel Engines - What is the Limit," February,
1998, SAE 980174.
7. 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.
8. 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.
9. Gray, A. and Ryan, T., "Homogenous Charge Compression Ignition (HCCI) of Diesel Fuel,"
May, 1997 SAE 971676.
10. 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.
11. Stanglmaier, R. and Roberts, C. "Homogenous Charge Compression Ignition (HCCI):
Benefits, Compromises, and Future Engine Applications," SAE 1999-01-3682.
12. "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.
13. "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.
14. Miller, R. et. al, "Design, Development and Performance of a Composite Diesel Particulate
Filter," March 2002, SAE 2002-01-0323.
4-142
-------
Technologies and Test Procedures for Low-Emission Engines
15. Hori, S. and Narusawa, K. "Fuel Composition Effects on SOF and PAH Exhaust Emissions
from DI Diesel Engines," SAE 980507.
16. "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.
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. Hawker, P., et. al., Effect of a Continuously Regenerating Diesel Particulate Filter on Non-
Regulated Emissions and Particle Size Distribution, SAE 980189.
19. Application of Diesel Particulate Filters to Three Nonroad Engines - Interim Report, January
2003. Copy available in EPA Air Docket A-2001-28.
20. "Nonroad Diesel Emission Standards - Staff Technical Paper", EPA Publication EPA420-R-
01-052, October 2001. Copy available in EPA Air Docket A-2001-28.
21. 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.
22. "Particulate Traps for Construction Machines, Properties and Field Experience," 2000, SAE
2000-01-1923.
23. Letter from Dr. Barry Cooper, Johnson Matthey, to Don Kopinski, US EPA, Air Docket A-
2001-28.
24. 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.
25. 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.
26. Highway Diesel Progress Review, United States Environmental Protection Agency, June
2002, EPA 420-R-02-016, Air Docket A-2001-28.
27. "Nonroad Diesel Emissions Standards Staff Technical Paper", EPA420-R-01-052, October
2001, Air Docket A-2001-28.
28. Allansson, et al, European Experience of High Mileage Durability of Continuously
Regenerating Diesel Particulate Filter Technology. SAE 2000-01-0480.
29. 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.,
4-143
-------
Draft Regulatory Impact Analysis
"Class 8 Trucks Operating On Ultra-Low Sulfur Diesel With Particulate Filter Systems: A Fleet
Start-Up Experience," SAE 2000-01-2821.
30. 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.
31. 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.
32. 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.
33. 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.
34. 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.
35. 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.
36. Flynn, P. et al, "Minimum Engine Flame Temperature Impacts on Diesel and Spark-Ignition
Engine NOx Production," SAE 2000-01-1177, March 2000.
37. Stanglmaier, Rudolf and Roberts, Charles "Homogenous Charge Compression Ignition
(HCCI): Benefits, Compromises, and Future Engine Applications". SAE 1999-01-3682.
38. 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.
39. 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.
40. Kawanami, M., et. al., Advanced Catalyst Studies of Diesel NOx Reduction for Highway
Trucks, SAE 950154.
41. 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 USA Dept. of Energy, Clean Air for
Europe and Japan Clean Air, October 9-10, 2002. Copy available in EPA Air Docket A-2001-28.
42. Koichiro Nakatani, Shinya Hirota, Shinichi Takeshima, Kazuhiro Itoh, Toshiaki Tanaka, and
Kazuhiko Dohmae, "Simultaneous PM and NOx Reduction System for Diesel Engines.", SAE
4-144
-------
Technologies and Test Procedures for Low-Emission Engines
2002-01-0957, SAE Congress March 2002.
43. 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.
44. Gregory, D. et al., "Evolution of Lean-NOx Traps on PFI and DISI Lean Burn Vehicles",
SAE 1999-01-3498.
45. McDonald, J., et al., "Demonstration of Tier 2 Emission Levels for Heavy Light-Duty
Trucks," SAE 2000-01-1957.
46. Brogan, M, et. al., Evaluation of NOx Adsorber Catalysts Systems to Reduce Emissions of
Lean Running Gasoline Engines, SAE 962045.
47. Gregory, D. et al., "Evolution of Lean-NOx Traps on PFI and DISI Lean Burn Vehicles",
SAE 1999-01-3498.
48. 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.
49. Brogan, M, et. al., Evaluation of NOx Adsorber Catalysts Systems to Reduce Emissions of
Lean Running Gasoline Engines, SAE 962045.
50. Gregory, D. et al., "Evolution of Lean-NOx Traps on PFI and DISI Lean Burn Vehicles",
SAE 1999-01-3498.
51. Highway Diesel Progress Review, United States Environmental Protection Agency, June
2002, EPA 420-R-02-016, Air Docket A-2001-28.
52. Kato, N. et al, "Thick Film ZrO2 NOx Sensor for the Measurement of Low NOx
Concentration," February 1998, SAE 980170.
53. Kato, N. et al, "Long Term Stable NOx Sensor with Integrated In-Connector Control
Electronics," March 1999, SAE 1999-01-0202.
54. 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.
55. Diesel Emission Control - Sulfur Effects (DECSE) Program Phase II Summary Report: NOx
Adsorber Catalysts, October 2000. Copy available in Air Docket A-2001-28.
56. 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.
57. Jobson, E. et al, "Research Results and Progress in LeaNOx U - A Cooperation for Lean
NOx Abatement " SAE 2000-01-2909.
4-145
-------
Draft Regulatory Impact Analysis
58. Asanuma, T. et al, "Influence of Sulfur Concentration in Gasoline on NOx Storage -
Reduction Catalyst," SAE 1999-01-3501.
59. Guyon, M. et al, "NOx-Trap System Development and Characterization for Diesel Engines
Emission Control," SAE 2000-01-2910.
60. Dou, Danan and Bailey, Owen, "Investigation of NOx Adsorber Catalyst Deactivation,"
SAE 982594.
61. Guyon, M. et al, "Impact of Sulfur on NOx Trap Catalyst Activity - Study of the
Regeneration Conditions", SAE 982607.
62. Dearth, et al, "Sulfur Interaction with Lean NOx Traps: Laboratory and Engine
Dynamometer Studies", SAE 982595.
63. Guyon, M. et al, "NOx-Trap System Development and Characterization for Diesel Engines
Emission Control," SAE 2000-01-2910.
64. Dou, D and Bailey, O.,"Investigation of NOx Adsorber Catalyst Deactivation," SAE 982594.
65. Dearth, et al, "Sulfur Interaction with Lean NOx Traps: Laboratory and Engine
Dynamometer Studies", SAE 982595.
66. Dearth, et al, "Sulfur Interaction with Lean NOx Traps: Laboratory and Engine
Dynamometer Studies", Figure 5 SAE 982595.
67. Dearth, et al, "Sulfur Interaction with Lean NOx Traps: Laboratory and Engine
Dynamometer Studies", SAE 982595.
68. Dou, D and Bailey, O.,"Investigation of NOx Adsorber Catalyst Deactivation," SAE 982594.
69. Heck, R. and Farrauto, R. Catalytic Air Pollution Control - Commercial Technology, page
64-65. 1995 Van Nostrand Reinhold Publishing.
70. Heck, R. and Farrauto, R. Catalytic Air Pollution Control - Commercial Technology,
Chapter 6. 1995 Van Nostrand Reinhold Publishing.
71. Asanuma, T. et al, "Influence of Sulfur Concentration in Gasoline on NOx Storage -
Reduction Catalyst," SAE 1999-01-3501.
72. Diesel Emission Control - Sulfur Effects (DECSE) Program Phase II Summary Report: NOx
Adsorber Catalysts, October 2000. Copy available in Air Docket A-2001-28.
73. Tanaka, H., Yamamoto, M., "Improvement in Oxygen Storage Capacity," SAE 960794.
74. Yamada, T., Kobayashi, T., Kayano, K., Funabiki M., "Development of Zr Containing TWC
Catalysts", SAE 970466.
75. McDonald, Joseph, and Lee Jones, U.S. EPA, "Demonstration of Tier 2 Emission Levels for
Heavy Light-Duty Trucks," SAE 2000-01-1957.
4-146
-------
Technologies and Test Procedures for Low-Emission Engines
76. Dearth, et al, "Sulfur Interaction with Lean NOx Traps: Laboratory and Engine
Dynamometer Studies", SAE 982595.
77. 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 H-G-137.
78. Reyes and Cutshaw, SCONOx Catalytic Absorption System, December 8, 1998,
www.glet.com Air Docket A-99-06 item U-G-147.
79. Danziger, R. et. al. 21,000 Hour Performance Report on SCONOX, 15 September 2000 EPA
Docket A-99-06 item IV-G-69.
80. 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
(i.e., would use a conventional 3-way catalyst, indicated by lambda symbol = 1). Copy available
in Air Docket A-2001-28.
81. 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.
82. 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.
83. 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.
84. 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.
85. 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.
86. 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.
87. 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.
88. Diesel Emission Control Sulfur Effects (DECSE) Program - Phase I Interim Data Report No.
1, August 1999. Copy available in Air Docket A-2001-28.
89. 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.
4-147
-------
Draft Regulatory Impact Analysis
90. Diesel Emission Control Sulfur Effects (DECSE) Program - Phase I Interim Date Report No.
3: Diesel Fuel Sulfur Effects on Paniculate Matter Emissions, November 1999. Copy available
in Air Docket A-2001-28.
91. 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.
92. Diesel Emission Control - Sulfur Effects (DECSE) Program Phase II Summary Report: NOx
Adsorber Catalysts, October 2000. Copy available in Air Docket A-2001-28.
93. Diesel Emission Control - Sulfur Effects (DECSE) Program Phase II Summary Report: NOx
Adsorber Catalysts, October 2000. Copy available in Air Docket A-2001-28.
94. 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..
95. 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 USA Dept. of Energy, Clean Air for
Europe and Japan Clean Air, October 9-10, 2002. Copy available in EPA Air Docket A-2001-28.
96. "Demonstration of Advanced Emission Control Technologies Enabling Diesel-Powered
Heavy-Duty Engines to Achieve Low Emission Levels", Manufacturers of Emissions Controls
Association, June 1999 Air Docket A-2001-28.
97. 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.
98. 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.
99. Engelhard DPX catalyzed diesel paniculate filter retrofit verification,
www.epa.gov/otaq/retrofit/techlist-engelhard.htm. a copy of this information is available in Air
Docket A-2001-28.
100. 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.
101. "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
102. 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.
4-148
-------
Technologies and Test Procedures for Low-Emission Engines
103. 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.
104. "Summary of Conference Call between US EPA and Deutz Corporation on September 19,
2002 regarding Deutz Diesel Particulate Filter System", EPA Memorandum to Air Docket A-
2001-28.
105. "Particulate Traps for Construction Machines: Properties and Field Experience" J.
Czerwinski et. al., Society of Automotive Engineers Technical Paper 2000-01-1923.
106. "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#n-A-12.
107. 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.
108. 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.
109. See Chapter 14, Section 4 of "Turbocharging the Internal Combustion Engine", N. Watson
and M.S. Janota, published by John Wiley and Sons, 1982.
110. 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.
111. 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.
112. EPA Memorandum "2002 Model Year Certification Data for Engines <50 Hp", William
Charmley, copy available in EPA Air Docket A-2001-28"
113. 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.
114. 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 HI, 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.
4-149
-------
Draft Regulatory Impact Analysis
115. "Effects of Injection Pressure and Nozzle Geometry on DI Diesel Emissions and
Performance," Pierpont, D., and Reitz, R., SAE Paper 950604, 1995.
116. 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.
117. 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.
118. 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.
119. "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.
120. 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, Docket Item # II-A-26.
121. "Demonstration of Advanced Emission Control Technologies Enabling Diesel-Powered
Heavy-duty Engines to Achieve Low Emission Levels: Interim Report Number 1 - Oxidation
Catalyst Technology and "Reduction of Diesel Exhaust Emissions by Using Oxidation
Catalysts", P. Zelenka et. al., Society of Automotive Engineers paper 902111, October 1990.
122. "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.
123. 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.
124. 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.
125. 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.
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 "Summary of Model Year 2001 Certification data for Nonroad Tier 1
Compression-ignition Engines with rated power between 0 and 50 horsepower", William
4-150
-------
Technologies and Test Procedures for Low-Emission Engines
Charmley, copy available in EPA Air Docket A-2001-28, docket item II-B-08.
128. "Effects of Injection Pressure and Nozzle Geometry on DI Diesel Emissions and
Performance," Pierpont, D., and Reitz, R., SAE Paper 950604, 1995.
129. 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.
130. 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.
131. "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.
132. "Demonstration of Advanced Emission Control Technologies Enabling Diesel-Powered
Heavy-duty Engines to Achieve Low Emission Levels: Interim Report Number 1 - Oxidation
Catalyst Technology.
133. Letter from Marty Barris, Donaldson Corporation, to Byron Bunker US EPA, March 2000.
A copy is available in Air Docket A-2001-28.
134. Hawker, P. et al, "Experience with a New Particulate Trap Technology in Europe," SAE
970182.
135. Hawker, P. et al, "Experience with a New Parti culate Trap Technology in Europe," SAE
970182.
136. Allansson, et al., "European Experience of High Mileage Durability of Continuously
Regenerating Filter Technology," SAE 2000-01-0480.
137. Letter from Dr. Barry Cooper, Johnson Matthey, to Don Kopinski, US EPA.A copy is
available in Air Docket A-2001-28.
138. Telephone conversation between Dr. Barry Cooper, Johnson Matthey, and Todd Sherwood,
EPA, Air Docket A-99-06.
139. Letter from Dr. Barry Cooper to Don Kopinski US EPA. A copy is available in Air Docket
A-2001-28.
140. Dou, Danan and Bailey, Owen, "Investigation of NOx Adsorber Catalyst Deactivation,"
SAE 982594.
141.Guyon, M. et al, "Impact of Sulfur on NOx Trap Catalyst Activity - Study of the
Regeneration Conditions", SAE 982607.
4-151
-------
Draft Regulatory Impact Analysis
142. 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.
143. Nonroad Test Cycle Development, Starr, M., Southwest Research Institute Contractor
report for the United States Environmental Protection Agency, September 1998
144. Nonroad Data Analysis and Composite Cycle Development, Webb, C., Southwest Research
Institute contractor report to the United States Environmental Protection Agency, September
1997
145. Memorandum from Kent Helmer to Cleophas Jackson, "National Excavator Fleet
Population Estimate", (Docket A-2001-28)
146.Bin Analysis of Nonroad Diesel Transient Duty Cycles; Hoffman, G., Dyntel Corporation;
Ann Arbor, MI, March 2003
147. See also memorandum to Docket "Maximum Speed Determination Procedure", Docket A-
2001-28.
148. This report may be found on and downloaded from the EPA-OTAQ website at
http://www.epa.gov/otaq/marine.htm. Follow links to the "December 29th, 1999 Marine Final
Rulemaking (FRM)" for the "Summary and Analysis of Comments" document.
149. Please see url: http://www.epa.gov/oms/regs/nonroad/equip-hd/cycles/nrcycles.htm
150. Memorandum from Kent Helmer to Cleophas Jackson, "In-house Testing of Arc Welder
Application Cycles for Record, February 7th, 2003", Docket A-2001-28.
151. Summary Note of Regression Statistics on Contract-testing of the Arc Welder Cycles on
various Dynamometer-mounted Engines Sent to Engine Manufacturers Association, Docket A-
2001-28.
152.Memorandum to Docket from Kent Helmer to Cleophas Jackson "Engine Control in
Transient Operations on a Dynamometer Test Bench", Docket A-2001-28.
153. Memorandum from Kent Helmer to Cleophas Jackson, "In-house Testing of Arc Welder
Application Cycles for Record, February 7th, 2003", Docket A-2001-28.
154. ISO Report on NRTC Cycle Development "Final Report on NRTC test Procedure, Summer
2002" Docket A-2001-28.
155. Memorandum From Kent Helmer to Cleophas Jackson "Cold Start Analysis of PEMs-SPOT
Data for 13 Nonroad Construction Equipment Units", docket A-2001-28.
156. Memorandum From Kent Helmer to Cleophas Jackson "Cold Start Analysis of PEMs-SPOT
Data for 13 Nonroad Construction Equipment Units", docket A-2001-28.
4-152
-------
Technologies and Test Procedures for Low-Emission Engines
157. Memorandum from Kent Helmer to Cleophas Jackson, "Applicability of EPA's NRTC
cycle to the US Nonroad Diesel Population", (Docket A-2001-28).
158. "Changes Considered for Part 1065-Test Procedures and Equipment," memorandum to
docket A-2001-28.
4-153
-------
CHAPTER 6: Estimated Engine and Equipment Costs
6.1 Methodology for Estimating Engine and Equipment Costs 6-1
6.2 Engine-Related Costs 6-4
6.2.1 Engine Fixed Costs 6-4
6.2.1.1 Engine and Emission Control Device R&D 6-4
6.2.1.2 Engine-Related Tooling Costs 6-13
6.2.1.3 Engine Certification Costs 6-17
6.2.2 Engine Variable Costs 6-20
6.2.2.1 NOx Adsorber System Costs 6-24
6.2.2.2 Catalyzed Diesel Particulate Filter Costs 6-29
6.2.2.3 CDPF Regeneration System Costs 6-34
6.2.2.4 Diesel Oxidation Catalyst (DOC) Costs 6-36
6.2.2.5 Closed-Crankcase Ventilation (CCV) System Costs 6-38
6.2.2.6 Variable Costs of Conventional Technologies for Engines Below 75 Horsepower and over 750
Horsepower 6-39
6.2.3 Engine Operating Costs 6-44
6.2.3.1 Operating Costs Associated with Oil Change Maintenance for New and Existing Engines . . . 6-45
6.2.3.2 Operating Costs Associated with CDPF Maintenance for New CDPF-Equipped Engines .... 6-49
6.2.3.3 Operating Costs Associated with Fuel Economy Impacts on New Engines 6-50
6.2.3.4 Operating Costs Associated CCV Maintenance on New Engines 6-55
6.3 Equipment-Related Costs 6-55
6.3.1 Equipment Fixed Costs 6-56
6.3.1.1 Equipment Redesign Costs 6-56
6.3.1.2 Costs Associated with Changes to Product Support Literature 6-59
6.3.1.3 Total Equipment Fixed Costs 6-60
6.3.2 Equipment Variable Costs 6-63
6.3.3 Potential Impact of the Transition Provisions for Equipment Manufacturers 6-64
6.4 Summary of Engine and Equipment Costs 6-66
6.4.1 Engine Costs 6-66
6.4.1.1 Engine Fixed Costs 6-66
6.4.1.2 Engine Variable Costs 6-66
6.4.1.3 Engine Operating Costs 6-67
6.4.2 Equipment Costs 6-68
6.4.2.1 Equipment Fixed Costs 6-68
6.4.2.2 Equipment Variable Costs 6-69
6.5 Costs for Example Pieces of Equipment 6-69
6.5.1 Summary of Costs for Some Example Pieces of Equipment 6-69
6.5.2 Method of Generating Costs for Our Example Pieces of Equipment 6-70
-------
Estimated Engine and Equipment Costs
CHAPTER 6: Estimated Engine and Equipment Costs
This chapter discusses the various engine and equipment cost elements considered for the
proposed emission standards and presents the total engine and equipment related costs we have
estimated for compliance with the proposed new standards. First, in Section 6.1, a brief outline
of the methodology used to estimate the engine and equipment cost impacts is presented. Next,
in Sections 6.2 and 6.3, the projected costs of the individual technologies expected to be used to
comply with the proposed standards are presented, 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 presents all engine, equipment, and operating costs in a concise
format. Section 6.5 then presents cost estimates for several example pieces of equipment. A
complete presentation of the aggregate cost of compliance for engines and equipment is
presented in Chapter 8 of this Draft RIA.
Note that we do not present any sensitivity analysis here. An analysis of sensitivity is
presented in Chapter 9 where we present monetized benefits and social costs. 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, because these are voluntary programs that,
while we fully expect industry to use them to reduce compliance costs, they are not required to
do so; all compliance costs presented here are for proposed regulatory requirements. Unless
noted otherwise, all costs presented here are in 2001 dollars.
6.1 Methodology for Estimating Engine and Equipment Costs
This analysis makes a number of simplifying assumptions regarding how manufacturers
would comply with the proposed standards. First, in each horsepower category, we assume a
single technology recipe as discussed in detail in Chapter 4 of this Draft RIA. However, we
expect that each manufacturer would evaluate all possible technology avenues to determine the
one or ones that best balance costs while ensuring compliance. In addition, we fully expect
manufacturers to make use of both the averaging, banking, and trading (ABT) program for
engine manufacturers and the transition program for equipment manufacturers (TPEM) as a way
to deploy varying degrees of emission control technologies on different engines and equipment.
As noted, for developing cost estimates, we have assumed that the industry does not use either
the TPEM or ABT programs, both of which offer the opportunity for significant cost reductions.
Given these simplifying assumptions, we believe that the cost projections presented here provide
a conservative cost estimate that probably overestimates the costs of the different approaches
toward compliance that manufacturers may ultimately take.
For smaller nonroad engines - those under 75 horsepower - many of the technologies we
expect would be needed for compliance would be applied for the first time. Therefore, we have
sought input from a large section of the regulated community regarding the future costs that
would be incurred to apply these technologies to diesel engines. Under contract from EPA, ICF
6-1
-------
Draft Regulatory Impact Analysis
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 would be required for compliance. These costs form the basis for our estimated costs for
"traditional" engine technologies such as EGR and fuel injection systems.1
Costs for exhaust emission control devices (e.g., catalyzed diesel particulate filters (CDPF),
NOx adsorbers, and diesel oxidation catalysts (DOC)) were estimated using the methodology
used in our 2007 HD highway diesel rulemaking. In that rulemaking effort, ICF Consulting,
under contract to EPA, provided surveys to nine engine manufacturers seeking their estimates of
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 for advanced
emission control technologies we expected would be applied in order to meet the proposed 2007
heavy-duty diesel highway standards.2 These costs were then further refined by EPA based upon
input from members of the Manufacturers of Emission Controls Association. Because the
exhaust emission control technologies expected for compliance with the proposed nonroad
standards 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 are using that analysis
as the basis for our cost estimates here.
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 upon 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 upon estimates from suppliers on expected total
costs to the manufacturers (inclusive of markups).3 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 marked up three percent to account for the cost of capital
tied up in inventory. This approach to estimating manufacturer and dealer markups to better
reflect the value added at each stage of the cycle was adopted by EPA based on industry input.4
EPA has also identified various factors that would 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.5 This
analysis incorporates the effects of this learning curve as described in Section 6.2.2 of this
chapter.
Fixed costs for engine R&D are estimated to be incurred over the five-year period preceding
introduction of the engine. Fixed costs for tooling and certification are estimated to be incurred
6-2
-------
Estimated Engine and Equipment Costs
one year ahead of initial production. Fixed costs for equipment R&D/redesign 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 costs are increased by seven percent for every year before the start of production. Engine
fixed costs are then "recovered" with a five-year amortization at the same rate 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 during the five years following 100 percent compliance.A Equipment fixed costs
are recovered with a 10-year amortization at the same seven percent rate; 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, potential 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 simplistic overview of the methodology used to estimate engine and equipment costs
would be 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 costs of recovery represent our
estimate of fixed costs associated with the proposal. 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 horsepower
categories in this cost analysis and by other factors (e.g., the engine prices we have estimated)
as discussed in more detail below. Because we do not know how manufacturers would
actually 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 of this Draft RIA.
• For engine variable costs (i.e., emission control and associated hardware), we first estimate
the cost per piece of technology. As described in detail in Section 6.2.2, emission control
hardware costs tend to be directly related to engine characteristics - e.g., exhaust emission
control devices are sized according to engine displacement so that costs vary by
displacement; fuel injection systems vary in cost according to how many fuel injectors are
required so that costs vary by number of cylinders. Therefore, we are able to determine a
variable cost equation as a function of engine displacement or as a function of the number of
A
We have estimated a "recovered" cost for all engine and equipment fixed costs to present a per unit analysis of the
cost of the proposal. In general, in environmental economics, it would be more conventional to simply count the total
cost of the program (i.e., opportunity costs) in the year they occur. However, this approach would not directly estimate a
per unit cost since fixed costs occur prior to implementation of the standards and, therefore, there are not yet any units
certified as complying with the new standards to which the fixed costs can be attributed. 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 rate of return to the manufacturer.
6-3
-------
Draft Regulatory Impact Analysis
cylinders. We then consider each unique engine's baseline technology package using a
database of all nonroad equipment sold in the United States (U.S.).6 That database lists
engine characteristics for every one of over 7,000 pieces of equipment sold in the US and
provides the sales of each piece of equipment. Using the current engine characteristics of
each engine, the projected technology package for that engine, and the variable cost equations
described in section 6.2, we calculate a variable cost for the engine in each of the over 7,000
pieces of equipment sold in the US. This variable cost per engine is then multiplied by that
engine's projected sales in each year for the years following implementation of the new
standards. 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 of this draft RIA.
• Note that the cost per ton calculation is never impacted by how many horsepower categories
we use in our cost analysis. We sometimes break up the fleet into more horsepower
categories than would seem reasonable given the structure of the proposed standards. We do
this for a couple of reasons: (1) phase-ins of standards and/or different levels of baseline
versus proposed standards sometimes force such breakouts; and, (2) greater stratification (i.e.,
breaking up the 75 to 175 horsepower range and the 175 to 750 horsepower range) provides a
better picture for use in our estimate of potential recovery of fixed costs. Importantly, the
number of horsepower categories used does not impact the total costs estimated as a result of
the proposed standards, and these total costs are the 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 would be reduced by several factors, as described below. All costs are
presented in 2001 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 of this Draft RIA represent those technologies we
believe will be used to comply with the proposed Tier 4 emission standards. These technologies
are also part of an ongoing research and development effort geared toward compliance with the
2007 heavy-duty diesel highway emission standards. Those engine manufacturers making R&D
expenditures toward compliance with highway emission standards will have to undergo 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 with minimum negative impacts on
6-4
-------
Estimated Engine and Equipment Costs
engine performance, durability, and fuel consumption. However, many nonroad engine
manufacturers are not part of the ongoing R&D effort toward compliance with highway
emissions 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 with highway manufacturers, emission
control device manufacturers, and the independent engine research laboratories conducting
relevant R&D. Despite these opportunities for learning, we would expect the R&D expenditures
for these nonroad-only manufacturers to be somewhat higher than for those manufacturers
already conducting R&D in response to the HD2007 rule.
We are projecting that several technologies will be used to comply with the proposed Tier 4
emission standards. We are projecting that NOx adsorbers and CDPFs would be the most likely
technologies applied by industry to meet our proposed emissions standards for >75 horsepower
engines and, for engines between 25 and 75 horsepower, that CDPFs would be used in 2013 to
meet the proposed PM standard. The fact that these technologies are being developed for
implementation in the highway market prior to the implementation dates in today's proposal, and
the fact that engine manufacturers would have several years before implementation of the
proposed Tier 4 standards, ensures that the technologies used to comply with the nonroad
standards would undergo significant development before reaching production. This ongoing
development could 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 given the current state of
development. Similarly, we anticipate that the continuing effort to improve the emission control
technologies will include innovations that allow lower-cost production. Finally, we believe that
manufacturers would 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, in order to meet the proposed standards, industry would introduce a
combination of primary technology upgrades. Achieving very low NOx emissions would require
basic research on NOx emission control technologies and improvements in engine management
to take advantage of the exhaust emission control system capabilities. The manufacturers are
expected to take a systems approach to the problem of optimizing the engine and exhaust
emission control system to realize the best overall performance. Since most research to date with
exhaust emission control technologies has focused on retrofit programs, there remains room for
significant improvements by taking such a systems approach. 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 requiring both a CDPF and a
NOx adsorber (i.e., >75 horsepower), we have attributed two-thirds of the R&D expenditures to
NOx control, and one-third to PM control.
In the 2007 highway rule, we estimated that each engine manufacturer would expend $35
million for R&D toward a successful implementation of catalyzed diesel paniculate filters
6-5
-------
Draft Regulatory Impact Analysis
(CDPF) and NOx adsorbers. For their nonroad R&D efforts on engines requiring CDPFs and
NOx adsorbers (i.e., >75 horsepower), engine manufacturers selling into the highway market
would 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, horsepower and torque
characteristics are often different so some effort will have to be expended to accommodate those
differences. Therefore, for these manufacturers, we have estimated that they would incur an
R&D expense 10 percent of that incurred for the highway rule, or $3.5 million. This $3.5 million
R&D expense would allow for the transfer of R&D knowledge from their highway experience to
their nonroad engine product line. For reasons noted above, two-thirds of this R&D is attributed
to NOx control and one-third to PM control.
For those manufacturers that sell engines only into the nonroad market, and where those
engines require a CDPF and a NOx adsorber, we believe that they will incur an R&D expense
nearing but not equaling that incurred by highway manufacturers for the highway rule. Nonroad
manufacturers would 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). This learning could be
done via seminars, conferences, and contact with highway manufacturers, emission control
device manufacturers, and the independent engine research laboratories conducting relevant
R&D. Therefore, we have estimated an expenditure of 70 percent of that spent by highway
manufacturers in their highway efforts. This lower number—$24.5 million versus $35 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.
Note that the $3.5 million and $24.5 million estimates represent our estimate of the average
R&D expected by manufacturers. These estimates would be different for each manufacturer -
some higher, some lower - depending on product mix and the ability to transfer knowledge from
one product to another.
For those engine manufacturers selling engines that would require CDPF-only R&D (i.e., 25
to 75 horsepower engines in 2013), we have estimated that the R&D they would incur would 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 would not be
appropriate. Using this estimate, the R&D incurred by manufacturers selling any engines into
both the highway and the nonroad markets would be $1.2 million, and the R&D for
manufacturers selling engines into only the nonroad market would be roughly $8 million. All of
this R&D is attributed to PM control.
For those engine manufacturers selling engines that would require DOC-only or some engine-
out modification R&D (i.e., to meet the PM standard on <75 horsepower engines in 2008), we
have estimated that the R&D they would incur would 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
6-6
-------
Estimated Engine and Equipment Costs
these devices have been around for years and because they require no special fueling strategies or
operating conditions to operate properly. Nonetheless, to be conservative we have estimated that
the R&D incurred by manufacturers selling any engines into both the highway and nonroad
markets would be roughly $600,000, and the R&D for manufacturers selling engines into only
the nonroad market would be roughly $4 million. Because these R&D expenditures are strictly
for meeting a PM standard, all of this R&D is attributed to PM control.
All of these R&D estimates are outlined in Table 6.2-1.
Table 6.2-1
Estimated R&D Expenditures by Type of Manufacturer
Totals per Manufacturer over Five Years
For proposed standards starting in
year
Horsepower Range
Manufacturer sells into both
highway and nonroad markets
Manufacturer sells into only 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
CDPF&NOx
Adsorber Engines
2011 & 2012
hp>75
$3,500,000
$24,500,000
33%
67%
R&D for CDPF-
only Engines
2013
25
-------
Draft Regulatory Impact Analysis
would 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 j oint 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 would conduct no R&D themselves and
would rely on their joint venture partner. This is true unless the parent company has no engine
sales in the horsepower 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 horsepower categories that Caterpillar does
not. As a result, we have attributed R&D costs to Caterpillar for conducting R&D that would
benefit Perkins engines. We have identified nine manufacturers to whom we have attributed no
R&D because of a joint partner agreement.8 Some of these (e.g., Perkins) we have attributed
R&D costs to their parent for the engines they will sell, and some are essentially the same
company as their parent (e.g., 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 would 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 their revenues.0 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. sales volumes, we believe it is unlikely that they would conduct the
necessary R&D themselves. Instead, they would probably license the technology from another
manufacturer which would 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. Because the
determination of which manufacturers would and would not invest in R&D is based on projected
sales data, we cannot share the manufacturer names. It is important to note that the total
projected sales for all three engine manufacturers was 77 engines in the 2002 model year.
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 a partner 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.
/-i
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 would 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; therefore, it would not result in increased costs associated with the proposed standards.
6-8
-------
Estimated Engine and Equipment Costs
Lastly, some certifying manufacturers do not appear to actually make engines. Instead, they
purchase engines from another engine manufacturer and then certify it as their own. We have
identified eight such certifying manufacturers and have attributed no R&D to these eight.0
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 R&D expenditures are shown
in Table 6.2-2.
Table 6.2-2
Estimated Industrywide R&D Expenditures for the Proposed Nonroad Tier 4 Standards"
Expenditures during Years:
Horsepower
Total Industry-wide R&D
Expenditures
R&D for PM
R&DforNOx
DOC/engine-out
R&Db
2003-2007
075hp
$118.0
$38.9
$79.0
CDPF-only R&Db
2008-2012
25
-------
Draft Regulatory Impact Analysis
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 US 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. Of the countries
expected to have nonroad emission standards of similar stringency to our proposed standards,
U.S. GDP constitutes 42 percent of the total.E Therefore, we have attributed 42 percent of the
R&D costs to U.S. sales.
We have weighted R&D recovery according to estimated revenues for engines sold in each
horsepower category. For example, CDPF&NOx Adsorber R&D benefits all engines above 75
horsepower. However, engines above 175 horsepower must introduce the new technologies in
2011, while engines from 75 to 175 horsepower would introduce it a year later. As a result,
R&D costs are assumed to be recovered on > 175 horsepower engines between 2011 and
2015/2018 and on 75 to 175 horsepower engines between 2012 and 2016/2018. 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
would 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 6.2-3.
Using this methodology, we have estimated the total R&D expenditures attributable to the
proposed standards at $7 to $33 million per year depending on the year, with an average of $18
million per year and a total of $199 million. Total R&D recovery on U.S. sales is estimated at
$279 million. All estimated R&D costs are shown in Table 6.2-4. Note that the engine sales
numbers shown in Table 6.2-4 are discussed in greater detail in Chapter 8 of this Draft RIA
where we present aggregate costs to society.
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 $7.8B; Australia's GDP was $0.4B; Canada's GDP was $0.7B; Japan's GDP was $4.7B; and the U.S.
GDP was $9.9B; for a total GDP of $23.5B (www.worldbank.org).
6-10
-------
Estimated Engine and Equipment Costs
Table 6.2-3
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
$2,800
$2,800
$5,000
$5,000
$10,000
$30,000
$125,000
$125,000
Revenue Weighted Recovery of R&D in the Indicated Years
PM
NOx
2008-2012
N/A
22%
46%
32%
100%
2011-2015
2011-2018
26%
44%
15%
15%
100%
2012-2016
2012-2018
11%
17%
19%
32%
10%
11%
100%
2013-2017
N/A
59%
41%
100%
6-11
-------
Table 6.2-4
Estimated R&D Costs Incurred (Non-Annualized) and Recovered (Annualized) - expressed in $2001
Thousands of dollars, except per engine values
2003 2004 2005 2006 2007 2008 2009 2010
in
V
•^
o
Q.
o
r3
R
Q.
8
8
jj-
0
IS
12
Q.
f
8
8
Q.
f
t
§
Q.
f
8
S
V
J-
o
o
g
R
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
131,507 135,623 139,739 143,855 147,971
$1,581 $1,581 $1,581 $1,581 $1,581
143,496 147,001 150,506 154,011 157,516
$3,294 $3,294 $3,294 $3,294 $3,294
100,051 102,097 104,142 106,188 108,234
$2,326 $2,326 $2,326 $2,326 $2,326
73,162 74,662 76,161 77,660 79,159
$825
$838
119,303 121,625 123,946 126,267 128,588
$1,350
$1,371
66,093 67,507 68,921 70,335 71,749
$1,625 $1,487
$1,650 $1,509
35,403 35,839 36,275 36,711 37,147
$2,687 $2,459
$2,728 $2,496
2,902 2,952 3,002 3,052 3,102
$904 $827
$917 $839
2,938 2,989 3,040 3,091 3,142
$915 $837
$928 $850
$7,201 $7,201 $7,201 $13,331 $14,986
$6,223 $7,903
$7,201 $7,201 $7,201 $19,555 $22,888
152,087
$2,218
$15
161,021
$5,304
$4,620
$29
110,279
$3,746
$3,262
$30
80,659
$825
$838
130,909
$1,350
$1,371
73,163
$1,487
$1,509
37,583
$2,459
$2,496
3,152
$827
$839
3,193
$837
$850
$16,835
$7,903
$24,737
$10,100
$10,100
156,203
$2,218
$14
164,526
$5,304
$4,620
$28
112,325
$3,746
$3,262
$29
82,158
$825
$1,676
133,230
$1,350
$2,741
74,577
$1,487
$3,019
38,019
$2,459
$4,992
3,202
$827
$1,679
3,244
$837
$1,699
$16,835
$15,805
$32,640
$10,100
$10,100
160,319
$2,218
$14
168,031
$5,304
$4,620
$27
114,371
$3,746
$3,262
$29
83,657
$825
$1,676
135,551
$1,350
$2,741
75,991
$1,487
$3,019
38,455
$2,459
$4,992
3,252
$827
$1,679
3,295
$837
$1,699
$16,835
$15,805
$32,640
$10,100
$10,100
164,435
$2,218
$13
171,536
$5,304
$4,620
$27
116,416
$3,746
$3,262
$28
85,157
$628
$1,016
137,872
$1 ,027
$1 ,662
77,405
$1 ,830
$2,279
$2,314
$59
38,891
$3,026
$3,769
$3,826
$195
3,302
$1,018
$1 ,268
$1 ,287
$774
3,346
$1 ,030
$1 ,283
$1 ,302
$773
$10,704
$9,582
$20,286
$18,698
$8,729
$27,427
168,551
$2,218
$13
175,041
$5,304
$4,620
$26
118,462
$3,746
$3,262
$28
86,656
$838
$1,158
$1,175
$27
140,193
$1,371
$1 ,894
$1 ,922
$27
78,819
$1 ,509
$2,085
$2,117
$53
39,327
$2,496
$3,449
$3,501
$177
3,352
$839
$1,160
$1,177
$697
3,397
$850
$1,174
$1,191
$696
$9,050
$7,903
$16,952
$21,018
$11,084
$32,102
172,667
178,546
$7,439
$42
120,507
$5,254
$44
88,155
$838
$1,158
$1,175
$26
142,514
$1,371
$1 ,894
$1 ,922
$27
80,233
$1 ,509
$2,085
$2,117
$52
39,763
$2,496
$3,449
$3,501
$175
3,402
$839
$1,160
$1,177
$687
3,448
$850
$1,174
$1,191
$686
$7,903
$7,903
$23,611
$11,084
$34,695
176,783
182,051
$7,439
$41
122,553
$5,254
$43
89,654
$1,158
$2,350
$39
144,836
$1 ,894
$3,845
$40
81 ,647
$2,085
$4,234
$77
40,199
$3,449
$7,002
$260
3,452
$1,160
$2,355
$1,018
3,499
$1,174
$2,383
$1,016
$23,611
$22,168
$45,779
180,899
185,556
$7,439
$40
124,599
$5,254
$42
91,154
$1,158
$2,350
$38
147,157
$1,894
$3,845
$39
83,061
$2,085
$4,234
$76
40,635
$3,449
$7,002
$257
3,502
$1,160
$2,355
$1,004
3,550
$1,174
$2,383
$1,002
$23,611
$22,168
$45,779
185,015
189,061
$7,439
$39
126,644
$5,254
$41
92,653
$880
$1,425
$25
149,478
$1,440
$2,331
$25
84,475
$2,567
$30
41 ,071
$4,245
$103
3,552
$1,428
$402
3,601
$1,445
$401
$15,013
$13,439
$28,452
189,131 193,247 193,247
192,566 196,071 196,071
$7,439
$39
128,690 130,736 130,736
$5,254
$41
94,152 95,652 95,652
$1,175 $1,175
$12 $12
151,799 154,120 154,120
$1,922 $1,922
$13 $12
85,889 87,303 87,303
$2,117 $2,117
$25 $24
41,507 41,943 41,943
$3,501 $3,501
$84 $83
3,602 3,652 3,652
$1,177 $1,177
$327 $322
3,652 3,703 3,703
$1,191 $1,191
$326 $322
$12,693
$11,084 $11,084
$23,777 $11,084
193,247
$7,905
$0
$11,088
$0
196,071
$42,988
$0
$60,294
$0
130,736
$30,359
$0
$42,581
$0
95,652
$3,929
$7,718
$5,510
$10,825
154,120
$6,428
$12,627
$9,015
$17,711
87,303
$7,572
$15,554
$10,620
$21,815
41,943
$12,522
$25,722
$17,563
$36,077
3,652
$4,211
$8,651
$5,907
$12,133
3,703
$4,262
$8,755
$5,978
$12,279
$120,177
$79,027
$199,204
$168,555
$110,839
$279,394
-------
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 2007 highway rule, we estimated approximately $1.6
million per engine line for tooling costs associated with CDPF/NOx adsorber systems. For the
proposed nonroad Tier 4 standards, we have estimated that nonroad-only manufacturers would
incur the same $1.6 million per engine line requiring a CDPF/NOx adsorber system and that
these costs would be split evenly between NOx control and PM control. We have estimated the
same tooling costs as estimated in the 2007 highway rule because we expect these Tier 4 engines
would use the same technologies as the 2007 highway rule (i.e., a CDPF and a NOx adsorber).
For those systems requiring only a CDPF, we have estimated one-half that amount, or $800,000
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 $400,000 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 $1.6 million baseline discussed above. For those engines requiring a CPDF/NOx
adsorber system (i.e., those >75 horsepower) we have adjusted that $1.6 million baseline by 50
percent. We believe this 50 percent adjustment is reasonable since many nonroad engines over
75 horsepower are produced on the same engine line with their highway counterparts. For such
lines, essentially no tooling costs would be incurred. For engine lines without a highway
counterpart, the $1.6 million tooling cost would be applicable. For highway manufacturers
selling into both the highway and the nonroad markets, we have assumed a 50/50 split of nonroad
engine product lines (i.e., 50 percent with highway counterparts and 50 percent without) and,
therefore, a 50 percent factor applied to the $1.6 million baseline. These tooling costs would be
split evenly between NOx control and PM control. For those engine lines requiring only a CDPF
(i.e., those between 25 and 75 horsepower), we have estimated the same tooling cost as used for
nonroad-only manufacturers, or $800,000. Similarly, the tooling costs for DOC and/or engine-
out engine lines has been estimated to be $400,000. We have used the same tooling costs as the
nonroad-only manufacturers for the <75 horsepower engines 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 have projected that engines in the 25 to 50 horsepower range would apply EGR systems
to meet the proposed NOx standards for 2013. For these engines, we have included an additional
tooling cost of $40,000 per engine line, consistent with the EGR-related tooling cost estimated
for 50-100 horsepower engines in our Tier 2/3 rulemaking where the same NOx standards was
required. This tooling cost is applied equally to all engine lines in that horsepower range
regardless of the markets into which the manufacturer sells. We have applied this tooling cost
equally because engines in this horsepower range do not tend to have highway counterparts.
Because EGR systems are expected to be added to engines between 25 and 50 horsepower to
meet the proposed NOx standard, tooling costs for EGR systems are attributed solely to NOx
control.
6-13
-------
Draft Regulatory Impact Analysis
Tooling costs per engine line and type of manufacturer are summarized in Table 6.2-5.
Table 6.2-5
Estimated Tooling Expenditures per Engine Line by Type of Manufacturer
Horsepower Range
For proposed standards starting in
Manufacturer sells into both highway
& nonroad markets
Manufacturer sells into only the
nonroad market
% Allocated to PM
% Allocated to NOx
DOC/engine-out
Engines
075
2011/2012
$800,000
$1,600,000
50%
50%
EGR Engines
25750 hp, half of the tooling costs are incurred one year ahead of
2011 and the other half are incurred one year ahead of 2014 due to the 50/50/50/100 percent
phase-in that begins in 2011. The costs are then recovered over an eight year period due to this
phase-in. A seven percent interest rate is used to account for the time value of money.
6-14
-------
Estimated Engine and Equipment Costs
Using this methodology, we estimate the total tooling expenditures attributable to the
proposed standards at $67 million. Total tooling recovery on U.S. sales is estimated at $81
million. All estimated tooling costs are shown in Table 6.2-6.
6-15
-------
Draft Regulatory Impact Analysis
Table 6.2-6
Estimated Tooling Costs Incurred (Non-Annualized) and Recovered (Annualized) — expressed in $2001
Thousands of dollars, except per engine values
8
&
S
Q.
II
if)
CM
if)
IL
II
O
if)
O
o
Q.
II
if)
if)
Q.
1
8
O
$
Q-
II
if)
O
8
Q.
II
O
8
O
if)
i.
ii
o
§
Q.
1
A
Q.
3
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
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 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
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
Estimated US Sales
PM Costs Incurred
NOx Costs Incurred
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 2010
147,971 152,087 156,203 160,319
$3,365
$821 $821 $821
$5 $5 $5
157,516 161,021 164,526 168,031
$3,756
$916 $916 $916
$6 $6 $5
108,234 110,279 112,325 114,371
$2,652
$647 $647 $647
$6 $6 $6
79,159 80,659 82,158 83,657
128,588 130,909 133,230 135,551
71,749 73,163 74,577 75,991
$10,665
$10,665
37,147 37,583 38,019 38,455
$5,879
$5,879
3,102 3,152 3,202 3,252
$475
$475
3,142 3,193 3,244 3,295
$253
$253
$9,773 $17,271
$17,271
$9,773 $34,543
$2,384 $2,384 $2,384
$2,384 $2,384 $2,384
2011
164,435
$821
$5
171,536
$916
$5
116,416
$647
$6
85,157
$2,685
$2,685
137,872
$4,392
$4,392
77,405
$2,601
$2,601
$67
38,891
$1,434
$1,434
$74
3,302
$116
$116
$70
3,346
$62
$62
$37
$7,077
$7,077
$14,154
$6,596
$4,212
$10,808
2012
168,551
$821
$5
175,041
$4,148
$506
$916
$5
118,462
$2,929
$647
$5
86,656
$655
$655
$15
140,193
$1,071
$1,071
$15
78,819
$2,601
$2,601
$66
39,327
$1,434
$1,434
$73
3,352
$116
$116
$69
3,397
$62
$62
$36
$7,077
$506
$7,583
$8,322
$5,938
$14,260
2013
172,667
178,546
$1,012
$123
$6
120,507
$714
$6
88,155
$655
$655
$15
142,514
$1,071
$1,071
$15
80,233
$2,601
$2,601
$65
39,763
$1,434
$1,434
$72
3,402
$116
$116
$68
3,448
$253
$253
$62
$62
$36
$253
$253
$506
$7,664
$6,062
$13,726
2014
176,783
182,051
$1,012
$123
$6
122,553
$714
$6
89,654
$655
$655
$15
144,836
$1,071
$1,071
$15
81,647
$2,601
$2,601
$64
40,199
$1,434
$1,434
$71
3,452
$116
$116
$67
3,499
$123
$123
$70
$7,726
$6,123
$13,849
2015
180,899
185,556
$1,012
$123
$6
124,599
$714
$6
91,154
$655
$655
$14
147,157
$1,071
$1,071
$15
83,061
$2,601
$2,601
$63
40,635
$1,434
$1,434
$71
3,502
$116
$116
$66
3,550
$123
$123
$69
$7,726
$6,123
$13,849
2016 2017 2018 2019
185,015 189,131 193,247 193,247
189,061 192,566 196,071 196,071
$1,012 $1,012
$123 $123
$6 $6
126,644 128,690 130,736 130,736
$714 $714
$6 $6
92,653 94,152 95,652 95,652
$655
$655
$14
149,478 151,799 154,120 154,120
$1,071
$1,071
$14
84,475 85,889 87,303 87,303
41,071 41,507 41,943 41,943
3,552 3,602 3,652 3,652
3,601 3,652 3,703 3,703
$62 $62 $62
$62 $62 $62
$34 $34 $33
$3,514 $1,788 $62
$1,911 $185 $62
$5,425 $1,973 $123
Total
$3,365
$0
$4,104
$0
$7,903
$506
$9,638
$616
$5,582
$0
$6,806
$0
$2,685
$2,685
$3,274
$3,274
$4,392
$4,392
$5,356
$5,356
$10,665
$10,665
$13,006
$13,006
$5,879
$5,879
$7,169
$7,169
$475
$475
$579
$579
$506
$506
$616
$616
$41,451
$25,107
$66,558
$50,548
$30,616
$81,164
6-16
-------
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.
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.8 To this
we have added the proposed certification fee of $2,156 per new engine family.9 This cost,
$62,156 per engine family was used for <75 horsepower engines certifying to the 2008 standards.
For 25 to 75 horsepower engines certifying to the 2013 standards, and for >75 horsepower
engines certifying to their proposed standards, we have added costs to cover the proposed test
procedures for nonroad diesel engines (i.e., the transient test and the NTE); these costs were
estimated at $10,500 per engine family. These certification costs—whether it be the $62,156 or
the $72,656 per engine family—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 horsepower rating of each. We grouped those horsepower ratings into the nine
horsepower ranges shown in Table 6.2-7. We have chosen these nine horsepower categories for
a couple of reasons: (1) phase-ins of standards and/or different levels of baseline versus proposed
standards force such breakouts; and, (2) greater stratification (i.e., breaking up the 75 to 175
horsepower range and the 175 to 750 horsepower 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 the proposed engine standards. Table 6.2-7 shows the
number of engine families in each horsepower range and the year for which they would have to
be certified to new standards, along with the total certification expenditures for those standards.
The cost expenditures shown in Table 6.2-7 would be incurred one year prior to the years
shown in the table. The years shown in the table coincide with the years for which the new
standards begin thereby forcing the certification of engines. Half of the 175 to 750 horsepower
engine families certified for 2011 must again be certified in 2014 when the NOx phase-in
becomes 100 percent. Half of the >750 horsepower engine families get certified in 2011 and the
remaining half get certified in 2014 due to the 50/50/50/100 percent PM & NOx phase-ins. For
the 25 to 50 horsepower engine families in 2013, half of the certification costs are attributed to
PM while half are attributed to NOx due to the proposal to add new PM and NOx standards for
those engines in that year; all of the certification costs for 50 to 75 horsepower engine families
are attributed to PM because only a new PM standard would be implemented in that year for
those engines.
Note that these certification costs should be considered conservative because they assume all
engines are certified because of the proposed standards. In reality, some engines would have
been certified due to factors independent of the proposed standards. Such engines would have
incurred certification costs regardless of any new standards.
6-17
-------
Draft Regulatory Impact Analysis
Table 6.2-7
Number of Engine Families, Estimated Certification Costs in $2001,
and Allocation of Certification Costs
Horsepower Range
0750a
Total families
Total Cert Costs
($MM)
% Allocated to PM
% Allocated to NOx
For Proposed Emissions Standards Starting in the Year
2008
102
132
88
322
$20.0
100%
0%
2011
102
64
9
20
195
$14.2
50%
50%
2012
55
73
128
$9.3
50%
50%
2013
132
132
$9.6
50%
50%
88
88
$6.4
100%
0%
2014
51
32
5
88
$6.4
0%
100%
20
20
$1.5
50%
50%
28
37
64
$4.7
0%
100%
a Forty engine families were certified in the >750 hp range, but only half would be certified in the indicated years due to
the proposed phase-in schedule.
To estimate recovery of certification expenditures, we have attributed the expenditures to
engines sold in the specific horsepower range and spread the recovery of costs over U.S. sales
within that category. Expenditures are incurred one year prior to the emission standard for which
the certification is conducted, and are then recovered over a five year period following the
certification. A seven percent interest rate is used to account for the time value of money. We
have spread these certification costs over only the U.S.-sold engines because the certification
conducted for the U.S. is not presumed to fulfill the certification requirements of other countries.
Total certification expenditures were estimated at $72 million. Recovery of certification costs
was estimated at $88 million. All estimated certification expenditures and the recovery of those
expenditures are shown in Table 6.2-8.
6-18
-------
Estimated Engine and Equipment Costs
Table 6.2-8
Estimated Certification Costs Incurred (Non-Annualized) and Recovered (Annualized) - expressed in $2001
Thousands of dollars, except per engine values
in
Q_
.c
o
o
Id
Q.
j;
8
in
Q_
1}
s
o
o
Q.
II
LO
t
Q.
II
S
s
CO
Q.
II
t=
S
CD
Q.
j;
§
§
Q.
"iT
O
S
f
IS
Q.
5
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
2007 2008 2009 2010
147,971 152,087 156,203 160,319
$6,340
$1,546 $1,546 $1,546
$10 $10 $10
157,516 161,021 164,526 168,031
$8,205
$2,001 $2,001 $2,001
$12 $12 $12
108,234 110,279 112,325 114,371
$5,470
$1,334 $1,334 $1,334
$12 $12 $12
79,159 80,659 82,158 83,657
128,588 130,909 133,230 135,551
71,749 73,163 74,577 75,991
$3,705
$3,705
37,147 37,583 38,019 38,455
$2,325
$2,325
3,102 3,152 3,202 3,252
$327
$327
3,142 3,193 3,244 3,295
$727
$727
$20,014 $7,084
$7,084
$20,014 $14,168
$4,881 $4,881 $4,881
$4,881 $4,881 $4,881
2011
164,435
$1,546
$9
171,536
$2,001
$12
116,416
$1,334
$11
85,157
$1,998
$1,998
137,872
$2,652
$2,652
77,405
$904
$904
$23
38,891
$567
$567
$29
3,302
$80
$80
$48
3,346
$177
$177
$106
$4,650
$4,650
$9,300
$6,609
$1,728
$8,337
2012
168,551
$1,546
$9
175,041
$4,795
$4,795
$2,001
$11
118,462
$6,394
$1,334
$11
86,656
$487
$487
$11
140,193
$647
$647
$9
78,819
$904
$904
$23
39,327
$567
$567
$29
3,352
$80
$80
$48
3,397
$177
$177
$104
$11,189
$4,795
$15,984
$7,743
$2,862
$10,605
2013
172,667
178,546
$1,170
$1,170
$13
120,507
$1,559
$13
88,155
$1,998
$487
$487
$11
142,514
$2,652
$647
$647
$9
80,233
$3,705
$904
$904
$23
39,763
$2,325
$567
$567
$29
3,402
$327
$80
$80
$47
3,448
$727
$727
$177
$177
$103
$727
$11,734
$12,461
$5,591
$4,031
$9,622
2014
176,783
182,051
$1,170
$1,170
$13
122,553
$1,559
$13
89,654
$487
$975
$16
144,836
$647
$1,294
$13
81,647
$904
$1,807
$33
40,199
$567
$1,134
$42
3,452
$80
$159
$69
3,499
$354
$354
$203
$5,768
$6,893
$12,661
2015
180,899
185,556
$1,170
$1,170
$13
124,599
$1,559
$13
91,154
$487
$975
$16
147,157
$647
$1,294
$13
83,061
$904
$1,807
$33
40,635
$567
$1,134
$42
3,502
$80
$159
$68
3,550
$354
$354
$200
$5,768
$6,893
$12,661
2016
185,015
189,061
$1,170
$1,170
$12
126,644
$1,559
$12
92,653
$487
$975
$16
149,478
$647
$1,294
$13
84,475
$904
$11
41,071
$567
$14
3,552
$80
$22
3,601
$177
$177
$98
$4,040
$5,165
$9,206
2017
189,131
192,566
$1,170
$1,170
$12
128,690
$1,559
$12
94,152
$487
$5
151,799
$647
$4
85,889
$904
$11
41,507
$567
$14
3,602
$80
$22
3,652
$177
$177
$97
$2,906
$4,031
$6,937
2018 2019
193,247 193,247
196,071 196,071
130,736 130,736
95,652 95,652
$487
$5
154,120 154,120
$647
$4
87,303 87,303
$904
$10
41,943 41,943
$567
$14
3,652 3,652
$80
$22
3,703 3,703
$177
$177
$96
$177
$2,862
$3,039
Total
$6,340
$0
$7,731
$0
$13,000
$4,795
$15,853
$5,848
$11,863
$0
$14,467
$0
$1,998
$3,996
$2,437
$4,873
$2,652
$5,304
$3,234
$6,468
$3,705
$7,411
$4,519
$9,037
$2,325
$4,650
$2,835
$5,670
$327
$654
$399
$797
$1,453
$1,453
$1,772
$1,772
$43,664
$28,263
$71,927
$53,246
$34,466
$87,712
6-19
-------
Draft Regulatory Impact Analysis
6.2.2 Engine Variable Costs
Engine variable costs are those costs for new hardware required to meet the proposed
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 horsepower 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 horsepower 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 horsepower range unless the range is kept very
small. To state an average variable cost for a range such as 175 to 300 horsepower is far less
precise than what we present here. Using the equations presented here, we have estimated the
engine variable costs for some specific example pieces of equipment; these estimates can be
found in Section 6.5 of this Draft RIA.
The discussion here contains both near term and long term cost estimates. We believe there
are factors that would 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.10
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"). The
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.11 The distribution of these progress ratios is
shown in Figure 6.2-1. Except for one company that saw increasing costs as production
6-20
-------
Estimated Engine and Equipment Costs
continued, every study showed cost savings of at least five percent for every doubling of
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.
6-21
-------
Figure 6.2-1
Distribution of Progress Ratios
15
10
o>
ET
CD
Distribution of Progress Ratios
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).
Source: Button and Thomas, 1984.
-------
Estimated Engine and Equipment Costs
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% decrease in cost (Lieberman 1984, Zimmerman
1982). The effect of learning is more difficult to decipher in the computer chip industry (Gruber
1992).
EPA believes the use of 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 in
only limited application. Therefore, these are 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, EPA
believes 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.
EPA believes a similar opportunity exists for the new control systems which will integrate the
function of the engine and the 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 fuel systems including common rail
systems - they will be applied in some new ways in response to the proposed 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,
over time, 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
horsepower range of engines. This way, learning is applied at the start of 2013 for >175
horsepower engines and in 2014 for 75 to 175 horsepower engines 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 proposed standards follows implementation 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 in highway serves as the
baseline level of learning for nonroad. 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 would occur 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-
6-23
-------
Draft Regulatory Impact Analysis
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 would be less
opportunity for lowering production costs. For that reason, we applied only one learning curve
effect. The situation is similar for nonroad engines. Because the technologies will be, by the
time they are introduced into the market, existing technologies, 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.12 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.
6.2.2.1 NOx Adsorber System Costs
The NOx adsorber system that we are anticipating would be applied for Tier 4 would be the
same as that used for highway applications. In order for the NOx adsorber to function properly, a
systems approach that includes a reductant metering system and control of engine A/F ratio is
also necessary. Many of the new air handling and electronic system technologies developed in
order to meet the Tier 2/3 nonroad 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 will likely 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 2007 HD Highway 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 6.2-9 and represent costs to the
engine manufacturers inclusive of supplier markups. The manufacturer costs shown in Table
6.2-9 (as well as Tables 6.2-11 and 6.2-16 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
6-24
-------
Estimated Engine and Equipment Costs
emission control technologies based on input from catalyst manufacturers. Specifically, we were
told that device manufacturers could not markup the cost of the individual components within
their products because those components consist of basic commodities (e.g., precious metals
used in the catalyst could not be arbitrarily marked up because of their commodity status).
Instead, manufacturing entities could only markup costs 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, of what is essentially 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
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.13 Also included is
our estimate of warranty costs for the NOx adsorber system.
6-25
-------
Draft Regulatory Impact Analysis
Table 6.2-9. 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 ($2001)
9hp
0.39
0.59
$3
$14
$16
$3
$1
$9
2
$28
$42
$17
$105
$9
$4
$119
$4
$106
$228
$204
$3
$4
$113
$3
$100
$216
$193
$174
33 hp
1.50
2.25
$13
$53
$62
$11
$1
$9
2
$28
$42
$17
$208
$17
$8
$233
$7
$134
$374
$326
$6
$8
$222
$7
$127
$355
$310
$273
76 hp
3.92
5.88
$33
$139
$163
$28
$1
$9
2
$28
$42
$17
$432
$34
$17
$483
$14
$195
$692
$593
$11
$17
$460
$14
$185
$659
$564
$489
150hp
4.70
7.05
$39
$167
$195
$34
$1
$9
2
$28
$42
$17
$504
$39
$20
$564
$17
$214
$795
$679
$13
$20
$537
$16
$204
$757
$647
$558
250 hp
7.64
11.46
$64
$271
$318
$55
$1
$13
2
$28
$42
$17
$780
$60
$31
$872
$26
$291
$1,189
$1,009
$20
$31
$832
$25
$277
$1,134
$962
$825
503 hp
18.00
27.00
$151
$638
$748
$129
$1
$19
2
$28
$56
$22
$1,764
$132
$71
$1 ,967
$59
$468
$2,494
$2,089
$44
$71
$1 ,879
$56
$446
$2,381
$1,994
$1,684
660 hp
20.30
30.45
$170
$720
$844
$145
$1
$19
2
$28
$56
$22
$1,977
$148
$79
$2,204
$66
$507
$2,778
$2,323
$49
$79
$2,105
$63
$483
$2,652
$2,218
$1 ,871
1000hp
34.50
51.75
$290
$1,223
$1,434
$246
$1
$19
2
$28
$56
$22
$3,291
$247
$132
$3,670
$110
$749
$4,529
$3,773
$82
$132
$3,505
$105
$715
$4,325
$3,603
$3,026
6-26
-------
Estimated Engine and Equipment Costs
We have estimated the cost of this system based on information from several reports.14'15'16
The individual estimates and assumptions used to estimate the cost for the system are
documented in the following subsections.
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.17 The respondents indicated that the catalyst volume for a NOx adsorber catalyst could
range from 1.5 times the engine displacement to as much as 2.5 times the engine displacement
based on today's washcoating technology. Based on current lean burn gasoline catalyst designs
and engineering judgement, 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.18
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 2007 Highway rule. This cost estimate was based upon a
relationship developed for current heavy-duty gasoline catalyst substrates.19 We have converted
that value to $5.60 ($2001) using the PPI for Motor Vehicle Parts and Accessories, Catalytic
Convenor s.20
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).21 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 the other components which
are typically 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 6.2-9 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% of that total and rhodium representing 10%. 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.22
6-27
-------
Draft Regulatory Impact Analysis
NOx Adsorber Alkaline Earth Metal - Barium
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% for
transition cones, plus 20% for scrappage (material purchased but unused in the final product) and
a price of $ 1.04 per pound for 18 gauge stainless steel as estimated in a contractor report to EPA
and converted into $2001.23
NOx Adsorber Direct Labor
The direct labor costs for the catalyst are estimated based upon 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.24
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 horsepower engine shown in
Table 6.2-9 would be:
[($3 + $14 + $16 + $3 + $1 + $9)(2.5) + ($50)(4hours)](3%) = $9.45
NOx Adsorber Manufacturer and Dealer Carrying Costs
The manufacturer's carrying cost was estimated at 4% 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% of the
incremental cost, again reflecting primarily the cost of capital tied up in extra inventory.25
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 in Section 6.2.2.3, below. Important to note here is that the DOC costs
6-28
-------
Estimated Engine and Equipment Costs
shown in Table 6.2-9 are lower in the long term because of the lower warranty claim rate - 3
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 6.2-9, we calculated a linear
regression to determine the NOx adsorber system cost as a function of engine displacement. This
way, the function could 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 6.2-10. Note that Table
6.2-9 shows NOx adsorber system costs for engines below 75 horsepower. We do not anticipate
any engines below 75 horsepower will apply NOx adsorber systems to comply with the proposed
standards. Nonetheless, the costs shown were used to generate the equations shown in Table 6.2-
10. 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 6.2-10
NOx Adsorber System Costs as a Function of
Engine Displacement (x represents engine displacement in liters)
Near Term Cost Function
Long Term Cost Function
$105(x) + $181
$84(x) + $159
R2=0.9998
R2=0.9997
Table 6.2-10 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
6.2-10 include costs for a clean-up DOC; results generated using the DOC cost estimation
equations presented in Table 6.2-14 should not be added to results generated using the equations
in Table 6.2-10 to determine NOx adsorber system costs.
6.2.2.2 Catalyzed Diesel Particulate Filter Costs
As with the NOx adsorber system, the CDPF system that we are anticipating would be
applied for Tier 4 would be the same as that used for highway applications, except that we are
projecting that some form of active regeneration system would be employed as a backup to the
6-29
-------
Draft Regulatory Impact Analysis
passive regeneration capability of the CDPF. In order for the CDPF to function properly, a
systems approach that includes a reductant metering system and control of engine A/F ratio is
also necessary. Many of the new air handling and electronic system technologies developed in
order to meet the Tier 2/3 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 as was
used in our 2007 FID Highway rulemaking (although here, for nonroad engines, we have
included costs for a regeneration system that was not part of the cost estimate in the 2007 HD
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 6.2-11 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.26
6-30
-------
Estimated Engine and Equipment Costs
Table 6.2-11. Catalyzed Diesel Paniculate 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
$37
$14
$11
$7
$48
$0
2
$28
$56
$22
$195
$12
$8
$215
$6
-$48
$174
$139
$4
$8
$207
$6
-$48
$166
$132
$106
Catalyzed
33 hp
1.50
2.25
$143
$53
$42
$7
$48
$0
2
$28
$56
$22
$372
$25
$15
$411
$12
-$48
$376
$301
$8
$15
$395
$12
-$48
$359
$287
$230
Diesel
76 hp
3.92
5.88
$375
$139
$109
$7
$48
$0
2
$28
$56
$22
$756
$54
$30
$840
$25
-$48
$817
$654
$18
$30
$804
$24
-$48
$780
$624
$499
Parti culate
150hp
4.70
7.05
$449
$167
$130
$7
$48
$0
2
$28
$56
$22
$880
$63
$35
$978
$29
-$48
$959
$768
$21
$35
$936
$28
-$48
$916
$733
$586
Filter (CDPF) Costs
250 hp
7.64
11.46
$730
$271
$212
$11
$48
$0
2
$28
$56
$22
$1,350
$98
$54
$1,502
$45
-$48
$1,499
$1,199
$33
$54
$1 ,436
$43
-$48
$1 ,432
$1,145
$916
503 hp
18.00
27.00
$1,721
$638
$499
$15
$48
$0
2
$28
$56
$22
$2,998
$222
$120
$3,340
$100
-$48
$3,393
$2,714
$74
$120
$3,192
$96
-$48
$3,240
$2,592
$2,074
($2001)
660 hp
20.30
30.45
$1,940
$720
$563
$15
$96
$0
4
$28
$112
$45
$3,490
$253
$140
$3,882
$116
-$48
$3,951
$3,161
$84
$140
$3,714
$111
-$48
$3,777
$3,022
$2,417
1000hp
34.50
51.75
$3,298
$1,223
$956
$15
$96
$0
4
$28
$112
$45
$5,744
$422
$230
$6,396
$192
-$48
$6,540
$5,232
$141
$230
$6,114
$183
-$48
$6,250
$5,000
$4,000
6-31
-------
Draft Regulatory Impact Analysis
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.27 The respondents indicated that the
particulate filter volume could 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 upon the maximum
allowable flow restriction for the engine. Genetically, the filter size is inversely proportional to
its resistance to flow (a larger filter is less restrictive than a similar smaller filter). In the
HD2007 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
aimed at improving filter porosity control to give a better trade-off between flow restrictions and
filtering efficiency.
CDPF Substrate
In the HD2007 rule, we estimated that CDPFs would consist of a cordierite filter costing $30
per liter. For nonroad applications, we have assumed the use of silicon carbide filters costing
double that amount, or $60 per liter. This cost is directly proportional to filter volume, which is
proportional to engine displacement. This $60 value is then converted to $2001 using the PPI for
Motor Vehicle Parts and Accessories, Catalytic Converters.28 The end result being a cost of $64
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 Housing
The material cost for the can housing is estimated based on the CDPF volume plus 20% for
transition cones, plus 20% for scrappage (material purchased but unused in the final product) and
a price of $ 1.04 per pound for 18 gauge stainless steel as estimated in a contractor report to EPA
and converted into $2001.29
6-32
-------
Estimated Engine and Equipment Costs
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 $45.30 A PPI adjusted cost of $48 per sensor
has been used in this analysis.
CDPF Direct Labor
Consistent with the approach for NOx adsorber systems, the direct labor costs for the CDPF
are estimated based upon 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.31
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% 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% of the incremental cost, again reflecting primarily the cost of
capital tied up in extra inventory.32
Savings Associated with Muffler Removal
CDPF retrofits today are 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."33 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 $48. The $48 estimate
is an average for all engines, the actual savings would be higher for some and lower for others.
CDPF System Cost Estimation Function
Using the example CDPF costs shown in Table 6.2-11, we calculated a linear regression to
determine the CDPF system cost as a function of engine displacement. This way, the function
6-33
-------
Draft Regulatory Impact Analysis
could 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 6.2-12.
Table 6.2-12
CDPF System Costs as a Function of
Engine Displacement (x represents engine displacement in liters)
Near term Cost Function
Long term Cost Function
$150(x) + $71
$114(x)+$54
R2=0.9998
R2=0.9998
The near term and long term costs shown in Table 6.2-12 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 (e.g., 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. The results of
that cost analysis are detailed in the report entitled, "Electronic Systems and EGR Costs for
Nonroad Engines," which is contained in the docket for this rule.34 The cost estimates developed
by ICF for a CDPF regeneration system are summarized in Table 6.2-13.
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 6.2-14.
Table 6.2-13.
CDPF Regeneration System - Costs to the Manufacturer
ICF Estimated Regeneration System Costs to Manufacturers ($2001)
Horsepower
Displacement (L)
CDPF Regeneration System Costs
20
1
$260
35
2
$274
80
3
$287
150
6
$376
250
8
$398
400
10
$420
650
16
$514
1000
24
$654
6-34
-------
Estimated Engine and Equipment Costs
Table 6.2-14.
CDPF Regeneration System - Costs to the User
EPA Estimate of CDPF Regeneration System
Horsepower
Displacement (L)
CDPF Regeneration System Costs
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
1
$260
$23
$10
$293
$9
$302
$8
$10
$278
$8
$286
$229
35
2
$274
$24
$11
$308
$9
$317
$8
$11
$292
$9
$301
$241
80
3
$287
$25
$11
$323
$10
$333
$8
$11
$307
$9
$316
$253
Costs ($2001)
150
6
$376
$31
$15
$422
$13
$435
$10
$15
$401
$12
$413
$331
250
8
$398
$33
$16
$447
$13
$460
$11
$16
$425
$13
$437
$350
400
10
$420
$34
$17
$471
$14
$485
$11
$17
$448
$13
$462
$369
650
16
$514
$42
$21
$576
$17
$593
$14
$21
$548
$16
$565
$452
1000
24
$654
$52
$26
$733
$22
$755
$17
$26
$698
$21
$719
$575
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 6.2-14 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 6.2-14 to be representative of
the costs that would be incurred on an engine with an indirect injection (TDI) fuel system. For a
direct injection (DI) fuel system, we expect that many of the functional capabilities for which
costs were generated would 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 would be lower than those for an IDI engine, we have estimated that the regeneration
system costs for a DI engine would be one-half those presented in Table 6.2-14.
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.
CDPF Regeneration System Cost Estimation Function
Using the example regeneration system costs shown in Table 6.2-14, we calculated a linear
regression to determine the CDPF regeneration system cost as a function of engine displacement.
This way, the function could be applied to the wide array of engines in the nonroad fleet to
6-35
-------
Draft Regulatory Impact Analysis
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 6.2-15.
Table 6.2-15
CDPF Regeneration System Costs as a Function of
Engine Displacement (x represents engine displacement in liters)
IDI Engine
DI Engine
Near term Cost Function
Long term Cost Function
Near term Cost Function
Long term Cost Function
$20(x) + $289
$15(x) + $219
$10(x) + $144
$7(x) + $110
R2=0.9912
R2=0.9912
R2=0.9912
R2=0.9912
Note that these costs - either the IDI or the DI costs, depending on the type of engine - would
be incurred for any engine adding a CDPF. The near term and long term costs shown in Table
6.2-15 change due to the different warranty claim rates and the application of a 20 percent
learning curve effect.
6.2.2.4 Diesel Oxidation Catalyst (DOC) Costs
The NOx adsorber regeneration and desulfation functions may produce undesirable by-
products in the form of momentary increases in HC emissions or in odorous hydrogen sulfide
(H2S) emissions. In order to control these potential products, we have assumed that
manufacturers may choose to apply a diesel oxidation catalyst (DOC) downstream of the NOx
adsorber technology. The DOC would serve a "clean-up" function to oxidize any HC and H2S
emissions to more desirable products. As discussed below, for our cost analysis we have also
estimated that engines <75 horsepower would add a DOC to comply with the 2008 PM
standards, not to serve a "clean-up" function but rather to serve as the primary means of emission
control.
Our estimates of DOC costs are shown in Table 6.2-16. The individual component costs for
the DOC were estimated in the same manner as for the NOx adsorber systems and CDPF
systems, as discussed above. However, no learning effects were applied to DOCs because we
believe that DOCs have been manufactured for a long enough time period such that learning has
already taken place.
6-36
-------
Estimated Engine and Equipment Costs
Table 6.2-16.
Diesel Oxidation Catalyst (DOC) Costs
Horsepower
Average Engine Displacement (Liter)
Material and Component Costs
Catalyst Volume (liter)
Substrate
Washcoating and Canning
Platinum (5 g/ft3)
Catalyst Can Housing
Direct Labor Costs
Estimated Labor hours
Labor Rate ($/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 ($2001)
9hp
0.39
0.39
$2
$63
$1
$5
0.5
$28
$14
$6
$91
$8
$4
$103
$3
$106
$3
$4
$97
$3
$100
33 hp
1.50
1.50
$8
$78
$5
$5
0.5
$28
$14
$6
$115
$10
$5
$130
$4
$134
$3
$5
$123
$4
$127
76 hp
3.92
3.92
$22
$110
$12
$5
0.5
$28
$14
$6
$168
$14
$7
$189
$6
$195
$5
$7
$180
$5
$185
150hp
4.70
4.70
$26
$120
$14
$5
0.5
$28
$14
$6
$185
$15
$7
$208
$6
$214
$5
$7
$198
$6
$204
250 hp
7.64
7.64
$43
$159
$24
$7
0.5
$28
$14
$6
$252
$20
$10
$282
$8
$291
$7
$10
$269
$8
$277
503 hp
18.00
18.00
$101
$214
$55
$16
0.5
$28
$14
$6
$406
$32
$16
$454
$14
$468
$11
$16
$433
$13
$446
660 hp
20.30
20.30
$114
$227
$63
$18
0.5
$28
$14
$6
$440
$35
$18
$492
$15
$507
$12
$18
$469
$14
$483
1000hp
34.50
34.50
$193
$302
$106
$30
0.5
$28
$14
$6
$651
$50
$26
$728
$22
$749
$17
$26
$694
$21
$715
6-37
-------
Draft Regulatory Impact Analysis
DOC Cost Estimation Function
Similar to what was done for NOx adsorber systems and CDPFs, we used the example costs
shown in Table 6.2-16 to determine a cost function with engine displacement as the dependent
variable. This way, the function could 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 6.2-17. Note that the NOx adsorber cost
estimation equations shown in Table 6.2-10 include costs for a clean-up DOC; results generated
using the DOC cost estimation equations presented in Table 6.2-17 should not be added to results
generated using the equations in Table 6.2-10 to determine NOx adsorber system costs.
Table 6.2-17
DOC Costs as a Function of
Engine Displacement (x represents engine displacement in liters)
Near term Cost Function
Long term Cost Function
$19(x) + $117
$18(x) + $110
R2=0.9943
R2=0.9943
6.2.2.5 Closed-Crankcase Ventilation (CCV) System Costs
Consistent with our HD2007 rule, we are proposing to eliminate the exemption that allows
turbo-charged 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 turbo charged engines would force manufacturers to rely on engineered closed
crankcase ventilation systems that filter oil from the blow-by gases prior to 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 6.2-18. These costs are incurred only by turbo-charged engines.
6-38
-------
Estimated Engine and Equipment Costs
Table 6.2-18.
Closed Crankcase Ventilation (CCV) System Costs
Horse powsr
Average Engine Displacement (Liter)
Cost to Manufacturer
V\ferranty 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
V\ferranty 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
Closed Crankcase Ventilation (CCV) System Costs ($2001)
9hp
0.39
$29
$5
$1
$35
$1
$36
$2
$1
$32
$1
$26
33 hp
0.93
$30
$5
$1
$36
$1
$37
$2
$1
$33
$1
$27
76 hp
3.92
$36
$6
$1
$43
$1
$44
$2
$1
$39
$1
$32
150 hp
4.7
$37
$6
$1
$44
$1
$46
$2
$1
$40
$1
$33
250 hp
7.64
$43
$6
$2
$50
$2
$52
$2
$2
$46
$1
$38
503 hp
18
$62
$8
$2
$72
$2
$75
$3
$2
$67
$2
$55
660 hp
20.3
$67
$8
$3
$77
$2
$80
$3
$3
$72
$2
$59
1000hp
34.5
$94
$10
$4
$107
$3
$111
$3
$4
$101
$3
$83
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 6.2-19. Note that
these costs would be incurred only by turbo-charged engines.
Table 6.2-19
CCV Costs as a Function of
Engine Displacement (x represents engine displacement in liters)
Near term Cost Function
Long term Cost Function
$2(x) + $35
$2(x) + $25
R2=l
R2=l
6.2.2.6 Variable Costs of Conventional Technologies for Engines Below 75 Horsepower
and over 750 Horsepower
For the smaller horsepower categories, we have projected a different technology mix to
enable compliance due to the different proposed standards. From a cost perspective, we have
projected that engines would comply by either adding a DOC or by making some engine
modifications resulting in engine-out emission reductions. Presumably, the manufacturer would
choose the least costly approach that provided the necessary emission reduction. If engine-out
modifications are less costly than a DOC, our estimate here is conservative. If the DOC proves
to be less costly, then our estimate is representative of what most manufacturers would do.
Therefore, we have assumed that, beginning in 2008, all engines below 75 horsepower add a
DOC. Note that this is a conservative estimate in that we have assume this cost for all engines
6-39
-------
Draft Regulatory Impact Analysis
when, as discussed in Chapter 4, some engines <75 horsepower already meet the proposed PM
standards. Our cost estimates for DOCs are presented above in Section 6.2.2.4.
As discussed in Chapter 4 of this Draft RIA, we have also projected that some engines in the
25 to 75 horsepower range would 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 diesel injection (IDI) engines in this horsepower range, we believe that
manufacturers would 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 horsepower range, we believe that all engines would add cooled EGR to meet
the NOx standards proposed for that horsepower category. This is also true for engines >750
horsepower (note that engines >750 horsepower are also assumed to add the previously discussed
exhaust emission control technologies - i.e., a NOx adsorber system, a CDPF system, and some
sort of CDPF regeneration system).
All of these engines - those <75 horsepower and those >750 horsepower - are assumed to
add CCV systems where those engines are turbocharged. The costs for CCV systems were
presented in Section 6.2.2.5 above.
6.2.2.6.1 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. The results of cost analysis are detailed in the report entitled, "Electronic Systems and
EGR Costs for Nonroad Engines," which is contained in the docket for this rule.35 Table 6.2-20
presents the costs to manufacturers as estimated by ICF for fuel injection systems.
6-40
-------
Estimated Engine and Equipment Costs
Table 6.2-20
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
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
Costs ($2001)
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 horsepower range (represented in Table 6.2-20 by the 80
horsepower engine) are assumed to have electronic rotary fuel injection systems as a baseline
configuration while smaller engines are assumed to have mechanical fuel injection. 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 is the reason for the large difference in fuel system costs for the 80
horsepower engine relative to the 20 and 35 horsepower engines.
The costs shown in Table 6.2-20 show consistency for all elements across the horsepower
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 6.2-21, which also builds on the
manufacturer costs shown in Table 6.2-21 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 in this manner so that a cost equation could be generated that would apply to all engines.
Unlike the other cost equations we have generated, the cost equation for fuel systems uses the
number of injectors (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.
6-41
-------
Draft Regulatory Impact Analysis
Table 6.2-21
Incremental Fuel System Costs - Costs to the User
EPA Estimated Incremental Fuel System Costs for Dl Enqlnes ($2001)
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
perlniector Remaining System
$64 $543
$8 $44
$3 $22
$74 $608
$2 $18
$77 $627
$3 $15
$3 $22
$69 $579
$2 $17
$71 $597
$57 $477
35
3
per Iniector Remaining System
$64 $543
$8 $44
$3 $22
$74 $608
$2 $18
$77 $627
$3 $15
$3 $22
$69 $579
$2 $17
$71 $597
$57 $477
80
4
per Injector Remaining System
$55 $150
$7 $14
$2 $6
$64 $170
$2 $5
$66 $175
$2 $5
$2 $6
$60 $161
$2 $5
$61 $166
$49 $132
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 horsepower DI engines.
Note also that, in determining aggregate variable costs for fuel injection systems, we have
attributed half of the costs to the proposed 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 25-75 hp range will
already be equipped with electronic fuel systems independent of the standards contained in this
Tier 4 proposal. This is due to the natural progression of electronic fuel systems currently
available in larger power engines into some of the smaller power engines. During our
discussions with some engine companies, they have indicated that the electronic fuel system
technologies they intend to use to comply with the existing Tier 3 standards in the 50-100 hp
range. These manufacturers have informed us that these electronic fuel systems will also be sold
on engines in the 25-50 hp range for those engine product lines which are built on a common
platform as engines above 50 hp. In addition, there are a number of end-user benefits associated
with electronic fuel systems. These include better torque response, lower noise, easier servicing
via on-board diagnostics, and better engine starting ability. 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. The results of cost analysis are detailed in the report entitled, "Electronic Systems and
EGR Costs for Nonroad Engines," which is contained in the docket for this rule.36 The
incremental manufacturer costs for cooled EGR systems are shown in Table 6.2-22.
6-42
-------
Estimated Engine and Equipment Costs
Table 6.2-22
Cooled EGR System - Costs to Manufacturers
ICF Estimated Cooled EGR System Costs to Manufacturers ($2001)
Horsepower
Displacement (L)
EGR Cooler
EGR Bypass
Electronic EGR Valve
EGR Total Cost to Manufacturer
20
1
$36
$15
$14
$65
35
2
$63
$16
$15
$94
1000
24
$289
$30
$88
$407
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
6.2-23.
Table 6.2-23
Cooled EGR System - Costs to the User
EPA Estimated Cooled EGR Costs ($2001)
Horsepower
Displacement (L)
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
1
$65
$8
$3
$75
$2
$78
$3
$3
$70
$2
$72
$58
35
2
$94
$10
$4
$108
$3
$111
$3
$4
$101
$3
$104
$83
1000
24
$407
$34
$16
$457
$14
$471
$11
$16
$434
$13
$447
$358
Note that we are projecting that only engines in the 25 to 50 horsepower range (in 2013) and
engines >750 horsepower will need to add cooled EGR (consistent with the NOx phase-in from
2011 to 2014) to comply with the proposed standards. All of the costs associated with these
systems have been attributed to compliance with the proposed standards (i.e., we have not
attributed any costs to user benefits).
6-43
-------
Draft Regulatory Impact Analysis
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. 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 6.2-24.
Table 6.2-24
Costs for Conventional Technologies as a
Function of the Indicated Parameter (x represents the dependent variable)
Technology
Fuel System Costs - DI Only
Near Term
Long Term
Near Term
Long Term
Cooled EGR System
Near Term
Long Term
Applicable Hp Range
25<=hp<50
25<=hp<50
50<=hp<75
50<=hp<75
25<=hp<50; >750hp
25<=hp<50; >750hp
Dependent
Variable
# of cylinders
displacement
Equation
$77(x) + $627
$57(x) + $477
$66(x) + $175
$49(x) + $132
$17(x) + $69
$13(x) + $51
R2
a
a
0.9986
0.9986
"Not applicable, because a linear regression was not used.
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 proposed Tier 4 emissions 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 of this Draft RIA. The incremental cost for low-
sulfur fuel is described in Chapter 7 of this Draft RIA 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 are expressed in terms of cents/gallon of fuel consumed.
In Section 6.5 of this Draft RIA, we present these lifetime operating costs as a net present value
(NPV) in 2001 dollars for several example pieces of equipment.
A note of clarification should be made here. In Chapter 8 of this Draft RIA, we present
aggregate operating costs. Every effort is made to be clear what costs are related to increased
costs for low sulfur fuel and what costs are related to maintenance costs and/or savings. The
operating costs discussed in this section are only the latter of these - maintenance related costs
and/or savings. Increased costs associated with the lowering of sulfur in nonroad diesel fuel are
6-44
-------
Estimated Engine and Equipment Costs
discussed in detail in Chapter 7 of this Draft RIA. The cent per gallon costs presented in Chapter
7, along with the cent per gallon costs and savings present here, are then combined with
projected fuel volumes to generate the aggregate costs of our proposed fuel program.
Total operating costs, other than fuel, 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; the
change in fuel costs associated with the incrementally higher costs for low sulfur fuel, 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 fuel consumption increase 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 would be far less corrosive than is current
nonroad diesel fuel. Less corrosion would mean a slower acidification rate (i.e., less
degradation) of the engine lubricating oil and, therefore, more operating hours between needed
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 would reduce engine wear and oil
degradation to the existing nonroad diesel fleet as well as locomotive and marine engines, and
that a further reduction to 15 ppm sulfur would result in even greater reductions to the nonroad
fleet. This reduction in wear and oil degradation would provide a savings to users of this
equipment. The cost savings would also be realized by the owners of future nonroad engines that
are subject to the standards in today's proposal. As discussed below, these maintenance savings
have been estimated to be greater than 3 cents per gallon for the use of 15 ppm sulfur fuel when
compared to the use of today's unregulated nonroad diesel fuel.
We have identified a variety of benefits from the low-sulfur diesel fuel. These benefits are
summarized in Table 6.2-25.
6-45
-------
Draft Regulatory Impact Analysis
Table 6.2-25.
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 would be substantial. Additional savings would be realized in 2010
when the 15 ppm sulfur fuel would be introduced
We estimate the single largest savings would be the impact of lower sulfur fuel on oil change
intervals. We have estimated the oil change interval extension that would be realized by the
introduction of 500 ppm sulfur fuel in 2007, as well as the additional oil extension that would be
realized with the introduction of 15 ppm sulfur nonroad diesel 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 which today's nonroad engines may 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 to
today's relatively high nonroad diesel fuel sulfur levels, we estimate the use of 500 ppm sulfur
fuel would enable an oil change interval extension of 31 percent, while 15 ppm sulfur fuel would
enable an oil change interval extension of 35 percent relative to today's products.37
We present here a fuel cost savings attributed to the oil change interval extension in terms of
a cents per gallon operating cost. We estimate that an oil change interval extension of 31
percent, as would be enabled by the use of 500 ppm sulfur fuel in 2007, results in a weighted fuel
operating costs savings of 3.0 cents per gallon for the nonroad fleet. We project an additional
weighted cost savings of 0.3 cents per gallon for the oil change interval extension which would
be enabled by the use of 15 ppm sulfur beginning in 2010. Thus, for the nonroad fleet as a
whole, beginning in 2010, nonroad equipment users can realize an operating cost savings of 3.3
6-46
-------
Estimated Engine and Equipment Costs
cents per gallon compared to today's engine. For a typical 100 horsepower nonroad engine, this
represents a net present value lifetime savings of more than $500. Table 6.2-26 shows the
calculation of cent per gallon savings for various horsepower segments of the nonroad fleet.
6-47
-------
Table 6.2-26. Oil Change Maintenance Savings for Existing and New Nonroad, Locomotive, and Marine Engines ($2001)
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 II
0-25
0.408
7.1
18
524
0.41
1.75
250
327.5
$50.00
$3.49
$18.00
$71.49
106
139
$0.67
$0.51
$0.160
337.5
$50.00
$3.49
$18.00
$71.49
139
143
$0.51
$0.50
$0.015
$0.175
2.4%
25-50
0.408
7.1
37
579
0.44
3.59
250
327.5
$50.00
$7.18
$18.00
$75.18
234
306
$0.32
$0.25
$0.076
337.5
$50.00
$7.18
$18.00
$75.18
306
316
$0.25
$0.24
$0.007
$0.083
5.1%
50-75
0.408
7.1
67
707
0.44
6.50
250
327.5
$50.00
$13.00
$18.00
$81.00
424
555
$0.19
$0.15
$0.045
337.5
$50.00
$13.00
$18.00
$81.00
555
572
$0.15
$0.14
$0.004
$0.050
14.0%
75-175
0.38996
7.1
113
696
0.47
10.96
250
327.5
$50.00
$21.92
$18.00
$89.92
729
955
$0.12
$0.09
$0.029
337.5
$50.00
$21.92
$18.00
$89.92
955
984
$0.09
$0.09
$0.003
$0.032
26.3%
175-300
0.367
7.1
223
525
0.56
21.63
250
327.5
$50.00
$43.26
$35.00
$128.26
1614
2114
$0.08
$0.06
$0.019
337.5
$50.00
$43.26
$35.00
$128.26
2114
2179
$0.06
$0.06
$0.002
$0.021
23.0%
300-600
0.367
7.1
381
585
0.56
36.96
250
327.5
$50.00
$73.91
$35.00
$158.91
2757
3612
$0.06
$0.04
$0.014
337.5
$50.00
$73.91
$35.00
$158.91
3612
3722
$0.04
$0.04
$0.001
$0.015
17.7%
600-750
0.367
7.1
717
931
0.55
69.55
250
327.5
$50.00
$139.10
$35.00
$224.10
5096
6676
$0.04
$0.03
$0.010
337.5
$50.00
$139.10
$35.00
$224.10
6676
6880
$0.03
$0.03
$0.001
$0.011
4.1%
750+
0.367
7.1
1263
921
0.54
122.51
250
327.5
$100.00
$245.02
$70.00
$415.02
8813
11546
$0.05
$0.04
$0.011
337.5
$100.00
$245.02
$70.00
$415.02
11546
11898
$0.04
$0.03
$0.001
$0.012
7.5%
Locomotive
0.367
7.1
1263
921
0.54
122.51
250
327.5
$100.00
$245.02
$70.00
$415.02
8813
11546
$0.05
$0.04
$0.01 1
337.5
$100.00
$245.02
$70.00
$415.02
11546
11898
$0.04
$0.03
$0.001
$0.012
Marine
0.367
7.1
1263
921
0.54
122.51
250
327.5
$100.00
$245.02
$70.00
$415.02
8813
11546
$0.05
$0.04
$0.01 1
337.5
$100.00
$245.02
$70.00
$415.02
11546
11898
$0.04
$0.03
$0.001
$0.012
Notes to table 6.2-26:
(1) Oil change intervals are from William Charmley memo to docket.38
(2) Labor costs are from ICF Consulting under contract to EPA.39
(3) Oil use estimates are based on sump volumes scaled to engine displacement and, as such, they show differences for each horsepower category. The labor and filter costs are
average costs across a broad range of horsepower sizes and, as such, may overstate the cost for some engines while understating the costs for others.
-------
Estimated Engine and Equipment Costs
Table 6.2-26 shows oil change maintenance intervals for both the 500 ppm fuel and the 15
ppm fuel. The existing and new nonroad fleets would realize the savings associated with the 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. The locomotive and marine fleet would realize the
savings associated with the 500 ppm fuel for the years 2007 and beyond. The oil change
maintenance savings for locomotive and marine engines associated with the 15 ppm fuel are
shown in Table 6.2-26 for informational purposes only; these values are used only in our analysis
of alternative program options presented in Chapter 12 of this Draft RIA. Note that the weighted
values of 3.0 cents per gallon and 3.3 cents per gallon are calculated by weighting the cent per
gallon for each horsepower category by the fuel use weighting shown in the table.
The savings shown in Table 6.2-26 would 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 savings. We have not estimated the value of the savings from the
other benefits listed in Table 6.2-25 and, therefore, we believe the 3.3 cents per gallon savings is
conservative as it only accounts for the impact of low sulfur fuel on oil change intervals.
Operating costs associated with oil change maintenance are attributed 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 were
discussed in Chapter 4.1.1.3.4. To be conservative, we have used a maintenance interval of
3,000 hours for engines below 175 horsepower and 4,500 hours for engines above 175
horsepower, 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 horsepower and $260 per event for engines above 600 horsepower.
The calculations for CDPF maintenance are shown in Table 6.2-27'. Weighting the savings
shown by the fuel use weightings shown in the table, we can calculate these costs as 0.6 cents per
gallon which would be incurred only by new engines equipped with a CDPF.40
Operating costs associated with CDPF maintenance are attributed only to PM control.
6-49
-------
Draft Regulatory Impact Analysis
Table 6.2-27
CDPF Maintenance Costs for New CDPF-Equipped Engines ($2001)
PM Filter Maintenance Costs
Rated Power
BSFC
Fuel Density
Population Weighted Avg. Horsepower
Population Weighted Avg. Activity
Population Weighted avg. Load Factor
Filter Maintenance Interval
Filter Maintenance Cost Materials
Filter Maintenance Labor
Total Filter Maintenance Cost per event
Fuel Use Between Maintenance Interval
Maintenance Cost
Fuel Use Weightings
Units
hp
Ibm/hp-hr
Ibm/gallon
hp
hrs/year
% full load
hours
$/event
$/event
$/event
gallons/period
$/gallon
% total
Nonroad Enaines
0-25
0.408
7.1
18
524
0.41
3,000
$0
$65
$65
1,272
$0.051
2.4%
25-50
0.408
7.1
37
579
0.44
3,000
$0
$65
$65
2,807
$0.023
5.1%
50-75
0.408
7.1
67
707
0.44
3,000
$0
$65
$65
5,082
$0.013
14.0%
75-175
0.38996
7.1
113
696
0.47
3,000
$0
$65
$65
8,751
$0.007
26.3%
175-300
0.367
7.1
223
525
0.56
4,500
$0
$65
$65
29,048
$0.002
23.0%
300-600
0.367
7.1
381
585
0.56
4,500
$0
$65
$65
49,629
$0.001
17.7%
600-750
0.367
7.1
717
931
0.55
4,500
$0
$130
$130
91 ,728
$0.001
4.1%
750+
0.367
7.1
1263
921
0.54
4,500
$0
$260
$260
158,642
$0.002
7.5%
Labor costs are from ICF Consulting under contract to EPA.
6.2.3.3 Operating Costs Associated with Fuel Economy Impacts on New Engines
6.2.3.3.1 What Would the Fuel Economy Impacts Be ?
The high efficiency emission control technologies expected to be applied to meet the PM
standards for engines greater than 25 horsepower and the NOx standards for engines greater than
75 horsepower 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 that the integration of the engine and exhaust emission control
systems into a single synergistic emission control system will lead to nonroad engines which 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 timeframe for this improvement to be
realized, we have conservatively included this impact in our cost estimates for the full period of
the program.
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.F This approach results in filtering efficiencies for diesel
PM greater than 90 percent but requires additional pumping work to force the exhaust through
Typically, the filtering media is a porous ceramic monolith or a metallic fiber mesh. We refer to it as a "filter
trap" in Table 6.2-11.
6-50
-------
Estimated Engine and Equipment Costs
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 can typically not 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.42'43 We have estimated that the average impact of this increased
pumping work will be equivalent to an increase fuel consumption of approximately one
percent.44
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 a number of
typical in-use operating cycles in Chapter 4.1.3 of this draft RIA. 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. To address this situation,
we believe that some manufacturers will need to employ active backup regeneration systems that
provide supplemental heat to initiate regeneration as discussed in Chapter 4.1 of this Draft RIA
and, as explained in Section 6.2.2.3, we are costing active regeneration systems for all engines
using a CDPF system. We have estimated a cost for active regeneration systems for all engines
even though CDPF systems on many nonroad engines are expected to regenerate passively. We
have done this because we think that 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 which in-
use experience passive regeneration. Examples of active PM filter systems today, 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.45
Given the clean diesel fuel proposed in this rulemaking, the ability to use catalytic coatings to
promote soot oxidation and the fact that many kinds of nonroad equipment are expected to be
operated in a manner such that passive regeneration will occur, we believe that the average fuel
economy impact of the backup regeneration systems will be no larger than one percent.
We have projected that engines in the horsepower category from 25 hp to 75 horsepower 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 horsepower 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.
Therefore, we 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 be realized by
engines in this category from 25 hp to 75 hp. We believe that it is likely that in the long term this
impact will be recovered through continuing technology refinement as has historically happened.
However, to be conservative in our cost analysis, we have included this two percent impact for
the entire duration of the program.
6-51
-------
Draft Regulatory Impact Analysis
For engines in the horsepower category below 25 hp we have projected no need to use CDPF
technologies to comply with the proposed PM standard. Therefore, no fuel consumption impact
from the CDPF is estimated for this category.
We believe all engines in the horsepower categories above 75 hp will use integrated NOx and
PM control technologies to comply with the emission standards proposed today. The advanced
catalyst based emission control technology that we project industry will use to comply with the
proposed NOx standard offers the opportunity to improve fuel economy as described in the
following section. Based on those projected improvements, we have estimated that the net
impact on fuel consumption for engines greater than 75 hp due to the CDPF technology and the
NOx technology to be one percent. Future technology improvements are likely to recover this
fuel consumption impact; however, to be conservative in our cost analysis, we have assumed that
a one percent fuel consumption impact persists for the period of the emission control program.
6.2.3.3.1.2 NOx Control and Fuel Economy
NOx adsorbers are expected to be the primary NOx control technology introduced in order to
provide the reduction in NOx emissions for engines greater than 75 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. In order 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 on-
highway diesel engines can be made to operate in this way, it is anticipated that fuel economy
while operating under these conditions will be worse than normal. This increase in fuel
consumption can be minimized by carefully controlling engine air-to-fuel (A/F) ratios using the
control systems we anticipate will be used to meet the Tier 3 emission standards. The lower the
engine A/F ratio, the lower the amount of fuel which must be added in order to give rich
conditions. In the ideal case where the engine A/F ratio is at stoichiometry, and additional fuel is
required only as a NOx reductant the fuel economy penalty is virtually zero. We are projecting
that practical limitations on engine A/F control will mean that the NOx adsorber release and
reduction cycles will lead to a one percent decrease in the engine fuel economy.46 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, NOx adsorbers are poisoned by sulfur
in the fuel even at the low sulfur levels proposed today. As discussed in chapter 4 of this Draft
RIA, the sulfur poisoning of the NOx adsorber 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.47 This
rich operation will, like the NOx regeneration event, require an increase in the fuel consumption
6-52
-------
Estimated Engine and Equipment Costs
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 on-
highway and nonroad diesel engines. With a 15 ppm fuel sulfur cap, we are projecting that fuel
consumption for desulfation would 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.
While NOx adsorbers require non-power producing consumption of diesel fuel in order to
function properly and, therefore, have an impact on fuel economy, they are not unique among
NOx control technologies in this way. In fact NOx adsorbers are likely to have a very favorable
NOx to fuel economy trade-off when compared to our projected Tier 3 NOx control
technologies, 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.48
Nonroad Tier 2 diesel engines rely primarily on charge-air-cooling and injection timing
control (retarding injection timing) in order to meet the Tier 2 NOx+NMHC emission standard.
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 when compared to timing control, we have
forecast that less reliance on timing control will be needed in Tier 3, when compared to Tier 2.
Therefore, fuel economy will not be changed even at this lower NOx level. Similarly for the 25-
50 hp engines which would need to meet a 3.3 g/bhp-hr Tier 4 NOx emission limit under today's
proposal, we believe that there will be no change in fuel consumption due to the NOx standard.
NOx adsorbers have a significantly more favorable NOx to fuel economy trade-off when
compared to cooled EGR or timing retard.49 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
6-53
-------
Draft Regulatory Impact Analysis
timing.0 Therefore, we 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
would, 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 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 to today's technologies. This will result in both 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 in the
horsepower categories above 75 hp.
6.2.3.3.1.3 Fuel Economy Impacts for Engines without Advanced Emission Control
Technologies (engines <25 horsepower)
The emission standard proposed today for engines below 25 hp does not change the NOx
emission standard from the current Tier 2 level. The PM standard, however, is reduced by
almost 50%. We believe that this significant PM reduction will be realized 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 could 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 could 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. Therefore, no
change in fuel consumption is estimated in our cost analyses for engines with rated power below
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 per gallon. To that, we have added the incremental cost per gallon
G EPA has 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 which estimates the total consumption as approximately 2.5%
of total fuel consumption. Additionally the contractor's estimate of NOx adsorber efficiency
ranges from 80-90 percent, while EPA believes over 90 percent control is possible as discussed
fully in Chapter 4 of this draft RIA.
6-54
-------
Estimated Engine and Equipment Costs
for 15 ppm fuel where appropriate. These incremental fuel costs are discussed in Chapter 7 of
this Draft RIA as 4.8 cents per gallon. This increased operating cost - 60 cents plus 4.8 cents - is
applied to only those gallons of fuel consumed in engines equipped with technologies for which a
fuel economy impact would be realized. For 25 to 50 horsepower engines, where we estimate a
two percent impact, the incremental cost would be 1.3 cents per gallon (2%*64.8 cents/gallon).
For >75 horsepower engines, where we estimate a one percent fuel economy impact, the
incremental cost would be 0.65 cents per gallon.
Operating costs associated with fuel economy impacts are attributed only to PM control.
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 are shown in Table 6.2-28. On
a weighted basis, these costs are 0.2 cents per gallon and would be incurred only by new engines
equipped with a CDPF.
Operating costs associated with CCV maintenance are attributed evenly to NOx and PM
control.
Table 6.2-28
Closed Crankcase Ventilation System
Maintenance Costs for New Turbo-Charged Engines ($2001)
CCV Maintenance Costs
Rated Power
BSFC
Fuel Density
Population Weighted Avg. Horsepower
Population Weighted Avg. Activity
Population Weighted avg. Load Factor
CCV Filter Replacement Interval
CCV Filter Replacement Cost
Filter Maintenance Labor
Total Filter Maintenance Cost per event
Fuel Use Between Maintenance Interval
Turbcharged Fleet Fraction
Maintenance Cost
Fuel Use Weightings
Units
hp
Ibm/hp-hr
Ibm/gallon
hp
hrs/year
% full load
hours
$/event
$/event
$/event
gallons/period
[%]
$/gallon
% total
0-25
0.408
7.1
18
524
0.41
675
$8
$0
$8.00
286
0%
$0.028
0.0%
25-50
0.408
7.1
37
579
0.44
675
$8
$0
$8.00
631
5%
$0.013
0.2%
50-75
0.408
7.1
67
707
0.44
675
$8
$0
$8.00
1,143
41%
$0.007
5.7%
75-175
0.38996
7.1
113
696
0.47
675
$8
$0
$8.00
1,969
41%
$0.004
10.7%
175-300
0.367
7.1
223
525
0.56
675
$10
$0
$9.60
4,357
73%
$0.002
16.9%
300-600
0.367
7.1
381
585
0.56
675
$12
$0
$12.00
7,444
100%
$0.002
17.7%
600-750
0.367
7.1
717
931
0.55
675
$24
$0
$24.00
13,759
100%
$0.002
4.1%
750+
0.367
7.1
1263
921
0.54
675
$48
$0
$48.00
23,796
100%
$0.002
7.5%
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 2000,50 there are
6-55
-------
Draft Regulatory Impact Analysis
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
proposed transition program for equipment manufacturers (TPEM) is to reduce the potential for
anomalously high costs for individual equipment models by providing significant additional time
(up to 7 years) for developing less costly designs or to align the changes with an already
scheduled redesign. To present a conservative estimate of equipment-related costs, we have
assumed that the industry does not use the TPEM program which, we believe, offers the
opportunity for significant cost reductions. However, in Section 6.3.3 of this Draft RIA we
present an analysis of the potential cost savings of the TPEM program.
6.3.1 Equipment Fixed Costs
6.3.1.1 Equipment Redesign Costs
The projected modifications to equipment resulting from the proposed standards relate to
packaging of the exhaust emission control hardware expected to be added by engine
manufacturers to their new engines (see Section 6.2 for cost estimates of new emission control
hardware). As noted in Section 6.2, the additional emission control hardware is proportional in
size to engine displacement by a 4:1 ratio (1.5x engine displacement for both the CDPF and the
NOx adsorber, and Ix 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. We expect that some redesigns would be major in scale, while others
would be minor in scale. For example, in some cases, the redesign would simply be bolting the
new devices onto the existing design, but in most cases we expect devices to be designed into the
piece of equipment such that their presence would not be obvious to the casual observer.
Additionally, a redesign to accommodate a DOC (Ix engine displacement) should be less
intensive than a redesign to accommodate a CDPF/NOx adsorber system. Lastly, for >75
horsepower engines where proposed NOx standards are phased-in, we assume that the redesign
effort for those final NOx phase-in pieces of equipment (i.e., when the phase-in goes from 50
percent to 100 percent) would be less costly than the first redesign effort.
6.3.1.1.1 Schedule of Equipment Redesigns
The proposal contains a variety of emission compliance dates for the range of nonroad diesel
engines; these dates are as shown in Table 6.3-1. For this analysis, because we are assuming no
use of the TPEM program, we assume that the timing of equipment redesigns would correlate to
the implementation of the proposed engine standards assuming no use of the engine ABT
program. This results in a redesign schedule as shown in Table 6.3-1. We have noted what
percentage of equipment models would be redesigned in years for which proposed engine
standards would be implemented. The table also notes what percentage are major redesign
efforts and what percentage are minor efforts. We also note what percentage of the redesign
costs are allocated to PM and what percentage to NOx.
6-56
-------
Estimated Engine and Equipment Costs
Table 6.3-1
Equipment Redesign Assumptions for Equipment Manufacturers
Horsepower
0750hp
Engine
Standard
Dates
2008
2008
2013
2008
2013
2012
2014
2011
2014
2011
2014
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
50% PM
50% NOx
50% PM
50% NOx
Percent of Equipment
Models Undergoing
Minor Redesign
100%
100%
100%
50%
50%
Percent of Equipment
Models Undergoing
Major Redesign
100%
100%
100%
100%
50%
50%
Note that we have assumed that all equipment redesigns for the 75 to 750 horsepower range
are major in the first year of proposed engine 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 have done this because we believe that equipment manufacturers would
expend less effort to redesign those pieces equipment needing to add only the NOx adsorber (in
those years where NOx phase-ins change from 50 percent to 100 percent) for three reasons: (1)
these models would already have been redesigned for the CDPF system and would already
incorporate the necessary electronic systems into their design; (2) equipment manufacturers
would, presumably, have gained experience during the major redesign phase that should make
the minor redesign phase more efficient; and, (3) manufacturers 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. For engines over 750 horsepower, we have
projected that 50 percent of the engines would be redesigned to incorporate a CDPF/NOx
adsorber system in 2011 with the remaining 50 percent being modified in 2014. These
projections are consistent with the phase-in of the proposed standards; both redesign efforts are
assumed to be major since we assume that the NOx phase-in engines/equipment would be the
same as the PM phase-in engines/equipment.
Our equipment redesign cost estimates were developed based on our meetings and
6-57
-------
Draft Regulatory Impact Analysis
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.
6.3.1.1.2 Costs of Equipment Redesigns
While developing our equipment redesign cost estimates for the proposed 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 2/3 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 increase heat rejection
and require 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 (e.g., 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 those engines
which don't have electronic fuel systems will require them, and equipment manufacturers will
now need to integrate aftertreatment systems (as compared to charge-air-coolers, turbochargers,
larger radiators and fans).
A number of the companies we met with in the past year provided us with 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).
In addition to the companies we met with in the past year, we also received redesign cost
estimates from a number of equipment companies during the Tier2/3 rulemaking regarding their
projected costs for the Tier 2 standards (these data are confidential business information). The
information provided to EPA 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, what we learned was those 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 judgement, to develop the
redesign cost estimates presented in Table 6.3-2. This table presents fixed cost per motive and
non-motive equipment model (motive equipment is that with some form of propulsion system
while non-motive equipment has none, e.g., air compressors, generator sets, hydraulic power
units, irrigation sets, pumps and welders) for each horsepower group. In general, non-motive
equipment has fewer design demands than does motive equipment - no operator line-of-sight
6-58
-------
Estimated Engine and Equipment Costs
demands, fewer serviceability constraints, and almost no impact (collision) concerns. As a result,
we have estimated a lower redesign cost for non-motive equipment relative to motive equipment.
Table 6.3-2
Estimated Equipment Redesign Costs Per Model
Horsepower
0750hp
Motive
$50,000
$50,000
$187,500
$350,000
$350,000
$500,000
$500,000
$750,000
$750,000
$750,000
Non-Motive
$50,000
$50,000
$75,000
$100,000
$100,000
$100,000
$100,000
$100,000
$100,000
$100,000
Using the PSR database we were able to determine the number of equipment models and the
type of equipment model (motive versus non-motive). We distinguished motive from non-
motive using our Nonroad Model definition of stationary applications. Non-motive applications
include air compressors, generator sets, pumps, hydraulic power units, irrigation sets, and
welders. All other applications are considered motive.
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 would
be about 100 hours - 75 hours of junior engineering time, and 20 hours of senior engineering
time, and 5 hours of clerical time - which would be about $10,000. We projected that the level
of effort needed by equipment manufacturers to modify support literature for each non-motive
application product line would be about 50 hours (distributed similarly), which is equivalent to
about $5,000. Table 6.3-3 contains the total costs per power category for changes to support
literature.
6-59
-------
Draft Regulatory Impact Analysis
Table 6.3-3
Costs Associated with Changes to Product Support Literature ($1,OOP's)
Horsepower
0750hp
Motive models
561
705
496
722
1289
1222
677
127
117
Motive Cost
$5,610
$7,050
$4,960
$7,220
$12,890
$12,220
$6,770
$1,270
$1,170
Non-motive
models
159
169
138
146
223
227
178
0
0
Non-motive
cost
$795
$845
$690
$730
$1,115
$1,135
$890
$0
$0
Total Cost
$6,405
$7,895
$5,650
$7,950
$14,005
$13,355
$7,660
$1,270
$1,170
6.3.1.3 Total Equipment Fixed Costs
The annual equipment fixed costs for each horsepower category are shown in Table 6.3-4.
As was done for engine fixed costs, we have attributed only a portion of the equipment fixed
costs to sales within the United States. We have done this because we believe that these efforts
would be needed to sell equipment not only in the US, but also in Australia, Canada, Japan, and
the countries of the European Union. Therefore, as was discussed in more detail in section
6.2.1.1, we have attributed 42 percent of the equipment fixed costs to U.S. sales.
The analysis projected that the costs would be incurred over a two year period prior to the
first year of the emission standards. The costs were then amortized over 10 years at a seven
percent rate beginning with the first year of the engine standard to reflect the time value of
money. The 10 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.
Per unit fixed costs are shown in Table 6.3-5 and use our projections of engine growth as
presented in Table 8-1.
6-60
-------
Table 6.3-4
Recovered (Annualized) Equipment Fixed Costs per Horsepower Category ($2001, in thousands of dollars)
Year
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
Total
0750hp
$0
$0
$0
$1 ,693
$1 ,693
$1 ,693
$3,387
$3,387
$3,387
$3,387
$3,387
$3,387
$3,387
$1 ,693
$1 ,693
$1 ,693
$33,867
Total
$4,852
$4,852
$4,852
$47,257
$74,730
$84,064
$102,804
$102,804
$102,804
$102,804
$97,952
$97,952
$97,952
$55,547
$28,074
$18,740
$1 ,028,036
-------
Table 6.3-5
Recovered Equipment Fixed Cost per Unit ($2001)
Year
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
0750hp
Sales
3,193
3,244
3,295
3,346
3,397
3,448
3,499
3,550
3,601
3,652
3,703
3,754
3,805
3,856
3,907
3,958
$/unit
$0
$0
$0
$506
$498
$491
$968
$954
$940
$927
$915
$902
$890
$439
$433
$428
-------
Estimated Engine and Equipment Costs
Costs per unit vary from year to year due to proposed standard phase-ins. The rapid decline
in per unit costs during the final two or three years for >75 horsepower engines is because the
latter redesign work - to accommodate the final year of the NOx phase-in - is considered a minor
and less costly redesign, as was discussed above.
6.3.2 Equipment Variable Costs
In addition to the incrementally higher cost of new engines estimated in section 6.2.1 and
6.2.2, equipment manufacturers would 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 contain 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 the new aftertreatment devices.
Here, we estimate the cost for additional sheet metal that could be used to shroud or
otherwise encase aftertreatment system within the confines of the hood or other body cladding on
a piece of equipment. The amount of 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 model in the PSR database
which incorporated an engine over 75hp (1.5 times engine displacement for CDPF and the same
for 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 $1.10 per square foot for hot rolled steel. The cost
for each model was multiplied by the total sales for that model using the 2000 sales information
in the 2002 PSR Sales Database. The total costs were summed for each power group and then
divided by the total sales for the power group for a sales weighted average cost. These costs
were then added to variable cost estimates for brackets and bolts required to secure the
aftertreatment devices within the equipment, other such miscellaneous items including
weldments, plastics, castings, gaskets, seals, and hoses, as well as the labor required to install the
new aftertreatment devices. A twenty-nine percent markup for overhead and profit is also
included in the final cost estimate as shown in Table 6.3-6.
6-63
-------
Draft Regulatory Impact Analysis
Table 6.3-6
Equipment Variable Costs"
Horsepower
0750hp
Year
2008
2013
2013
2012
2012
2011
2011
2011
2011
Bolts
$0
$4
$4
$20
$20
$20
$40
$40
$80
Sheet Metal
$0
$0
$0
$3
$3
$5
$6
$9
$14
Labor
$0
$10
$10
$20
$20
$29
$59
$59
$78
Subtotal
$0
$14
$14
$42
$43
$54
$105
$108
$173
29% Markup
$0
$4
$4
$12
$12
$16
$30
$31
$50
Total
$0
$18
$18
$55
$55
$70
$135
$139
$223
a Some equipment types have strict surface temperature requirements for exhaust components. Air gapping and water
jacketing systems are on such engines and would likely be extended to include the area of the aftertreatment. Such costs
are not included in this analysis for these costs would only apply to specialized equipment (<1%). However, costs have
been calculated in a memo to the docket (docket A-2001-28).
As shown in Table 6.3-6, we have estimated equipment variable costs for less than 25
horsepower equipment to be $0 under the assumption that an added DOC would replace the
existing muffler and make use of the same bracket/bolt/labor used for the muffler. This is also
assumed for engines in the 25 to 75 horsepower range during the years 2008 through 2012 when
only a DOC is being used by the engine manufacturer for compliance; additional bolts and labor
costs are added for the addition of a CDPF beginning in 2013.H 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.
6.3.3 Potential Impact of the Transition Provisions for Equipment Manufacturers
As discussed in Section VII.B of the preamble, we have proposed to extend the Transition
Provisions for Equipment Manufacturers (TPEM) which were developed in the 1998 nonroad
rule into the proposed Tier 4 program (with a number of modifications as discussed in Section
VII.B of the preamble). The TPEM is an important component of our proposal because of the
flexibility it provides for equipment manufacturers. However, as explained earlier, because the
program is optional, we have not included an estimate of the potential impacts of the program on
the overall costs of our proposed Tier 4 program. Nevertheless, in this section we discuss why
the TPEM program can have a substantial impact reducing equipment manufacturer costs.
Note that, for costing purposes, we have assumed that a DOC is used on all <75 horsepower engines to comply
with the 2008 standards although test data suggests that some engines may not need to add a DOC because they would
already meet the proposed standards.
6-64
-------
Estimated Engine and Equipment Costs
The TPEM can reduce equipment manufacturer costs in two ways. First, the proposed Tier 4
TPEM program would allow equipment manufacturers to continue to sell a limited number of
equipment with non-Tier 4 engines even after the Tier 4 standards go into effect. Therefore, any
engine price increase associated with the proposed Tier 4 standards would not be incurred by the
equipment manufacturer or by the end user during the time frame the manufacturers make us of
the TPEM. Second, the TPEM program allows manufacturers to schedule equipment design
cycles so that the normal redesign cycle can overlap 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 number of small
volume equipment 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 contained in
an industry-wide database (the Power Systems Research database).51 This analysis looked at
each equipment manufacturers product offerings (e.g., different equipment models) by power
category and the estimated 2000 U.S. sales of each equipment model. We used this database to
analyze how equipment manufacturers could make use of the proposed TPEM program to
maximize the number of equipment models which could take advantage of the TPEM to delay
any equipment redesign associated with the proposed Tier 4 standards until the eighth year of the
program (as discussed in Section VII.B of the preamble, we have proposed to allow the TPEM
program to last until seven years after the Tier 4 standards are implemented.). We specifically
analyzed the proposed 80 percent allowance and the small volume option we have requested
comment on (as discussed in the preamble). The results are shown in Table 6.3-7.
Table 6.3-7
Potential Impact of TPEM Program on Equipment Models and Sales
Equipment Models/
Equipment Sales
% of all equipment models
which could use TPEM for
full-seven years
Percent of equipment sales
which could use TPEM for
fi]H-s,pvpn vpnr^
Engine Power Category
<25hp
56%
7%
25< hp <70a
61%
, J 0% ,
70a750 hp
80%
91%
All Power
Categories
66%
1 0%
believe the results of this analysis would have been significantly different if the power outpoint was reduce at 75hp.
This analysis indicates that if fully utilized by equipment manufacturers, 66 percent of all of
the nonroad diesel equipment models could 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 of their equipment models which used 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 6.3-7 indicates that for the 25 - 75 hp
category, 61 percent of all equipment models in this power range could take advantage of the
TPEM to delay an equipment redesign for seven years. It is important to note that while the
6-65
-------
Draft Regulatory Impact Analysis
TPEM can substantially reduce equipment redesign costs, it would be 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 which can be impacted by the TPEM (also shown in Table 6.3-
7) is estimated to be no higher than ten percent for no more than seven years.
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.
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 were
discussed in detail in Section 6.2.1. The total estimated engine fixed costs are summarized in
Table 6.4-1.
Table 6.4-1
Summary of Engine Fixed Costs (millions)
R&D
Tooling
Certification
Total
Incurred Costs
$199
$67
$72
$338
Recovered Costs
$279
$81
$88
$448
6.4.1.2 Engine Variable Costs
Engine variable costs were 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. These equations are summarized in Table 6.4-2. Note that not all equations were used
for all engines; equations were used in the manner shown in Table 6.4-2. We have calculated the
aggregate engine variable costs and present them in Chapter 8 of this Draft RIA. The net present
value of these variable costs between the years 2004 through 2036 is $13.9 billion.
6-66
-------
Estimated Engine and Equipment Costs
Table 6.4-2
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
Common Rail Fuel
Injection
(mechanical fuel
system baseline)
Common Rail Fuel
Injection
(electronic rotary fuel
system baseline)
Time Frame3
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
$105(x) + $180
$84(x) + $158
$150(x) + $71
$114(x) + $54
$20(x) + $289
$15(x) + $219
$10(x) + $144
$7(x) + $110
$19(x) + $117
$18(x) + $110
$2(x) + $35
$2(x) + $25
$17(x) + $69
$13(x) + $51
$77(x) + $627
$57(x) + $477
$66(x) + $175
$49(x) + $132
Dependent
Variable (x)
Displacement1"
Displacement
Displacement
Displacement
Displacement
Displacement
Displacement
# of cylinders/
injectors
# of cylinders/
injectors
How Used
>75hp engines according to
phase-in of NRT4 NOx std.
>25hp engines according to
NRT4 PM std.
IDI engines adding a CDPF
DI engines adding a CDPF
<25hp engines beginning in 2008;
25-75hp engines 2008 thru 2012
All turbo-charged engines when
they first meet a proposed PM std.
25-50 hp engines beginning in
2013
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 & 2; Long term = years 3+. Explanation of near term and long term can be found in Section
6.1.
b Displacement refers to engine displacement in liters.
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 of this Draft RIA.
6-67
-------
Draft Regulatory Impact Analysis
Table 6.4-3
Engine Operating Costs Associated with the Proposed Fuel Program
(cents/gallon of fuel consumed)
Horsepower
category
0750hp
Locomotive/Marine
Oil Change
Savings
(17.5)
(8.3)
(5.0)
(3.2)
(2.1)
(1.5)
(1.1)
(1.2)
(1.1)
CDPF
Maintenance
0.0
2.3
1.3
0.7
0.2
0.1
0.1
0.2
0.0
CCV
Maintenance
0.0
1.3
0.7
0.4
0.2
0.2
0.2
0.2
0.0
CDPF
Regeneration3
0.0
1.30
1.30
0.65
0.65
0.65
0.65
0.65
0.0
Net Operating
Costsb
(17.5)
(3.4)
(1.7)
(1.5)
(1.1)
(0.6)
(0.2)
(0.2)
(1.1)
a A one or two percent fuel consumption increase, a 60 cent/gallon baseline fuel price, and a 4.8 cent/gallon
incremental fuel cost.
b The incremental costs for the proposed low sulfur fuel are not included here. Fuel costs are presented in Chapter
7 of this Draft RIA.
Engines that make up the existing fleet would realize the oil change savings shown in Table
6.4-3 while incurring none of the other operating costs because these engines would not be
equipped with a CDPF system or be adding a CCV system. 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 pieces of equipment and present those in Section 6.5. Aggregate operating costs - the
annual total costs - are presented in Chapter 8 of this Draft RIA.
6.4.2 Equipment Costs
6.4.2.1 Equipment Fixed Costs
Equipment fixed costs were discussed in detail in Section 6.3.1. Table 6.4-4 shows estimated
equipment fixed costs associated with the proposed program. These costs include costs for
equipment redesign and generation of new product support literature.
6-68
-------
Estimated Engine and Equipment Costs
Table 6.4-4
Summary of Equipment Fixed Costs (millions)
Redesign
Product Literature
Total
Incurred Costs
$678
$19
$697
Recovered Costs
$999
$29
$1,028
6.4.2.2 Equipment Variable Costs
Equipment variable costs are discussed in detail in Section 6.3.2. Table 6.4-5 shows our
estimated per unit equipment variable costs. This table is a repeat of Table 6.3-6.
Table 6.4-5
Equipment Variable Costs per Unit
Horsepower
0750hp
Year
2008
2013
2013
2012
2012
2011
2011
2011
2011
Bolts
$0
$4
$4
$20
$20
$20
$40
$40
$80
Sheet Metal
$0
$0
$0
$3
$3
$5
$6
$9
$14
Labor
$0
$10
$10
$20
$20
$29
$59
$59
$78
Subtotal
$0
$14
$14
$42
$43
$54
$105
$108
$173
29% Markup
$0
$4
$4
$12
$12
$16
$30
$31
$50
Total
$0
$18
$18
$55
$55
$70
$135
$139
$223
We have calculated the aggregate equipment variable costs in Chapter 8 of this Draft RIA.
Those costs show the annual total variable costs we have estimated for our proposal. The net
present value of these variable costs between the years 2004 through 2036 is $498 million.
6.5 Costs for Example Pieces of Equipment
6.5.1 Summary of Costs for Some Example Pieces of Equipment
To better illustrate the engine and equipment cost impacts we are estimating for today's
proposed standards, we have chosen several example pieces of equipment and presented the
estimated costs for them. Using these examples, we can calculate the costs for a specific piece of
equipment in several horsepower ranges and better illustrate the cost impacts of today's proposed
standards. These costs along with information about each example piece of equipment are
6-69
-------
Draft Regulatory Impact Analysis
shown in Table 6.5-1. Costs presented are near term and long term costs for the final standards
to which each piece of equipment would comply. Long term costs are only variable costs and,
therefore, represent costs after all fixed costs have been recovered. Included in the table are
estimated prices for each piece of equipment to provide some perspective on how our estimated
control costs relate to existing equipment prices.
Table 6.5-1
Near Term and Long Term Costs for Several Example Pieces of Equipmenta
($2001, 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
$170
$3,500
-$90
$940
Skid/Steer
Loader
33 hp
1.5
3
natural
DI
$760
$1,100
$13,500
$40
$2,680
Backhoe
76 hp
3.9
4
turbo
DI
$1,210
$1,680
$50,000
$370
$7,960
Dozer
175 hp
10.5
6
turbo
DI
$2,590
$3,710
$235,000
$1,550
$77,850
Ag
Tractor
250 hp
7.6
6
turbo
DI
$2,000
$2,950
$130,000
$1,320
$23,750
Dozer
503 hp
18
8
turbo
DI
$4,210
$6,120
$575,000
$4,950
$77,850
Off-
Highway
Truck
1000 hp
28
12
turbo
DI
$6,780
$10,100
$700,000
$12,550
$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. "Estimated Price of New Nonroad Example Equipment," memorandum from Zuimdie Guerra to docket A-2001 -
28.52
c. Present value of lifetime costs.
6.5.2 Method of Generating Costs for Our Example Pieces of Equipment
To facilitate the readers ability to duplicate this example analysis for other pieces of
equipment, this section will briefly describe the necessary steps to create the cost analysis based
on the information contained in this Draft RIA.
The first step required to develop an estimate of our projected cost for control under the
proposed 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
6-70
-------
Estimated Engine and Equipment Costs
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 combustion system used by the engine (i.e., indirect-injection, IDI, or direct injection,
DI);
• model year of production; and,
• the horsepower category of the engine.
With this information, and the data tables contained in this Draft RIA, an estimate of the
compliance costs can be made.
As an example, here we will estimate the cost of compliance for the 76hp backhoe in the year
2012. Table 6.5-1 shows the near term cost to be $1,680 and the long term cost to be $1,210.
The first step is to define our engine characteristics as shown in Table 6.5-2.
Table 6.5-2
Engine and Equipment Characteristics of an Example Cost Estimate
76 hp Backhoe Example
Model Year
Displacement (liters)
Cylinder (number)
Aspiration
Combustion System
Horsepower 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.
6-71
-------
Draft Regulatory Impact Analysis
Table 6.5-3
Fixed Costs for an Example Cost Estimate
2012 76hp Backhoe Example
Engine R&D
Engine Tooling
Engine Certification
Equipment Fixed
Total Fixed Costs
$27
$15
$11
$90
$143
Table 6.2-4 Engine R&D Costs (per engine)
Table 6.2-6 Engine Tooling Costs (per engine)
Table 6.2-8 Engine Certification Costs (per engine)
Table 6.3-5 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 6.4-2. The table includes all of the different variable cost
components for different size ranges of engines meeting different proposed 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 which do not apply for a particular example (see Table 6.5-4 below).
6-72
-------
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
$105(x) + $180
$84(x) + $158
$105(3.9)+$180 =
$590
$150(x) + $71
$114(x) + $54
$150(3.9)+$71=
$656
$20(x) + $289
$15(x) + $219
not applicable
$10(x) + $144
$7(x) + $110
$10(3. 9)+$ 144=
$183
$19(x) + $117
$18(x) + $110
not applicable
$2(x) + $35
$2(x) + $25
$2(3.9)+$35=
$43
$17(x) + $69
$13(x) + $51
not applicable
$77(x) + $627
$57(x) + $477
not applicable
$66(x) + $175
$49(x) + $132
not applicable
Dependent
Variable (x)
Displacement11
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
>75hp engines according to
phase-in of NRT4 NOx std.
In 2012 a 76 hp engine in the
NOx phase-in set would require a
NOx adsorber
>25hp 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 (DI) combustion system
not an indirect injection (IDI)
DI engines adding a CDPF
The example engine is a DI
engine and has a CDPF
<25hp engines beginning in 2008;
25-75hp engines 2008 thru 2012
Example engine rated power is
greater than 75 hp
All turbo-charged engines when
they first meet a proposed PM std.
The example engine is
turbocharged
25-50 hp engines beginning in
2013
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 &
6.1.
b Displacement refers to
2; Long term = years 3+. Explanation of near term and long term can be found in Section
engine displacement in liters.
-------
Draft Regulatory Impact Analysis
Summing the applicable variable costs estimated in table 6.5-4 gives a total engine variable
cost for the 76hp Backhoe example of $1472. The equipment variable costs are presented in
table 6.4-3 and are referenced by engine power category. For the 76hp example here, the
estimated equipment variable costs are $55.
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 estimate here. The resulting total is $1670 ($143 + $1472 + $55, note that
rounding may result in slightly different results). Typically we have presented these total cost
estimates to the nearest ten dollars.
6-74
-------
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."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.
3. "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.
4. "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.
5. "Learning Curves in Manufacturing," Linda Argote and Dennis Epple, Science, February 23,
1990, Vol. 247, pp. 920-924.
6. Power Systems Research, OELink Sales Version, 2002.
7. 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.
8. Nonroad Diesel Final Rule, 63 FR 56968, October 23, 1998.
9. Certification Fees Proposed Rule, 67 FR 51402, August 7, 2002.
10. "Learning Curves in Manufacturing," Linda Argote and Dennis Epple, Science, February 23,
1990, Vol. 247, pp. 920-924.
11. "Treating Progress Functions As Managerial Opportunity", J.M Dutton and A. Thomas,
Academy of Management Review, Rev. 9, 235, 1984. Copy available in EPA Air Docket A-
98-32, docket item number U-D-13.
12. Nonconformance Penalty Final Rule, 67 FR 51464, August 8, 2002.
13. "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.
14. Estimated Economic Impact of New Emission Standards for Heavy-Duty On-Highway
Engines, March 1997, EPA 420-R-97-009.
15. "Cost Estimates for Heavy-Duty Gasoline Vehicles," Arcadis Geraghty & Miller, September
1998, EPA Air Docket A-2001-28.
6-75
-------
Draft Regulatory Impact Analysis
16. "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.
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.
18. McDonald and Bunker, "Testing of the Toyota Avensis DPNR at U.S. EPA-NVFEL," SAE
2002-01-2877, October 2002.
19. "Cost Estimates for Heavy-Duty Gasoline Vehicles," Arcadis Geraghty & Miller, September
1998, EPA Air Docket A-2001-28.
20. 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 Converters,
Series Id: PCU3714#503.
21. "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.
22. Johnson Matthey Platinum Today, www.platinum.matthey.com/prices .
23. "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.
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.
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.
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.
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.
28. 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 Converters,
Series Id: PCU3714#503.
6-76
-------
Estimated Engine and Equipment Costs
29. "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.
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.
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.
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.
33. Czerwinski, Jaussi, Wyser, and Mayer, "Particulate Traps for Construction Machines
Properties and Field Experience," SAE 2000-01-1923, June 2000.
34. "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.
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. "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.
38. "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.
39. "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.
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. "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.
6-77
-------
Draft Regulatory Impact Analysis
42. 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.
43. LeTavec, C., et al, "Year-Long Evaluation of Trucks and Buses Equipped with Passive
Diesel Particulate Filters," March 20002, SAE 2002-01-0433.
44. "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.
45. Johnson, T., "Diesel Emission Control: 2001 in Review," March 2002, SAE 2002-01-0285.
46. "Regulatory Impact Analysis: Heavy-Duty Engine and Vehicle Standards and Highway
Diesel Fuel Sulfur Control Requirements," December 2000, EPA420-R-00-026.
47. Dou, D. and Bailey, O., "Investigation of NOx Adsorber Catalyst Deactivation" SAE982594.
48. Herzog, P. et al, NOx Reduction Strategies for DI Diesel Engines, SAE 920470, Society of
Automotive Engineers 1992 (from Figure 1).
49. Zelenka, P., et al., "Cooled EGR - A Key Technology for Future Efficient HD Diesels", SAE
980190.
50. Power Systems Research, OELink Sales Version, 2002.
51. Power Systems Research, OELink Sales Version, 2002.
52. "Estimated Price of NewNonroad Example Equipment," memorandum from Zuimdie Guerra
to docket A-2001-28.
6-78
-------
CHAPTER 5: Fuel Standard Feasibility
5.1 Blendstock 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 5-6
5.2.1 Introduction to Diesel Fuel Sulfur Control 5-6
5.2.2 Conventional Hydrotreating 5-7
5.2.2.1 Fundamentals of Distillate Hydrotreating 5-8
5.2.2.2 Meeting a 15 ppm Cap with Distillate Hydrotreating 5-12
5.2.2.3 Low Sulfur Performance of Distillate Hydrotreating 5-17
5.2.3 Phillips S-Zorb Sulfur Adsorption 5-19
5.2.4 Linde Isotherming 5-22
5.2.5 Chemical Oxidation and Extraction 5-25
5.2.6 FCC Feed Hydrotreating 5-25
5.3 Feasibility of Producing 500 ppm Sulfur Nonroad Diesel Fuel in 2007 5-26
5.3.1 Expected use of Desulfurization Technologies for 2007 5-26
5.3.2 Leadtime Evaluation 5-27
5.3.2.1 Tier 2 Gasoline Sulfur Program 5-28
5.3.2.2 15 ppm Highway Diesel Fuel Sulfur Cap 5-29
5.3.2.3 Leadtime Projections for Production of 500 ppm NRLM Diesel Fuel 5-31
5.3.2.4 Comparison with the 500 ppm Highway Diesel Fuel Program 5-34
5.3.2.5 Small Refiners 5-35
5.4 Feasibility of Distributing 500 ppm Sulfur Non-Highway Diesel Fuel in 2007 and 500 ppm Locomotive and
Marine Diesel Fuel in 2010 5-35
5.4.1 The Diesel Fuel Distribution System Prior to the Implementation of the Proposed 500 ppm Sulfur
Program: 5-36
5.4.2 Summary of the Proposed 500 ppm Sulfur Standards 5-36
5.4.3 Limiting Sulfur Contamination 5-38
5.4.4 Potential Need for Additional Product Segregation 5-39
5.5 Feasibility of Producing 15 ppm Sulfur Nonroad Diesel Fuel in 2010 5-44
5.5.1 Expected use of Desulfurization Technologies for 2010 5-44
5.5.2 Leadtime Evaluation 5-47
5.6 Feasibility of Distributing 15 ppm Sulfur Nonroad Diesel Fuel in 2010 5-48
5.6.1 The Diesel Fuel Distribution System Prior to the Implementation of the Proposed 15 ppm Nonroad
Diesel Sulfur Program 5-48
5.6.2 Summary of the Proposed 15 ppm Nonroad Diesel Sulfur Standard 5-48
5.6.3 Limiting Sulfur Contamination 5-49
5.6.4 Potential need for Additional Product Segregation Due to the Implementation of the Proposed 15 ppm
Sulfur Specification for Nonroad Diesel Fuel 5-50
5.7 Impacts on the Engineering and Construction Industry 5-52
5.7.1 Design and Construction Resources Related to Desulfurization Equipment 5-52
5.7.2 Number and Timing of Revamped and New Desulfurization Units 5-53
5.7.3 Timing of Desulfurization Projects Starting up in the Same Year 5-61
5.7.4 Timing of Design and Construction Resources Within a Project 5-62
5.7.5 Projected Levels of Design and Construction Resources 5-63
5.8 Supply of Nonroad, Locomotive, and Marine Diesel Fuel (NRLM) 5-68
5.9 Desulfurization Effect on Other Non-Highway Diesel Fuel Properties 5-74
5.9.1 Fuel Lubricity 5-74
5.9.2 Volumetric Energy Content 5-77
5.9.3 Fuel Properties Related to Storage and Handling 5-79
5.9.4 Cetane Index and Aromatics 5-79
5.9.5 Other Fuel Properties 5-80
5.10 Feasibility of the Use of a Marker in Heating Oil from 2007-2010 and in Locomotive and Marine Fuel from
2010-2014 5-81
Appendix 5A: EPA's Legal Authority for Proposing Nonroad, Locomotive, and Marine Diesel Fuel Sulfur Contr6188
-------
Fuel Standard Feasibility
CHAPTER 5: Fuel Standard Feasibility
In this chapter, we present the methodology used to develop the costs which would result
from the proposed fuel program, as well as the projected costs themselves. In Section 5.1, we
estimate the volumes of diesel fuel which would be affected by the 500 and 15 ppm sulfur caps
in various phases of the proposed fuel program. In Section 5.2, we evaluate a wide variety of
distillate desulfurization technologies which refiners could potentially use to meet 500 and 15
ppm sulfur caps. In Section 5.3, we formally assess the technical feasibility of the 500 ppm
sulfur cap in 2007, including the sufficiency of the leadtime provided refiners. In Section 5.4, we
assess the feasibility of distributing the 500 ppm sulfur fuel which would be required in 2007. In
Section 5.5, we formally assess the technical feasibility of the 15 ppm sulfur cap in 2010,
including the sufficiency of the leadtime provided refiners. In Section 5.6, we assess the
feasibility of distributing the 15 ppm sulfur fuels which would be required in 2010. Finally, in
Section 7.6, we project the possible impacts of the proposal on diesel fuel prices.
5.1 Blendstock 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 which 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 U.S. is about one weight
percent.2 Most of the sulfur in crude oil is in the heaviest boiling fractions. Since most of the
refinery blend stocks 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 blend stocks 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
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 U.S.
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 abroad 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 number 2 distillate which
comprises the majority of diesel fuel and heating oil.
5-1
-------
Draft Regulatory Impact Analysis
Figure 5.1-1
Diagram of a Typical Complex Refinery
Natural
Gas
Coker
Refineries differ from the model in the preceding diagram depending on the characteristics of
the crude oils refined, and their product slate. For example:
- 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 in the
manufacture of heating oil, kerosene and jet fuel. Much of the distillate product from the
hydrocracker is often blended into jet fuel rather than diesel fuel and current highway
regulations generally require that a refinery have a hydrotreater which usually would not be
necessary if the refinery produced only heating oil.
On an aggregate basis, most of the distillate manufactured in the U.S. 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 small
fraction of diesel fuel volume comes from a coker conversion unit or other units which crack
5-2
-------
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 is
currently accomplished in fixed-bed hydrotreaters that operate at moderate pressures (500-800+
psi)3. Nearly all of 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 -
3000 psi, although most operate at 1500 - 3000 psi, which naturally produces distillate fuel with
sulfur levels of about 100 ppm.
To comply with applicable sulfur standards, which range from 2000-5000 ppm, or the 40
cetane standard for non-highway 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 will be 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 contained in Table 5.1-1.
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
-
-
-------
Draft Regulatory Impact Analysis
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.
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, for a
couple of reasons. First, some of the desulfurization technologies affect these other fuel
properties. Second, as will be discussed further below, some sulfur compounds are more
difficult to treat than others. Refiners could potentially try to shift these more difficult
compounds to fuels which face less stringent sulfur standards. Their ability to do this depends in
part on the effect of such shifts on non-sulfur 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 their applicable industry standards.
Data on the current distillation characteristics, API gravity, pour point, natural cetane level,
and aromatics content of diesel fuel blendstocks are contained in the 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
no
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.
5-4
-------
Fuel Standard Feasibility
The American Society for Testing Materials (ASTM) has established consensus standards
which apply to #2 non-highway diesel fuel, as well as for #2 distillate fuel (e.g., heating oil).8
The specifications which are most relevant to desulfurization are summarized in Table 5.1-3.
Table 5.1-3
ASTM Specifications which 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
#2 Diesel Fuel (Non-
highway)
540
640
None
46 to -0.4
5000
40
#2 Fuel Oil/Heating Oil
540
640
0.876 (30.0)
21.2
5000
#2 Marine Distillate
(DMA)
I
0.890 (27.5)
21.2
40
Comparing Tables 5.1-2 and 5.1-3 shows that the average properties of current non-highway
distillate are within the specifications, and for some properties, well within specifications. 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, it should be
noted that we did not assume that refineries would be shifting blendstocks between various
distillate fuels in order to reduce the compliance costs associated with the proposed NRLM diesel
fuel sulfur standards. Instead, we projected the use of desulfurization techniques which would be
sufficient to meet the proposed 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 which currently produce non-highway distillate fuel to reduce
NRLM diesel fuel sulfur. The ability to shift blendstocks between fuels to reduce costs would be
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 would be 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 would be to shift these difficult to treat streams and portions of streams to heating oil,
which must 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
5-5
-------
Draft Regulatory Impact Analysis
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.
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, therefore, is 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 could potentially reduce their
diesel sulfur levels in 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 compliance with the 500 ppm sulfur cap
standard, and certainly fall short of the 15 ppm sulfur cap standard. Thus, changing to a sweeter
crude oil was not considered to be viable for complying with the proposed nonroad, locomotive
and marine diesel sulfur standards.
Another method to reduce diesel fuel sulfur, but much more significantly, is to chemically
remove sulfur from the hydrocarbon compounds which comprise diesel fuel. This is usually
accomplished through catalyzed reaction with hydrogen at moderate to high temperature and
pressure. A couple of specific examples of this process are hydrotreating and hydrocracking. A
modified version of hydrotreating which operates solely in the liquid state was announced
recently. Another process was announced recently which 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 pressure processes being developed which don't rely on hydrotreating,
such as biodesulfurization, and 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 those refinery streams which 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 which would
be required by the proposed 2007 500 ppm sulfur cap standard would occur through chemical
removal via conventional hydrotreating. For complying with the proposed 15 ppm cap standard
for nonroad diesel fuel which would be required in 2010, we expect that it would be met
5-6
-------
Fuel Standard Feasibility
primarily through adsorption and liquid phase hydrotreating, which are emerging advanced
desulfurization technologies. Thus, 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 which will complement the primary sulfur reduction achieved via
hydrotreating and adsorption. These other methods, such as FCC feed hydrotreating,
biodesulfurization, and chemical oxidation will be discussed following the primary discussion of
distillate hydrotreating, liquid phase hydrotreating and adsorption. Another means for aiding the
desulfurization of diesel fuel, particularly to comply with the 15 ppm cap 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 cap standard for any of the desulfurization
technologies, we provide a discussion of undercutting below in this subsection.
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 a number of purposes. For example, naphtha (gasoline like material which itself
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 which 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 in order to
improve the yield and quality of high value products, such as gasoline and distillate, from the
FCC unit. Refineries currently producing highway diesel fuel which must meet a 500 ppm cap
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 nonhighway distillate. EPA expects 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 of the
contaminants are removed and olefms and aromatics are saturated into paraffins and naphthenes.
Outside the U.S., this process is commonly used to produce distillate from heavier, less
marketable refinery streams. In the U.S. 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 U.S. to "upgrade" petroleum streams which are otherwise too heavy or too low in
5-7
-------
Draft Regulatory Impact Analysis
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 which encourages aromatic saturation is used in lieu of one
that simply encourages contaminant removal.
To meet the 500 ppm and the 15 ppm diesel fuel sulfur cap standards, EPA expects refiners
to focus as much as possible on sulfur removal. Other contaminants, such as nitrogen and
metals, are already sufficiently removed by existing refinery processes. While saturation of
aromatics generally improves cetane, the cetane numbers of todays nonroad, locomotive and
marine diesel fuels are already sufficient to comply with the ASTM standards which apply.
Thus, refiners would 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, and hydrotreating to remove sulfur should not degrade quality, except
possibly lubricity, as discussed in Section C. 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. Should a
refiner choose to do so, it would be to improve profitability,0 and not related to meeting the 15
ppm sulfur cap standard.
5.2.2.1 Fundamentals of Distillate Hydrotreating
Essentially all distillate hydrotreater designs follow the same broad format. Liquid distillate
fuel is heated to temperatures of 300-380°C, pumped to pressures of 500-700 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 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. Itself, this would cause a build up of hydrogen
sulfide and ammonia in the system, since it would have 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
c Refiners can choose to "upgrade" heavy refinery streams which 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.
5-8
-------
Fuel Standard Feasibility
higher levels of hydrogen sulfide and ammonia in the reactor, but it avoids the cost of building
and operating a scrubber.
Desulfurization processes in use today in the U.S. generally use only one reactor, due to the
need to only desulfurize diesel fuel to 500 ppm or slightly lower. However, a second reactor can
be used, particularly to meet lower sulfur levels. 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.
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, 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 is and lowest at the outlet where the hardest to remove
sulfur is. The 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 control (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 that 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 which 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 essentially flat in nature and the carbon atoms bound
5-9
-------
Draft Regulatory Impact Analysis
to the sulfur atom hinder the approach of the sulfur atom to the catalyst surface. Despite this,
today's catalysts are very effective in desulfurizing dibenzothiophenes, as long as only hydrogen
is attached to the carbon atoms bound directly to the sulfur atom.
However, distillate fuel can contain dibenzothiophenes which 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
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 SRLGO.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 6 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 will be discussed later below.
Because moderate sulfur reduction is often all that is required in current distillate
hydrotreating, catalysts have been developed which 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.
5-10
-------
Fuel Standard Feasibility
Other catalysts have been developed which 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 approach the catalyst surface much more easily. 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).
There are a number important issues which should be highlighted about using the
hydrogenation pathway for desulfurization. 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 which results in the desulfurization of a sterically hindered compound. If the
mono aromatics compounds are also saturated, there would only be 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, would improve significantly as
more of the aromatic compounds 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 can
be controlled temperature and pressure.
The vendors also point out a number of reasons why the cycle length of the catalysts which
catalyze hydrogenation reactions, which would 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 would have a cycle
length which is as long as or longer as 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 diesel fuel 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 would then allow the
5-11
-------
Draft Regulatory Impact Analysis
product to be sold as highway diesel fuel. The much higher sales price of the highway diesel fuel
compared to 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 around 15 ppm 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 which indicate 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 strongly inhibits desulfurization reactions, as will be discussed
further in the next section. It can also recombine with non-sulfur containing hydrocarbon
compounds at the end of the reactor or even in subsequent piping, essentially 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 SRGO and LCO
at 150 ppm sulfur and found that essentially all of 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 SRLGO 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 in 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
5-12
-------
Fuel Standard Feasibility
Using distillate hydrotreating to meet a 15 ppm sulfur cap on diesel fuel has been
commercially demonstrated, as will be discussed below. 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 would also be
producing locomotive and marine diesel fuel which would have to meet a 500 ppm cap and
heating oil which only needs to meet a 5000 ppm cap, which would have less stringent sulfur
requirements and could, 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 would face a more difficult task of
complying with a 15 ppm cap, than those processing primarily straight run distillate.
To understand the types of modifications which can be made 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.
Rate of = k x C" x Pffia + k x Cm x pffib
Desulfurization (1 + Kms x pms) (1 + KF x CF)
Per Catalyst
Surface Area
Where:
k, Kms and KF are various rate constants, which only vary 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 which
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
5-13
-------
Draft Regulatory Impact Analysis
aromatics inhibition leads them to favor CoMo catalysts even for desulfurizing feeds with a high
concentration of sterically hindered compounds.
These relationships essentially identify the types of changes which could be made to 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 U.S. 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 year 2000.16
KF 752 can be considered to be typical of an Akzo Nobel catalyst of the 1992-93 timeframe,
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 U Active Reaction .Sites. Type II refers to a specific kind of catalyst site
which is particular 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, would produce 405, 270 and 160ppm sulfur with KF 756, KF757, and
KF 842, respectively.
In 2001, Akzo Nobel announced a new highly active catalyst named Nebula which offers a
different way in which coatings are used for catalysts. A typical catalyst is composed of two
parts: an active coating which contains 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
5-14
-------
Fuel Standard Feasibility
sulfur standard by simply replacing its existing catalyst with Nebula and providing significantly
more hydrogen (which may possibly require the addition of a hydrogen plant). Nebula is a new
catalyst that could avoid some or much of the capital investment that would otherwise be
required for meeting the 15 ppm sulfur standard.
Haldor-Topsoe has also developed a more active catalyst. 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
SRLGO, 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.
UOP projects a similar reduction in sulfur due to an improved catalyst. They estimate that a
hydrotreater producing 500 ppm sulfur distillate today (20% LCO, 10% light coker gas oil) could
produce 280 ppm sulfur distillate with a 50 percent more active catalyst.18
Over the last four years, Criterion Catalyst Company announced two new lines of catalysts.
One is called Century, and the other is called Centinel.19 These two lines of catalysts are reported
to be 45 - 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 through
better dispersion of the active metal on the catalyst substrate.
Another catalyst vendor shared some information about its catalyst development program
which involves advances in the geometry of its substrate. These advances have resulted in
significant improvements in the contact of diesel fuel with the catalyst. The vendor also shared
that it is combining its substrate technology with other reactor enhancements to further increase
the contact between diesel fuel and the catalyst and hydrogen. Preliminary tests suggest that this
combination could improve the catalyst activity by at least a minimum of two.
Thus, by itself, changing to a more active catalyst 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 4 years. Vendors have informed
EPA that the cost of these advanced catalysts is very modest relative to less active catalysts. This
will help to reduce the reactor size needed.
The second type of improvement which can be made to improve 500 ppm hydrotreating 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 recycle hydrogen can reduce sulfur levels from
roughly 285 to 180 ppm in an existing hydrotreater.
5-15
-------
Draft Regulatory Impact Analysis
The third type of improvement which can be made to current distillate hydrotreating is to
improve vapor-liquid contact. Akzo Nobel estimates that an improved vapor-liquid distributor
can reduce the temperature necessary to meet a 50 ppm sulfur level by 10 °C, which in turn
would 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 would lower the
temperature needed to achieve the same sulfur removal rate by eight to nine °C. Or 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 project 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 SRLGO and LCO with its advanced TK-574
CoMo catalyst will reduce sulfur from 120 ppm to 40 ppm.20 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 six 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 only decreases the necessary reactor
temperature by six to eight °C. Or, temperature can be maintained and the final sulfur level
reduced by 35-45 percent.
Seventh, reactor volume can be increased. UOP projects that doubling reactor volume would
reduce sulfur from 120 to 30 ppm.
5-16
-------
Fuel Standard Feasibility
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 are probably indicative of improvements which can be made in many cases, it is
not likely that all of the improvements mentioned are applicable to any one unit; the degree of
improvement could either be greater than, or less than the benefits that are indicated.
Therefore, many refiners may have to implement one additional technical change listed by
UOP to be able to meet the 15 ppm cap 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 a higher pressure to be used in the new second
reactor. 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 of
these units are already operating in refineries but may have to be expanded or enlarged.
Overall, Akzo-Nobel projects that current hydrotreaters can be modified short of a revamp to
achieve 50 ppm sulfur. Acreon/IFP/Procatalyse is less optimistic, believing that more than a
catalyst change will be necessary to meet this sulfur level.21 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.22 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 reductions needed to meet a 15 ppm cap
would require additional measures. To assess the degree that these measures would be needed, it
is useful to examine the commercial and pilot plant performance of distillate hydrotreaters 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 which indicate
the capability of various hydrotreating technologies to reduce distillate sulfur levels to very low
levels. While many reports of existing commercial operations are available which 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., 10-50 ppm) are associated with also reducing aromatics
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 cap standard. Other
European diesel fuel must also meet a 10 ppm sulfur cap standard.
5-17
-------
Draft Regulatory Impact Analysis
Another study projected the technology and resulting cost to reduce sulfur for EPA, the
Engine Manufacturers Association retained Mathpro for this study. The projections of this study
will be discussed in the next chapter on the economic impacts of this rule. 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 cap standard was
established for highway diesel fuel in the U.S. 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 advancements by
catalyst manufacturers demonstrating the feasibility of producing diesel fuel which meets these
levels through pilot plant testing and some commercial demonstrations should be considered a
first-generation of technology, with new and continual advancements expected over time.
Starting with SynTechnology, as of August 2, 1999, there were 24 units either in operation or
in the process of being constructed. The purpose for each of those units ranged from
desulfurization to desulfurization plus dearomatization to mild hydrocracking. Of particular
interest here is a revamp of an existing two reactor distillate hydrotreater at the Lyondell-Citgo
refinery in Texas. The revamped unit was designed to process a low-cost feed very heavily
weighted towards cracked material (65-70 percent LCO and LCGO). One existing reactor was
converted to SynSat Technology, while the other was used simply as a flash drum. A new first-
stage reactor was added. Both reactors were designed to operate in a co-current fashion. Pilot
plant studies predicted average sulfur and aromatics levels of seven ppm and 31 volume percent,
respectively, based on feed sulfur and aromatics levels of 11,900 ppm and 53 volume percent,
respectively. The unit exceeded expectations in the case of sulfur, producing an average sulfur
level of less than five ppm from a feed sulfur level of 13,800 ppm. The actual aromatic level
achieved was above the target by four volume percent, but the feed aromatic level was five
volume percent higher than expected. Thus, the net reduction in aromatic content in terms of
volume percent was still higher than found in the pilot plant. ABB Lummus and Criterion
indicate that their catalyst technology is sufficiently flexible to focus on the deep desulfurization
with or without the significant aromatics reduction seen here. This is reflected in their projection
of the technology needed to meet a 15 ppm sulfur cap which is discussed in the next chapter.
While this two-stage unit initially produced less than 5 ppm product, it does not do so
consistently. The primary purpose of the unit is to increase cetane so that the product can be
blended directly into diesel fuel. The primary sulfur reduction requirement is to protect the noble
metal catalyst in the second stage reactor. This generally requires that the product from the first
stage be less than 50 ppm. Thus, if the cetane specification is being met at less severe sulfur
reduction conditions, there is no incentive to reduce sulfur any further than necessary for catalyst
protection. In addition, the unit is seeing a heavier feedstock than designed, and the
desulfurization reactor is being operated at a lower temperature than designed to increase the
cycle lengths.
IFF, in conjunction with various catalyst manufacturers, offers its Prime D technology for
deep desulfurization, aromatics saturation and cetane improvement.23 Using a NiMo catalyst,
IFP's Prime D process can produce distillate sulfur levels of 10 ppm from SRLGO and of less
5-18
-------
Fuel Standard Feasibility
than 20 ppm from distillate containing 20-100 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 ASAT
catalyst, which uses platinum and palladium.24 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, essentially eliminate nitrogen and reduce aromatics to two to five volume percent.
Hydrogen consumption was 800-971 standard cubic feet per barrel (SCFB).
Akzo Nobel recently presented a summary of the commercial experience of about a years
worth of operations of their STARS catalyst for desulfurizing diesel fuel at the BP-Amoco
refinery in Grangemouth, UK.25 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-loaded0 into the reactor
to produce 45,000 barrels per day diesel fuel at 10 - 20 ppm (to meet the 50 ppm cap standard).
From the data, it was clear to see that as the space velocity changed, the sulfur level changed
inversely 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.
These studies indicate the commercial feasibility of producing diesel fuel with 10 ppm or less
sulfur. The primary issue remaining is to commercially demonstrate that the 15 ppm cap
standard can be met using the desulfurization/hydrogenation method without saturating much of
the aromatics in diesel fuel, especially with a feedstock blend which contains a substantial
amount of cracked material. The ease or difficulty 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 off-highway diesel fuel and heating oil. However, our desulfurization
technology feasibility analysis did not considered the occurrence of feedstock shifting as
necessary for refiners to meet the proposed diesel sulfur standards.
5.2.3 Phillips S-Zorb Sulfur Adsorption
A prospective diesel desulfurization process was announced by Phillips Petroleum in late
2001.26 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
that instead of reacting the sulfur with hydrogen over a catalyst to form H2S, the S-Zorb process
adsorbs the sulfur molecule onto a sorbent for later removal. At a pressure of 275 - 500 pounds
per square inch gauge (psig) and at a temperature of 700 - 800 Fahrenheit and in the presence of
D Dense loading is a process of packing a certain volume of catalyst into a smaller space than conventional
catalyst loading.
5-19
-------
Draft Regulatory Impact Analysis
hydrogen in the S-Zorb reactor, the sulfur atom of the sulfur-containing compounds adsorbs onto
the sorbant. The catalyst activity of the sorbent next cleaves the sulfur atom from the sulfur-
containing hydrocarbon. To prevent the accumulation of sulfur on the catalyst, the catalyst is
continually removed from the reactor. From the reactor, the sorbent is moved over to a receiving
vessel by an inert lift gas, which likely is nitrogen, and then in the receiving vessel the lift gas
and the entrained diesel fuel is removed from the sorbent. The sorbent next drops down into a
lockhopper which 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 4-5 years between shutdowns. Because untreated distillate can
contain several percent sulfur, Phillips believes that its S-Zorb process for diesel could get
overwhelmed by the amount of sulfur which is adsorbing onto the catalyst. However, some
refiners who run sweet crudes and produced low sulfur non-highway diesel volumes (from
straight run diesel and hydrocrackate diesel) may have lower uncontrolled nonroad sulfur levels.
These refiners may be able to use the S-Zorb process to lower their nonroad diesel sulfur. Thus,
the S-Zorb process may not be able to economically treat all untreated distillate streams which
are high in sulfur, and would be best suited to treat distillate containing 500 ppm sulfur or less.
Phillips has been involved in sorbent technology at least as far back as 1980 which is when
the company filed a patent application for sorbent technology. However, it seems apparent that
Phillips did not develop its S-Zorb technology until much later as it filed for a patent for a
technology for circulating sorbent during March of 1997. The purpose of the March, 1997 patent
was to remove hydrogen sulfide. The technical focus of that patent was both the sorbent
chemistry and the sorbent handling technology. The catalyst content was specified to be alumina,
silica, zinc oxide and metal oxide (probably in the form of nickel oxide) and the sorbent size was
specified to be in the range of 20 to 500 micrometers. Then in August of 1999, Phillips filed a
patent for using its sorbent recirculating technology to desulfurize cracked gasoline and diesel
fuels. The sorbent change as specified in the patent was to add substantially reduced valence
nickel to enable the removal of sulfur from the targeted refinery streams. Then Phillips filed a
patent in May of 2001 for an improved catalyst to desulfurize cracked gasoline and diesel fuel.
The change was to add a calcium compound which increases the porosity of the sorbent and
increases the resistance to deactivation. The latest patent also listed the potential candidates for
metal promotors to include cobalt, nickel, iron, manganese, copper, molybdenum, silver,
tungsten tin and vanadium, or mixtures of these metal oxides, with valences of 2 or less.
Phillips' S-Zorb diesel desulfurization process has been demonstrated in a pilot plant which
started up in early 2002. This pilot plant has provided Phillips data on how the unit would
operate processing varying formulations of diesel fuel or diesel fuel blendstocks. The pilot plant
testing data which has been released by Phillips has shown that diesel fuels blended with LCO
can be desulfurized down 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. These testing results are summarized in chapter 7 where we use the data to
5-20
-------
Fuel Standard Feasibility
develop the inputs for our refinery cost model. Phillips is constructing a commercial unit to
demonstrate their S-Zorb diesel desulfurization unit. Phillips is completing the engineering
phase of this project to design this unit and is expected to begin the construction for an estimated
start-up date during early 2004.
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 two 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 upcoming Tier 2
gasoline sulfur program. Starting in 2004 many refiners will need to 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. Phillips shared with us in late 2002 that they have licensed their S-Zorb
for gasoline processing for installation in 9 refineries. That the Borger S-Zorb gasoline
desulfurization unit has operated as designed and that there are 9 new S-Zorb gasoline units
planned to start up demonstrates that there is agreement within the refining industry that the S-
Zorb process works.
Much of the refining industry's trust with the S-Zorb gasoline desulfurization unit is likely to
apply to S-Zorb for diesel fuel as well. First, the sorbent has been shown to be effective at
adsorbing sulfur, releasing the sulfur when it is burned and then at being regenerated for reuse.
Also, the S-Zorb unit has been shown to be able to move the sorbent out of the reactor into a
number of different vessels for handling and treatment and then recycling back to the reactor.
The part of diesel fuel desulfurization which cannot be demonstrated with the S-Zorb gasoline
desulfurization unit is how effectively the sorbent would be able to adsorb and cleave the sulfur
molecule from the sulfur-containing molecules of diesel fuel. However, that part of the S-Zorb
diesel fuel desulfurization unit should be able to be demonstrated with testing in the pilot plant.
Phillips can even test the diesel fuels from specific refineries in their pilot plant to help design
the S-Zorb unit for those refineries. Thus, Phillips is marketing its diesel fuel desulfurization
unit even before their diesel fuel desulfurization commercial demonstration unit has started up.
Most refiners, however, are very conservative and would not be willing to only rely on pilot
plant testing for a unit which would likely cost tens of millions of dollars, and without its proper
operation they might not be able to operate. Thus, they would want to see a particular technology
operating as a commercial unit for a significant period of time, such as two years, before trusting
that the technology is reliable.
Since the process has never been demonstrated commercially on diesel fuel, the commercial
demonstration unit will go a long way toward proving to refiners that the Phillips process works
as designed. In particular, the sulfur compounds in diesel fuel are different, usually more
refractory, than those in gasoline. Phillips reports, though, that the absorption catalyst more
readily desulfurizes the sterically hindered sulfur compounds than the thiophenes (single ring
compounds which contain sulfur) which must be desulfurized in gasoline. This suggests the
5-21
-------
Draft Regulatory Impact Analysis
possibility that S-Zorb for diesel may actually desulfurize diesel fuel more easily than gasoline.
Phillips projects that they will have an S-Zorb diesel desulfurization commercial unit up and
running during the first quarter of 2004. After hearing Phillips' timeline for developing the S-
Zorb diesel desulfurization process, and understanding the scrutiny by refiners for new
desulfurization processes, it seems that refiners may consider this process too risky for
complying with the 500 ppm cap in 2007, especially since the demonstration unit will be a
revamp of an existing hydrotreating unit for producing 15 ppm diesel fuel. However, after seeing
the commercial unit producing 15 ppm sulfur for the Highway Program for several years, we
believe that this process will be a serious contender for 2010 for nonroad.
5.2.4 Linde Isotherming
A professor at the University of Arkansas has applied some ingenuity in reaction chemistry to
diesel desulfurization. After conceiving of this process, he started a company named Process
Dynamics and then took the new technology to Linde for development and eventual licensing.
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
only needs to be designed to handle a liquid, instead of the two phase reactors designed for
conventional hydrotreating. Since 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 be
desulfurized several times as it is being treated.
The Linde Isotherming process has a couple 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 Linde, the kinetics of
conventional hydrotreating are mass transfer limited which is the rate at which gaseous hydrogen
can transfer into the liquid phase. Linde makes this point by the following reaction equations.
For conventional hydrotreating the following two equations apply:
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.
P*H2 = Partial pressure of hydrogen at the catalyst.
and
5-22
-------
Fuel Standard Feasibility
rs = ks T[S][PXH2] (rate of desulfurization at the catalyst site)
Where:
rs = rate of reaction of sulfur.
ks = reaction rate constant for sulfur removal.
P*H2 = partial pressure of hydrogen at the catalyst.
T = temperature in degrees absolute.
[S] = concentration of sulfur.
If the desulfurization rate of reaction (rs) is much slower than the rate at which hydrogen can
dissolve into diesel fuel (rg), then there would probably not be any benefit for the Linde
Isotherming process. However, according to Linde, the rate of reaction for desulfurization is
faster than the rate of mass transfer, thus, the rate of reaction for diesel hydrotreating is limited by
the mass transfer of hydrogen into diesel fuel. Thus, the Linde process increases the rate of
reaction by dissolving the hydrogen needed for the reaction into the liquid phase prior to sending
this liquid to the reactor. The faster rate of reaction is indicated by the fact that the Linde
desulfurization process which can desulfurize a unhydrotreated distillate comprised of a typical
mix of distillate blendstocks down to about 500 ppm at a space velocity of 8 hour"1. Conversely,
conventional hydrotreating requires a space velocity of about 2 hour"1 to accomplish the same
task.
There are a two important benefits to the Linde process because it has a higher space velocity.
One benefit is that the amount of catalyst needed for the Linde process is lower. 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 Linde Isotherming
deslfurization process. The lower catalyst volume required by Linde Isotherming costs
proportionally less because the Linde 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 to larger reactors.
The Linde 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 Linde 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 no external heating
5-23
-------
Draft Regulatory Impact Analysis
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 Linde 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 it remains in solution, there is no residual hydrogen gas to recycle.
Compressors operating at the pressures that diesel fuel desulfurization occurs are expensive,
long time delivery items. Thus, by omitting the recycle gas compressor and using smaller
reactors, the Linde desulfurization process saves substantial capital costs compared to
conventional hydrotreating which likely 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 Linde process requires a smaller footprint compared to conventional hydrotreating
While aspects of the Linde Isotherming desulfurization process for diesel fuel desulfurization
are novel compared to conventional diesel desulfurization, many aspects of the process are the
same. Much of the list of required equipment is the same for the Linde process as for
conventional hydrotreating. Table 5.2-1 shows both the similarities and differences between the
two.
Table 5.2-1
Major Equipment Needed for Linde 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
Linde 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
Linde has accumulated some data on the Isotherming desulfurization process from testing
which they have done with their pilot plant. Linde started up a pilot plant in late 2001. Recently,
Linde 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
5-24
-------
Fuel Standard Feasibility
fuel 15 ppm sulfur cap standard which begins in mid 2006. The unit was started up in September
of 2002 and the Linde engineers have been working with the refinery engineers to optimize the
unit for the refinery.
5.2.5 Chemical Oxidation and Extraction
Another desulfurization technology based on chemical oxidation is being pursued by Unipure
and Petrostar.27 For theses 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 catalyzed peroxyacetic acid. With an oxygen atom attached
to the sulfur atom, the sulfur-containing hydrocarbon molecules becomes polar and hydrophilic
and then move into the aqueous phase. These sulfone compounds can either be desulfurized or
be converted to a surfactant which 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.
Petrostar has a bench scale pilot plant up and running and they intend to demonstrate their
technology next with a commercial demonstration unit. Unipure is in the process of setting up a
pilot plant and it is expected to be up and running by mid 2003.
Early in 2003, Lyondell 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 process is similar in that
the process uses oxidation compounds to oxidize the sulfur compounds to sulfones and then
removes relies on extraction to remove the sulfones. 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).
The oxidant is degraded t-butyl alcohol during the conversion of sulfur species to sulfones. The
T-butyl alcohol by product can be converted to MTBE or isooctane or used as fuel in the refinery.
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 demonstrated pilot
plant success desulfurizing 500 ppm diesel fuel to below 10 ppm and plans to have a commercial
plant operable by late 2003 to early 2004. Lyondell's technology was announced too late be
incorporated into this analysis but we intend to stay abreast as they develop this new technology .
5.2.6 FCC Feed Hydrotreating
At the beginning of this section, it was mentioned that sulfur could be removed from
distillate material early or late in the refining process. Early in the process, the most practical
place to remove sulfur is prior to the FCC unit. The FCC unit primarily produces gasoline, but it
also produces a significant quantity of LCO.
Many refineries already have an FCC feed hydrotreating unit. The LCO from these refineries
should contain a much lower concentration of sterically hindered compounds than refineries not
5-25
-------
Draft Regulatory Impact Analysis
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. 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 cap standards. Thus, this study will rely
on distillate hydrotreating as the primary means with which refiners would meet the 15 ppm
sulfur cap. For those refiners who would choose FCC feed hydrotreating, their costs would be
presumably lower than distillate hydrotreating and the costs estimated in the next chapter can
then be considered to be somewhat conservative in this respect.
5.3 Feasibility of Producing 500 ppm Sulfur Nonroad 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
proposed 2007 sulfur cap standard, and also to estimate the cost of complying with the proposed
sulfur standard (see Chapter 7), we needed to project the mix of technologies that would be
available and used for compliance. We considered several different factors for projecting the
mix of technologies which would be used. First and foremost, we considered the time which
refiners will have to choose a new technology which is important because of the relatively short
timeline allowed for compliance. Second, we considered whether the technology would 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 to other technologies. If a refiner
finds that technology is lower cost than another, it is more likely to use that technology. We also
considered whether the technology is available from an industry-trusted vendor which has proven
itself to the industry by providing other successful refining technologies and particularly if the
vendor has proven itself in the U.S. 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 proposed 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. Since this rulemaking is expected to be promulgated sometime in 2004, this would
provide refiners three full years to comply with the 500 ppm sulfur standard if it is promulgated
5-26
-------
Fuel Standard Feasibility
as predicted. Because refiners need about 3 years to complete the mentioned steps to have a
working new unit, there is little time to shop around for a new desulfurization technology which
is just beginning to prove itself. A thorough review of a newer technology can take months, thus
if refiners do not have this extra time, they will tend toward the technologies with which they are
familiar. See the next section for a more detailed discussion about the leadtime issues for the
2007 standard.
Of the various technologies which 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 cap standard. In the U.S.
there are about 90 distillate hydrotreaters with virtually all of them being conventional
hydrotreaters operating since 1993 or before. The one exception is a Linde Isotherming
commercial demonstration unit which started up recently at a Giant refinery in New Mexico.
Phillips S-Zorb, the two oxidation and extraction and biodesulfurization technologies are all
being demonstrated by pilot plants only. Of those being demonstrated by a pilot plant, Phillips is
expected to start a highway diesel fuel commercial demonstration unit in early 2004. 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^ In 2004,
biodesulfurization, oxidation and extraction are not expected to have units operating at all.
Phillips may have a diesel fuel desulfurization unit operating by then, but certainly it will not be
operating for two years. While Phillips has a gasoline desulfurization operating now, refiners
may be skeptical that it truly demonstrates the technology for diesel desulfurization. The Linde
desulfurization unit which 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 lack of an excess of leadtime is the
central issue of whether refiners will choose between conventional hydrotreating and other
advanced desulfurization technologies for 2007. Refiners would not have the luxury of many
months needed to carefully consider the advanced technologies which are still in the
development and demonstration stage, so we believe that this issue is the most critical affecting
refiners choice of desulfurization technologies for 2007. For these reasons, we believe that
refiners would default to what they know will work, which is conventional desulfurization.
Since there are multiple vendors which can provide the preliminary engineering design and any
followup support for conventional hydrotreating, these vendors would not be overly taxed and
would be able to serve the refiners which will need to put in desulfurization units for 2007.
5.3.2 Leadtime Evaluation
Refiners must have sufficient leadtime to design, construct and start up desulfurization
technology to meet the 500 ppm standard if this standard is to be implemented smoothly and
E Refiners want low maintenance refining units because they have cut back their engineering staff to reduce
their refining costs to improve their margins and thus will seek units which are consistent with that strategy.
5-27
-------
Draft Regulatory Impact Analysis
without undue economic impacts. If one or more refiners were unable to comply on time, it
would have major repercussions for the refiner and potentially for the national fuel supply. If
some refiners who were planning on producing 500 ppm NRLM fuel could not do so on time
and could not buy credits, they would have to sell their high sulfur distillate as heating oil, export
this distillate or temporarily cease production. As discussed below in Section 5.8, heating oil
would 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 obviously 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 in June 1,
2007. Thus, providing adequate leadtime for refiners to design, construct and prove out the
necessary new hydrotreaters is critical to avoiding serious economic harm to both the refining
and NRLM industries.
We project that refiners would use conventional hydrotreating to meet the 500 ppm standard
beginning on June 1, 2007. Of the 42 refineries projected to produce 500 ppm NRLM diesel fuel
beginning in 2007, 13 are projected to do so with either no or minor modifications to their
highway diesel fuel hydrotreaters. These refineries produce a relatively small volume of non-
highway diesel fuel compared to their highway production. The remaining 29 refineries would
need to design and construct a new hydrotreater to produce 500 ppm NRLM fuel.F This is
roughly 20% of all U.S. refineries producing transportation fuels today. 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.
EPA has conducted two leadtime assessments for the refining industry in the past 4 years.
One was conducted for the Tier 2 gasoline sulfur program.0 The other was conducted as part of
our review of progress being made towards compliance with the 15 ppm sulfur, highway diesel
fuel program.11 The results of both of these assessments are reviewed below and then applied to
the proposed NRLM sulfur control program.
5.3.2.1 Tier 2 Gasoline Sulfur Program
Chapter IV of the Final RIA for the Tier 2 gasoline sulfur program presented the following
table which contains the results of its leadtime assessment
F Without the proposed small refiner provisions, an additional 20 refineries would have to produce 500 ppm
NRLM fuel by June 1, 2007, including 19 refineries owned by small refiners.
G 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.
H "Highway Diesel Progress Review," U.S., EPA, June 2002, EPA420-R-02-016.
5-28
-------
Fuel Standard Feasibility
Table 5.3-1
Leadtime 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
Timea
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 Timea
0.5
1.0-1.25
1.25-2.0
1.5-2.25
2.5-3.5
2.75-3.75
a Several of the steps shown can overlap.
b Projected to begin before Tier 2 gasoline final rule.
This table contains leadtime projections for two distinctly different approaches to gasoline
sulfur control. The first, naphtha hydtrotreating, is more closely related to conventional distillate
hydrotreating. In fact, a number of 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 would likely have been 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 diesel fuel. The Tier 2 gasoline sulfur standards are fully phased in for most
refiners by 2006. Thus, it is highly unlikely that a refiner would just begin considering FCC feed
hydrotreating as the result of this NRLM rule. We will therefore only focus on the portion of the
table which addresses the leadtime 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 will be examined in detail. However, it would be useful first to review EPA's
leadtime projections related to the 15 ppm highway diesel fuel cap.
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 leadtime required for each individual step of the process. That rule provided 5.5 years of
leadtime between promulgation and initial implementation. This amount of leadtime
5-29
-------
Draft Regulatory Impact Analysis
significantly exceeded that considered necessary to design and construct desulfurization
equipment. This amount of leadtime 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.
EPA 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 which
refiners had already completed and the length of time needed for those still remaining. The tasks
considered were essentially 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 that the refiner was going to
continue marketing highway diesel fuel and if so, whether he 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
essentially all gasoline sold in the U.S. meets highway quality standards, and refiners have no
other market for their gasoline feedstocks.
The results of the leadtime review are presented in the table below.
Table 5.3-2
Leadtime Assessment: Progress Review of 15 ppm Highway Diesel Fuel Cap
Project Stage
Planning and Front End Engineering3
Detailed Engineering and Permits
Procurement and Construction
Commissioning and Start-Up
Time Allotted
0 9S-? vpnrs
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
a 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 leadtimes needed to fabricate specialized
equipment.
Because the progress review was conducted over a year after the rule was promulgated, EPA
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
5-30
-------
Fuel Standard Feasibility
determine if they had indeed begun those steps which should have been started by the date of the
assessment.
5.3.2.3 Leadtime Projections for Production of 500 ppm NRLM Diesel Fuel
We utilized the information for gasoline and highway diesel analyses to project the leadtime
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
would 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 would be too small in their area to support their entire
production of high sulfur fuel. For those with a real choice, this step would 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 would 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, 3-6 months to arrive at
a corporate decision to participate 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 which supply
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, a refiner would likely send samples of their various distillate streams to the firms
marketing desulfurization technology to determine how well each technology removed 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 0.5-. 1.0 years would be required to evaluate the
various naphtha desulfurization technologies which were or soon to be available. This extensive
period of time was deemed 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 have 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
prior to investing significant amounts of capital in these technologies. We believed that at a
minimum, refiners should have 6 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 which
was optimum for their situation. Thus, there was really no upper limit to the amount of time for
this step.
5-31
-------
Draft Regulatory Impact Analysis
The scoping and screening task which refiners would face with respect to the 500 ppm
NRLM sulfur cap is both different and similar to the situation refiners faced with the Tier 2
gasoline program. The NRLM program would differ, because refiners had to choose between a
wide variety of gasoline desulfurization technologies to comply with the Tier 2 sulfur standards.
In contrast, above, we project that conventional hydrotreating would likely be the dominant
choice for desulfurizing diesel fuel to 500 ppm in 2007. The similarity would exist, because
refiners would have to consider how they are going to comply with the 15 ppm nonroad diesel
fuel cap in 2010 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 years, we project that the
refiners building new conventional hydrotreating units for 2007 would plan these units to be
easily revamped in 2010 to produce 15 ppm nonroad diesel fuel. Therefore, the specific
conventional hydrotreating design selected for 2007 would have to mesh with their plan for 2010.
At minimum, this would involve the 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. The selection of operating pressure is likely the most time critical,
because of the long lead times involved in procuring pressure vessels. Some time for vendors to
assess the performance of their 15 ppm technologies via pilot plants testing on specific refiners'
diesel fuel samples should be included.
Fortunately, this process has been underway for some time involving refiners' highway diesel
fuels. By mid-2004, this process should be essentially complete. 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 not only involves 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 3-6 months for strategic planning and the 3-6 months for
scoping and screening can go on concurrently.
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 olefm saturation. Octane
is not an issue with distillate hydrotreating. In general, the design of a grassroots distillate
hydrotreater would be more complex than that of a revamp. However, here the revamp is to
produce 15 ppm diesel fuel, a much more challenging task than producing 500 ppm diesel fuel.
5-32
-------
Fuel Standard Feasibility
Thus, 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 0.75-1.0
year, as the choice of distillate hydrotreating is clear.
Regarding permitting, EPA has taken a number of steps to help state/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 3-9 months would be a more realistic range, as opposed to the 3-12 months which was
projected in the Tier 2 Final RIA. There, we identified the 12 month period as being a worse
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. Fewer firms fabricate higher pressure reactors and compressors. 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
6-9 months to design and 12-15 months to construct a distillate hydrotreater. Cumulatively, the
two steps would take 1-1.25 years from the time at which permits were obtained.
This range is about 3 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. This estimate was reasonable in the review of
progress being made towards compliance with the 2007 highway diesel fuel rule due to the
extensive amount of leadtime still available. For this to be typical, all refiners planning to
produce 15 ppm highway diesel fuel would already be constructing their new or revamped
hydrotreaters. Clearly this is not the case, yet refiners consider 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 3
months for start up for naphtha, FCC feed and highway diesel fuel hydrotreater s. 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.
5-33
-------
Draft Regulatory Impact Analysis
Table 5.3-3
Leadtime 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
Assuming that the final rule is signed in April of 2004, this analysis indicates that some
refiners should be able to produce 500 ppm NRLM fuel as early as July 2006. This coincides
with the 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 the implementation of the 500 ppm
NRLM cap. Should any refiners be 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 proposed 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 10 years of experience producing 500 ppm diesel fuel
commercially. This should only shorten the time required to prepare for the standard relative to
1993. The 500 ppm highway diesel rulemaking was promulgated in August 1990 and took effect
on October 1, 1993.28 Thus, that rulemaking provided 3 years and two months of leadtime,
nearly identical to that provided by the NRLM proposal. Some price spikes occurred with the
implementation of 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 U.S. was one of the first
nation's to require 500 ppm diesel fuel, so little commercial experience was available upon
which to base designs. Today, refiners and technology vendors have over 10 years of
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 that being affected by today's proposed
rule, causing greater stress on the engineering and construction industries than today's proposed
program would cause.
5-34
-------
Fuel Standard Feasibility
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, construciton 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 leadtime provided by today's proposal would be sufficient.
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 which
included assessing the time required for small refiners to obtain loans for capital investments.
The simple survey revealed that small refiners would 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.29 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 that the contractors which design and install refinery
processing units will likely focus first on completing the more expensive upgrade projects for
large refiners. Thus, the design and construction of desulfurization hardware for the small refiner
would take longer as well. The three additional years being provided small refiners should be
sufficient to compensate for these factors. This additional leadtime should provide not only
enough time for these small refiners to construct equipment, but to also allow these refiners more
time 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 Distributing 500 ppm Sulfur Non-Highway Diesel Fuel in
2007 and 500 ppm Locomotive and Marine Diesel Fuel in 2010
There are two considerations with respect to the feasibility of distributing non-highway diesel
fuels meeting the proposed 500 ppm sulfur standard. The first pertains to whether sulfur
contamination can be adequately managed throughout the distribution system so that fuel
delivered to the end-user does not exceed the specified 500 ppm maximum sulfur concentration.
The second pertains to the physical limitations of the system to accommodate any additional
segregation of product grades. These considerations are evaluated in the following Sections 5.4.3
and 5.4.4 of this Draft RIA.
5-35
-------
Draft Regulatory Impact Analysis
5.4.1 The Diesel Fuel Distribution System Prior to the Implementation of the Proposed
500 ppm Sulfur Program:
Prior to 1993, most number 2 distillate fuel was produced to essentially 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 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. Based on the most recent EIA data, roughly 15 percent of highway fuel is
spillover, representing nearly a third of non-highway consumption.
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 would be present
in 40 percent of the fungible fuel distribution system including the Northeast, parts of the
Midwest and in and adjacent to the concentration of refineries in PADD 3.
5.4.2 Summary of the Proposed 500 ppm Sulfur Standards
The proposed sulfur standards generally cover all the diesel fuel that is used in mobile
applications but is not already covered by the previous standards for highway diesel fuel. This
fuel is defined primarily by the type of engine which it is used to power, nonroad, locomotive,
and marine diesel engines. In shorthand, this fuel will be referred to as NRLM fuel.
NRLM fuels typically include:
1) Any number 1 and 2 distillate fuels used in or intended to be used in land-based nonroad,
locomotive or marine diesel engines,
2) Any number 1 distillate fuel (e.g., kerosene) added to such number 2 diesel fuel, e.g., to
improve its cold flow properties and
5-36
-------
Fuel Standard Feasibility
3) Any other fuel used in or blended with diesel fuel for use in nonroad, locomotive, or
marine diesel engines that has comparable chemical and physical characteristics.
The proposed sulfur standards would not apply to:
1) Number 1 distillate fuel used to power jet aircraft,
2) Number 1 or number 2 distillate fuel used for other purposes, such as to power stationary
diesel engines or for heating,
3) Number 4 and 6 fuels (e.g., bunker or residual fuels, IFO Heavy Fuel Oil Grades 30 and
higher, ASTM DMB and DMC fuels), and
4) Any fuel used to power equipment for which a national security exemption has been
approved.
As in the recent highway diesel rule, in those cases where the same batch of kerosene is
distributed for two purposes (e.g., as kerosene to be used for heating and to improve the cold
flow of number 2 nonroad diesel fuel), that batch of fuel would have to meet the standards being
proposed today for nonroad diesel fuel. An alternative compliance approach would be to
produce and distribute two distinct kerosene fuels. Under such an approach, one batch would
meet the proposed sulfur standards and could be blended into number 2 NRLM diesel fuel. The
other batch would only have to meet any applicable specifications for heating fuel.
We are proposing to restrict the sulfur content of NRLM fuel nationwide to no more than 500
ppm beginning in 2007. These provisions mirror controls on highway diesel fuel to 500 ppm in
1993.30 Refiners and importers could comply with the proposed requirement by either producing
NRLM fuel at or below 500 ppm, and/or by obtaining credits under the proposed averaging
banking and trading (ABT) provisions. The 2007 deadline for meeting the proposed 500 ppm
NRLM sulfur standard would not apply to refineries covered by special hardship provisions for
small refiners. In addition, a different schedule might apply for any refineries that might be
approved under the proposed general hardship provisions.
We are proposing that high sulfur NRLM diesel fuel which remains after June 1, 2007 due to
the small refiner and fuel ABT provisions could be commingled with 500 ppm NRLM diesel fuel
after it has been dyed to the IRS specifications until June 1, 2010. Thus, at some points in the
distribution system, NRLM fuel higher than the 500 ppm standard would remain until it is
precluded from production beginning June 1, 2010. The proposed 15 ppm sulfur standard for
nonroad diesel fuel would take effect in June 1, 2010.
Under the proposed 500 ppm NRLM program, heating oil would be allowed to have its sulfur
level remain uncontrolled; limited only by various state regulations. Thus, while NRLM is
commonly distributed today with heating oil, after implementation of the proposed sulfur
standards, these two grades of fuel would have to be distributed separately. To ensure that high-
sulfur diesel fuel manufactured for the heating oil market would not be used in nonroad,
locomotive, or marine applications, we are proposing that heating oil be injected with a fuel
marker before it leaves refinery. After June 1, 2010, the fuel standards situation is simplified
5-37
-------
Draft Regulatory Impact Analysis
considerably and the fuel program structure can therefore also be simplified. We are proposing
that after June 1, 2010 high sulfur diesel fuel no longer be permitted to be used in any NRLM
equipment. The only high sulfur distillate remaining in the market should be heating oil.
Heating oil would have to be kept segregated and preventing its use in NRLM equipment could
be enforced on the basis of sulfur level, avoiding the need for a unique marker to be added to
heating oil. Thus, we are proposing that the marker requirement for heating oil expire after June
1,2010.
After June 1, 2010, we are proposing that locomotive and marine diesel fuel would be
allowed to remain at the 500 ppm level. Under the proposed small refiner hardship and fuel
credit provisions, we would also allow the continued production and use of 500 ppm nonroad
diesel fuel for a limited time past June 1, 2010. To ensure that adequate quantities of 15 ppm
diesel are produced, we are proposing the use of a marker to segregate locomotive and marine
diesel fuel from 500 ppm nonroad diesel fuel beginning June 1, 2010. Since use of the marker in
heating oil is no longer required, we are proposing that the same marker used for heating oil from
June 1, 2007 through 2010 be the marker used in locomotive and marine diesel fuel beginning
June 1, 2010. We propose that the marker would be required to be added at the refinery gate just
as visible evidence of the red dye is required today, and fuel containing the marker would be
prohibited from use in any nonroad application.
Beginning June 1, 2014, after all small refiner and credit provisions have ended, the 500 ppm
locomotive and marine standard could be enforced based on sulfur level throughout the
distribution system and at the end-user. Therefore, there would no longer be any need for a
marker. Consequently, we are proposing that after May 31, 2014 there would be no marker
requirement and the different grades of fuel, 15 ppm, 500 ppm, and high sulfur would merely
have to be kept segregated in the distribution system.
We are proposing that the current requirement that non-highway distillate fuels be dyed at the
refinery gate be made voluntary effective June 1, 2006. The IRS requirement that non-highway
fuel be dyed prior to sale to consumers to exempt it from excise taxes will still apply.
There are two considerations with respect to the feasibility of distributing non-highway diesel
fuels meeting the proposed 500 ppm sulfur standard. The first pertains to whether sulfur
contamination can be adequately managed throughout the distribution system so that fuel
delivered to the end-user does not exceed the specified 500 ppm maximum sulfur concentration.
The second pertains to the physical limitations of the system to accommodate any additional
segregation of product grades. These considerations are evaluated in the following Sections 5.4.1
and 5.4.2 of this Draft RIA.
5.4.3 Limiting Sulfur Contamination
With respect to limiting sulfur contamination during distribution, the physical hardware and
distribution practices for non-highway diesel fuel do not differ significantly from those for
current highway diesel fuel. Therefore, we do not anticipate any new issues with respect to
5-38
-------
Fuel Standard Feasibility
limiting sulfur contamination during the distribution of non-highway 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.
5.4.4 Potential Need for Additional Product Segregation
During the first step of today's program, we anticipate that 500 ppm non-highway diesel
engine fuel would be distributed in fungible batches with 500 ppm highway diesel fuel through
the distribution system to the terminal level. When the second step of the proposed program
would require nonroad diesel fuel to meet a 15 ppm sulfur standard all highway fuel would also
be required to meet a 15 ppm standard. Thus a large fraction of the potential 500 ppm diesel fuel
pool would disappear. Since marked 500 ppm locomotive and marine diesel fuel would be a
relatively small volume product, we anticipate that in most parts of the distribution system it
would not be carried as a separate product in the fungible distribution system. Therefore we
anticipate that most shipments of 500 ppm locomotive and marine fuel would be from refinery
racks or other segregated shipments directly into end-user tankage. Any diesel fuel supplied off
the fungible supply system for locomotive and marine uses would therefore likely be spillover
from 15 ppm supply.
The proposed non-highway sulfur program would require the use of a unique fuel marker in
heating oil to differentiate it from other non-highway diesel engine fuels that would be subject to
today's proposed sulfur standards. Under the proposed program, heating oil would be injected
with a marker at the refinery and segregated throughout the fuel distribution system to the end-
user. The heating oil marker requirement would expire after 2010, to be replaced with the
requirement that 500 ppm diesel fuel destined for locomotive and marine use contain the marker
previously used in heating oil. The presence of the marker raises the potential for additional
product segregation needs in both 2007 and 2010.
The proposed application of different sulfur standards to portions of the non-highway
distillate pool based on end-use also 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 non-highway 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.
We anticipate that the significant cost of desulfurizing non-highway diesel engine fuel to
meet today's proposed sulfur standards would provide a strong incentive for the fuel distribution
5-39
-------
Draft Regulatory Impact Analysis
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 implementation of the current EPA requirement in 1993 that highway diesel
fuel have a sulfur content of less than 500 ppm.
The Internal Revenue Service (IRS) requirement that diesel fuel used in non-highway engines
be dyed before it leaves the terminal to indicate its non-taxed status also raises concerns about
the potential need for additional product segregation under the proposed NRLM sulfur program.
Fuel that meets highway diesel specifications but is destined for the non-highway 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 non-highway 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 non-highway diesel fuel that differ in no other respect than the
presence of dye (both after the implementation of the 15 ppm highway diesel requirements in
2007, and the proposed requirements for NRLM fuel).
We expect that after the implementation of the proposed NRLM standards most bulk plant
operators would request that the terminal (or refinery rack) dye 500 ppm fuel destined for sale
into the non-highway market, so that they continue their current practice of offering untaxed
diesel to their non-highway customers. This raises issues of available tankage.
The following discussion evaluates the potential need for additional product segregation for
each segment of the distribution system from the refinery through to the end-user due to the
implementation of the proposed 500 ppm non-highway diesel sulfur standard.
Refineries:
Due to economies of scale involved in desulfurization, we expect that most individual
refineries would choose to manufacture a single or perhaps in some case two sulfur grades of
diesel fuel. We do not anticipate that individual refineries would produce substantial quantities
of all the different diesel fuel sulfur grades (15 ppm highway fuel, 500 ppm, and heating oil).
We do not anticipate the need for additional product segregation at refineries. As discussed
above, today's proposal would allow highway and nonroad diesel fuels to be shipped fungibly
until NRLM fuel is dyed for IRS excise tax purposes. Therefore, today's proposed sulfur
standards for NRLM diesel fuel would not require refiners to put in new product storage tanks to
handle these fuels. The proposed marker requirements for heating oil from 2007-10 and for
locomotive and marine diesel fuel from 2010-14 would also not cause the need for additional
product segregation at the refinery. We expect that refiners would inject the marker into the fuel
stream as it leaves their facility. Since the dye requirement for these fuels is removed at the
refinery gate, they should be able to modify their existing additive injection hardware to satisfy
5-40
-------
Fuel Standard Feasibility
this need. The ability of a refinery to sell diesel fuel directly from the would mean that they
could market dyed and marked fuel with out the need for additional tankage. The dye and
marker could be injected as the fuel is loaded into the tank truck from the rack.
A limited number of refiners would be allowed to produce high-sulfur NRLM until 2010.
We expect that such fuel would be distributed via segregated means from the refinery to the end-
user. Thus, we do not expect that such fuel would result in the need for additional tankage.
There will be no physical differences between 500 ppm highway and 500 ppm NRLM
produced by refiners. The distinction between the two fuels is only made 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.
Pipelines:
Under today's proposal, pipeline operators would ship only a single 500 ppm diesel fuel to be
later directed to either the highway or NRLM market. We project that only the 40 percent of
pipelines that the highway diesel rule projected would carry 500 ppm highway diesel fuel would
be the pipelines that elect to 500 ppm diesel fuel after the implementation of the proposed NRLM
diesel fuel program. Therefore, we do not expect that the proposed 500 ppm sulfur standards
would necessitate the need for additional product segregation in the pipeline distribution system.
There is no physical separation between product batches shipped by pipeline. When the
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 product downgrade. When the
mixture must be removed for reprocessing, it is referred to as transmix. Given that the pipelines
that carry 500 ppm diesel fuel would be able to combine batches of 500 ppm non-highway diesel
fuel with batches of 500 ppm highway diesel fuel, we do not expect that today's program would
result in an 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 500
ppm NRLM shipped by pipeline would represent a sizeable fraction of the total 500 ppm diesel
fuel volume.
We also do not expect that the marker requirement for heating oil would result in an
increased need for product segregation in the pipeline or an increase in product downgrade or
transmix volumes. After the implementation of the proposed 500 ppm standard for nonroad,
locomotive, and marine fuel, we project that significant volumes of heating oil would continue to
be present only in the fuel distribution system that supplies the Northeast, limited adjoining parts
of the Midwest, and the Pacific Northwest.
We believe that only in these areas, would the demand for heating oil be sufficiently large 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. Therefore, heating oil would
5-41
-------
Draft Regulatory Impact Analysis
not be present in pipeline systems that supply areas outside of the Northeast, limited adjoining
parts of the Midwest, and the Pacific Northwest. The pipelines that we project would handle
heating oil after the implementation of today's proposal are those that we projected would also be
carrying 500 ppm highway diesel fuel from 2006-10. Therefore, these pipelines would already
have facilities to also carry 500 ppm NRLM in 2007 (in the pipeline there is no physical
distinction between 500 ppm highway and 500 ppm NRLM diesel fuel). Consequently, we do
not expect that the heating oil marker requirement would result in additional product segregation
by pipeline.
We anticipate that in some cases high sulfur fuel will be sold directly from refinery racks
throughout the country. In addition, some terminals outside of these areas may market limited
quantities of high-sulfur diesel fuel that was generated in the pipeline during the distribution of
15 ppm diesel fuel. We expect that such fuel would be marketed directly from the terminal to the
end user. The limited additional tankage at the terminal was accounted for under the highway
program.
The situation for pipeline operators after 2010 when the marker must be used in locomotive
and marine fuel would be somewhat different, but is still not expected to result in any new
product segregation needs. Under today's proposal, all nonroad diesel fuel would be required to
meet a 15 ppm sulfur standard in 2010 except for limited quantities of small refiner and credited
fuel that could remain at 500 ppm for a limited additional time. We expect that this nonroad fuel
which remains at 500 ppm after 2010 would be distributed by the refiner to the end-user directly.
Therefore, its presence in the distribution system would not result in the need for additional
product segregation. The highway diesel program also requires that all highway diesel fuel meet
a 15 ppm sulfur standard beginning in 2010. Consequently, the only 500 ppm diesel fuel
possibly remaining in the fungible distribution system would be marked 500 ppm locomotive and
marine diesel fuel. We expect that pipelines that carried 500 ppm diesel fuel prior to 2010 would
be the only pipelines that might choose to carry marked 500 ppm locomotive and marine diesel
fuel. Therefore, the equipment that had been used to handle unmarked 500 ppm diesel fuel prior
to 2010 would be switched to handling marked 500 ppm diesel fuel after 2010. Due to the
reduction in the total potential 500 ppm diesel pool beginning in 2010, it is likely that a number
of pipelines will no longer find it economical to carry 500 ppm as well as 15 ppm diesel fuel.
We are projecting that most pipelines would elect not to carry 500 ppm diesel fuel and would
carry only 15 ppm diesel fuel after 2010. This could result in some overall simplification of the
diesel distribution system. Another factor that mitigates any potential need for additional product
segregation as a result of the marker requirement for locomotive and marine diesel fuel is that
locomotive and marine diesel fuel is often distributed through a segregated distribution system.1
Based on the above discussion, we anticipate that the locomotive and marine diesel fuel marker
requirement would not result in an increased need for product segregation in the pipeline or an
increase in product downgrade or transmix volumes.
1 In addition, we understand that marine diesel fuel is often shipped by barge from the refiner to the end user.
This is also the case for locomotive diesel fuel when there is an opportunity for waterborne transportation.
5-42
-------
Fuel Standard Feasibility
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 today's
proposed rule on terminal operators closely parallels the preceding discussion on the potential
impacts on pipeline operators. The proposed allowance that highway and non-highway diesel
fuel meeting the same sulfur specification could be shipped fungibly until non-highway diesel
fuel must be dyed to indicate its non-tax status obviates the need for additional product
segregation at the terminal that might otherwise result from today's proposed sulfur standards.
We expect that terminal operators would store non-highway and highway diesel fuel meeting the
same sulfur specification in the same tank.
We do not anticipate that the proposed marker requirement for heating oil would require any
additional storage tanks. As discussed above, in most of the country, we do not anticipate
heating oil would continue to be carried as a separate grade in the fungible distribution system
after the implementation of the proposed NRLM sulfur standards. As a result, 500 ppm fuel
could take the place of the current tank of high sulfur fuel. In the areas where we project heating
oil would continue to be distributed, 500 ppm highway fuel is also projected to be distributed.
Consequently, marked heating oil can remain in its high sulfur tank, and the existing 500 ppm
highway tank can service both highway and NRLM uses.
Bulk Plants:
Bulk plants are secondary distributors of refined petroleum products. They typically receive
fuel from terminals and distribute fuel in bulk by truck to end users. Consequently, while for
highway fuel, bulk plants often serve the role of a fuel distributor, delivering fuel to retail
stations, for nonroad fuel, they often serve the role of the retailer, delivering fuel directly to the
end-user. Bulk plants represent the one point in the distribution system where we anticipate
some additional tankage would likely be added as a result of today's proposal. However, we
project that only a small subset of the bulk plants would be faced with the choice of adding
additional tankage. In most areas of the country, a distinct grade of heating oil would no longer
be carried, and bulk plant operators could simply switch the tank that they previously devoted to
high sulfur service to 500 ppm NRLM and heating oil service in 2007.
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
in order to distribute dyed 500 ppm NRLM diesel fuel in addition to dyed and marked heating
oil. In this context demanifolding refers to the process of separating a single storage tank on a
delivery tank truck 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 is too high, or may not have the space or capability to
add new tank. However, such bulk plants would 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 work out
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
5-43
-------
Draft Regulatory Impact Analysis
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
would 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
would be located in the areas of the country where heating oil would be continued to be carried
by the fungible distribution system after the implementation of the proposed NRLM standards
and where 500 ppm fuel would also be carried. Of these, we expect no more than 1,000 would
choose to install a new tank.31
We do not anticipate that bulk plants would invest to carry locomotive and marine fuel as a
separate grade in 2010. Therefore, unless a bulk plant had existing tankage available or supplied
a majority of its fuel to locomotive and marine uses, this grade would likely be limited to refinery
and terminal distribution. This is how the bulk of the distribution of locomotive and marine
diesel fuel occurs today.
Based on the above discussion, we believe that the potential impacts of today's proposed rule
on the distribution system due to the need for additional product segregation would be minimal
and easily accommodated by industry. Please see 7.3 of this Draft RIA for a discussion of the
increased distribution costs associated with the need for additional segregation at bulk plants.
5.5 Feasibility of Producing 15 ppm Sulfur Nonroad Diesel Fuel in 2010
5.5.1 Expected use of Desulfurization Technologies for 2010
Like the 500 ppm sulfur standard for 2007, we considered a number of different criteria to
project which desulfurization technologies which would be used to comply with a 15 ppm
nonroad sulfur cap standard for 2010. The criteria we considered included: 1) the time which
refiners will have to choose a new technology, 2) whether the technology would be available for
2010 and if the technology is available, how proven it is, 3) whether the technology is cost-
competitive by comparing it to other technologies, 4) whether the technology is available from
an industry-trusted vendor which has proven itself to the industry by providing other successful
refining technologies, particularly if the vendor has proven itself in the U.S., and 5) whether the
vendor has the capability to meet the demands of the industry.
Unlike the 2007 standard, refiners would have plenty of time to evaluate the various
desulfurization technologies and to choose which one would be best suited for their particular
application. As stated above, we believe that this rule would be promulgated sometime in early
2004, thus, refiners will have 6 years between when the rule is promulgated and when the rule
takes effect. Therefore, refiners would not be constrained in any way when making their
decisions so this particular issue did not figure into our choice of the technologies which they
would use.
Next, we considered whether a technology would be expected to be available for 2010. Of
5-44
-------
Fuel Standard Feasibility
course, conventional hydrotreating is available as it has been used in a number of applications to
comply with a very stringent sulfur standard like a 15 ppm sulfur standard as described above. 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 prior to making a decision on what technology to use. Phillips is targeting
to have their diesel fuel commercial demonstration unit up and running in early 2004 and they are
expecting to have numerous gasoline desulfurization units starting up in 2004 as well. The
operation of these units for two or more years prior to having to make their decisions for 2010
would give refiners confidence that these units can operate effectively over a significant period of
time.
Linde 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, however, refiners would like to see the technology
demonstrated over the 500 ppm to 15 ppm sulfur reduction interval as well. With the 15 ppm
highway diesel fuel sulfur standard taking effect in 2006, Linde should be able to demonstrate its
technology for the 500 ppm to 15 ppm sulfur reduction interval. Thus, refiners that would be
seeking to comply with the proposed 15 ppm sulfur nonroad standard should be able to see at
least one, and probably more, examples of the Linde Isotherming process operating to desulfurize
diesel fuel down to 15 ppm.
The oxidation and extraction technologies by Petrostar and Unipure do not have units
operating now, but are projecting to have commercial demonstration units operating by 2006.
However, an oxidation and extraction unit which begins operation in 2006 would not provide
two years of operations for interested nonroad refiners prior to when they will have to choose
their technology for 2010. Similarly, biodesulfurization is not expected to have a commercial
demonstration unit operating before 2006.
Another issue which refiners would consider is the cost of installing and operating these
various technologies. Biodesulfurization has not yet developed detailed desulfurization costs for
their process. 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. Petrostar, however, has not yet
provided us with desulfurization information. Phillips also has provided us with diesel fuel
desulfurization cost information from their pilot plant which is backed up by the success which
they have had with their commercial gasoline desulfurization unit (see Chapter 7.2). That
technology seems to be less expensive than conventional hydrotreating, it appears to be suited
primarily for desulfurizing low sulfur diesel fuel down to very low sulfur values rather than for
desulfurizing higher sulfur feedstocks. As a result, its primary usefulness would be for refiners
revamping from compliance with 500 ppm in 2007 to 15 ppm in 2010. Finally, Linde provided
us diesel fuel desulfurization cost information which is based on their pilot plant and their
engineering cost estimates for the commercial demonstration unit at the Giant refinery. The
Linde process seems to be less expensive than conventional hydrotreating and is capable of
5-45
-------
Draft Regulatory Impact Analysis
desulfurizing high sulfur feedstocks down to 15 ppm (see Chapter 7.2).
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.
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.
Linde licenses several different technologies now including sulfur and olefms recovery,
natural gas processing, hydrogen production, reforming, air separation, and of course the
Isotherming process for desulfurizing diesel fuel. Linde has a large engineering and design
department which has been active for over 30 years, and now Linde has an alliance with Roddy
Engineering for additional engineering support. Thus, Linde is capable of supporting its
desulfurization technology for a significant penetration into the U.S. refining industry.
The oxidation and extraction technologies are being developed by two separate entities, one
being Unipure and the other Petrostar. Unipure is associated with Texaco and Mustang
engineering. Thus, Unipure potentially has both research and development and engineering
support for its technology. Petrostar is affiliated with DeGussa Catalysts which can provide
research and development support. Neither of these technologies have yet been licensed for
desulfurizing diesel fuel.
After evaluating the various criteria for each technology and comparing across technologies,
we developed a projection for the mix of technologies which would be used in 2010 for
complying with the 15 ppm cap standard. Since refiners will have plenty of time to sort through
the various technologies, we believe that the leadtime issue would 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 oxidation
and extraction and biodesulfurization technologies as these technologies have not announced that
their technology is available for licensing yet, and are not expected to have a commercial
demonstration unit operating for at least two years. Thus, while the Petrostar, Unipure
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, Phillips S-Zorb and Linde Isotherming. Obviously,
5-46
-------
Fuel Standard Feasibility
conventional hydrotreating will have the most refining experience due to refiners' experience
earlier on and also due to production of 15 ppm highway fuel for 2006. However, both S-Zorb
and Isotherming are expected to have one or more diesel fuel desulfurization commercial
demonstration units operating for over two years. Both the S-Zorb and the Isotherming
hydrotreating processes are expected to be lower in cost than conventional hydrotreating
providing a strong incentive to refiners which are seeking to lower their refining margins. Also
both Phillips and Linde have research and development and engineering capacity to support their
processes, although not the same level of support as the multiple conventional hydrotreating
firms. After comparing these various criteria, we decided that the lower cost of S-Zorb and
Isotherming would be the most important driver for these technologies. Thus, we believe that S-
Zorb and Isotherming would each be used to a greater extent than conventional hydrotreating.
We project that S-Zorb and Isotherming would each capture 40 percent of the nonroad
desulfurization market by 2010, while conventional hydrotreating would capture the remaining
20 percent of the nonroad desulfurization market.
It was also necessary to estimate the technology mix for other potential years for a 15 ppm
sulfur cap standard per the various other alternative fuel options being considered. The relative
cost of these technologies is not estimated to change, however, the degree to which refiners have
confidence in each of these technologies would change over time. In the years before 2010,
refiners would not be expected to place as much trust with S-Zorb and Isotherming because there
would be less time for these technologies to be proven to refiners. In 2009 we project that S-
Zorb and Isotherming would each capture 30 percent of the nonroad desulfurization market.
Similarly, we project that S-Zorb and Isotherming each capture 20 percent of the nonroad
desulfurization market in 2008. For 2010, we project that S-Zorb and Isotherming would each
capture 40 percent of the desulfurizaiton market. Finally, in 2012 and later, we project that S-
Zorb and Isotherming would each capture 50 percent of the nonroad desulfurization markets.
5.5.2 Leadtime Evaluation
More leadtime would be required to meet a 15 ppm diesel fuel cap than a 500 ppm cap. The
additional time would primarily involve the scoping and screening step, as the technology to
achieve a 15 ppm sulfur cap is just being demonstrated on a commercial scale and a number of
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 the 15 ppm cap
provides an additional three years of leadtime. Therefore, the amount of leadtime available for
the 15 ppm cap on nonroad diesel fuel should 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 proposed for nonroad diesel fuel. The time periods listed in Table 5.3-3 indicate
that refiners would have to start their process designs 2.0-2.75 years before first producing 15
ppm diesel fuel and complete these process designs 1.5-2.25 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
5-47
-------
Draft Regulatory Impact Analysis
ppm nonroad diesel fuel with 15-24 months of experience by highway diesel fuel refiners before
initiating their process design. Given that catalyst cycles last 2-3 years, refiners could observe
the performance of catalysts used to produce 15 ppm highway diesel fuel for one half to two
thirds of a full cycle before having to begin their process design for nonroad While most of the
units producing highway diesel fuel in 2006 are expected to use conventional hydrotreating, as
discussed above, we also expect both Linde Isotherming and Phillips' SZorb processes to be used
to commercially produce 15 ppm diesel fuel by the end of 2004. Thus, refiners planning for
2010 would be able to observe these newer processes for more than 3 years prior to their
selection of vendor and technology. This should be sufficient to overcome any uncertainty about
their performance. Overall, the available leadtime should allow all refiners to take advantage of
the operating performance of the highway units and minimize their costs.
5.6 Feasibility of Distributing 15 ppm Sulfur Nonroad Diesel Fuel in 2010
The same two criteria apply regarding the evaluation of the feasibility of distributing 15 ppm
sulfur nonroad diesel fuel as apply regarding the feasibility of distributing 500 ppm sulfur non-
highway diesel fuel: limiting sulfur contamination, and the potential need for additional product
segregation. However, concerns related to limiting contamination during the distribution of 15
ppm nonroad diesel fuel are more substantial given that industry is just now in the process of
learning how to accomplish the task of distributing 15 ppm diesel fuel in the fungible distribution
system in preparation for compliance with the 15 ppm sulfur specification for highway diesel fuel
in 2007. These considerations are evaluated in the following 5.6.3 and 5.6.4 of this Draft RIA.
5.6.1 The Diesel Fuel Distribution System Prior to the Implementation of the Proposed 15
ppm Nonroad Diesel Sulfur Program
Refer to 5.4 of this Draft RIA for a discussion of the diesel fuel distribution system prior to
the implementation of the proposed 500 ppm NRLM sulfur program. Section 5.4 also contains a
discussion of the potential effects on the distribution system of the implementation of the 500
ppm NRLM program in 2007 and the continuance of the 500 ppm sulfur standard for locomotive
and marine diesel fuel past 2010. The discussion in section 5.4 provides the baseline against
which the potential effects on the distribution system from the implementation of the proposed
15 ppm nonroad diesel sulfur standard in 2010 are evaluated.
5.6.2 Summary of the Proposed 15 ppm Nonroad Diesel Sulfur Standard
We are proposing to restrict the sulfur content of nonroad diesel fuel nationwide to no more
than 15 ppm beginning in 2010. This proposed requirement mirrors the 15 ppm sulfur
requirement for highway diesel fuel which will become effective in 2006.32 As with the 500 ppm
NRLM standard that we are proposing, refiners and importers could comply with the proposed
15 ppm nonroad standard by either physically producing 15 ppm fuel or by obtaining sulfur
credits. Also similar to the proposed 500 ppm NRLM standard, the deadlines for meeting the 15
ppm nonroad sulfur standard would not apply to refineries covered by special hardship
provisions for small refiners. In addition, a different schedule might apply for any refineries that
5-48
-------
Fuel Standard Feasibility
might be approved under the proposed general hardship provisions. Only 15 ppm diesel fuel
would be permitted for use in 2011 and later model year nonroad equipment. As discussed in
5.4, locomotive and marine diesel fuel would continue to be subject to the proposed 500 ppm
sulfur standard after 2010.
In order to allow for a smooth and orderly transition of diesel fuel in the distribution system
to 15 ppm, we are proposing that parties downstream of the refineries be allowed a small amount
of additional time to turnover their tanks to 15 ppm. We are proposing that at the terminal level,
nonroad diesel fuel would be required to meet the 15 ppm sulfur 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 sulfur standard by September 1, 2010. The
proposed transition schedule for compliance with the 15 ppm standard at refineries, terminals,
and secondary distributors are the same as those allowed under the recently promulgated highway
diesel fuel program.
5.6.3 Limiting Sulfur Contamination
In the highway diesel rule, EPA 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
contamination in the system (e.g., tank bottoms, deal 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.33
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 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 (AOPL), the
Independent Fuel Terminal Operators Association (TFTOA), 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.34 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
5-49
-------
Draft Regulatory Impact Analysis
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 nonroad
diesel fuel will be resolved well in advance of the proposed 2010 implementation date 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
disappearance of heating oil from much of the distribution system outside of the North East,
adjoining parts of the Midwest, and North West would tend to lessen the opportunity for sulfur
contamination.
We do not anticipate that there would be a substantial increase in the number of off-
specification 15 ppm diesel fuel batches in the distribution system due to sulfur contamination.
To the extent that there are off-specification batches of nonroad (and highway) diesel fuel, the
500 ppm locomotive and marine diesel fuel markets could provide a market for off-spec product
that could be important to during the transition to 15 ppm nonroad diesel fuel in 2010.
5.6.4 Potential need for Additional Product Segregation Due to the Implementation of the
Proposed 15 ppm Sulfur Specification for Nonroad Diesel Fuel
Two of the three factors discussed in 5.4 of this Draft RIA regarding the potential need for
additional product segregation due to the implementation of the proposed 500 ppm NRLM
standard in 2007 also apply with respect to the potential impact of the proposed 15 ppm standard
for nonroad diesel fuel in 2010: 1) the application of a different sulfur standard to a portion of the
non-highway distillate pool based on end-use, and 2) the Internal Revenue Service (IRS)
requirement that diesel fuel used in non-highway engines be dyed prior to sale to consumers to
indicate its non-taxed status before it leaves the terminal. The potential impact on product
segregation of the proposed marker requirement was discussed in 5.4of this Draft RIA within the
context of the proposed 500 ppm sulfur specification for NRLM fuel in 2007 and for locomotive
and marine diesel fuel in 2010. The implementation of the proposed 15 ppm sulfur standard
would not alter the conclusions we reached in 5.4 regarding the potential impacts on the
proposed marker requirements
The following discussion evaluates the potential need for additional product segregation in
each segment of the distribution system from the refinery through to the end-user due to the
implementation of the proposed 15 ppm sulfur standard for nonroad diesel fuel.
5-50
-------
Fuel Standard Feasibility
Refineries:
Due to economies of scale involved in desulfurization, we expect that most individual
refineries would choose to manufacture a single or perhaps in some case two sulfur grades of
diesel fuel. We do not anticipate that individual refineries would produce substantial quantities
of all the different sulfur grades (15 ppm, 500 ppm locomotive and marine, and heating oil). We
do not anticipate the need for additional product segregation at refineries. As discussed above,
we do not anticipate that there would be any physical differences between 15 ppm manufactured
for the highway market and 15 ppm diesel fuel manufactured for the non-highway market.
Today's proposal would allow 15 ppm diesel fuels intended for the highway and nonroad
markets to be shipped fungibly until NRLM fuel is dyed for IRS excise tax purposes. Therefore,
today's proposed 15 ppm sulfur standards for nonroad diesel fuel would not require refiners to
put in new product storage tanks.
A limited number of refiners would be allowed to produce 500 ppm nonroad diesel fuel until
2010. However, we expect that such fuel would be distributed via segregated means from the
refinery to the end-user. Thus, we do not expect that such fuel would result in the need for
additional tankage.
Pipelines:
Under today's proposal, pipeline operators would ship only one 15 ppm diesel fuel.
Therefore, we do not expect that the proposed 15 ppm nonroad diesel sulfur standards would
necessitate the need for additional product segregation in the pipeline distribution system (i.e.
there would be no increase in the number of different diesel fuel grades carried by the pipeline
system relative to 2007). Due to the large reduction in the potential 500 ppm diesel pool that
would accompany the implementation of the proposed 15 ppm nonroad diesel sulfur standard, we
expect that 500 ppm diesel fuel would all but disappear from the fungible pipeline distribution
system. This could result in a simplification of in the number of fuel grades carried in certain
parts of the fungible distribution system.
We also project that today's program would not result in an increase the volume of product
downgrade or transmix. To the contrary, similar to the situation associated with shipping batches
of 500 ppm diesel fuel by pipeline until 2010, there may be some opportunity for improved
efficiency (i.e. a reduction in downgrade and transmix volumes) because of the increase in 15
ppm batch sizes shipped by pipeline.
Terminals:
Under the proposed sulfur program we expect that terminal operators would maintain storage
facilities for a single 15 ppm diesel fuel. Only when 15 ppm fuel leaves the refinery would it be
segregated into two distinct products due the addition of dye to nonroad diesel fuel per the IRS
requirements to indicate its non-taxed status. Therefore, we do not expect that the
implementation of the proposed 15 ppm nonroad sulfur standard would result in the need for
additional product segregation at terminals.
5-51
-------
Draft Regulatory Impact Analysis
Bulk Plants:
When the 15 ppm standard for nonroad diesel fuel would be implemented in 2010, we expect
that bulk plant operators would switch the tank that previously contained 500 ppm NRLM to
dyed 15 ppm nonroad service in 2010.35 Therefore, we do not anticipate the need for additional
product segregation at bulk plants due to the implementation of the proposed 15 ppm nonroad
sulfur specification.
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 proposed 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.
standard and the proposed nonroad diesel sulfur programs. 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 these specific job categories.
To accomplish this task, we first estimated the level of design and construction resources
related to revamped and new desulfurization equipment. We next projected the number of
revamped and new desulfurization units which would be needed under the proposed NRLM fuel
program, as well as under a couple of alternative programs also considered. 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 2003 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.36 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.37 These estimated job hours are summarized in Table 5.7-1.
5-52
-------
Fuel Standard Feasibility
Table 5.7-1
Design and Construction Factors for Desulfurization Equipment
Number of Pieces of Equipment per Refinery
Gasoline3
New
Hydrotreater
60
Highway and
Nonroad
Diesel
Treaters
New
Hydrotreater
60
Highway and
Nonroad
Diesel
Treaters
Revamp
Existing
Hydrotreater
30
Job hours per piece of equipment3 III III
Front End Design
Detailed Design
Direct and indirect construction
300
1200
9150
300
1200
9150
150
600
4575
1 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.
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.38 We updated the projections for the 2007 highway diesel program per the
analysis presented in Section 7.2.2.1. These projections are shown in Table 5.7-2 below.
Table 5.7-2
Number of Gasoline and Highway Diesel Desulfurization Units Becoming Operational"39
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
74
2007
5
2008
3
2009
4
2010
6
40
a Units become operational on January 1st for gasoline desulfurization and June 1st for highway diesel desulfurization 1
units.
The next step was to estimate the types of equipment modifications necessary to meet the
proposed NRLM fuel requirements. This was a complex task, due to the close integration of the
highway and NRLM fuel programs and the fact that refiners' relative production of highway and
5-53
-------
Draft Regulatory Impact Analysis
high sulfur distillate fuel varies dramatically. Therefore, we broke refiners of high sulfur
distillate into three categories and assessed their need for new or revamped desulfurization
equipment separately. The categories as 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),
mix refiners (producers of high sulfur distillate fuel not falling into one of the other categories).
Table 5.7-3 presents the results of our analysis of the 62 refineries which are projected to
produce either 500 or 15 ppm NRLM diesel fuel under the proposed program. The methodology
used to determine that these 62 refineries would produce NRLM diesel fuel is described in
Section 7.2.
5-54
-------
Table 5.7-3
Types of Equipment Modifications Needed Under the Proposed NRLM Fuel Program
Fuel Type
15 ppm Highway Diesel
Hydrotreater Modifications
High Sulfur Diesel
Hydrotreater Modifications
Year and
Fuel Control
2006
2010
Total
2007
500 ppm fuel
(total of 42)
2010
500 ppm fuel
(total of 20)
2010
1 5 ppm fuel
(total of 25)
2014
1 5 ppm fuel
(total of 12)
Highway
Refiners
Units
8
2
10
7
3
5
3
Mix 2006 Refiners"
New
Units
7(4)
Revamp
Units
12
None
19
12(1)
4(3)
2
0
0
0
6(1)
5(3)
3
0
0
0
Mix 20 10 Refiners"
New
Units
11(9)
Revamp
Units
8
None
19
12(4)
0
0
0
0
7(5)
0
4(3)
0
0
8(2)
0
High Sulfur
Refiners
Units
0
0
Revamp
14
8
6
0
0
0
0
4
0
1 Numbers in parentheses are a subset for each category and represent mix refineries that currently have no highway diesel fuel hydrotreater.
-------
Draft Regulatory Impact Analysis
As shown in the Table 5.7-3, we project that 10 highway refineries, 38 mix refineries and 14
high sulfur refineries are projected to produce NRLM diesel fuel in 2007 and beyond. Refineries
in the first two categories also produce highway diesel fuel. We further sub-divide refineries in
these two categories according to whether they are projected to produce 15 ppm highway diesel
fuel in 2006 or 2010, because the timing of their modifications to their highway diesel fuel
hydrotreater can affect what modifications are necessary to produce 500 ppm or 15 ppm NRLM
diesel fuel. As shown, of the 10 highway refineries, we project that 8 will revamp or replace
their current hydrotreater in 2006, while the other two will do so in 2010. Of the 38 mix
refineries, we project that half will revamp or replace their current hydrotreater in 2006, while the
other half will do so in 2010. No current high sulfur refineries are projected to produce 15 ppm
highway diesel fuel in either 2006 or 2010. It should be noted that the 48 highway and mix
refineries shown in Table 5.7-3 are not all the refineries producing highway diesel fuel today or
in 2006 and beyond. The 48 refineries are those which are projected to produce some highway
diesel fuel in 2006 and beyond, as well as NRLM fuel in 2007 and beyond.
Regarding the highway refineries, our cost analysis projects that 7 would produce 500 ppm
NRLM fuel in 2007. Five of these refineries would further desulfurize their 500 ppm NRLM
diesel fuel to 15 ppm in 2010, while three new highway refineries would produce 500 ppm
NRLM diesel fuel for the first time in 2010. Finally, in 2014, an additional three refineries
would further desulfurize their 500 ppm NRLM diesel fuel to 15 ppm in 2014, leaving 2 highway
refineries producing 500 ppm NRLM diesel fuel in the long term.
As mentioned above, the highway refineries produce relatively small quantities of high sulfur
distillate today (i.e., less than 5% of total no. 2 distillate production). Thus, we project that these
refineries could incorporate their high sulfur distillate into the design and construction of their
highway hydrotreaters with no additional engineering or construction requirements. Section
7.2.2 describes the type of hydrotreater modifications which are projected for highway refineries
to enable the production of low sulfur NRLM diesel fuels.
Moving to the mix refineries, their treatment depends on when they are projected to produce
15 ppm highway diesel fuel and whether or not they would do so by revamping their current
hydrotreater or construct the new hydrotreater. Of the 19 mix refineries which are projected to
produce 15 ppm highway diesel fuel in 2006 ("2006 mix refineries"), we project that 7 would
construct a new hydrotreater. We project that 4 of these 7 refineries would need a new
hydrotreater because available data indicate that they do not currently have a no. 2 distillate
hydrotreater. We assumed that 20% of the remaining 15 refineries (3 refineries) would need a
new hydrotreater. This is consistent with the analysis and assumptions for the 2007 highway
rule, where we estimated that 20% of all refineries producing highway diesel fuel would need a
new hydrotreater. Discussions with industry continue to confirm the reasonableness of this
assumption. (The other 80% are projected to be able to revamp their current hydrotreater to
produce 15 ppm diesel fuel.) The same procedure was applied to the 19 mix refineries projected
to produce 15 ppm highway diesel fuel initially in 2010 ("2010 mix refineries"). The only
difference was that 9 of these refineries apparently do not currently have a no. 2 distillate
hydrotreater, therefore necessitating that a new one be built.
5-56
-------
Fuel Standard Feasibility
In 2007, we project that 15 2006 mix refineries would begin producing 500 ppm NRLM
diesel fuel. Twelve of these are projected to need to construct a new hydrotreater to do so, while
three do not. These three refineries are those which built new hydrotreaters in 2006 to produce
15 ppm highway diesel fuel and which also currently have a highway diesel fuel hydrotreater.
These three refineries could use their current highway diesel fuel hydrotreater to produce 500
ppm NRLM diesel fuel. Four additional 2006 mix refineries would begin producing 500 ppm
NRLM diesel fuel in 2010, with all needing to construct new hydrotreaters to do so.
We project that 13 of the 19 2006 mix refineries would produce 15 ppm nonroad diesel fuel,
beginning either in 2010 or 2014. All but two of these units would be revamps of units built in
2007. The two new 15 ppm units would be at refineries projected to produce 500 ppm NRLM
diesel fuel in 2007 with the current highway diesel fuel hydrotreater.
In 2007, we project that 12 2010 mix refineries would begin producing 500 ppm NRLM
diesel fuel. All of these refineries are projected to need to construct a new hydrotreater to do so,
because they will need their existing hydrotreater to continue producing 500 ppm highway diesel
fuel through 2009. Seven additional 2010 mix refineries would begin producing 500 ppm
NRLM diesel fuel in 2010. All seven of these refineries are projected to incorporate the
desulfurization of NRLM diesel fuel into their plans for producing 15 ppm highway diesel fuel in
2010. Due to the significant volume of NRLM fuel involved, we project that the resources
needed to add the desulfurization of NRLM fuel to their plans would constitute a revamp of a
desulfurization unit.
We project that 12 of the 19 2010 mix refineries would produce 15 ppm nonroad diesel fuel,
beginning either in 2010 or 2014. The 10 refineries beginning 15 ppm nonroad fuel production
in 2010 are projected to not require significant amounts of additional design and construction
resources, as these units were designed in 2007 to be easily revamped in 2010 to produce 15 ppm
fuel. The 4 refineries beginning 15 ppm nonroad fuel production in 2014 are projected to require
a modest amount of additional design and construction resources (revamp level), as these
refineries built new desulfurization capacity in 2010 to produce 500 ppm NRLM fuel at the same
time that they began production of 15 ppm highway diesel fuel. Therefore, we projected that
they would likely require some additional engineering and construction resources to produce 15
ppm nonroad fuel in 2014.
Moving to the high sulfur refineries, we project 8 such refineries would begin producing 500
ppm NRLM diesel fuel in 2007 and 6 more refineries in 2010. All of these refineries are
projected to need to construct a new hydrotreater to produce 500 ppm NRLM fuel, because their
existing hydrotreating capacity is likely only capable of producing 2000-5000 ppm sulfur levels.
In 2010, we project that 4 of these refineries producing 500 ppm NRLM diesel fuel in 2007
would revamp their units to produce 15 ppm nonroad diesel fuel.
We repeated this analysis for two of the alternative NRLM fuel programs considered in
developing this proposed rule: 1) the proposed program plus extension of the 15 ppm cap to
locomotive and marine diesel fuel in 2010 (two step 15 ppm NRLM) and 2) a one step program
5-57
-------
Draft Regulatory Impact Analysis
consisting of the final standards as the proposal, but with all the standards occurring in 2008 (one
step in 2008). The breakdown of desulfurization equipment modifications required under the
two step 15 ppm NRLM program are summarized in Table 5.7-4. There are no differences
between this program and the proposal with respect to the production of 500 ppm fuel in 2007 in
2010. However, due to the further control of locomotive and marine diesel fuel to 15 ppm in
2010, additional new and revamped units would be needed in 2010.
We again repeated this analysis for the one step NRLM fuel program in 2008. The results are
shown in Table 5.7-5. The key difference here is that most new and revamped units occur in
2008. Also, we project more revamped units and fewer new units for 2010 mix refineries as we
project that these refineries would combine their plans to produce 15 ppm highway and nonroad
diesel fuel.
Table 5.7-6 summarizes the results of Tables 5.7-3 through 5.7-5.
5-58
-------
Table 5.7-4
Types of Equipment Modifications Needed Under the
Two Step Alternative Program with 15 ppm NRLM Diesel Fuel in 2010
Fuel Type
15 ppm Highway Diesel
Hydrotreater Modifications
High Sulfur Diesel
Hydrotreater Modifications
Year and
Fuel Control
2006
2010
Total
2007
500 ppm fuel
(total of 42)
2010
500 ppm fuel
(total of 20)
2010
1 5 ppm fuel
(total of 43)
2014
1 5 ppm fuel
(total of 12)
Highway
Refiners
Units
8
2
10
7
3
7
3
Mix 2006 Refiners"
New
Units
7(4)
Revamp
Units
12
None
19
12(1)
4(3)
2
0
0
0
13(1)
5(3)
3
0
0
0
Mix 20 10 Refiners"
New
Units
11(9)
Revamp
Units
8
None
19
12(4)
0
0
0
0
7(5)
0
4(3)
0
0
15(2)
0
High Sulfur
Refiners
Units
0
0
Revamp
14
8
6
0
0
0
0
6
0
1 Numbers in parentheses are a subset for each category and represent mix refineries that currently have no highway diesel fuel hydrotreater.
-------
Table 5.7-5
Types of Equipment Modifications Needed Under the
One Step Alternative Program with 15 ppm Nonroad and 500 pmm L&M Fuel in 2010
Fuel Type
15 ppm Highway Diesel
Hydrotreater Modifications
High Sulfur Diesel
Hydrotreater Modifications
Year and
Fuel Control
2006
2010
Total
2008
500 ppm fuel
(total of 13)
2008
1 5 ppm fuel
(total of 30)
2012
500 ppm fuel
(total of 12)
2012
1 5 ppm fuel
(total of 7)
Highway
Refiners
Units
8
3
11
0
7
2
1
Mix 2006 Refiners"
New
Units
7(4)
Revamp
Units
12
None
19
4
10(1)
1
(3)
0
0
0
0
1
0
0
0
Mix 20 10 Refiners"
New
Units
11(9)
Revamp
Units
10
None
21
0
0
0
0
5(2)
7(2)
4(2)
3(3)
0
0
0
0
High Sulfur
Refiners
Units
0
0
Revamp
13
3
6
5
0
0
0
0
0
1 Numbers in parentheses are a subset for each category and represent mix refineries that currently have no highway diesel fuel hydrotreater.
-------
Fuel Standard Feasibility
Table 5.7-6
Number and Timing of NRLM Desulfurization Units
Program
Proposed Two
Step Program
Proposed Two
Step Program
with 15 ppm
Locomotive and
Marine Fuel in
2010
One Step
NRLM Program
in 2008
Type of Treater
No Treaters
Modification
Revamp
Treaters
New Treaters
Total Units
No Treaters
Modification
Revamp
Treaters
New Treaters
Total Units
No Treaters
Modification
Revamp
Treaters
New Treaters
Total Units
2007
10
0
32
42
10
0
32
42
2008
7
12
24
43
2009
2010
16
17
12
45
25
26
11
62
2011
2012
3
7
9
19
2013
2014
3
9
0
12
3
16
0
19
5.7.3 Timing of Desulfurization Projects Starting up in the Same Year
A worst case assumption would be that all of 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,
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. Thus, with respect to units starting up in a given year, we assumed that the design and
construction of these units would be spread out throughout the year, with 25 percent of the units
5-61
-------
Draft Regulatory Impact Analysis
starting up per quarter. Given this assumption, we developed the breakdowns of personnel
requirements by month for a given project shown in Table 5.7-7.
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. The results of this analysis are summarized in Table 5.7-7.
Table 5.7-7
Distribution of Personnel Requirements Throughout the Project
Duration per project
Duration for projects starting up in a
given calendar year
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
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
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 figures shown in Table 5.7-7 were taken from a similar analysis performed in support of
the 2007 highway diesel fuel program. The fraction of total hours expended each month
5-62
-------
Fuel Standard Feasibility
estimated in Table 5.7-7 was derived based on the following. Per Moncrief and Ragsdale, front
end design typically takes six months to complete.40 If 25 percent of the refineries scheduled to
start of 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 roughly two-thirds of that during the
peak middle months. The same process was applied to the other two job categories. Finally, we
assumed that personnel were able to actively work 1877 hours per year, or at 90 percent of
capacity assuming a 40 hour work week. The reader is referred to the Final RIA for the 2007
highway diesel rule for a more detailed description of the methodology used.
5.7.5 Projected Levels of Design and Construction Resources
Applying the above factors, we projected the maximum number of personnel needed in any
given month for each type of job for the desulfurization projects related to the Tier 2 gasoline,
highway diesel fuel and NRLM diesel fuel programs combined . The results are shown in Table
5.7-8. 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 recent
employment levels 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).
5-63
-------
Draft Regulatory Impact Analysis
Table 5.7-8
Maximum Monthly Demand for Personnel
Front-End Design
Detailed Engineering
Construction
Tier 2 Gasoline Sulfur Program Plus Highway Diesel Fuel Program
Number of Workers
Current Workforce a
630 (Apr 03)
33%
2,223 (Apr 04)
23%
14,614(Nov04)
9%
With Proposed Two Step NRLM Program
Number of Workers
Current Workforce a
630 (Apr 03)
33%
2,223 (Apr 04)
23%
17, 176 pec 04)
11%
With Proposed Two Step NRLM Program with 15 ppm NRLM in 2010
Number of Workers
Current Workforce a
630 (Apr 03)
33%
2,223 (Apr 04)
23%
17,076 pec 04)
11%
With One Step NRLM Program in 2008
Number of Workers
Current Workforce a
630 (Apr 03)
33%
2,223 (Apr 04)
23%
14,6 14 pec 04)
9%
' 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 5.7-8, the proposed NRLM diesel fuel program does not impact
the maximum monthly personnel requirements for either front end design or detailed engineering
design. Maximum use of construction personnel is increased slightly, by 2% in November of
2004. This appears to be a minor impact. 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 program and the 2006 implementation date for the 15 ppm
highway diesel fuel standard.
The alternative two step NRLM program with a 15 ppm cap on locomotive and marine diesel
fuel would have the same impact, since the difference between this alternative and the proposal
occurs in 2010, after the peak impacts occurs. The alternative one step NRLM fuel program in
2008 avoids any impact on the peak resource need due to its starting one year later.
Tables 5.7-9, 5.7-10 and 5.7-11 present a summary of the average personnel demand for the
three job categories in each year.
5-64
-------
Fuel Standard Feasibility
Table 5.7-9.
Annual Front End Engineering Personnel Demand
Calendar
Year
2003
2004
2005
2006
2007
2008
2009
2010
2011
2012
2013
2014
Gasoline +
Highway Diesel
Baseline
534
83
32
57
231
23
0
0
0
0
0
0
Plus
Two Step
Nonroad to 15
ppm in 2010
549
344
64
67
398
42
0
2
37
4
0
0
Plus
Two Step to 15
ppm in 2010
549
344
64
67
444
48
0
4
65
8
0
0
Plus
One Step to 15
ppm in 2008
534
100
325
9
231
29
102
18
0
0
0
0
5-65
-------
Draft Regulatory Impact Analysis
Table 5.7-10
Annual Detailed Engineering Personnel Demand
Calendar Year
2003
2004
2005
2006
2007
2008
2009
2010
2011
2012
2013
2014
Gasoline +
Highway Diesel
Baseline
1166
1656
372
345
593
757
46
0
0
0
0
0
Plus
Two Step Nonroad
to 15 ppm in 20 10
1166
1988
1207
407
806
1292
84
0
46
117
9
0
Plus
Two Step to 15
ppm in 2010
1166
1988
1207
407
842
1383
92
0
83
209
15
0
Plus
One Step to 15
ppm in 2008
1166
1656
682
1128
651
757
175
326
24
0
0
0
5-66
-------
Fuel Standard Feasibility
Table 5.7-11
Construction Worker Personnel Demand
Calendar Year
2003
2004
2005
2006
2007
2008
2009
2010
2011
2012
2013
2014
Gasoline +
Highway Diesel
Baseline
4,914
12,462
7,653
249
579
4,948
3,246
0
0
0
0
0
Plus
Two Step
Nonroad to 15
ppm in 2010
4,914
12,743
12,800
4,179
759
8,246
5,764
0
40
724
553
0
Plus
Two Step to 15
ppm in 2010
4,914
12,742
12,800
4,179
790
8,810
6,194
0
70
1,287
983
0
Plus
One Step to 15
ppm in 2008
4,914
12,462
7,916
5,074
4,264
4,948
3,356
2,010
1,535
0
0
0
The impact of the nonroad programs on the maximum monthly demand for front end design
is not increased from the 2000 highway rule determinations. Thus, 33 percent of available front
end personnel U.S. resources are required for the nonroad programs which is not different than
the maximum predicted impact for the highway diesel rule. The annual front end demand for
personnel in Table 5.7-9 reveals that the front end resource demands are spaced over many years
with an initial peak in years 2003-04 and a second sub peak in 2006-07. The level of front end
resource demand drops off dramatically after years 2003 and 2004. Detailed engineering annual
demands for nonroad has a maximum peak in years 2003-05 and a second sub peak in years
2006-08. Neither of the peaks represent a significant percentage of available detailed resources
and furthermore are not higher than demands determined for the highway diesel program. The
nonroad programs contribute to the second peak in front end engineering and detailed
engineering in demands in 2006-07, but we believe the yearly time spread in peak resource
demand will provide an ample period for E&C industry to respond to nonroad implementation.
The maximum monthly impact on construction services is not significant at eleven percent of
available industry which is not considerably increased over highway diesel requirements, see
Table 5.7-11. Thus, we believe the construction industry should be able to provide services for
the nonroad programs.
Thus, we believe that the E&C industry is capable of supplying the refining industry with the
equipment necessary to comply with our proposed nonroad diesel fuel programs. We believe
5-67
-------
Draft Regulatory Impact Analysis
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)
EPA has developed the proposed fuel program to minimize its impact on the supply of
distillate fuel. For example: we have proposed to transition the fuel sulfur level down to 15 ppm
in two steps, providing an estimated 6 years of leadtime for the final step; up to 10 years for
small refiners. We are proposing to provide flexibility to refiners through the availability of
banking and trading provisions and we are proposing hardship provisions for qualifying refiners.
In order to evaluate the effect of this proposal on supply, EPA evaluated four possible cases: 1)
whether today's proposed standards could cause refiners to remove certain blendstocks from the
fuel pool, 2) whether the proposed standards could require chemical processing which loses fuel
in the process, 3) whether the cost of meeting the proposed standards could lead some refiners to
leave that market, and 4) whether the cost of meeting the proposed 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 2007
highway diesel rule, we do not expect this proposed rule to cause any supply shortages of
nonroad, locomotive and 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 being demonstrated as refiners in both the U.S. and Europe are producing
selected batches of number 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 will be addressed below. 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
5-68
-------
Fuel Standard Feasibility
blocking the sulfur atomj. 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, it
is 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 cracked stocks via distillation and avoid the need to desulfurize
them. However, this would likely require the construction of a distillation column and
significant operating costs in the form of heat input. Another 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. Thus, shifting heavy LCO to residual fuel would
involve a significant long term reduction in revenue (and profits). Thus, we do not believe that
many refiners would attempt to reduce the cost of desulfurizing diesel fuel in this way.
It is more feasible to shift some or all of the LCO stream to the heating oil pool. It is unlikely
to be shifted to locomotive and marine (LM) diesel fuel due to their 40 minimum cetane
specification and the very low cetane level of LCO. Straight run distillate could be shifted from
heating oil to diesel fuel to compensate for the volume. Thus, little if any volume loss of NRLM
diesel fuel should result. However, even this approach would require the refiner to maintain
separate inventories of NRLM diesel fuel and heating oil, which may require additional tankage.
Of course, the refiner would need to have access to a significant heating oil market after 2007.
In our cost projections, we projected that individual refineries would produce either 15 ppm,
500 ppm or high sulfur distillate to avoid additional tankage and maximize economies of scale
for the desulfurization equipment. Thus, we did not assume that refiners could reduce costs by
shifting 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. Nor
should such exchanges reduce the supply of NRLM diesel fuel.
Processing Losses: We evaluated whether the proposed standards could require chemical
processing which 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 which is 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 which are used to produce other finished products including
'Meeting a 500 ppm cap standard can be met without desulfurizing much or any of the sterically hindered
compounds.
5-69
-------
Draft Regulatory Impact Analysis
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 would 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 proposed standards (i.e., the
reduction in sulfur content from 3400 ppm to 500 ppm and from 500 ppm to 15 ppm). For the
first step of the proposed fuel program and that portion of the diesel fuel pool which would
remain at the 500 ppm level indefinitely, the impacts would only be 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 only projected to be used as a second step to
reduce the fuel to 15 ppm sulfur after it has been reduced from 3400 ppm to 500 ppm using
conventional hydrotreating technology. We project that the Linde process might reduce the
conversion to light products for the second step by 55 percent, while the Phillips SZorb® process
reportedly would not convert any diesel to light products.
Exit the NRLM Diesel Fuel Market: We evaluated whether the compliance costs associated
with this rulemaking could 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
could 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, a refiner could
either focus entirely on the 500 ppm LM markets or shift part or all of its supply to heating oil.
The result would be a potential oversupply of heating oil beginning in 2007 and LM fuel and
heating oil beginning in 2010. We expect such an oversupply of these fuels would 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 could 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 U.S. We believe that these low market differentials
combined with the additional transportation costs would 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
5-70
-------
Fuel Standard Feasibility
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
which 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.
There are six reasons why we believe a similar conclusion can be drawn from an analysis of
today's proposed rule:
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 2007 highway
diesel rule and the proposed sulfur caps on NRLM fuel have been implemented. Within
that same timeframe, we expect few, if any, of the common carrier pipelines, except
perhaps those serving the Northeast, will 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 prior to the implementation of the 15 ppm sulfur standard for nonroad diesel
fuel.
3. The volume of fuel affected by the 15 ppm nonroad diesel fuel standard in 2010 would be
only one-seventh of that affected by the 2007 highway diesel program. This dramatically
reduces the required capital investment.
4. Both 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 U.S. and Europe are operating at a lower percentage of their
capacity than U.S. refineries.K Capacity utilization rates at U.S. refineries are well over
90 percent. Historically, if refinery utilization rates approached their maxima, it was
K Europe currently imports diesel fuel and is expected to continue to do so. However, European sulfur caps will
be equivalent to those in the U.S. 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.
5-71
-------
Draft Regulatory Impact Analysis
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 could, and most likely would, 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
would 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 U.S. Consequently, we do not believe U.S. refiners will 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 marginsL 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.
Once refiners have made their investments to meet the proposed NRLM diesel fuel standards,
or have decided to produce high sulfur heating oil, we expect that the various distillate markets
would operate very similar to today's markets. When fully implemented in 2014, there will be
three distillate fuels in the market, 15 ppm highway and nonroad diesel fuel, 500 ppm locomotive
and marine diesel fuel and high sulfur heating oil. The market for 500 ppm locomotive and
marine diesel fuel is much smaller than the other two, particularly considering that it is
nationwide and the heating oil market is geographically concentrated. Therefore, the vast
majority of refiners are expected to focus on producing either 15 ppm or high sulfur distillate,
which is similar to today, where 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. Refiners with only 500 ppm desulfurization
capability can supply two markets. Those refiners only capable of producing high sulfur
distillate would not be able to participate either the 15 or 500 ppm markets. However, this is not
different from today. 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.
LThe 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).
5-72
-------
Fuel Standard Feasibility
Refinery Closure: There are a number of reasons why we do not believe that refineries
would completely close down under this proposed rule. One reason is that we have included a
provisions in the proposed regulations for adjustments to 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
today's proposed requirements. However, those refiners would also presumably be in the best
position to apply for special hardship provisions, presuming that 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 proposed fuel program 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, EPA can request
detailed financial documents 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
usually apply to more than just the operation of a single refinery.
Another factor is the need for capital investments other than for this proposed rule. EPA 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.
5-73
-------
Draft Regulatory Impact Analysis
Some studies in this area point to fuel pricing over the past 15 years or so and conclude that
prices will only increase 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 over the past 15 years and until
recently, there has been a steady decline in the number of refineries operating in the U.S.
However, this may have been due to a couple of circumstances specific to that time period. One,
refinery capacity utilization was less than 80 percent in 1985. Two, at least regarding gasoline,
the oxygen mandate for reformulated gasoline caused an increase in gasoline supply despite low
refinery utilization rates. While this led to healthy financial returns for oxygenate production, it
did not help refining profit margins.
Today, refinery capacity utilization in the U.S. is generally considered to be at its maximum
sustainable rate. There are no regulatory mandates on the horizon which will increase production
capacity significantly, even if ethanol use in gasoline increases substantially.M Consistent with
this, refining margins have been much better over the past two and a half years than during the
previous 15 years and the refining industry itself is projecting good returns for the foreseeable
future.
Conclusions: Therefore, consistent with our findings made during the 2007 highway diesel
rule, we do not expect this proposed rule to cause any supply shortages of nonroad, locomotive
and 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
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
M Both houses of the U.S. Congress are considering bills which 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.
5-74
-------
Fuel Standard Feasibility
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 the implementation of the federal
500 ppm sulfur highway diesel program and the state of California's diesel program circa 1993.41
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 be performed 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, the implementation of the proposed sulfur standards would
likely require that some action be taken to maintain the lubricity of non-highway diesel fuel.
The potential impacts on fuel lubricity from adoption of NRLM sulfur standards that we are
proposing are associated solely with the additional refinery processing that would be necessary to
meet these standards. Although we are proposing to extend the cetane index/aromatics content
specification to NRLM diesel fuel, we do not expect that this would have a significant impact on
fuel lubricity. EPA requires 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 in general is 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 proposed requirement would have an actual impact only on a limited number of
refiners and there would be little overall impact on other diesel fuel qualities (including fuel
lubricity) associated with producing fuel to meet the proposed cetane/aromatic requirement.
Blending small amounts of lubricity-enhancing additives increases the lubricity of poor-
lubricity fuels to acceptable levels. These additives are available in today's market, are effective,
and are in widespread use around the world. Several commenters on our final rule setting a
15ppm 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 would be necessary. However, more testing may be required to determine the
necessary level of biodiesel for fuels not yet being produced, such as the 15ppm fuel being
proposed today.
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 be applicable to all diesel fuel used in any type of engine. Although ASTM has not yet
adopted specific protocols and standards, refiners that supply the US market have been treating
5-75
-------
Draft Regulatory Impact Analysis
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.42 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 the 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 would provide adequate customer protection from engine failures due
to 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 engine fuels, we project that the potential decrease in the lubricity of non-
highway diesel fuel that might result from the adoption of the proposed sulfur standards would 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.43 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 used to produce
highway diesel fuel today would be used to reduce the sulfur content of non-highway diesel
engine fuel to meet the 500 ppm sulfur standard during the first step of the proposed program.
Therefore, we estimate that there would only a marginal increase in the use of lubricity additives
in NRLM diesel fuel meeting the proposed 500 ppm sulfur standard for 2007.
5-76
-------
Fuel Standard Feasibility
The highway diesel program projected that hydrotreating would 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 implementation date for the proposed 15 ppm standard for nonroad
diesel fuel would allow the use of advanced technologies to remove sulfur from 80 percent of the
affected nonroad diesel pool. The use of such developing desulfuriztion processes is discussed
in 5.5 of this Draft RIA. 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 proposed 15 ppm standard for
nonroad diesel fuel, we assumed that the potential impact on fuel lubricity of the new
desulfurization processes would 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 would have to be treated with lubricity additives. The
cost associated with the increased use of lubricity additives in 500 ppm NRLM diesel fuel in
2007 and in 15 ppm nonroad diesel fuel in 2010 is discussed in chapter 7 of this Draft RIA.
5.9.2 Volumetric Energy Content
Some of the desulfurization processes that we project would be used to meet the proposed
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 that 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 that
the potential reduction in VEC which might result from the proposed sulfur standards would
affect the ability of refiners to supply sufficient quantities of non-highway diesel fuel. The
potential impacts on diesel supply are discussed in 5.8 of this Draft RIA.
However, since a greater volume of fuel must be consumed in the engine to produce the same
amount of power, a larger volume of fuel would need to be distributed to meet the same level of
demand. The potential increase in the distribution costs associated with a reduction in non-
highway diesel VEC is discussed in 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 non-highway
diesel fuel with those used to produce highway diesel fuel is contained in 5.2 of this Draft RIA.
Based on this comparison, we believe a comparable level of severity in the desulfurization
process would be required to produce non-highway diesel fuel meeting a given sulfur
specification as would 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 confidential information that we used
to estimate the accompanying reduction in VEC. Using this information, we estimated that
hydrodesulfurization of non-highway diesel fuel to meet a 500 ppm sulfur standard would result
in a reduction in volumetric energy content of 0.7 percent.
5-77
-------
Draft Regulatory Impact Analysis
The proposed 15 ppm sulfur standard for nonroad diesel fuel would not be implemented until
2010. The additional lead-time would allow a number of refiners to take advantage of several
less-expensive desulfurization technologies currently under development to meet the proposed 15
ppm nonroad diesel sulfur standard in addition to hydrodesulfurization (see section 5.3). The
new desulfurization technologies also have less of an impact on diesel fuel volumetric energy
content than does hydrodesulfurization. Using the mix of desulfurization technologies we
project would be available, we estimate that desulfurizing nonroad diesel fuel from 500 ppm to
15 ppm in 2010 as proposed would reduce the volumetric energy content by an additional 0.35
percent. Thus, reducing the sulfur content of nonroad diesel fuel from the current maximum
5,000 ppm sulfur cap to the proposed 15 ppm cap on sulfur content is estimated to result in a 1.1
percent reduction in VEC. The following table (5.9-1) provides a summary of the projections we
used to estimate the impact of the proposed sulfur standards on VEC, including : 1) the
percentage of the applicable non-highway diesel fuel pool that we expect would 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 Proposed Sulfur Standards
Desulfurization Process3
Hydrodesulfurization
S-Zorb Sulfur Adsorption
Linde
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
NR, L, & Mb
500 ppm
in 2007
100 %
NA
NA
-
NR
15 ppm
in 2010
20%
40%
40%
-
Reduction in Volumetric Energy Content
Associated with a Given Desulfurization
Process
Reduction in Sulfur Content
HSC to 500 ppm
0.7%
NA
NA
0.7%
500 ppm to 15 ppm
0.7 %
0.1%
0.4 %
0.4 %
a See section 5.3 of this Draft RIA regarding the use of hydrodesulurization , the Phillips S-Zorb Sulfur Adsorption
process, and the Linde Isotherming process to meet the proposed 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 3400 ppm.
It is important to remember that the anticipated reduction in VEC discussed above would
only apply to those gallons of nonroad diesel fuel that currently have a high sulfur content. Due
to logistical constraints in the fuel distribution system, much of the fuel used in non-highway
engines meets highway diesel fuel standards (see section 7.1 of this Draft RIA). The costs
related to the reduction in non-highway diesel fuel VEC that would accompany the adoption of
the proposed sulfur standards are discussed in section 7.3 of this Draft RIA.
5-78
-------
Fuel Standard Feasibility
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 would not result in an
increase in the use of diesel performance additives. Since that time, we have identified no new
information which would alter that projection. Consequently, our estimate of the increase in
additive use that would result from the adoption of today's proposed rule parallels that under the
highway program. We estimate that the use of lubricity additives would increase, and that the
use of other additives would be unaffected.
5.9.4 Cetane Index and Aromatics
We are proposing that nonroad, locomotive and marine diesel fuel would need to comply
with the current highway diesel fuel requirements for cetane index or aromatics. Thus, these
non-highway diesel fuels would have to meet either a 40 minimum cetane index, or a 35
maximum aromatics limit. In this subsection, we present information on what these properties
are currently for nonhighway 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 during
the years 1997 through 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 1999 -
2001 reports. 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, or 6 percent, were under the cetane index value of 40 and
potentially would not comply with the proposed cetane index minimum of 40. However, those
that were below the 40 cetane index proposed 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 non-highway diesel fuel to comply with the 500 ppm cap
standard in 2007 and then again to comply with the 15 ppm cap standard in 2010, they would be
expected to experience an increase in the cetane levels of their non-highway 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 calculate 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 are applicable to non-highway
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
5-79
-------
Draft Regulatory Impact Analysis
following table. As described in Chapter 7 of the Draft RIA, much of the high sulfur diesel pool
is already hydrotreated (on the order of 50 percent in some PADDs) and would 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 that refiners would
only have to hydrotreat half of their existing high sulfur pool to produce 500 ppm sulfur fuel,
while the upper value reflects that refiners would 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
which provided us the desulfurization information.
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
Linde Isotherming
+2 to +4
+2
+4 to +6
Phillips S-Zorb
Very Small
Very Small
Small
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 cap standard, and 1 to 2 numbers for the 15
ppm sulfur cap 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 suggests, then
the desulfurization of that pool to comply with the 500 ppm sulfur standard, which is expected to
be accomplished using conventional desulfurization technology, is expected to increase the
cetane index to a value above the 40 minimum, thus refiners are not expected to be constrained
by the a cetane index requirement.
Aromatics would also be expected to decrease, although this decrease is expected to occur
mostly through the saturation of polynuclear aromatics to monoaromatics. The biggest decrease
in aromatics is expected by conventional hydrotreating and Linde Isotherming. Phillips S-Zorb
probably only reduces aromatics a minimal amount.
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 in the cost analysis for
compliance with the 500 ppm sulfur cap 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 for compliance with
the 15 ppm sulfur cap standard.
5-80
-------
Fuel Standard Feasibility
Conventional desulfurization and Linde Isotherming are expected to affect the distillation
temperature of non-highway 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) would experience a 5 degree fahrenheit decrease. Consistent with that, API gravity
would be expected to increase by 4 numbers, thus, density would experience a commensurate
decrease. Linde Isotherming is expected to impact the distillation temperature less than
conventional hydrotreating due to the lower API gravity increase caused by Linde compared to
conventional hydrotreating. Phillips S-Zorb would likely not impact the distillation temperature.
5.10 Feasibility of the Use of a Marker in Heating Oil from 2007-2010 and
in Locomotive and Marine Fuel from 2010-2014
We are proposing that the solvent yellow 124 marker be used in heating oil at a concentration
of 6 milligrams per liter from June 1, 2007 through June 1, 2010. The marker would be required
to be added to heating oil at the refinery gate just as visible evidence of the red dye is required
today. Beginning June 1, 2010, the same marker at the same concentration would be required to
be added to locomotive and marine diesel fuel until June 1, 2014. After June 1, 2014, our
proposal would not require the use of a marker. Any fuel with a marker concentration of greater
than 0.1 mg per liter would be precluded from use in NRLM equipment prior to 2010 and NR
equipment after 2010.
Following is a discussion of our evaluation of the feasibility of the use of yellow solvent 124
as the specified fuel marker under the proposed NRLM fuel program and our rationale for
selecting solvent yellow 124. The potential impacts of the proposed marker requirements on the
fuel distribution system are contained in section 5.4 of this draft RIA. The costs associated with
the proposed marker are discussed in section 7.3 of this draft RIA.
The qualification criteria for a marker under the proposed NRLM program include:
1) Solubility in diesel fuel under the range of conditions experienced in the distribution
system from the refinery to the end-user.
2) Not naturally present in diesel fuel
3) Chemical stability under the range of conditions that can be experienced during storage
and distribution of diesel fuel
4) Difficult to remove from fuel or obscure presence to avoid detection
5) Presence in fuel is positively identifiable using laboratory and field tests
6) Detectable in very small concentrations to reveal mixtures of marked and unmarked fuels
7) Economic acceptability, ready availability, and ease of application
8) No increased public health risk
There are a number of types of dyes and markers. Visible dyes are most common, are
typically least expensive, and are easily detected in the field. Laboratory tests are often also
available for such dyes to quantify the concentration of the dye present in fuel. This is the case
with red dye 164 which is required by the U.S. IRS to be present in non-taxed diesel fuel at a
5-81
-------
Draft Regulatory Impact Analysis
minimum level that is spectrally equivalent to 3.9 pounds per thousand barrels (ptb) of the IRS-
specified standard solvent red 26 dye. The longtime presence of a number of visible dyes in fuels
means that their acceptability under the above qualification criteria has been well established.
However, using a second diesel dye for segregation of heating oil based on visual
identification is problematic. Most dye colors that provide a strong visible trace in fuels are
already in use for different fuel applications. More importantly, mixing two fuels containing
different strong dyes can result in interference between the two dyes rendering identification of
the presence of either dye difficult. Yet, the mixing of nonroad diesel fuel into heating oil for
eventual sale as heating oil would be an acceptable and often an economically desirable practice.
Furthermore, to avoid interfering with the IRS tax code, it would be advantageous to maintain the
current red color. Based on these considerations, we believe that the use of a second strong dye
to visibly segregate heating oil from NRLM is not practicable.
Fuel markers that do not depend on a visible trace for detection are beginning to see more use
in branded fuels. Invisible markers are typically somewhat more expensive than visible markers.
Soluble dyes have also been used at a concentration too low to allow reliable detection of their
presence visually but that does allow detection by other means. When a dye is used in this
fashion it is more appropriate to refer to it as a marker, since its functionality is not associated
with the slight color it may impart to the fuel. Fuel markers typically have a simple method to
detect the presence of the marker in the field and a more rigorous method the quantify the
concentration of the marker in the fuel which must be conducted in the laboratory. Such
laboratory methods are favored for developing strong evidence of noncompliance for use in
enforcement actions.
Depending on the marker type, detection in the field is accomplished 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 more suitable for use in the field than others. For
example, some are more suited for laboratory use due to the complexity of the detection process
or concerns regarding the toxicity of the reagents used to reveal the presence of the marker.
Ideally, after conducting a field test for the presence of the marker and finding the fuel to be
compliant, the inspector returns the fuel sample to the fuel batch or otherwise ensures that it is
used for the intended fuel purpose. This practice avoids the difficulty associated with disposing
of the fuel sample. For most types of field tests for markers, however, this practice is not be
possible. The introduction of the reagent to the test fuel sample typically makes returning the
returning the fuel sample for its intended use impossible, and it must be disposed of by other
means. The toxicity of the by-products from testing can also be a concern. Chemical-based field
tests are typically inexpensive. However, if such tests produce toxic by-products, the cost of
disposing of such by-products can be significant. In addition, there are public health concerns
related to the potential improper disposal of such by-products.
Near-infra-red and ultra-violet flourescent markers can be easily detected in the field using a
small device that requires only brief training for the operator and leaves that sample unaffected.
Therefore, concerns regarding test reagents and by-products are not an issue and the fuel sample
5-82
-------
Fuel Standard Feasibility
can be returned to the fuel batch. However, the devices used in the field detection of such
markers can be more expensive..
There are also more exotic markers available such as based on immunoassay, and isotopic or
molecular enhancement. Such markers typically can only be detected by laboratory analysis and
are more expensive than the markers discussed above. Because of the lack of a easy field test,
we believe that further consideration of the use of such markers for the proposed purpose is not
warranted.
The Euromarker:
Effective in August 2002, the European Union (EU) enacted a marker requirement for diesel
fuel that is taxed at a lower rate (which applies in all of the EU member states).44 The marker
selected by the EU is N-ethyl-N-[2-[l-(2-methylpropoxy)ethoxyl]-4-phenylazo]-benzeneamine.45
This compound is also referred to as solvent yellow 124 or the Euromarker. The treatment level
required by the EU is 6 milligrams per liter. Despite its name, solvent yellow 124 does not
impart a strong color to diesel fuel when used at the proposed concentration. The EU allows its
member states to choose which visible dye to use in lower-taxed fuel in addition to the
Euromarker. A number of countries in the EU use a red dye.46 The Euromarker imparts a slight
orange shade to fuel that is dyed red. However, experience of the EU members has shown the
fuel containing red dye and the Euromarker is still recognizable as red dyed diesel fuel.47 The
specific type of red dye used in Europe is not the same type used in the U.S. Nevertheless, we
believe that the experience of EU member states that the Euromarker does not interfere with the
identification of the presence of strong red dyes in diesel fuel is sufficiently predictive of its
potential impact on the color that the IRS red dye impart to diesel fuel. Therefore, we do not
expect that the presence of solvent yellow 124 in diesel fuel that contains the IRS-specified red
dye would interfere with the use of the red dye by IRS to identify non-taxed fuels.
Solvent yellow 124 is substantially similar to diesel fuel and is registered under EPA's Fuel
and Fuel Additive program which evaluates an additive's suitability for use based on the
potential effects on human health and vehicle emissions performance. In addition, extensive
evaluation and testing of the Euromarker was conducted by the European Commission. This
included combustion testing which showed no detectable difference between the emissions from
marked and unmarked fuel. We also understand that Norway specifically evaluated the use of
distillate fuel containing the Euromarker for heating purposes and determined that the presence
of the Eurmarker did not cause an increase in harmful emissions from heating equipment. Based
on the European experience with the Euromarker, we do not expect that there would be concerns
regarding the compatibility of the Euromarker in the U.S. fuel distribution system or for use in
motor vehicle engines and other equipment such as in residential furnaces. The European Union
intends to review the use of Euromarker after December 2005, or 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 the Euromarker depending on the European
Union's findings.
5-83
-------
Draft Regulatory Impact Analysis
Fuel additives are typically required to be tested for their suitability for use under the unique
conditions present in aircraft engines and fuel supply systems before they are allowed to be
present in jet fuel. Due to safety concerns, jet fuel is held to very strict standards regarding the
allowable presence of contaminants and additives that are not specifically allowed for use in jet
fuel. For example, the Department of Defense (DoD) maintains a zero-tolerance for any
contamination of jet fuel with the red dye required by the IRS (and EPA). Given their past
experience with red dye, DoD raised concerns regarding the extent to which jet fuel might
become contaminated with solvent yellow 124 due to the presence of solvent yellow 124-
containing fuels and jet fuel in the U.S. common carrier pipeline distribution system, and whether
any such contamination might be cause for concern.
We do not believe that there any significant pathways for such contamination to take place
other than by potential human error. In addition, the fact that the fuel distribution industry in the
U.S. has been successful in managing contamination of jet fuel with red dye indicates that the
potential contamination of jet fuel with the solvent yellow 124 can also be successfully managed
in the US fuel distribution system. Therefore, we believe that our proposed use of solvent yellow
124 should not pose a significant risk to the maintenance of jet fuel purity.
There is currently no official procedure recognized by the European Union to quantify the
presence of the Euromarker in distillate fuels. The most commonly accepted method used in the
European Community is based on the chemical extraction of the Euromarker using hydrochloric
acid solution and cycloxane, and the subsequent evaluation of the extract using a visual
spectrometer to determine the concentration of the Euromarker.48 This test is inexpensive and
easy to use for field inspections. However, the test involves reagents that require some safety
precautions and the small amount of fuel required in the test must be disposed of as hazardous
waste. Nevertheless, we believe that such safety concerns are manageable here in the U.S. just as
they are in Europe and that the small amount of waste generated can be handled along with other
similar waste generated by the company conducting the test, and that the associated effort/costs
would be negligible.
Similar to the approach proposed regarding the measurement of fuel sulfur content, we are
proposing a performance based procedure to measure the concentration of solvent yellow 124 in
distillate fuel. Under the performance-based approach, a given test method could be approved
for use in a specific laboratory or for field testing by meeting certain precision and accuracy
criteria. There would be no designated marker test method. Properly selected precision and
accuracy values potentially would allow multiple methods and multiple commercially available
instruments to be approved, thus providing greater flexibility in method and instrument selection
while also encouraging the development and use of better methods and instrumentation in the
future. For example, we are hopeful that with more time and effort a simpler test can be
developed that can avoid the use of reagents and the generation of hazardous waste that is by
product of the current commonly accepted method.
In developing the precision and accuracy criteria for the sulfur test method, EPA drew upon
the results of an interlaboratory study conducted by the American Society for Testing and
5-84
-------
Fuel Standard Feasibility
Materials (ASTM) to support ASTM's standardization of the sulfur test method. Unfortunately,
there has not been sufficient time for industry to standardize the test procedure used to measure
the concentration of solvent yellow 124 (Euromarker) in distillate fuels or to conduct an
interlaboratory study regarding the variability of the method. Nevertheless, the European Union
has been successful in implementing its Euromarker requirement while relying on the marker test
procedures which are currently available. As referenced above, the most commonly accepted
method used in the European Union is a visual spectrometer-based procedure. We are proposing
to use this procedure to establish the precision and accuracy criteria on which a marker test
procedure would be approved under the performance based approach..
There has been substantial experience in the use of the proposed reference market test
method since the August 2002 effective date of the European Union's Euromarker requirement.
However, EPA is aware of only limited summary data on the variability of the reference test
method from a manufacturer of the visible spectrometer apparatus used in the testing.49 The
stated resolution of the test method from in the materials provided by this equipment
manufacturer is 0.1 mg/L, with a repeatability of plus or minus 0.08 mg/L and a reproducibility
of plus or minus 0.2 mg/L. In the lack of more extensive data, we are proposing to use these
available data as the basis of our proposed precision and accuracy criteria as discussed below.
The referenced repeatability and reproducibility are terms related to test variability used by
ASTM in defining their voluntary consensus test standards. ASTM defines repeatability as the
difference between successive results obtained by the same operator with the same apparatus
under constant operating conditions on identical test materials that would, in the long run, in the
normal and correct operation of the test method be exceeded only in one case in 20.
Reproducibility is defined by ASTM as the difference between two single and independent
results obtained by different operators working in different laboratories on identical material that
would, in the long run, be exceeded only in one case in twenty.
The first qualification criterion, precision, refers to the consistency of a set of measurements
and is used to determine how closely analytical results can be duplicated based on repeat
measurements of the same material under prescribed conditions. To demonstrate the precision of
a given marker test method under the performance-based approach, a laboratory facility would
perform 20 repeat tests over several days on samples taken from a homogeneous supply of a
commercially available diesel fuel that contains the marker. Using a similar methodology to that
employed in deriving the proposed sulfur test procedure precision value, results in a precision
value for the marker test procedure of 0.043 mg/L. This value was derived as follows:
0.43 mg/L is equal to 1.5 times the standard deviation (0.029) where the standard deviation is
equal to the repeatability of the reference test method (0.08 mg/L) divided by 2.77. Since the
conditions of the precision qualification test admit more sources of variability than the conditions
under which ASTM repeatability is determined (longer time span, different operators,
environmental conditions, etc.) the repeatability standard deviation derived from the repeatability
value was multiplied by what we believe to be a reasonable adjustment factor, 1.5 to compensate
for the difference in conditions.
5-85
-------
Draft Regulatory Impact Analysis
We are concerned that the use of the 0.043 mg/L precision value derived above, because it is
based on very limited data, might preclude the acceptability of test procedures that would be
adequate for the intended regulatory use. In addition, the lowest measurement of marker
concentration that would have relevance under the proposed regulations is 0.1 mg per liter.
Consequently, we are proposing that the precision of a test procedure would need to be less than
0.1 mg/L for it to qualify.
The second criterion, accuracy, refers to the closeness of agreement between a measured or
calculated value and the actual or specified value. To demonstrate the accuracy of a given test
method under the performance-based approach, a laboratory facility would be required to
perform 10 repeat tests, the mean of which could not deviate from the Accepted Reference Value
(ARV) of the standard by more than 0.05 mg/L. We believe that the proposed accuracy level is
not overly restrictive, while being sufficiently protective considering that the lowest marker level
of regulatory significance would be 0.1 mg/L.
These tests would be performed using commercially available solvent yellow 124 standards.
Ten tests would be required using each of two different marker standards, one in the range of 0.1
to 1 mg/L and the other in the range of 4 to 10 mg/L of solvent yellow 124. We selected the two
ranges of the marker standards to cover the two marker concentrations that are of most regulatory
concern: 6 mg/L is the minimum marker concentration required in fuels that we are proposing
must contain the marker, while 0.1 mg/L is the maximum allowed concentration for fuel to be
considered as not containing the fuel marker for the purposes of the fuel use restrictions on which
the fuel marker requirements are based.
We believe that these precision and accuracy criteria would limit the allowed test procedures
to those capable of satisfying the intended use for enforcement and affirmative defenses to
presumptive liability purposes, while not being overly restrictive.
Solvent yellow 124 is marketed by several manufactures and is in current wide-scale use in
the European community. We anticipate that these manufactures would have sufficient lead-time
to increase their production of solvent yellow 124 to supply the increase in demand that would
result from the proposed marker provisions.
The proposed treatment rate would ensure adequate detection in the distribution system even
if diluted by a factor of 50. Removal of the marker is possible through an expensive laundering
process. However, we believe that there would be little economic incentive to attempting to the
remove the marker in the United States given that its removal would only allow the use of the
fuel in other nontaxed applications. Even if the marker were removed, the IRS red dye would
still be present to prevent the use of the fuel for highway (taxed) purposes.
Other Potential Candidate Fuel Markers:
We considered other potential markers that might be used to identify and segregate heating
oil from NRLM fuel. One of the potential alternatives that we identified is the Clir-Code®
marker system manufactured by ISOTAG Technologies Inc. The Clir-Code® marker system has
5-86
-------
Fuel Standard Feasibility
been used extensively in U.S. fuel and includes a field test that employs a hand-held near infra-
red detector which does not require the use of any reagents. EPA deferred proposing the use of
the Clir-Code® marker because we believe that the advantage of a simpler field test would not
compensate for the increased treatment cost relative to the use of the Euromarker
5-87
-------
Draft Regulatory Impact Analysis
Appendix 5A: EPA's Legal Authority for Proposing Nonroad,
Locomotive, and Marine Diesel Fuel Sulfur Controls
We are proposing 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 a
nonroad 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.
We are proposing controls on sulfur levels in nonroad 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 500
ppm for this fuel will lead to significant, cost-effective reductions in emissions of these
pollutants. 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 proposed emission standards. Chapter 4 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 significant, cost-
effective reduction in emissions of these pollutants. 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
5-88
-------
Fuel Standard Feasibility
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 proposed first step in the reduction in the sulfur level of nonroad, locomotive, and
marine diesel fuel to 500 ppm would achieve approximately a 90 percent reduction in the
emissions of SO2 and sulfate PM emissions from nonroad, locomotive, and marine diesel
engines compared to today's levels. The proposed second step of nonroad sulfur control to 15
ppm (and the control of locomotive and marine diesel fuel also being considered) would achieve
in excess of a 99 percent reduction in these pollutants. The rationale for the two-step approach to
fuel sulfur control is 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.
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.
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
5-89
-------
Draft Regulatory Impact Analysis
under section 202 prior to establishing controls or prohibitions on fuels or fuel additives under
section 211(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)(B) 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)(l)(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 proposing stringent new standards for nonroad diesel engines under section 213.
The nonroad, locomotive and marine diesel sulfur standards of 500 ppm proposed today
represent an appropriate exercise of the Agency's discretion under section 21 l(c)(l)(A). The 500
ppm level is consistent with current highway diesel fuel (until 2010) and adopting the same level
for nonroad, locomotive, and marine diesel fuels avoids costs associated with more grades of fuel
in the distribution system. The 500 ppm level also will achieve significant and cost-effective
environmental benefits, providing approximately 90 percent of the sulfate PM and SO2 benefits
associated with control to 15 ppm. It also allows for a short lead time for implementation,
enabling the environmental benefits to begin as soon as possible.
5A.2 Impact of Diesel Sulfur Emission Products on Emission-Control
Systems
EPA is also proposing to restrict 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. EPA believes that sulfur in nonroad diesel fuel would significantly impair the emission-
control technology of nonroad engines designed to meet the proposed emission standards. We
know that diesel sulfur has a negative impact on engine emission controls. This is not a new
development. As discussed in Chapter 4, we believe existing aftertreatment technologies will be
capable of achieving dramatic reductions in NOx and PM emissions from nonroad engines for
the 2009 time frame. 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, EPA believes the lead time between now and 2011 will provide
ample opportunity to adapt this technology for feasible operation on nonroad engines. EPA
believes these aftertreatment technologies would be in general use by 2009 and 2011,
respectively, with the diesel sulfur controls proposed in this rule.
5-90
-------
Fuel Standard Feasibility
These aftertreatment technologies are ineffective in reducing NOx and PM emissions and
incapable of being introduced widely into the marketplace at the nonroad diesel sulfur
concentrations typical today or less. Not only does their efficiency at reducing NOx and PM
emissions fall off dramatically at elevated fuel sulfur concentrations, but engine operation
impacts and permanent damage to the aftertreatment systems are also possible. In order to ensure
regeneration of the diesel paniculate filter at exhaust temperatures typical of nonroad diesel
engines as described in Chapter 4, we are expecting that significant amounts of 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 particulate 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 SOS,
forming a precursor to sulfate particulate matter. With respect to NOx aftertreatment, all of the
NOx aftertreatment technologies discussed in Chapter 4 that EPA believes will generally be
available to meet the proposed 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 absorber, 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. 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 guild 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.
5A.3 Sulfur Levels that Nonroad Engines Can Tolerate
As discussed in Chapter 4, there are three key factors which when 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
proposed emission standards. These factors, as discussed in more detail in Chapter 4, are the
implications sulfur levels in excess of 15 ppm would have on the efficiency and reliability of the
systems and their impact on the fuel economy of the engine.
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
proposed today, unless nonroad diesel fuel sulfur levels are below 15 ppm. When sulfur is kept
5-91
-------
Draft Regulatory Impact Analysis
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 could
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 the reliability of
these technologies can be raised to the point where they become feasible for successful use by
nonroad engines. 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, 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 nonroad
diesel fuel, the greater this adverse impact on fuel economy. As sulfur levels increase above 15
ppm the fuel economy impact transitions quickly 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 paniculate 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 believe that 15 ppm represents an upper threshold of
acceptable nonroad diesel fuel sulfur levels and are therefore proposing to cap in-use sulfur levels
there.
5A.4 Sulfur Sensitivity of Other Catalysts
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 proposed
nonroad standards would be significantly impaired by operation on high sulfur 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 proposed control (Chapter
5-92
-------
Fuel Standard Feasibility
3), an analysis of the costs and the technological feasibility of controlling sulfur to the proposed
levels (Chapter 7), and a cost-effectiveness analysis of the proposed sulfur control and nonroad
emission standards (Chapter 8). Under section 21 l(c)(2)(B), 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 proposed 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 standards proposed today 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, all of the aftertreatmrent technologies that could be used to meet
the PM or NOx standards are significantly impaired by the sulfur in diesel fuel. For PM control,
EPA is not aware of a PM aftertreatment technology that is capable of meeting the PM standard
adopted today and that would not 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 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 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. Significant and widespread changes to the fuel distribution system
infrastructure would have to be made and in place by then, and there is no practical expectation
that this would occur, with or without the low sulfur standard adopted today. 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, 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. In addition, even if
SCR were used to obtain the emission performance required by today's standards, it is not clear
that the vehicles would continue to maintain that level of performance in-use. Finally, for NOx
control, both NOx adsorbers and compact SCR are significantly impaired by sulfur in diesel fuel,
and both technologies would need very large reductions in sulfur from current levels to meet the
NOx standard adopted today. 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 require
significant additional or different EPA fuel control regulations before the technology could be
considered generally available.
5-93
-------
Draft Regulatory Impact Analysis
In sum, EPA believes that both PM and NOx aftertreatment technologies require the level of
sulfur control adopted today to meet the PM standards. There is no PM or NOx emission-control
device or system that would be in general use that does not need this level of sulfur control for
purposes of controlling PM. EPA also believes that the only NOx aftertreatment technologies
that would be considered generally available for use to meet the NOx standard, need the level of
sulfur adopted today to be considered generally available for use to meet the NOx standard.
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 low sulfur
nonroad diesel fuel. 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 today. 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 today based
on what would be technologically feasible with current or 500 ppm sulfur levels.
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 proposing to prohibit sulfur
in nonroad, locomotive or marine fuel, but rather to control the levels of sulfur in these diesel
fuels, this finding is not required prior to regulation. However, EPA does not believe that the
proposed 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 in the case of unleaded gasoline in the past where lead was providing a primary
function in 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. If it were not for the
fact that it costs money to remove sulfur from diesel fuel, it would have been removed 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 to suggest that as sulfur is removed from
diesel fuel the natural lubricity characteristics of diesel fuel may be reduced. Depending on the
5-94
-------
Fuel Standard Feasibility
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 and 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. As we move
forward with this rulemaking and its implementation we will continue to investigate this issue,
and welcome any comment on it.
5-95
-------
Draft 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", ASTMD 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.
5-96
-------
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."
IS.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.
19.Centinel Hydroprocessing Catalysts: A New Generation of Catalysts for High-Quality Fuels,
Criterion Catalysts and Technologies Company, October 2000.
20.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.
21. 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.
22.Wilson, R., "Cost Curves for Conventional HDS to Very Low Levels," February 2, 1999.
23. "Processes for Sulfur Management," IFF.
24.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.
25.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.
26.Kidd, Dennis, S-Zorb - Advances in Applications of Phillips S-Zorb Technology, Presented at
the NPRA Q & A meeting, October 2000.
27.Chapados, Doug, Desulfurization by Selective Oxidation and Extraction of Sulfur-Containing
Compounds to Economically Achieve Ultra-Low Proposed Diesel Fuel Sulfur Requirements,
Paper presented at the 2000 NPRA Annual Meeting.
28. 55 FR 34138, August 21, 1990.
29.Refming Industry Profile Study; EPA contract 68-C5-0010, Work Assignment #2-15, ICF
Resources, September 30, 1998.
5-97
-------
Draft Regulatory Impact Analysis
30.Fuel Quality Regulations for Highway Diesel Fuel Sold in 1993 and Later Calendar Years,
Final Rule, 55 FR 34120, August 21, 1990
31 .Memorandum to the docket entitled "Diesel Products Carried by Bulk Plants Under the
Proposed Non-Highway Sulfur Program"
32.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
33.Regulatory Impact Analysis for the Highway Diesel Final Rule, EPA Air Docket A-99-06
34.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.
35.Memorandum to the docket entitled "Diesel Products Carried by Bulk Plants Under the
Proposed Non-Highway Sulfur Program"
36. 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.
37. National Petroleum Council, " U.S. Petroleum Assuring Adequacy and Affordability of
Cleaner Fuels", June 2000 pages 118-133.
38. 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, EPA420-R-99-023.
39. 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, EPA420-R-99-023.
40. 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.
41. 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
lubricity in low sulfur fuels, EPA Air docket A-99-06.
42. Letter from L. Erlandsson, MTC AB, to Michael P. Walsh, dated October 16, 2000. Docket
A-99-06, item IV-G-42.
43. 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
5-98
-------
Fuel Standard Feasibility
nation. http://www.chevron.com/prodserv/fuels/bulletin/diesel/12%5F7%5F2%5Frf.htm
44. The EU marker legislation, 2001/574/EC, document C(2001) 1728, was published in the
European Council Official Journal, L203 28.072001.
45. 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.
46. European Refining and Marketing. Voume 1, Number 10, November 15, 2002, "Fuels
Refining and Marketing in Europe and the Former Soviet Union", pp!6, "Marking of Fuel",
Naptha Publications Ltd.
47. TOTALFINAELF, Safety and Environmental Update, "The Euromarker Arrives in the UK
(United Kingdom) - Your Questions Answered", July 2002
48.Memorandum to the docket entitled "Use of a Visible Spectrometer Based Test Method in
Detecting the Presence and Determining the Concentration of Solvent Yellow 124 in Diesel
Fuel"
49.Technical Data on Fuel/Dye/Marker & Color Analyzers, as downloaded from the Petroleum
Analyzer Company L.P. website at
http://www.petroleum-analyzer.com/product/PetroSpec/lit_pspec/DTcolor.pdf
5-99
-------
CHAPTER 7: Estimated Costs of Low-Sulfur Fuels
1.1 Nonroad Fuel Volumes 7-1
7.1.1 Overview 7-1
7.1.2 Diesel Fuel Demand by PADD for 2000 7-2
7.1.2.1 Highway Diesel Fuel Volumes and Highway Spillover 7-2
7.1.2.2 Land-Based Nonroad Fuel Volumes 7-7
7.1.2.3 Locomotive Diesel Fuel Demand 7-15
7.1.2.4 Marine Diesel Fuel Demand 7-16
7.1.2.5 Remaining Non-Highway Diesel Fuel Demand 7-17
7.1.2.6 Summary of Diesel Fuel Demand for 2000 7-18
7.1.3 Diesel Fuel Demand by PADD for 2008 7-20
7.1.3.1 2000-2008 Growth Factors 7-20
7.1.3.2 Division of Low-Sulfur Diesel Fuel into 15 ppm and 500 ppm Volumes 7-21
7.1.3.3 Summary of Diesel Fuel Demand for 2008 7-22
7.1.4. Annual Diesel Fuel Demand (2000-2040) and Associated In-Use Sulfur Levels 7-24
7.1.4.1 Annual Diesel Demand Volume Estimates 7-24
7.1.4.2 In-Use Diesel Sulfur Concentrations 7-25
7.1.4.3 Summary of Annual Diesel Fuel Demand and Sulfur Levels 7-30
7.1.5 Refinery Supply Volumes 7-41
7.2 Refining Costs 7-48
7.2.1 Methodology 7-48
7.2.1.1 Overview 7-48
7.2.1.2 Basic Cost Inputs for Specific Desulfurization Technologies 7-49
7.2.1.3 Composition of Distillate Fuel by Refinery 7-76
7.2.1.4 Summary of Cost Estimation Factors 7-82
7.2.1.5 How Refiners are Expected to Meet the Nonroad Sulfur Requirements 7-89
7.2.2 Refining Costs 7-103
7.2.2.1. 15 ppm Highway Diesel Fuel Program 7-103
7.2.2.2 Costs for Proposed Two Step Nonroad Program 7-105
7.2.2.3 15 ppmNonroad Diesel Fuel with Conventional Technology 7-116
7.2.2.4 Refining Costs for Alternative NRLM Fuel Programs 7-118
7.2.2.5 Capital Investments by the Refining Industry 7-121
7.2.2.6 Other Cost Estimates for Desulfurizing Highway Diesel Fuel 7-124
7.3 Cost of Distributing Non-Highway Diesel Fuel 7-130
7.3.1 Distribution Costs Under the 500 ppm Sulfur Non-Highway Diesel Fuel Program 7-130
7.3.1.1 Fuel Distribution-Related Capital Costs Under the 500 ppm Sulfur Non-Highway Diesel Fuel
Program 7-130
7.3.1.2 Distribution Costs Due to the Reduction in Fuel Volumetric Energy Content Under the Proposed
500 ppm Sulfur Diesel Fuel Program 7-131
7.3.1.3 Other Potential Distribution Costs Under the Proposed 500 ppm Sulfur Diesel Fuel Program 7-132
7.3.2 Distribution Costs Under the 15 ppm Sulfur Nonroad Diesel Fuel Program 7-132
7.3.2.1 Fuel Distribution-Related Capital Costs Under the 15 ppm Sulfur Nonroad Diesel Fuel Progr3nl32
7.3.2.2 Distribution Costs Due to the Reduction in Fuel Volumetric Energy Content Under the 15 ppm
Sulfur Nonroad Diesel Fuel Program 7-132
7.3.2.3 Other Potential Distribution Costs Under the 15 ppm Sulfur Nonroad Diesel Fuel Program 7-133
7.3.3 Cost of Lubricity Additives 7-134
7.3.4 Fuel Marker Costs 7-135
7.3.5 Distribution, Lubricity, and Marker Costs Under Alternative Sulfur Control Options 7-135
7.4 Net Cost of the Two-Step Nonroad Diesel Fuel Program 7-138
7.5 Potential Fuel Price Impacts 7-139
Appendix 7A: Estimated Total Off-Highway Diesel Fuel Demand and Diesel Sulfur Levels 7-144
Appendix 7B: Land-Based Nonroad Engine Growth Rate Based on Annual Energy Outlook 2002 7-153
-------
Estimated Costs of Low-Sulfur Fuels
CHAPTER 7: Estimated Costs of Low-Sulfur Fuels
7.1 Nonroad Fuel Volumes
7.1.1 Overview
This section describes how the estimates of diesel fuel demand for land-based nonroad
engines, locomotives, and marine vessels, which will be directly affected by the proposed rules,
were determined. Volumes are provided for various geographic regions of interest. The
discussion focuses on how these volumes were developed for 2000 and 2008, and then describes
how the estimates for other years were produced. This section also describes diesel fuel supply
volumes for 2008, which are used in the economic assessment.
Of course, only the amount of high-sulfur fuel used by land-based nonroad engines,
locomotives, and marine vessels will be directly affected by today's proposal. In this analysis,
the basic approach to estimating this fuel volume is to: 1) find the total diesel fuel demand in
each category, 2) determine the respective amount of this fuel which already meets the highway
fuel standards, and 3) subtract the low-sulfur volume from the total diesel fuel demand to yield
the volume of high-sulfur diesel in the category.
Estimating diesel fuel consumption for the engine categories covered by the proposal also
requires a basic understanding of the fueling practices for non-highway equipment. Generally,
these equipment types are capable of using either high-sulfur diesel fuel or low-sulfur fuel that
complies with the EPA highway diesel sulfur regulations. This latter fuel type may be used in
non-highway applications for a variety of reasons. First, some equipment may be refueled at
service stations where only low-sulfur, highway compliant fuel is available. Second, high-sulfur
fuel may not be available due to limitations in the distribution or storage systems in some areas
or during certain times of the year. Third, operators may choose to use low-sulfur diesel fuel
based on some real or perceived benefit such as improved engine durability.
The estimates of diesel fuel volumes used in this analysis are principally based on the Fuel
Oil and Kerosene Sales 2000 (FOKS) report, which is produced by the Energy Information
Administration (EIA).1 This report represents the most detailed, comprehensive distillate fuel
demand study available. The report contains estimates of distillate fuel sales for highway
vehicles and 10 non-highway end uses. Unfortunately, the values reported in FOKS for the non-
highway categories can not be used directly in this analysis, because it does not always report
fuel volumes into the specific equipment types or diesel fuel grades that will be affected by the
proposed rules.
As explained in detail in the next section, EPA in consultation with EIA identified six of the
broadly reported categories in the EIA FOKS report as being relevant to this analysis. In
addition, EPA found that EIA's railroad category contained distillate fuel used in both land-based
nonroad engines, e.g., rail maintenance equipment and locomotives. Finally, EPA identified
7-1
-------
Draft Regulatory Impact Analysis
EIA's vessel bunkering category as containing both recreational and commercial distillate fuel.
The categories and end uses of interest from the EIA FOKS report are generally shown in Table
7.1-1.
Table 7.1-1
Application of EIA FOKS End Use Categories to EPA Off-Highway Categories
FOKS Category
Farm
Other Off-Highway
Construction
Other
Industrial
Commercial
Oil Company
Military
Railroad
Vessel Bunkering (Marine)
EPA Proposal Categories
Land-Based Nonroad
X
X
X
X
X
X
X
X
Locomotives
X
Marine
X
Each of these topics is discussed in detail in the remainder of this section, along with the
resulting estimates of high-sulfur diesel fuel that would be affected by the proposed rules.
7.1.2 Diesel Fuel Demand by PADD for 2000
High-sulfur diesel fuel is calculated by subtracting the low-sulfur diesel fuel demand from the
total diesel fuel demand in the respective category. A common element in determining the
volume of low-sulfur fuel is the amount, or percentage of low-sulfur, highway compliant diesel
fuel that is spilled over into each of the non-highway end-use categories. Therefore, this section
begins by identifying the amount of spillover for the various end uses of interest, and progresses
to applying that information to estimate the volume of high-sulfur diesel fuel in each of the end-
use categories.
7.1.2.1 Highway Diesel Fuel Volumes and Highway Spillover
Spillover is defined as the total volume of low-sulfur, highway compliant fuel supplied into
the U.S. minus the volume of this fuel that is consumed (i.e., demand) by highway vehicles. The
of volume of highway compliant fuel supplied to each PADD is provided in the Petroleum
Supply Annual 2000, which is published by the Energy Information Agency.2 The values from
7-2
-------
Estimated Costs of Low-Sulfur Fuels
that report have been converted from barrels to gallons using a conversion factor of 42 gallons
per barrel. The volume of highway fuel demand is provided in the EIA FOKS report.
Table 7.1-2 shows the spillover volumes in each of the five PADDs based on the above
information.
Now that the total volume of low-sulfur diesel spillover is known, the next step in
determining the low-sulfur spillover percentage is to find the total volume of diesel fuel
consumed by all non-highway end-uses. The EIA FOKS report provides distillate sales numbers
for the various off-highway end-use categories that could contain spillover fuel. Some of the
distillate fuel grade categories contained in the report are quite broad in scope, making it difficult
to accurately determine only the fuel volumes that are clearly interchangeable with the diesel fuel
grades affected by the proposed rules. For example, certain end-use categories report distillate
fuel oil or total distillate. These specifications may contain incompatible fuel types such as No. 4
fuel oil that is used in commercial burner applications. When more specific fuel grade
information was unavailable, the volumes for these broader specifications are used to determine
the total "potential" volume of non-highway fuel consumption. Fortunately, the volumes of these
broad specification distillate fuels are relatively small compared to the total volumes of better
defined diesel fuel grades. A detailed table showing how the potential non-highway diesel fuel
volumes were determined is shown in Appendix 7A. The relevant fuel demand volumes are
summarized in Table 7.1-3.
7-3
-------
Table 7.1-2
Highway Diesel Fuel Spillover Volumes by PADD (million gallons)
Highway Diesel
Category
Supply
Demand
Spillover
1
11,257
10,228
1,029
2
12,939
11,141
1,799
3
6,947
5,644
1,303
4
2,213
1,475
738
5
5,892
4,643
1,250
5
AZ,NV,
OR, WA
NA
NA
NA
5
CA
2,633
2,633
0
5
AK
NA
NA
NA
5
HI
NA
NA
NA
NA = Spillover volume is not used to determine the spillover percentages for these areas as explained later in Section 7.1.2.1.
7-4
-------
Table 7.1-3
Potential Non-Highway Diesel Fuel Demand by PADD (million gallons)
End Use
Residential
Commercial
Industrial
Oil Company
Farm
Electric Utility
Railroad
Marine
Military
Construction
Other
Total
PADD
1
5,399
1,944
617
19
433
305
500
490
70
511
159
10,447
2
629
567
598
42
1,612
134
1,233
301
36
549
59
5,760
3
1
347
418
561
552
195
686
1,033
9
394
123
4,319
4
39
13
241
29
221
9
345
0
4
150
30
1,171
5
137
213
236
34
351
151
307
256
113
295
60
2,153
5
AZ, NV, OR,
WA
82
97
176
2
89
17
114
62
89
91
31
849
5
CA
7
87
45
6
254
8
189
101
7
194
22
921
5
AK
48
26
14
26
0
36
4
80
6
7
7
254
5
HI
0
3
1
0
8
90
0
13
11
3
0
129
7-5
-------
Draft Regulatory Impact Analysis
The low-sulfur spillover percentages for the PADDs are calculated by dividing the total
spillover volume in each PADD by the respective total potential non-highway demand volume.
We use the demand volumes for all non-highway categories for this calculation, in the absence of
information indicating that spillover fuel is used differentially in any non-highway end-use
categories. This implicitly assumes that each spillover gallon has an equal chance of being sold
for use in any non-highway application within each PADD.A The resulting low-sulfur spillover
fractions for each of the five PADDs are shown in Table 7.1-4.
Table 7.1-4
Highway Diesel Fuel Spillover Percentages by PADD
Diesel Fuel
Category
Highway Spillover
(million gallons)
Potential Off-
Highway
(million gallons)
Spillover (%)
PADD I
1,029
10,447
10
PADD II
1,799
5,760
31
PADD III
1,303
4,319
30
PADD IV
738
1,171
63
PADDV
1,250
2,153
58
For PADD 5, it was necessary to develop separate refining regions within PADD 5 for the
refinery cost analysis. For this reason, separate spillover percentages were estimated for these
separate PADD 5 subregions. This was accomplished by first estimating the spillover
percentages of states which are known to have specific spillover characteristics. The State of
California already regulates the sulfur content of both the highway and nonroad diesel fuel pools
to 500 ppm, thus very little of the diesel fuel is currently unregulated by the State. The tendency
is that as more of the fuel pool is regulated, the higher the percentage of spillover into the non-
highway diesel fuel pool as the distribution system has little tolerance for small volumes of high
sulfur fuels. This was confirmed by talking to a staff member within California's fuel regulatory
division of the Air Resources Board. Based on this conversation, California's spillover fraction
was estimated to be 100 percent. At the other end of the spectrum, Alaska's highway volume is
much smaller than the non-highway volume, thus, very little spillover is expected. Using PADD
1 as a guide, which has about 10 percent spillover and a higher ratio of highway to non-highway
diesel fuel, the spillover for Alaska was estimated to be half that of PADD 1, or 5 percent. The
spillover for Hawaii and the rest of PADD 5 (Washington, Oregon, Nevada and Arizona) was
back-calculated from volumes from these various states, their estimated spillover volumes and
the overall spillover percentage of the PADD which is 58 percent. The spillover percent for
Hawaii and the rest of PADD 5 was estimated to be 24 percent. The spillover percentages for
each of the geographic areas in PADD 5 are shown in Table 7.1-5.
A Different national average estimates for spillover by non-highway end use still result due to differences i
spillover percentages and fuel volumes among the non-highway applications between the PADDs.
7-6
-------
Estimated Costs of Low-Sulfur Fuels
* Different national average estimates for spillover by non-highway end use still result due to differences in spillover
percentages and fuel volumes among the non-highway applications between the PADDS.
Table 7.1-5
Highway Diesel Fuel Spillover Percentages for PADD 5 Subregions
Spillover (%)
PADDV
AZ, NV, OR, WA
24
PADDV
CA
100
PADDV
AK
5
PADDV
HI
24
The spillover percentages for PADDs 1-4 and the various subregions for PADD 5 are used in
the following sections to estimate the volume of high-sulfur diesel fuel which would be affected
by the proposed rules.
7.1.2.2 Land-Based Nonroad Fuel Volumes
As previously mentioned, the primary information source underlying our assessment of
nonroad fuel volumes is the Fuel Oil and Kerosene Sales Report, published annually by the
Energy Information Administration (EIA).1 The report presents results of a national statistical
survey of approximately 4,700 fuel suppliers, including refiners and large companies who sell
distillate fuels for end use (rather than resale). The sample design involves classification of fuel
suppliers based on sales volume (stratification), with subsamples in individual classes (strata)
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, kerosene and jet fuel. The
survey requests respondents to report estimates of fuel sold for eleven "end uses," that
correspond to broad economic sectors, such as "Industrial," "Construction" and "Farm," as
described below. (See Table 7.1-6).
Before publication, EIA takes measures to quality-assure survey results. Automated and
manual procedures serve to identify missing values, potential misreporting, and evaluate
"outlier" values. Diesel consumption for the on-highway end use is represented by estimates
published annually by the Federal Highway Administration (FHWA).3 EIA uses the FHWA data
because it is their perspective that EIA's sampling technique gives inadequate coverage of truck
stops. Finally, they perform an adjustment or "post-stratification," to bring total survey results
into agreement with total annual supply as reported in the Petroleum Supply Annual2 For this
step, "supply" refers to "product-supplied" to the end-use market, calculated as domestic
production plus imports less exports and stock changes, as calculated for each Petroleum
Administration for Defense District (PADD). The adjustment is calculated at the PADD level,
and applied uniformly to each state and end-use within each PADD.
The EIA FOKS report estimates volumes of distillate sold into end-uses or economic sectors.
It does not directly represent fuel consumption, or attempt to determine how fuel is used after it
is sold. Thus, sales estimates encompass all potential uses, including on-highway mobile
sources, non-road mobile sources, and stationary sources such as heating, cooling, crop drying
7-7
-------
Draft Regulatory Impact Analysis
and power generation. In deriving an estimate of total fuel consumption for nonroad engines, our
basic approach is to estimate a fraction of total sales in each end use that represents nonroad fuel
consumption. With the exception of the railroad and on-highway end uses, the resulting fractions
directly follow guidance from EIA staff.
We derived the nonroad fraction in each end use in two steps. Beginning with total fuel
volumes for a given fuel grade or grades, we estimate a proportion representing diesel fuel (as
opposed to heating oil), and of the diesel fuel portion, we estimate a second fraction assumed to
represent nonroad use. We describe nonroad diesel fuel consumption as estimated for each end
use category below.
Farm. For this end use, two fuel grades are reported, "diesel" and "distillate." We assume
that 100% of the diesel represents nonroad use, and 0% of the distillate, which represents other
uses, 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% of total sales is diesel fuel, and that 100% of the diesel
represents nonroad use.
Industrial. This end use is essentially equivalent to the manufacturing sector, and differs
from most others in that EIA reports sales for five individual fuel grades, which simplifies
estimation of nonroad diesel consumption. At the outset, we assume that sales of No. 2 fuel oil
and No. 4 distillate include no diesel fuel. These grades represent other uses in this category,
such as space heating, meaning that none of the fuel in these categories represents nonroad use.
Conversely, for No. 2 diesel (low and high sulfur), we assume that 100% of sales is diesel fuel,
and 100% of the diesel represents nonroad use. For the remaining category, No. 1 distillate,
diesel and fuel oil are not distinguished. Following guidance from EIA staff, we have estimated
that 40% of No. 1 distillate sales represent diesel fuel, that 100% of this diesel represents
nonroad use, and that the remainder represents No. 1 fuel oil used in other applications, such as
space heating.
Commercial. This end use is broadly equivalent to the service sector. As with the industrial
end use, distillate sales are also reported by fuel grade. However, the commercial and industrial
end uses differ in that the commercial category includes sales for on-highway use. Distillate
sales for use in motor vehicles include fuel supplied to school-bus and government fleets (local,
state and federal). These sales are classified as "commercial" sales because they are exempt from
fuel taxes, as is fuel for nonroad use in most jurisdictions. As in the industrial end use, we
assume that none of the No. 2 fuel oil or No. 4 distillate represents nonroad use of diesel fuel. In
addition, to account for the on-highway fuel consumption in this end use, we assume that none of
the low-sulfur No. 2 diesel represents nonroad use. As in industrial, we assign 100% of the high-
sulfur No. 2 diesel to nonroad use. After consultation with EIA staff, we have estimated that
40% of the No. 1 distillate is diesel fuel, and that 50% of this diesel represents nonroad use.
7-8
-------
Estimated Costs of Low-Sulfur Fuels
For most of the remaining end uses, individual fuel grades are not distinguished, necessitating
broader assumptions in estimation of nonroad fuel use.
Oil Company. Sales for this end use include fuel purchased for drilling and refinery
operations. We assume that 50% of the reported distillate is diesel fuel, and assign 100% of the
diesel to operation of nonroad equipment. We assume that the remainder represents other uses
such as underground injection under pressure to fracture rock.
Military. For the military end use, fuel sales are reported for diesel and distillate. We
assume that 85% of the diesel represents use in 'non-tactical' nonroad equipment, and that 0% of
the distillate represents nonroad use. We exclude some fuel 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. Again, we assume that the vast majority of fuel sales in the railroad end use
represents locomotive operation, however, based on guidance from a major railroad, we assume
that a small fraction of reported sales represent operation of nonroad equipment used by
railroads. Accordingly, we assign 1% of the railroad fuel sales to nonroad use, which corresponds
to "Railway Maintenance" equipment as represented in the NONROAD model.
In three of the remaining end uses, Electric Utility, Vessel Bunkering and Residential, we
assign no fuel to nonroad use.
On-Highway. As the name implies, this end use represents sales for use in motor vehicles on
roads and highways, and is represented in the survey by the volume reported by FHWA.3 Many
organizations own mixed fleets and purchase both highway and non-road diesel, for which reason
it is plausible to assume that some fraction of the fuel attributed by FHWA to on-highway use is
actually used in nonroad engines. Because owners can legally use undyed low-sulfur diesel in
nonroad equipment, convenience or economy may encourage owners who purchase undyed
diesel to use it in nonroad equipment. Additionally, some owners might find it expedient or
necessary to purchase at least some of their diesel in commercial outlets such as gas stations,
where dyed "offroad" diesel is less available.
However, to reassign a fraction of the on-highway fuel to nonroad use, it is not sufficient
simply to postulate that low-sulfur undyed diesel is used in nonroad engines. Additional
constraints must be met to ensure that the EIA survey has not included the fuel in another end-
use, and that FHWA has not accounted for the fuel by subtracting it from the on-highway total.
For purposes of this study, we believe that four conditions must apply to justify a presumption
that fuel sales assigned to on-highway use would have been used in nonroad engines: (1) The fuel
sales were taxed, i.e., sales of undyed "low-sulfur highway diesel," (2) The buyer does not claim
a tax credit or refund on the fuel sale(s), (3) The buyer uses the fuel in nonroad equipment, and
(4) The EIA survey has not already accounted for the fuel.
7-9
-------
Draft Regulatory Impact Analysis
The first condition is necessary because FHWA estimates on-highway fuel on the basis of
fuel tax receipts reported by the states. In general, sales of undyed diesel are subject to state and
federal sales and use taxes; however, the purchaser is eligible for a tax refund or credit in most
jurisdictions, if the fuel is used in offroad equipment. To account for this possibility, FHWA
subtracts tax refunds from total receipts, which should effectively remove undyed fuel purchased
for use in nonroad equipment from the on-highway total. However, it is probable that only a
fraction of owners who are eligible actually take advantage of fuel tax refunds or credits, because
they are unaware that the option is available or because they find the process inconvenient. Thus,
if the purchaser forgoes applying for a refund or credit, FHWA leaves the fuel in the on-highway
total (the second condition above), and if the fuel is actually used in nonroad equipment (the third
condition above), FHWA also misclassifies it as on-highway consumption.
To reclassify such fuel as nonroad consumption, it is also necessary to be confident that the
EIA survey has not effectively assigned it to another end use (the fourth condition above). During
quality-assurance, EIA attempts to identify and remove distillate sales intended "primarily for on-
highway use" Fractions of such sales used in nonroad engines would thus not be reflected in
estimates of distillate sales. Also, fuel purchased at truck stops or gas stations and subsequently
used in nonroad equipment would not appear in survey results, because the survey does not
attempt to represent sales from these retail outlets.
An example scenario meeting all four conditions stated above would represent sales of
undyed diesel at retail outlets, for which the purchaser claims no tax credit or refund, and uses
the fuel in nonroad equipment. We assume that such a scenario is not uncommon in the
construction or commercial end uses, in which operations can be decentralized, dispersed or
remote, and operators numerous and highly mobile, refueling when and where convenient. Such
a situation is especially likely for the growing fleet of diesel rental equipment where available
refueling sites are likely to be highway service stations and where volumes may not warrant
seeking tax refunds.
The Northeast States for Coordinated Air Use Management (NESCAUM) recently conducted
a survey of diesel fuel use in construction equipment in New England, under a grant funded by
EPA. 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 low-sulfur
(undyed) 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 would be eligible under state or federal law.
Based on EPA's analysis of the survey results, approximately 20% of all diesel fuel
purchased for use "in construction" was undyed diesel for which the purchaser had not applied
for a tax refund. For purposes of deriving a protective estimate, it was assumed that 50% of the
7-10
-------
Estimated Costs of Low-Sulfur Fuels
un-refunded fuel was purchased at gas stations or truck stops, amounting to 10% of total diesel
purchased for use "in construction equipment." In the context of the scenario described above,
the implication is that 10% of the total nonroad fuel consumption in the construction and
commercial end uses (FTOTAL) i§ undyed diesel misclassified as on-highway use (^FHWA)* or
~ -*-* TOTAL
At the same time, the nonroad fuel consumption in these end uses captured by the FOKS survey
comprises the remaining 90% of the total, or
F = 0 9F
1 FOKS V.71 TOTAL
These two relationships allow us to estimate the misclassified diesel fuel in terms of nonroad fuel
consumption estimated from the FOKS survey
(0.0
T-i
^
A Q FOKS
meaning that FJ^A can be estimated as ~0. IFFOKS.
We estimated the misclassified highway volume (FJHWA) at the national level and individually
for each PADD, using FOKS-derived estimates of nonroad diesel consumption in the
construction and commercial end uses for the nation and each PADD, respectively. Summing
across the nation, this estimate represents 230 million gallons or approximately 0.7% of the on-
highway total.
Table 7.1-6 presents national land-based nonroad fuel consumption for calendar year 2000,
by end use. At the national level, the table shows estimates of total sales in each end use, plus
fractions representing diesel fuel and nonroad consumption, and resulting fuel volumes
representing nonroad consumption.
We derived fuel consumption estimates for each PADD by applying the same distillate and
diesel fractions developed above to fuel sales for each PADD. To meet requirements for the
economic analysis, the states of California, Hawaii and Alaska are presented individually, with
the remaining states in PADD 5 treated as an aggregate. Tables 7.1-7 and 7.1-8 present fuel sales
and estimated nonroad fuel consumption for each PADD, with PADD 5 subdivided as described.
7-11
-------
Draft Regulatory Impact Analysis
Table 7.1-6
Land-Based Nonroad Distillate Use, National Estimates, Calendar Year 2000
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
(M gal)
3,080
89
1,900
431
357
39
54
810
889
1,576
198
64
1,061
475
685
180
54
793
3,071
2,081
33,130
6,086
118
57,217
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
(Mgal)
3,080
0
1,805
409
0
0
22
810
889
0
0
25
1,061
475
342
180
0
793
2,917
1,873
33,130
0
0
47,800
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)
3,080
0
1,805
409
0
0
22
810
889
0
0
13
0
475
342
153
0
0
29
0
229
0
0
8,254
7-12
-------
Table 7.1-7
Distillate Fuel Sales by PADD, Calendar Year 2000 (million gallons)
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
PADD
1
389
44
511
160
219
33
1
116
281
1,304
197
3
418
219
19
41
29
304
500
490
10,228
5,391
8
20,906
2
1,572
40
549
59
111
3
26
176
285
102
0.7
36
276
153
42
15
21
134
1,233
301
11,141
557
72
16,904
3
549
3
394
123
4
2
3
193
218
141
0
0.9
146
58
561
9
11
195
686
1,033
5.644
1
0.1
9.976
4
219
1
150
30
8
2
13
202
18
7
0
11
69
16
29
2
2
8
345
0.2
1,475
30
9
2,647
5
AZ, NV, OR,
WA
90
0.08
91
31
11
1
1
79
74
5
0.02
3
66
15
1
87
2
106
114
61
1,885
76
6
2,806
5
CA
254
0
194
22
0.3
0
0
43
2
3
0
0.4
79
5
6
7
0
8
189
101
2,633
7
0.2
3,553
5
AK
0.03
0.001
7
7
4
0
10
0.02
10
12
0.02
10
4
6
26
6
0.05
36
4
80
91
25
23
361
5
HI
8
0
3
0.04
0.05
0
0
1
0.6
0.05
0
0
3
3
0.05
11
0
0.9
0
13
34
0.009
0
78
7-13
-------
Table 7.1-8
Land-Based Nonroad Diesel Consumption for the Nation and by PADD, 2000 (million gallons)
End Use
Farm
Construction
Other/(Logging)
Industrial
Commercial
Oil Company
Military
Electric Utility
Railroad
On-Highway
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
Nation
3,080
0
1,805
409
0
0
22
810
889
0
0
13
0
475
342
153
0
0
29
229
8,254
PADD
1
389
0
485
151
0
0
0.5
116
281
0
0
0.5
0
219
10
35
0
0
5
71
1,762
2
1,572
0
522
56
0
0
10
176
285
0
0
7
0
153
21
13
0
0
12
68
2,895
3
549
0
375
116
0
0
1
193
218
0
0
0.2
0
58
280
8
0
0
7
43
1,849
4
219
0
143
29
0
0
5
202
18
0
0
2
0
16
15
2
0
0
3
16
669
5
AZ, NV, OR,
WA
89
0
86
30
0
0
0.5
79
74
0
0
0.5
0
15
0.7
74
0
0
1
10
461
5
CA
254
0
184
21
0
0
0
43
2
0
0
0.1
0
5
3
6
0
0
2
19
539
5
AK
0.03
0
6
6
0
0
4
0.02
10
0
0
2
0
6
13
5
0
0
0.04
1
54
5
HI
8
0
3
0.04
0
0
0
1
0.6
0
0
0
0
3
0.02
9
0
0
0
0.6
26
-------
Estimated Costs of Low-Sulfur Fuels
The high-sulfur diesel fuel volumes are estimated by applying the highway spillover
percentages to the results shown in Tables 7.1-4 and 7.1-5. Specifically, the spillover percentage
is applied to the volume of diesel fuel remaining after the reclassified highway volume (i.e.,
highway fuel actually used in nonroad engines) is subtracted from the total land-based nonroad
engine volume. This is done because the spillover fraction was developed from the total highway
demand before the transfer was made. Table 7.1-9 shows the derivation of the high-sulfur diesel
fuel volume for land-based nonroad engines.
Table 7.1-9
Land-Based High-Sulfur Diesel Fuel Demand by PADD, 2000 (million gal)
Diesel Fuel Category
Total Land-Based
Low-Sulfur Hwy Transfer
Total Less Hwy Transfer
Hwy Spillover Percentage
(%)
Land-Based Low Sulfur
Land-Based High Sulfur
PADD
1
1,762
71
1,691
10
168
1,523
2
2,895
68
2,827
31
882
1,945
3
1,849
43
1,806
30
545
1,261
4
669
16
653
63
410
243
5
AZ,NV,
OR, WA
461
10
451
24
107
344
5
CA
539
19
520
100
520
0
5
AK
54
1
53
5
3
50
5
HI
26
1
25
24
6
19
7.1.2.3 Locomotive Diesel Fuel Demand
The estimates of diesel fuel demand for locomotives are taken from the information
presented in Section 7.1.2.2. In summary, the locomotive estimates were developed by taking the
railroad distillate fuel values directly from the EIA FOKS report for the geographic areas of
interest, and multiplying them by 0.95, which is the fraction of distillate fuel that is assumed to
be diesel grade. This results in estimates of the diesel fuel demand for railroads. To find only
the volume of diesel fuel used by locomotives, the fraction of diesel fuel that is assumed to be
used by rail maintenance (i.e., 0.01) is subtracted from the diesel railroad volumes. The
estimates of high-sulfur diesel are determined by applying the highway spillover percentages to
the total locomotive fuel volumes.
The locomotive fuel demand estimates for 2000 are shown in Table 7.1-10 for the geographic
areas of interest.
7-15
-------
Draft Regulatory Impact Analysis
Table 7.1-10
Locomotive High-Sulfur Diesel Fuel Demand by PAAD, 2000 (million gallons)
Diesel Fuel Category
Total Locomotive
Hwy Spillover
Percentage (%)
Locomotive Low
Sulfur
Locomotive High
Sulfur
PADD
1
470
10
47
423
2
1,160
31
362
798
3
646
30
195
451
4
324
63
204
120
5
AZ,NV,
OR, WA
107
24
25
82
5
CA
178
100
178
0
5
AK
4
5
0
4
5
HI
0
24
0
0
7.1.2.4 Marine Diesel Fuel Demand
The estimates of diesel fuel demand for marine vessels were developed base on guidance
from EIA staff. Specifically, the demand volumes are estimated by taking the vessel distillate
values directly from the EIA FOKS report and multiplying it by 0.90, which is the fraction of
distillate fuel sales that is assumed to represent diesel fuel for that category. The estimates of
high-sulfur diesel are determined by applying the highway spillover percentages to the resulting
total marine fuel volumes.
The marine fuel demand estimates for 2000 are shown in Table 7.1-11 for the geographic
areas of interest.
Table 7.1-11
Marine High-Sulfur Diesel Fuel Demand by PAAD, 2000 (million gallons)
Diesel Fuel Category
Total Marine
Hwy Spillover
Percentage (%)
Marine Low Sulfur
Marine High Sulfur
PADD
1
441
10
47
397
2
271
31
36,285
7,186
3
930
30
281
649
4
0
63
0
0
5
(except
CA, HI,
AK)
55
24
13
42
5
CA
91
100
91
0
5
AK
72
5
4
68
5
HI
12
24
3
9
7-16
-------
Estimated Costs of Low-Sulfur Fuels
7.1.2.5 Remaining Non-Highway Diesel Fuel Demand
It is also necessary to estimate diesel fuel demand volumes for the remaining non-highway
end uses that may use diesel fuel in order to complete the economic analysis. By definition, this
category includes any application other than land-based nonroad engines, locomotives, or marine
vessels.
The demand for diesel fuel in this broad category is found in three steps. First, the overall
volumes of fuel consumed by all non-highway end-uses is determined from the EIA FOKS
report. These demand volumes were developed for the geographic areas of interest as presented
in Section 7.1.2.1, Table 7.1-3. Second, the demand volumes are adjusted to include the volume
of fuel reclassified from the highway vehicle category to the land-based nonroad engine category.
These volumes were derived in Section 7.1.2.2, Table 7.1-8. Third, and finally, diesel fuel
demands for remaining non-highway end uses are calculated by subtracting the combined
volumes for land-based nonroad engines, locomotives, and marine vessel (as previously
determined in Sections 7.1.2.2 through 7.1.2.4) from the adjusted diesel demand for all non-
highway end uses. The estimates of high-sulfur diesel are then found by applying the highway
spillover percentages to these other non-highway demand volumes. The results are shown in
Table 7.1-12.
Table 7.1-12
Other Off-Highway High-Sulfur Diesel Fuel Demand by PADD, 2000 (million gallons)
Diesel Fuel Category
Potential Off-
Highway
Highway Transfer
Adjusted Off-
Highway
Land-Based
Nonroad,
Locomotive, and
Marine
Other Off-Highway
Hwy Spillover
Percentage (%)
Other Off-Highway
Low Sulfur
Other Off-Highway
High Sulfur
PADD
1
10,447
71
10,518
2673
7,845
10
781
7,064
2
5,760
68
5,828
4326
1,502
31
469
1,034
3
4,319
43
4,362
3425
937
30
283
654
4
1,171
16
1,187
993
194
63
122
72
5
AZ,NV,
OR, WA
849
10
859
623
307
24
73
234
5
CA
921
19
940
808
131
100
131
0
5
AK
254
1
255
130
141
5
7
134
5
HI
129
1
130
38
7
24
2
5
7-17
-------
Draft Regulatory Impact Analysis
7.1.2.6 Summary of Diesel Fuel Demand for 2000
Table 7.1-13 summarizes the diesel fuel demand estimates for each of the geographic areas of
interest for 2000 based on the information in the preceding sections. In this table, the low-sulfur
demand volumes for land-based nonroad engines are found by applying the highway spillover
percentages to the total volumes for this category minus the reclassified highway gallons. The
reclassified highway spillover gallons are then added to these results to produce the total low-
sulfur volumes for land-based nonroad engines. Totals for the U.S. and the U.S. minus
California are also shown for completeness.
7-18
-------
Table 7.1-13
Summary of Diesel Fuel Demand for 2000 (million gallons)
Category
Revised Highway
^arid-Based Nonroad
^ocomotive
Marine
subtotal
NR, Loc, Marine)
3ther Non-Highway
TOTAL
Fuel Type
total
highS
total
lowS
highS
total
low S
highS
total
low S
highS
total
lowS
highS
total
lowS
highS
total
low S
hiehS
PADD
1
10,157
n/a
1,762
239
1,523
470
47
423
441
44
397
2,673
330
2,343
7,845
781
7,064
20,675
11,269
9.406
2
11,074
n/a
2,895
950
1,945
1,160
362
798
271
85
186
4,326
1,397
2,929
1,502
469
1,034
16,902
12,939
3 963
3
5,601
n/a
1,849
588
1,261
646
195
451
930
281
649
3,425
1,064
2,361
937
283
654
9,963
6,948
3.015
4
1,459
n/a
669
426
243
324
204
120
0
0
0
993
630
363
194
122
72
2,646
2,211
435
5
AZ, NV, OR,
WA
1,875
n/a
461
117
344
107
25
82
55
13
42
623
155
468
307
73
234
2,805
2,103
702
AK
90
n/a
54
4
50
4
0
4
72
4
68
130
7
123
141
7
134
361
105
257
HI
33
n/a
26
7
19
0
0
0
12
3
9
38
10
28
7
2
5
78
44
34
CA
2,614
n/a
539
539
0
178
178
0
91
91
0
808
808
0
131
131
0
3,553
3,553
0
U.S.
32,902
n/a
8,255
2,871
5,384
2,889
1,011
1,878
1,872
520
1,352
13,016
4,402
8,614
11,065
1,868
9,197
56,983
39,171
17.812
U.S. -CA
30,288
n/a
7,716
2,332
5,384
2,711
833
1,878
1,781
429
1,352
12,208
3,594
8,614
10,934
1,737
9,197
53,430
35,618
17.812
7-19
-------
Draft Regulatory Impact Analysis
7.1.3 Diesel Fuel Demand by PADD for 2008
Diesel fuel demand in 2008 is projected for each end use category by applying various growth
factors to the 2000 diesel fuel demand volumes shown in Table 7.1-13. This section shows how
the growth factors were determined and applied to each end-use category. Finally, the low-sulfur
diesel fuel estimates for 2008 are divided into separate volumes of 15 ppm and 500 ppm sulfur
concentrations (i.e., highway diesel fuel) in order to facilitate the air quality analysis.
7.1.3.1 2000-2008 Growth Factors
The growth factors for highway diesel fuel, locomotives, and other non-highway end uses
were developed from the Annual Energy Outlook 2002 (AEO2002) report, which is published by
the Energy Information Administration.4 The growth factor for land-based nonroad engines was
taken from estimates of diesel fuel consumption from the draft NONROAD2002 model. The
factor for marine diesel fuel was developed from information contained in the 1999 Final
Regulatory Impact Analysis for the Marine Diesel Emission Standards, which was published by
EPA.5 Each of the growth factors and their respective sources are shown in Table 7.1-14. The
derivation of the composite growth factor that was used for the other non-highway end use
category is shown in Table 7.1-15.
Table 7.1-14
2000-2008 Growth Factors by End-Use Category
End Use
Highway
Land-Based Nonroad
Locomotive
Marine
Other Non-Highway
2000-2008
Multiplicative
Growth Factor
1.238
1.229
1.083
1.090
1.074
% Simple
Annual Growth
Rate
2.98
2.87
1.04
1.13
0.93
Source/Comments
AEO2002, Table 7, Energy Use by Mode, Freight Trucks
(over 10,000 Ibs. GVWR)
Calculated from 2000 and 2008 Draft NONROAD2002
Model diesel fuel consumption outputs.
AEO 2002, Table 7, Energy Use by Mode, Railroad
Calculated from 2000 and 2008 CO emissions inventories
(as a surrogate for fuel consumption) as reported in the
Final Regulatory Impact Analysis, Control of Emissions
for Marine Diesel Engines, 1999.
See Table 7. 1-15. Primarily diesel fuel and heating oil.
7-20
-------
Estimated Costs of Low-Sulfur Fuels
Table 7.1-15
2000-2008 Composite Growth Factor for Other Non-Highway End-Uses
End Use
Commercial
Industrial
Farm
Construction
Composite
Average3
Energy
Consumption
(Quadrillion BTU)
0.42
1.18
0.528
0.285
Fraction of
Total
0.174
0.489
0.219
0.118
2000-2008
Multiplicative
Growth Factor
1.105
1.063
1.039
1.138
Consumption
Weighted
Multiplicative
Growth Factor
0.192
0.520
0.227
0.134
1.074
Source of Energy Consumption
AEO2002, Table 2, Commercial,
Distillate Fuel
AEO2002, Table 2, Industrial,
Distillate Fuel
AEO2002, Table 32, Agriculture,
Distillate Fuel
AEO2002, Table 32, Construction,
Distillate Fuel
a Growth in the residential heating oil end-use category was inadvertently excluded from the composite growth factor of
the other non-highway category. This will be added for the final rulemaking.
The 2000-2008 average annual growth rate of 2.87 percent for land-based nonroad engines,
presented in Table 7.1-14 above, is identical to that used in Chapter 3 for these engines. This
compares with an annual growth rate of 0.97 percent that was developed using data from
AEO2002 (See Appendix 7B.) The growth rates for locomotives and marine shown in Table 7.1-
14 are the same as the values used in Chapter 3 for these engines.
7.1.3.2 Division of Low-Sulfur Diesel Fuel into 15 ppm and 500 ppm Volumes
As previously noted, the highway diesel fuel spillover volume is divided into 15 ppm and 500
ppm sulfur levels to facilitate the air quality analysis. The 15 ppm sulfur pool is projected to
comprise 74 percent of the spillover volume, while 500 ppm sulfur pool is projected to comprise
26 percent of the spillover volume. The value is 74 percent 15 ppm diesel fuel because although
80 percent of each PADDs highway diesel fuel must be 15 ppm in 2006, highway diesel fuel
produced by small refineries is allowed to be exempt from having to comply in 2006, and they
comprise 5 percent of the national highway diesel fuel production volume. Then, the 75/25
relative volumes were adjusted to account for downgrading in the distribution system thus
resulting in the 74 and 26 percent values. When this volume methodology was created, the
highway plans for most of the small refiners were not known, so it was assumed that all of them
would take the delay option. However, we now know that some are taking the gasoline for diesel
fuel option which requires them to comply with the highway diesel fuel option in 2006, in return
for a three year delay with the Tier 2 gasoline sulfur standard. These small refineries will
therefore comply with the Highway Program sulfur requirements in 2006 and will make the
percentage of highway diesel fuel complying to the 15 ppm cap standard in 2006 closer to 80
percent. This will be updated for the final rulemaking.
7-21
-------
Draft Regulatory Impact Analysis
7.1.3.3 Summary of Diesel Fuel Demand for 2008
The diesel fuel demand estimates for 2008 are shown in Table 7.1-16.
7-22
-------
Table 7.1-16
Summary of Diesel Fuel Demand by PADD for 2008 (million gallons)
Category
Revised Highway
,and-Based Nonroad
,ocomotive
/[arine
lubtotal
MR, Loc, Marine)
)ther Off -Highway
'OTAL
Fuel Type
total
15 ggmdiesel
500_p_pm diesel
highS
total
15 ggmdiesel
500_p_pm diesel
highS
total
15 ggmdiesel
500_p_pm diesel
highS
total
15 ggmdiesel
500_p_pm diesel
highS
total
15 ggmdiesel
500_gpm diesel
highS
total
15 ggmdiesel
500_p_pm diesel
highS
total
15 ggmdiesel
500_p_pm diesel
highS
PADD
1
12,575
9,324
3,251
n/a
2,166
219
77
1,872
509
38
13
458
481
35
12
433
3,156
291
102
2,763
8,425
622
217
7,586
24,157
10,238
3,570
10,349
2
13,710
10,165
3,544
n/a
3,559
866
302
2,391
1,256
291
101
864
295
68
24
203
5,111
1,225
427
3,458
1,614
373
130
1,110
20,434
11,764
4,102
4,569
3
6,934
5,141
1,793
n/a
2,273
536
187
1,550
700
157
55
488
1,014
227
79
708
3,987
920
321
2,746
1,007
225
79
703
11,927
6,287
2,192
3,448
4
1,806
1,339
467
n/a
822
389
135
298
351
163
57
130
0
0
0
0
1,173
552
193
429
209
97
34
78
3,188
1,988
693
506
5
AZ, NV, OR,
WA
2,321
1,721
600
n/a
567
106
37
423
116
20
7
88
60
11
4
46
743
137
48
557
330
58
20
252
3,394
1,917
669
809
AK
111
83
29
n/a
66
3
1
62
4
0
0
4
78
3
1
75
149
6
2
141
151
6
2
144
412
95
33
284
HI
41
30
11
n/a
32
6
2
23
0
0
0
0
13
2
1
10
45
9
3
33
8
1
0
6
93
40
14
39
U.S. -CA
37,499
27,804
9,695
n/a
9,486
2,125
741
6,620
2,936
669
233
2,034
1,941
347
121
1,474
14,364
3,141
1,095
10,127
11,743
1,383
482
9,878
63,605
32,328
11,272
20,005
CA
3,236
3,236
0
n/a
663
662
0
0
193
193
0
0
99
99
0
0
955
955
0
0
141
141
0
0
4,332
4,332
0
0
U.S.
40,735
31,040
9,695
n/a
10,149
2,788
741
6,620
3,129
862
233
2,034
2,040
446
121
1,474
15,318
4,096
1,095
10,127
11,884
1,524
482
9,878
67,937
36,660
11,272
20,005
7-23
-------
Draft Regulatory Impact Analysis
7.1.4. Annual Diesel Fuel Demand (2000-2040) and Associated In-Use Sulfur Levels
The annual diesel fuel volumes and respective in-use sulfur concentrations for land-based
nonroad engines, locomotives, and marine vessels are estimated in this section. The estimates
of in-use diesel fuel sulfur levels are used in emissions inventory analysis. The diesel volumes
are used in the economic analysis. Some of these volume estimates are also used in the
emissions inventory analysis described in Chapter 3.
This section begins with a description of the methodology that is used to estimate the diesel
demand volumes for 2000-2040. Then the basic inputs for determining the in-use sulfur
concentration is discussed. Finally, the volumes and corresponding in-use sulfur levels for each
year are presented.
7.1.4.1 Annual Diesel Demand Volume Estimates
Diesel fuel volume estimates by year and by geographic area (nationwide, 49-state without
California, and 48-state without Alaska or Hawaii) and corresponding average sulfur levels by
year were calculated from the 2008 fuel use estimates presented in Section 7.1.3. The resulting
volumes and sulfur levels are presented below in Section 7.1.4.3. The demand estimates for each
of the other years were determined by extrapolating the 2008 values according to the nationwide
growth rates shown in Table 7.1-17 for land-based nonroad model equipment categories,
locomotives, and marine (commercial and recreational). The sources for these growth rates are
the same as described earlier in Table 7.1-14.
Although the growth rates in the two tables are consistent, they are not directly comparable.
The values in Table 7.1-14 are expressed as simple annual growth (i.e., the percentage change
from 2000-2008 divided by the number of years, in this case eight years). The values in Table
7.1-17 are expressed as a percentage change from the previous year, or year-to-year change.
7-24
-------
Estimated Costs of Low-Sulfur Fuels
Table 7.1-17
Nationwide Annual Growth Rates for Nonroad Diesel Fuel Use
Year
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
7.040
Nonroad
—
2.88
2.80
2.72
2.65
2.58
2.50
2.44
2.38
2.32
2.27
2.23
2.18
2.14
2.09
2.05
1.99
1.95
1.91
1.88
1.84
1.81
1.78
1.75
1.72
1.69
1.65
1.62
1.60
1.57
1.55
1.52
1.50
1.48
1.46
1.44
1.41
1.40
1.38
1.36
1 34
Locomotive
—
5.15
-1.63
1.74
1.38
1.38
0.97
0.97
0.44
0.69
0.72
1.70
0.45
0.27
0.28
0.45
1.02
0.57
0.52
0.56
0.33
0.89
0.89
0.89
0.89
0.89
0.89
0.89
0.89
0.89
0.89
0.89
0.89
0.89
0.89
0.89
0.89
0.89
0.89
0.89
0 89
Marine
—
1.08
1.08
1.08
1.09
1.09
1.09
1.10
1.10
1.10
1.11
1.11
1.11
1.12
1.12
1.12
1.13
1.13
1.13
1.14
1.14
1.15
1.15
1.16
1.16
1.16
1.17
1.17
1.18
1.18
1.19
1.19
1.20
1.20
1.21
1.21
1.22
1.23
1.23
1.24
1 7.4
7.1.4.2 In-Use Diesel Sulfur Concentrations
Table 7.1-18 shows the diesel sulfur levels that were used in generating the national in-use
average sulfur levels by year that are shown in Section 7.1.4.3.
7-25
-------
Draft Regulatory Impact Analysis
Table 7.1-18
Factors Used to Calculate In-use Sulfur Levels
Parameter
Average in-use sulfur level for fuel intended to be used in nonroad engines, prior to sulfur control
Average in-use fuel sulfur level for any fuel designed to meet a standard of 500 ppm
Average in-use fuel sulfur level for fuel designed to meet California's diesel fuel specifications
Average in-use fuel sulfur level for any fuel designed to meet a standard of 15 ppm
Nonroad spillover: Percentage of fuel consumed by nonroad engines that is actually produced to
meet highway fuel sulfur standards
Locomotive and marine spillover: Percentage of fuel consumed by locomotives and marine
vessels that is actually produced to meet highway fuel sulfur standards
Value
3400 ppm
340 ppm
120 ppm
11 ppm
34.9%
32.4%
Each of the sulfur levels and spillover assumptions is further described below.
High-Sulfur Diesel Fuel. The national average in-use sulfur level of uncontrolled nonroad,
locomotive, and marine diesel fuel is approximately 3400 ppm. This estimate is derived from
1996 through 2001 fuel survey data reported by the National Institute for Petroleum and Energy
Research (NIPER) and TRW Petroleum Technologies (TRW).6'7 These annual reports provide
measured sulfur concentrations and respective fuel volumes for multiple samples in several
geographic regions of the country. The information was used to estimate the national average in-
use sulfur level as follows. First, the geographic regions were assigned to each of the five
PADDs. Second, individual annual average sulfur levels in each PADD were calculated by
weighting the respective sulfur content of each sample by its fuel volume (i.e., volume
weighting). Third, an overall average sulfur level for each PADD was found by volume
weighting the individual annual average sulfur concentrations. Fourth, the national average
sulfur level was determined by volume weighting each PADD's overall average sulfur
concentration. The final results of this analysis are shown in Table 7.1-19. The details of the
method are provided in a memo to the Docket entitled, "Derivation of the National Average In-
use Sulfur Level of Uncontrolled Nonroad, Locomotive, and Marine Diesel Fuel".
7-26
-------
Estimated Costs of Low-Sulfur Fuels
Table 7.1-19
In-Use Sulfur Concentrations for High-Sulfur Diesel Fuel
PADD
1
2
3
4
5
U.S.
Year
1996
1997
1998
1999
2000
2001
Average
1996
1997
1998
1999
2000
2001
Average
1996
1997
1998
1999
2000
2001
Average
1996
1997
1998
1999
2000
2001
Average
1996
1997
1998
1999
2000
2001
Average
1996
1997
1998
1999
2000
2001
Volume
(million gallons)
7,637,500
6,000,000
4,637,500
4,275,000
9,025,000
4,937,500
36,512,500
2,825,000
2,775,000
1,275,000
2,912,500
10,412,500
5,212,500
25,412,500
3,137,500
3,637,500
3,137,500
4,637,500
3,887,500
1,775,000
20,212,500
412,500
275,000
275,000
275,000
275,000
275,000
1,787,500
1,912,500
3,550,000
1,550,000
1,550,000
—
—
8,562,500
15,925,000
16,237,500
10,875,000
13,650,000
23,600,000
12,200,000
Sulfur
3,423
2,663
3,998
3,474
3,653
3,055
3,384
3,600
2,740
1,818
1,717
2,939
3,936
2,999
4,539
3,945
5,004
4,177
4,361
4,298
4,366
4,100
1,000
3,400
2,000
2,600
2,340
2,691
3,002
2,268
3,077
2,065
—
—
2,541
3,641
2,849
3,886
3,148
3,442
3,596
7-27
-------
1
2
3
4
5
1996
1997
1998
1999
2000
2001
Average
1996
1997
1998
1999
2000
2001
Average
1996
1997
1998
1999
2000
2001
Average
1996
1997
1998
1999
2000
2001
Average
1996
1997
1998
1999
Average
7,637,500
6,000,000
4,637,500
4,275,000
9,025,000
4,937,500
36,512,500
2,825,000
2,775,000
1,275,000
2,912,500
10,412,500
5,212,500
25,412,500
3,137,500
3,637,500
3,137,500
4,637,500
3,887,500
1,775,000
20,212,500
412,500
275,000
275,000
275,000
275,000
275,000
1,787,500
1,912,500
3,550,000
1,550,000
1,550,000
92,487,500
3,423
2,663
3,998
3,474
3,653
3,055
3,384
3,600
2,740
1,818
1,717
2,939
3,936
2,999
4,539
3,945
5,004
4,177
4,361
4,298
4,366
4,100
1,000
3,400
2,000
2,600
2,340
2,691
3,002
2,268
3,077
2,065
3,401
500 ppm Low-Sulfur Diesel Fuel. The in-use sulfur level of diesel fuel meeting a 500 ppm
sulfur standard is 340 ppm. This in-use level, which is based on fuel survey data from NIPER,
the American Automobile Manufacturers Association (AAMA), and the American Petroleum
Institute (API) /National Petrochemical and Refiners Association (NPRA), is documented in the
Final Regulatory Impact Analysis for the emission standards and diesel fuel sulfur requirements
affecting 2007 and later heavy-duty highway engines and vehicles.8
California 500 ppm Low-Sulfur Diesel Fuel. The in-use sulfur level of diesel fuel meeting a
500 ppm sulfur standard in California is 120 ppm. A level of 140 ppm was previously estimated
in the Final Regulatory Impact Analysis for the emission standards and diesel fuel sulfur
requirements affecting 2007 and later heavy-duty highway engines and vehicles.8 However,
more recent in-use survey data shows a constantly decreasing sulfur level in California under this
-------
Estimated Costs of Low-Sulfur Fuels
standard. Therefore, it is estimated that California will experience an in-use sulfur level of 120
ppm for diesel fuel complying with the 500 ppm sulfur standard in that state.
11 ppm Low-Sulfur Diesel Fuel. It is estimated that refiners will produce diesel fuel with
approximately 7-8 ppm sulfur in order for all parties downstream of the refinery gate to meet the
15 ppm sulfur standard. The actual in-use level likely will be somewhere between 7 and 15 ppm.
In complex distribution segments, diesel fuel could have a sulfur level close to the 15 ppm sulfur
cap due to contamination that occurs throughout the distribution system. On the other hand,
simple distribution segments should not experience as much contamination and the resulting
sulfur level should not be as high. On average we expect the in-use sulfur level to be
approximately 10 ppm. For emissions inventory modeling purposes, 1 ppm sulfur is added to the
in-use fuel sulfur level to account for the combustion of lubricating oil in non-highway engines.
Therefore, an 11 ppm total sulfur concentration is used to evaluate the effects on emissions of a
fuel complying with a 15 ppm sulfur standard.
Spillover Percentages. The average spillover percentages for the land-based nonroad
engines and separately for locomotives and marine vessels are calculated by summing the
spillover volume for all the PADDs and dividing by the total volume of either land-based
nonroad or locomotives and marine volume for all the PADDs. This approach yields slightly
different spillover percentages for 50-state and 48-state cases, so as a simplifying assumption in
this analysis, the average of these two spillover percentages was used in both cases.
The estimated average in-use sulfur levels of the highway spillover diesel fuel are estimated
by applying the sulfur factors shown in Table 7.1-18 to the phase-in schedule for the highway
fuel sulfur standards, which were promulgated in 2001 [66 FR 5002]. The results are described
in Table 7.1-20.
Estimating the average in-use sulfur levels of non-highway diesel fuel also involves three
transitions when fuel sulfur levels are moving from uncontrolled to a proposed standard, or from
one proposed control level to the next. The sulfur levels for these transitions are calculated using
the information from Table 7.1-21 with the assumption that any fuel transition occurs in June of
the calendar year in which the new standard takes effect. Table 7.1-21 displays the resulting
transitional year non-highway fuel sulfur levels.
The information described above is used in the next section to calculate the resulting annual
in-use sulfur levels for each.
7-29
-------
Draft Regulatory Impact Analysis
Table 7.1-20
Average Sulfur Level for On-highway Fuel
Year
2005 and
earlier
2005 and
earlier
2006
2007
2008
2009
20 10 and
later
Average
sulfur (ppm)
340
300
165
69
69
69
11
Explanation
Nationwide average, excluding California, prior to introduction of 15ppm standard.
This is used in the 48-state and 50-state analyses.
Nationwide average, including California, prior to introduction of 15ppm standard.
Assumes 10% of nationwide highway diesel meets California's requirements. This is
used in the 49-state analysis.
15ppm standard applies beginning in June. Only 80% of the pool meets the 15ppm
standard.
Only 80% of the pool meets the 1 5ppm standard.
Only 80% of the pool meets the 1 5ppm standard.
Only 80% of the pool meets the 1 5ppm standard.
100% of the pool meets the 15ppm standard
Table 7.1-21
Average Sulfur Levels for Off-highway Fuel Sulfur Standard Transitions (ppm)
Prior to transition year
Transition year
After transition year
Uncontrolled to
SOOppm standard
3400
1615
340
SOOppm standard to
1 5ppm standard
340
148
11
Uncontrolled to 15ppm
standard
3400
1423
11
7.1.4.3 Summary of Annual Diesel Fuel Demand and Sulfur Levels
Tables 7.1-22 through 30 present the diesel demand volumes and average in-use sulfur levels
for each year, end use category, and area of interest (50-state, 49-state without California, and 48-
state without Alaska or Hawaii). The demand volumes are determined by applying the growth
rates from Table 7.1-17 to the 2008 demand volumes shown in Table 7.1-16. The average in-use
sulfur concentrations are found by combining the average sulfur levels for highway fuel shown in
Table 7.1-20 with the average sulfur levels for non-highway fuel from Table 7.1-21. The
spillover fractions given in Table 7.1-18 are used to properly weight the highway and non-
highway sulfur levels.
The column headings in the subsequent tables are defined as follows. "Affected Volume"
refers to the fuel produced to meet applicable nonroad fuel sulfur requirements. The term
7-30
-------
Estimated Costs of Low-Sulfur Fuels
"Spillover" refers to fuel produced to meet highway sulfur requirements, but which ends up being
used in nonroad equipment. The final columns labeled "Combo S ppm" show the fuel volume-
weighted average sulfur levels for the combination of the "Affected Volume" and the "Spillover
Volume." The final "Combo S ppm" columns are the sulfur levels that were used for the 50-state
and 48-state emissions inventory modeling. Separate 49-state (without California) emissions
modeling was not conducted. Note that the 50-state and 48-state Base and Control combination
sulfur levels have been set to the average of the 50 & 48-state values, since the difference was
negligible. Similarly, the Locomotive and Marine combination sulfur levels have been set to
their average to simplify the analysis.
7-31
-------
Draft Regulatory Impact Analysis
Table 7.1-22
50-State Nonroad Land-based Diesel Fuel Volumes and Sulfur Content
Year
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
7.040
Total
Volume
8,255
8,492
8,730
8,967
9,204
9,442
9,678
9,913
10,149
10,385
10,620
10,857
11,094
11,331
11,568
11,805
12,040
12,275
12,509
12,744
12,979
13,214
13,448
13,683
13,918
14,153
14,386
14,619
14,852
15,085
15,319
15,552
15,785
16,018
16,252
16,485
16,718
16,951
17,185
17,418
17651
Affected
Volume
5,384
5,539
5,694
5,849
6,004
6,158
6,312
6,466
6,620
6,773
6,927
7,082
7,236
7,391
7,545
7,700
7,853
8,006
8,159
8,312
8,465
8,619
8,772
8,925
9,078
9,231
9,383
9,535
9,687
9,839
9,992
10,144
10,296
10,448
10,600
10,752
10,904
11,056
11,209
11,361
11 513
BaseS
ppm
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
Control S
ppm
3400
3400
3400
3400
3400
3400
3400
1615
340
340
148
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
11
11
11
Spillover
Volume
2,871
2,953
3,036
3,118
3,201
3,283
3,365
3,447
3,529
3,611
3,693
3,776
3,858
3,940
4,023
4,105
4,187
4,269
4,350
4,432
4,513
4,595
4,677
4,758
4,840
4,922
5,003
5,084
5,165
5,246
5,327
5,408
5,489
5,570
5,652
5,733
5,814
5,895
5,976
6,057
6 138
Spillover S
ppm
300
300
300
300
300
300
165
69
69
69
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
11
11
11
11
Base
Combo S
ppnr*
2318
2318
2318
2318
2318
2318
2271
2237
2237
2237
2217
2217
2217
2217
2217
2217
2217
2217
2217
2217
2217
2217
2217
2217
2217
2217
2217
2217
2217
2217
2217
2217
2217
2217
2217
2217
2217
2217
2217
2217
7.7.17
Control
Combo S
ppnr*
2318
2318
2318
2318
2318
2318
2271
1075
245
245
100
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
11
11
11
a 50-state and 48-state Base and Control combination sulfur levels have been set to the average of 50 & 48-state values,
since the difference was negligible.
7-32
-------
Estimated Costs of Low-Sulfur Fuels
Table 7.1-23
50-State Locomotive Diesel Fuel Volumes and Sulfur Content
Year
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
7.040
Total
Volume
2825
2,970
2,922
2,973
3,014
3,055
3,085
3,115
3,129
3,150
3,173
3,227
3,242
3,251
3,260
3,274
3,308
3,327
3,344
3,363
3,374
3,404
3,434
3,465
3,496
3,527
3,559
3,590
3,622
3,655
3,687
3,720
3,753
3,787
3,821
3,855
3,889
3,924
3.959
3,994
4030
Affected
Volume
1836
1931
1,899
1,932
1,959
1,986
2,005
2,025
2,034
2,048
2,063
2,098
2,107
2,113
2,119
2,128
2,150
2,163
2,174
2,186
2,193
2,213
2,233
2,252
2,273
2,293
2,313
2,334
2,355
2,376
2,397
2,418
2,440
2,462
2,484
2,506
2,528
2,551
2.573
2,596
7 670
BaseS
ppm
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
Control S
ppm
3400
3400
3400
3400
3400
3400
3400
1615
340
340
340
340
340
340
340
340
340
340
340
340
340
340
340
340
340
340
340
340
340
340
340
340
340
340
340
340
340
340
340
340
340
Spillover
Volume
989
1,040
1,023
1,040
1,055
1,069
1,080
1,090
1,095
1,102
1,110
1,129
1,134
1,138
1,141
1,146
1,158
1,164
1,170
1,177
1,181
1,191
1,202
1,213
1,223
1,234
1,245
1,256
1,268
1,279
1,290
1,302
1,314
1,325
1,337
1,349
1,361
1,373
1.385
1,398
1 410
Spillover S
ppm
299
299
299
299
299
299
165
69
69
69
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
11
11
11
11
Base
Combo S
ppma
2396
2396
2396
2396
2396
2396
2352
2321
2321
2321
2302
2302
2302
2302
2302
2302
2302
2302
2302
2302
2302
2302
2302
2302
2302
2302
2302
2302
2302
2302
2302
2302
2302
2302
2302
2302
2302
2302
2302
2302
7307
Control
Combo S
ppma
2396
2396
2396
2396
2396
2396
2352
1114
252
252
233
233
233
233
233
233
233
233
233
233
233
233
233
233
233
233
233
233
233
233
233
233
233
233
233
233
233
233
233
233
733
a 50-state and 48-state Base and Control combination sulfur levels have been set to the average of 50 & 48-state values,
since the difference was negligible. Similarly, the Locomotive and Marine combination sulfur levels have been set to
their average to simplify the analysis.
7-33
-------
Draft Regulatory Impact Analysis
Table 7.1-24
50-State Marine Diesel Fuel Volumes and Sulfur Content
Year
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
7.040
Total
Volume
1870
1,890
1,911
1,931
1,953
1,974
1,996
2,018
2,040
2,063
2,086
2,109
2,132
2,156
2,180
2,205
2,230
2,255
2,280
2,306
2,333
2,359
2,387
2,414
2,442
2,471
2,499
2,529
2,559
2,589
2,620
2,651
2,683
2,715
2,748
2,781
2,815
2,850
2,885
2,920
7 957
Affected
Volume
1,350
1,365
1,380
1,395
1,410
1,426
1,442
1,458
1,474
1,490
1,507
1,523
1,540
1,557
1,575
1,593
1,610
1,629
1,647
1,666
1,685
1,704
1,724
1,744
1,764
1,785
1,805
1,827
1,848
1,870
1,892
1,915
1,938
1,961
1,985
2,009
2,033
2,058
2,084
2,110
7136
BaseS
ppm
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
Control S
ppm
3400
3400
3400
3400
3400
3400
3400
1615
340
340
340
340
340
340
340
340
340
340
340
340
340
340
340
340
340
340
340
340
340
340
340
340
340
340
340
340
340
340
340
340
340
Spillover
Volume
519
525
530
536
542
548
554
560
567
573
579
586
592
599
605
612
619
626
633
640
648
655
663
670
678
686
694
702
710
719
727
736
745
754
763
772
782
791
801
811
871
Spillover S
ppm
299
299
299
299
299
299
165
69
69
69
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
11
11
11
11
Base
Combo S
ppma
2396
2396
2396
2396
2396
2396
2352
2321
2321
2321
2302
2302
2302
2302
2302
2302
2302
2302
2302
2302
2302
2302
2302
2302
2302
2302
2302
2302
2302
2302
2302
2302
2302
2302
2302
2302
2302
2302
2302
2302
7307
Control
Combo S
ppma
2396
2396
2396
2396
2396
2396
2352
1114
252
252
233
233
233
233
233
233
233
233
233
233
233
233
233
233
233
233
233
233
233
233
233
233
233
233
233
233
233
233
233
233
733
a 50-state and 48-state Base and Control combination sulfur levels have been set to the average of 50 & 48-state values,
since the difference was negligible. Similarly, the Locomotive and Marine combination sulfur levels have been set to
their average to simplify the analysis.
7-34
-------
Estimated Costs of Low-Sulfur Fuels
Table 7.1-25
49-State Nonroad Land-based Diesel Fuel Volumes and Sulfur Contenta
Year
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
7.040
Total
Volume
7716
7938
8,160
8,382
8,603
8,825
9,046
9,266
9,486
9,707
9,927
10,149
10,370
10,591
10,813
11,034
11,254
11,473
11,693
11,912
12,131
12,351
12,570
12,790
13,009
13,228
13,446
13,664
13,882
14,100
14,318
14,536
14,754
14,972
15,190
15,408
15,626
15,844
16,062
16,280
16498
Affected
Volume
5,384
5,539
5,694
5,849
6,004
6,158
6,312
6,466
6,620
6,773
6,927
7,082
7,236
7,391
7,545
7,700
7,853
8,006
8,159
8,312
8,465
8,619
8,772
8,925
9,078
9,231
9,383
9,535
9,687
9,839
9,992
10,144
10,296
10,448
10,600
10,752
10,904
11,056
11,209
11,361
11 513
BaseS
ppm
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
Control S
ppm
3400
3400
3400
3400
3400
3400
3400
1615
340
340
148
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
11
11
11
Spillover
Volume
2,332
2,399
2,466
2,533
2,600
2,667
2,734
2,800
2,867
2,933
3,000
3,067
3,134
3,201
3,268
3,334
3,401
3,467
3,533
3,600
3,666
3,732
3,799
3,865
3,931
3,998
4,063
4,129
4,195
4,261
4,327
4,393
4,459
4525
4,590
4,656
4,722
4,788
4,854
4,920
4 986
Spillover S
ppm
340
340
340
340
340
340
186
77
77
77
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
11
11
11
11
Base Combo
S ppm
2475
2475
2475
2475
2475
2475
2429
2396
2396
2396
2376
2376
2376
2376
2376
2376
2376
2376
2376
2376
2376
2376
2376
2376
2376
2376
2376
2376
2376
2376
2376
2376
2376
2376
2376
2376
2376
2376
2376
2376
7.376
Control
Combo S
ppm
2475
2475
2475
2475
2475
2475
2429
1150
260
260
107
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
11
11
11
' 49-state analysis includes all states except California.
7-35
-------
Draft Regulatory Impact Analysis
Table 7.1-26
49-State Locomotive Diesel Fuel Volumes and Sulfur Contenta
Year
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
7.040
Total
Volume
2651
2,787
2,742
2,790
2,828
2,867
2,895
2,923
2,936
2,956
2,977
3,028
3,042
3,050
3,059
3,073
3,104
3,122
3,138
3,156
3,166
3,194
3,223
3,252
3,281
3,310
3,339
3,369
3,399
3,430
3,460
3,491
3,522
3,554
3,585
3,617
3,650
3,682
3,715
3,748
3 787.
Affected
Volume
1,836
1,931
1,899
1,932
1,959
1,986
2,005
2,025
2,034
2,048
2,063
2,098
2,107
2,113
2,119
2,128
2,150
2,163
2,174
2,186
2,193
2,213
2,233
2,252
2,273
2,293
2,313
2,334
2,355
2,376
2,397
2,418
2,440
2,462
2,484
2,506
2,528
2,551
2,573
2,596
7 670
BaseS
ppm
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
Control S
ppm
3400
3400
3400
3400
3400
3400
3400
1615
340
340
340
340
340
340
340
340
340
340
340
340
340
340
340
340
340
340
340
340
340
340
340
340
340
340
340
340
340
340
340
340
340
Spillover
Volume
815
856
842
857
869
881
889
898
902
908
915
930
935
937
940
944
954
959
964
970
973
982
990
999
1,008
1,017
1,026
1,035
1,044
1,054
1,063
1,073
1,082
1,092
1102
1,111
1,121
1,131
1,142
1,152
1 167
Spillover S
ppm
340
340
340
340
340
340
186
77
77
77
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
11
11
11
11
Base
Combo S
ppm
2460
2460
2460
2460
2460
2460
2413
2379
2379
2379
2359
2359
2359
2359
2359
2359
2359
2359
2359
2359
2359
2359
2359
2359
2359
2359
2359
2359
2359
2359
2359
2359
2359
2359
2359
2359
2359
2359
2359
2359
7359
Control
Combo S
ppm
2460
2460
2460
2460
2460
2460
2413
1142
259
259
239
239
239
239
239
239
239
239
239
239
239
239
239
239
239
239
239
239
239
239
239
239
239
239
239
239
239
239
239
239
739
1 49-state analysis includes all states except California.
7-36
-------
Estimated Costs of Low-Sulfur Fuels
Table 7.1-27
49-State Marine Diesel Fuel Volumes and Sulfur Contenta
Year
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
7.040
Total
Volume
1779
1,798
1,818
1,838
1,858
1,878
1,899
1,920
1,941
1,963
1,984
2,006
2,029
2,051
2,074
2,098
2,121
2,145
2,170
2,194
2,219
2,245
2,271
2,297
2,323
2,350
2,378
2,406
2,434
2,463
2,492
2,522
2,552
2,583
2,614
2,646
2,678
2,711
2,744
2,778
7 813
Affected
Volume
1,350
1,365
1,380
1,395
1,410
1,426
1,442
1,458
1,474
1,490
1,507
1,523
1,540
1,557
1,575
1,593
1,610
1,629
1,647
1,666
1,685
1,704
1,724
1,744
1,764
1,785
1,805
1,827
1,848
1,870
1,892
1,915
1,938
1,961
1,985
2,009
2,033
2,058
2,084
2,110
7 136
BaseS
ppm
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
Control S
ppm
3400
3400
3400
3400
3400
3400
3400
1615
340
340
340
340
340
340
340
340
340
340
340
340
340
340
340
340
340
340
340
340
340
340
340
340
340
340
340
340
340
340
340
340
340
Spillover
Volume
428
433
438
442
447
452
457
462
467
473
478
483
488
494
499
505
511
516
522
528
534
540
547
553
559
566
573
579
586
593
600
607
614
622
629
637
645
653
661
669
677
Spillover S
ppm
340
340
340
340
340
340
186
77
77
77
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
11
11
11
11
Base
Combo
S ppm
2663
2663
2663
2663
2663
2663
2626
2600
2600
2600
2584
2584
2584
2584
2584
2584
2584
2584
2584
2584
2584
2584
2584
2584
2584
2584
2584
2584
2584
2584
2584
2584
2584
2584
2584
2584
2584
2584
2584
2584
7584
Control
Combo S
ppm
2663
2663
2663
2663
2663
2663
2626
1245
277
277
261
261
261
261
261
261
261
261
261
261
261
261
261
261
261
261
261
261
261
261
261
261
261
261
261
261
261
261
261
261
761
1 49-state analysis includes all states except California.
7-37
-------
Draft Regulatory Impact Analysis
Table 7.1-28
48-State Nonroad Land-based Diesel Fuel Volumes and Sulfur Contenta
Year
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
7.040
Total
Volume
8,175
8,410
8,645
8,880
9,115
9,350
9,584
9,817
10,051
10,284
10,518
10,752
10,987
11,222
11,456
11,691
11,923
12,156
12,388
12,621
12,853
13,086
13,318
13,550
13,783
14,015
14,246
14,477
14,708
14,939
15,170
15,401
15,632
15,863
16,094
16,325
16,556
16,787
17018
17,249
17480
Affected
Volume
5,315
5,468
5,621
5,773
5,926
6,079
6,231
6,383
6,534
6,686
6,838
6,990
7,143
7,296
7,448
7,601
7,752
7,903
8,054
8,205
8,356
8,507
8,659
8,810
8,961
9,112
9,262
9,412
9,562
9,713
9,863
10,013
10,163
10,313
10,463
10,614
10,764
10,914
11,064
11,214
11 364
BaseS
ppm
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
Control S
ppm
3400
3400
3400
3400
3400
3400
3400
1615
340
340
148
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
11
11
11
Spillover
Volume
2,860
2,942
3,025
3,107
3,189
3,271
3,353
3,435
3,516
3598
3,680
3,762
3,844
3,926
4,008
4,090
4,172
4,253
4,334
4,415
4,497
4,578
4,659
4,741
4,822
4,903
4,984
5,065
5,146
5,227
5,307
5,388
5,469
5,550
5,631
5712
5,792
5,873
5,954
6,035
6 116
Spillover S
ppm
299
299
299
299
299
299
165
69
69
69
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
11
11
11
11
Base
Combo S
ppmb
2318
2318
2318
2318
2318
2318
2271
2237
2237
2237
2217
2217
2217
2217
2217
2217
2217
2217
2217
2217
2217
2217
2217
2217
2217
2217
2217
2217
2217
2217
2217
2217
2217
2217
2217
2217
2217
2217
2217
2217
7.7.17
Control
Combo S
ppmb
2318
2318
2318
2318
2318
2318
2271
1075
245
245
100
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
11
11
11
a 48-state analysis includes all states except Alaska and Hawaii.
b 50-state and 48-state Base and Control combination sulfur levels have been set to the average of 50 and 48-state values,
since the difference was negligible.
7-38
-------
Estimated Costs of Low-Sulfur Fuels
Table 7.1-29
48-State Locomotive Diesel Fuel Volumes and Sulfur Contenta
Year
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
Tn/in
Total
Volume
2821
2966
2,918
2969
3,010
3,051
3,081
3,111
3,124
3,146
3,169
3,223
3,237
3,246
3,255
3,270
3,303
3,322
3,340
3,358
3,369
3,399
3,430
3,460
3,491
3,522
3,554
3585
3,617
3,650
3,682
3,715
3748
3,782
3,815
3,849
3,884
3,918
3,953
3,989
A m/i
Affected
Volume
1,833
1,927
1,896
1,929
1,955
1,982
2,001
2,021
2,030
2,044
2,058
2,094
2,103
2,109
2,115
2,124
2,146
2,158
2,169
2,182
2,189
2,208
2,228
2,248
2,268
2,288
2,309
2,329
2,350
2,371
2,392
2,413
2,435
2,457
2,479
2,501
2,523
2,546
2,568
2,591
Tfil/1
BaseS
ppm
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
1Af\f\
Control S
ppm
3400
3400
3400
3400
3400
3400
3400
1615
340
340
340
340
340
340
340
340
340
340
340
340
340
340
340
340
340
340
340
340
340
340
340
340
340
340
340
340
340
340
340
340
1Af\
Spillover
Volume
988
1,039
1,022
1,040
1,054
1,069
1079
1,090
1,095
1102
1,110
1,129
1,134
1,137
1,140
1,146
1,157
1,164
1,170
1,177
1,181
1,191
1,202
1,212
1,223
1,234
1,245
1,256
1,267
1,279
1,290
1,302
1,313
1,325
1,337
1,349
1,361
1,373
1,385
1,398
1/1 in
Spillover S
ppm
299
299
299
299
299
299
165
69
69
69
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
11
11
11
1 1
Base
Combo S
ppmb
2396
2396
2396
2396
2396
2396
2352
2321
2321
2321
2302
2302
2302
2302
2302
2302
2302
2302
2302
2302
2302
2302
2302
2302
2302
2302
2302
2302
2302
2302
2302
2302
2302
2302
2302
2302
2302
2302
2302
2302
22Q2
Control
Combo S
ppmb
2396
2396
2396
2396
2396
2396
2352
1114
252
252
233
233
233
233
233
233
233
233
233
233
233
233
233
233
233
233
233
233
233
233
233
233
233
233
233
233
233
233
233
233
111
a 48-state analysis includes all states except Alaska and Hawaii.
b 50-state and 48-state Base and Control combination sulfur levels have been set to the average of 50 & 48-state values,
since the difference was negligible. Similarly, the Locomotive and Marine combination sulfur levels have been set to
their average to simplify the analysis.
7-39
-------
Draft Regulatory Impact Analysis
Table 7.1-30
48-State Marine Diesel Fuel Volumes and Sulfur Contenta
Year
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
Total
Volume
1786
1,805
1,825
1,845
1,865
1,886
1,906
1,928
1,949
1,970
1,992
2,014
2,037
2,059
2,082
2,106
2,130
2,154
2,178
2,203
2,228
2,254
2,279
2,306
2,333
2,360
2,387
2,415
2,444
2,473
2,502
2,532
2,562
2,593
2,624
2,656
2,689
2,722
2,755
2,789
2,824
Affected
Volume
1,273
1,287
1,301
1,315
1,330
1,344
1,359
1,374
1,389
1,405
1,420
1,436
1,452
1,468
1,485
1,501
1,518
1,535
1,553
1,570
1,588
1,607
1,625
1,644
1,663
1,682
1,702
1,722
1,742
1,763
1,784
1,805
1,827
1,849
1,871
1,894
1,917
1,940
1,964
1,989
2013
BaseS
ppm
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
3400
Control S
ppm
3400
3400
3400
3400
3400
3400
3400
1615
340
340
340
340
340
340
340
340
340
340
340
340
340
340
340
340
340
340
340
340
340
340
340
340
340
340
340
340
340
340
340
340
340
Spillover
Volume
513
518
524
530
535
541
547
553
560
566
572
578
585
591
598
605
611
618
625
632
640
647
654
662
670
677
685
693
702
710
718
727
736
745
754
763
772
781
791
801
811
Spillover S
ppm
299
299
299
299
299
299
165
69
69
69
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
11
11
11
11
Base
Combo S
ppmb
2396
2396
2396
2396
2396
2396
2352
2321
2321
2321
2302
2302
2302
2302
2302
2302
2302
2302
2302
2302
2302
2302
2302
2302
2302
2302
2302
2302
2302
2302
2302
2302
2302
2302
2302
2302
2302
2302
2302
2302
2302
Control
Combo S
ppmb
2396
2396
2396
2396
2396
2396
2352
1114
252
252
233
233
233
233
233
233
233
233
233
233
233
233
233
233
233
233
233
233
233
233
233
233
233
233
233
233
233
233
233
233
233
a 48-state analysis includes all states except Alaska and Hawaii.
b 50-state and 48-state Base and Control combination sulfur levels have been set to the average of 50 & 48-state values,
since the difference was negligible. Similarly, the Locomotive and Marine combination sulfur levels have been set to
their average to simplify the analysis.
7-40
-------
Estimated Costs of Low-Sulfur Fuels
7.1.5 Refinery Supply Volumes
After developing the 2008 volume estimates for the consumption of highway diesel fuel;
nonroad, locomotive and marine diesel fuel and other non-highway distillate fuel, it was
necessary to estimate the refinery supply volumes for each of these subpools to develop a
baseline for the refinery cost analysis for the proposed rule. The refinery supply volumes are
different from the consumption volumes because of the downgrade which occurs from the low
sulfur highway diesel fuel pool to the high sulfur non-highway diesel fuel pool during the
distribution between the refineries and the terminals. For the highway diesel rule promulgated in
2001, EPA estimated that downgrade would increase by 2.2 percent due to the 15 ppm highway
diesel fuel sulfur standard which takes effect in 2006. EPA also estimated that there is already a
2.2 percent downgrade due to the current 500 ppm sulfur which results in a total of 4.4 percent
downgrade from the highway diesel fuel pool to the non-highway distillate pool. The 4.4 percent
downgrade was applied equally in each PADD and the resulting volumes are representative after
any inter-PADD transfers have taken place. While the downgrade has not yet occurred in the
refinery supply table, the spillover volume is considered the same between the two tables as this
volume is an intended transfer from the highway to the nonroad diesel pool to avoid having to
invest capital investments in the distribution system.
The 4.4 percent highway downgrade is accounted for by dividing the highway diesel fuel
demand volume by 95.6 percent, and the downgraded highway diesel fuel was then added to the
high sulfur distillate fuel pool. The highway diesel fuel downgrade is presumed to all go to the
other non-highway distillate fuel (i.e., heating oil). This is a conservative estimate as it is likely
that much of this downgraded volume would be under 500 ppm and could be downgraded to the
500 ppm pools, either the 500 ppm highway, nonroad, locomotive and marine diesel pools from
2006 to 2010, or to the 500 ppm locomotive and marine pool after 2010. This assumption will
be reconsidered for the final rule.
The sulfur levels of the spillover volume for the refinery supply estimates were determined
differently from the fuel demand estimates, which assumed the same proportion of highway
15/500 ppm fuel in each non-highway subpool as in the highway pool. For the supply estimates,
we presumed all spillover into the NRLM diesel fuel is 15 ppm.
The result of the these adjustments in pool volumes to account for the downgrade in the
distribution system is summarized in Table 7.1-31. These are the diesel fuel volumes that were
used in the subsequent cost analysis.
7-41
-------
Table 7.1-31
Summary of Diesel Fuel Supply by PADD for 2008 (million gallons)
Category
levised Highway
^and-Based
'•Jonroad
^ocomotive
Vlarine
subtotal
NR, Loc, Marine)
Dther Off-
Tighway
TOTAL
Fuel Type
total j
15_p_pm diesel
500p_p_m diesel
highS
total j
15p_pm diesel
500p_p_m diesel
highS
total j
15p_pm diesel
500p_p_m diesel
highS
total j
15p_pm diesel
500p_p_m diesel
highS
total j
15p_pm diesel
500p_p_m diesel
highS
total j
15p_pm diesel
500p_p_m diesel
highS
total j
15p_pm diesel
500p_p_m diesel
hiehS
1 | 2 ! 3
1 1
13,000 _] 14,158 _j 7,156 j
9,647 J 10,186 j 5,042 _,
3,353 _] 3,972 _j 2,113 j
n/a ! n/a | n/a
2,166 J 3,559 J 2,273 _,
308 J 1,222 j 757 _,
_o_ I _o_ | _o_ i
1,858 j 2,337 j 1,517
509 j 1,256 J 700 j
53 J 410 j 221 j
_o_ I _o_ | _o_ i
456 j 846 j 479
481 j 295 j 1,014 j
50 J 96 j 320 j
_o_ I _o_ | _o_ i
431 j 199 j 693
3,156 J 5,111 j 3,987 _,
411 J 1,728 j 1,298 _,
_o_ I _o_ | _o_ i
2,745 j 3,382 j 2,689
8,001 j 1,165 j 785 _,
681 J 409 j 247 _,
217 _] 130 _j 79 j
7,103 j 627 j 460
24,157 J 20,434 J 1 1,927 j
10,739 J 12,323 j 6,587 _,
3,570 _] 4,102 _j 2,192 _,
9,848 ! 4,009 ! 3,148
4 ! 5 ! 5 | 5 ! U.S. -CA ! 5
! AZ,NV, OR, WA ! AK ! HI ! ! CA
1,859 _[ 2,398 _j 115 J 42 _j_ 38,728 _j 3,236 _,
1,199 J 1,750 J 84 J 29 _j_ 27,938 j 3,236 _,
659 _[ 648 _j 31 J 14 _j_ 10,790 _j 0 j
n/a ! n/a | n/a | n/a | n/a | n/a
822 J 567 j 66 j 32 J_ 9,486 j 663 _,
548 J 150 J 5 J 9 _j_ 2,999 j 663 _,
_o_ { P_ j _o_ j _o_ !_ o i _o_ ,
274 j 417 j 62 j 23 j 6,488 j 0
351 J 116 j 4 j 0 J_ 2,936 j 193 _,
231 J 29 J 0 J 0 _j_ 944 j 193 _,
_o_ [ p_ j _o_ j _o_ |_ o j _o_ ,
120 j 87 j 4 j 0 j 1,992 j 0
0 J 60 j 78 j 13 J_ 1,941 j 99 _,
0 J 15 J 4 J 3 _j_ 489 J 99 _,
_o_ [ p_ j _o_ j _o_ !_ o i _o_ ,
0 j 45 j 74 j 10 j 1,452 j 0
1,173 J 743 j 149 j 45 J_ 14,364 j 955 _,
779 J 194 J 9 J 12 _j_ 4,431 j 955 _,
_o_ [ p_ j _o_ j _o_ !_ o j _o_ ,
394 j 549 j 140 j 33 j 9,932 j 0
156 J 253 j 148 j 6 J_ 10,514 j 141 _,
106 J 63 J 6 J 1 _j_ 1,513 j 141 _,
34 _[ 20 _j 2 _j 0 _j_ 482 _j 0 j
16 j 169 j 140 j 4 j 8,518 j 0
3,188 J 3,394 j 412 j 93 J_ 63,605 j 4,332 _,
2,085 J 2,008 J 99 J 42 _j_ 33,882 j 4,332 _,
693 _[ 668 _j 33 _j 14 _j_ 11,272 _j 0 _,
410 ! 718 ! 280 ! 37 ! 18,451 ! 0
U.S.
L 41,964
L 31,174
L 10,790
n/a
L 10,149
L 3,661
L____°___.
6,488
L 3,129
L 1,136
L____°___.
1,992
L 2,040
L 588
L____°___.
1,452
L 15,318
L 5,386
L____°___.
9,932
L 10,654
L 1,654
L 482
8,518
L 67,937
L 38,214
L__LU72_
18,451
7-42
-------
Estimated Costs of Low-Sulfur Fuels
For using the supply volumes in Table 7.1-31 for estimating the cost-effectiveness of the
analysis contained in Chapter 8, it is necessary to project the supply volumes into future years for
the Proposed Two Step NRLM fuel program and the other options considered. This was done
using the applicable growth rates for each respective pool, considering the volume of small
refiner NRLM diesel fuel which is exempted from having to comply during the two exemption
periods (2007 to 2010, and 2010 to 2014), and considering the increased spillover of 500 ppm
and 15 ppm sulfur diesel fuel into the heating oil market. Spillover increases because refineries
in PADD 2 and 4 are not expected to be able to sell off their current high sulfur distillate pool
solely into the heating oil market.
The following table provide the base volumes used for estimating the total volume of 500
ppm diesel fuel and the total volume of 15 ppm diesel fuel affected by the Proposed Two step
Nonroad Program (this table is only shown for the Proposed Two Step fuel program).
Table 7.1-34 Future nonhighway Supply volumes for the U.S. outside of California for the Proposed Two Step NRLM Program
(MMGallons/yr)
Year
2000
2001
2002
2003
2004
2005
2006
2007
2008
2009
Jan-May '10
Jun-Dec'10
2011
2012
2013
Jan-May '14
Jun-Dec'14
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2031
2032
2033
2034
2035
2036
Nonroad
Total
Volume
7,716
7,938
8,160
8,382
8,603
8,825
9,046
9,266
9,486
9,707
9,927
9,927
10,149
10,370
10,591
10,813
10,813
1 1 ,034
1 1 ,254
1 1 ,473
11,693
11,912
12,131
12,351
12,570
12,790
13,009
13,228
13,446
13,664
13,882
14,100
14,318
14,536
14,754
14,972
15,190
15,408
15,626
HS Pool
and
Affected
Volume
5,277
5,429
5,580
5,732
5,884
6,036
6,186
5,654
5,788
5,923
6,057
5,893
6,025
6,156
6,288
6,419
7,545
7,700
7,853
8,006
8,159
8,312
8,465
8,619
8,772
8,925
9,078
9,231
9,383
9,535
9,687
9,839
9,992
10,144
10,296
10,448
10,600
10,752
10,904
Small
Refiner
Exempted
Volume
683
700
716
732
1034
1057
1080
1103
1126
Spillover
Volume
2,439
2,509
2,579
2,649
2,720
2,790
2,859
2,929
2,999
3,068
3,138
3,000
3,067
3,134
3,201
3,268
3,268
3,334
3,401
3,467
3,533
3,600
3,666
3,732
3,799
3,865
3,931
3,998
4,063
4,129
4,195
4,261
4,327
4,393
4,459
4,525
4,590
4,656
4,722
Locomotive
Total
Volume
2,651
2,787
2,742
2,790
2,828
2,867
2,895
2,923
2,936
2,956
2,977
2,977
3,028
3,042
3,050
3,059
3,059
3,073
3,104
3,122
3,138
3,156
3,166
3,194
3,223
3,252
3,281
3,310
3,339
3,369
3,399
3,430
3,460
3,491
3,522
3,553
3,585
3,617
3,649
HS Pool
and
Affected
Volume
,799
,892
,861
,893
,919
,946
,964
,770
,778
,790
,803
,951
,985
,994
,999
2,005
2,005
2,014
2,034
2,046
2,057
2,068
2,075
2,094
2,112
2,131
2,150
2,169
2,189
2,208
2,228
2,248
2,268
2,288
2,308
2,329
2,350
2,371
2,392
Small
Refiner
Exempted
Volume
214
215
216
218
Spillover
Volume
852
896
881
897
909
922
930
940
944
950
957
1,026
1,044
1,048
1,051
1,054
1,054
1,059
1,070
1,076
1,081
1,087
1,091
1,101
1,111
1,120
1,130
1,141
1,151
1,161
1,171
1,182
1,192
1,203
1,214
1,224
1,235
1,246
1,257
Marine
Total
Volume
,779
,798
,818
,838
,858
,878
,899
,920
,941
,963
,984
,984
2,006
2,029
2,051
2,074
2,074
2,098
2,121
2,145
2,170
2,194
2,219
2,245
2,271
2,297
2,323
2,350
2,378
2,406
2,434
2,463
2,492
2,522
2,552
2,582
2,613
2,644
2,675
HS Pool
and
Affected
Volume
1,331
1,345
1,360
1,375
1,390
1,405
1,421
1,282
1,296
1,310
1,325
1,416
1,431
1,447
1,463
1,480
1,480
1,496
1,513
1,530
1,548
1,565
1,583
1,601
1,620
1,638
1,657
1,677
1,696
1,716
1,736
1,757
1,778
1,799
1,820
1,842
1,864
1,886
1,908
Small
Refiner
Exempted
Volume
155
157
158
160
Spillover
Volume
448
453
458
463
468
473
478
484
489
494
500
569
575
581
588
595
595
601
608
615
622
629
636
643
651
658
666
674
682
690
698
706
714
723
731
740
749
758
767
Heating Oil
Other
Affected
Volume
(new
spillover)
500 ppm
0
0
0
0
0
0
0
636
642
648
654
556
561
567
572
577
577
583
588
593
599
605
610
616
622
627
633
639
645
651
657
663
669
676
682
688
695
701
708
Heating Oil
Other
Affected
Volume
(new
spillover)
15 ppm\
0
0
0
0
0
0
0
0
0
0
0
98
99
100
101
102
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
123
124
125
7-43
-------
Draft Regulatory Impact Analysis
The following table summarizes the estimated volume of 500 ppm diesel fuel and 15 ppm
diesel fuel affected by the Proposed Two step Nonroad Program by each year from 2007 to 2036.
The spillover of highway into the NRLM diesel pools and the total volume of NRLM diesel fuel
is also presented.
Table 7.1 -35 Future nonhighway Supply volumes for U.S. outside of California for
the Proposed Two Step NRLM Program (MMGallons/yr)
Totals (NR, Loc & Mar)
Year
2000
2001
2002
2003
2004
2005
2006
2007
2008
2009
Jan-May '10
Jun-Dec '10
2011
2012
2013
Jan-May '14
Jun-Dec '14
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2031
2032
2033
2034
2035
2036
Total
Volume
12,145
12,523
12,719
13,009
13,289
13,571
13,839
14,745
15,006
15,274
15,543
15,543
15,844
16,107
16,366
16,625
16,625
16,890
17,171
17,438
17,705
17,973
18,235
18,514
18,795
19,076
19,358
19,641
19,923
20,205
20,489
20,773
21,058
21,344
21,631
21,918
22,205
22,494
22,783
Affected Volume for 500 ppm
2007-2010
5,449
9,504
9,671
4,099
2010-2014
2,892
5,034
5,088
5,137
2,162
2014+
2,369
4,093
4,136
4,170
4,203
4,238
4,268
4,311
4,354
4,397
4,441
4,485
4,530
4,575
4,621
4,668
4,715
4,763
4,811
4,859
4,908
4,958
5,008
Affected Volume for 15 ppm
2010-2014
3,495
6,124
6,256
6,389
2,717
2014+
4,461
7,803
7,957
8,111
8,265
8,419
8,573
8,727
8,881
9,035
9,190
9,344
9,497
9,650
9,803
9,956
10,110
10,263
10,416
10,569
10,723
10,876
1 1 ,029
Spillover Volume
3,739
3,858
3,918
4,009
4,096
4,184
4,268
4,352
4,431
4,513
4,595
4,595
4,685
4,763
4,840
4,916
4,916
4,995
5,078
5,158
5,237
5,316
5,393
5,476
5,560
5,644
5,728
5,812
5,896
5,980
6,064
6,149
6,234
6,319
6,404
6,489
6,575
6,660
6,746
7-44
-------
Estimated Costs of Low-Sulfur Fuels
The following table summarizes the estimated volume of 500 ppm diesel fuel and 15 ppm
diesel fuel affected by the One step Nonroad Program (Option #1) by each year from 2008 to
2036. The spillover of highway into the NRLM diesel pools and the total volume of NRLM
diesel fuel is also presented.
Table 7.1-36 - Future nonhighway Supply volumes for the U.S. outside
of California for the One Step Program with Locomotive and Marine to
500 and NR to 15 in 2008 (MMGallons/yr)
Totals (NR, Loc & Mar)
Year
2000
2001
2002
2003
2004
2005
2006
2007
Jan-May '08
Jun-Dec '08
2009
2010
2011
Jan-May '12
Jun-Dec '12
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2031
2032
2033
2034
2035
2036
Total
Volume
12,145
12,523
12,719
13,009
13,289
13,571
13,839
14,109
14,364
15,006
15,274
15,543
15,844
16,107
16,107
16,372
16,631
16,896
17,177
17,445
17,711
17,980
18,241
18,521
18,802
19,083
19,365
19,648
19,930
20,213
20,496
20,781
21 ,066
21,351
21 ,638
21 ,925
22,213
22,502
22,783
Affected Volume for 500
ppm
2008-2012
2,251
3,892
3,927
3,981
1,671
2012+
2,343
4,043
4,071
4,102
4,145
4,179
4,213
4,248
4,278
4,321
4,363
4,407
4,451
4,495
4,540
4,586
4,632
4,679
4,726
4,773
4,822
4,870
4,919
4,969
5,012
Affected Volume for 15
ppm
2008-2012
3,293
5,775
5,905
6,035
2,569
2012+
4,280
7,493
7,648
7,804
7,958
8,112
8,266
8,420
8,574
8,728
8,882
9,036
9,191
9,345
9,498
9,651
9,804
9,958
10,111
10,264
10,417
10,571
10,724
10,877
1 1 ,029
Spillover Volume
3,739
3,858
3,918
4,009
4,096
4,184
4,268
4,352
4,431
4,431
4,511
4,591
4,682
4,760
4,760
4,836
4,912
4,991
5,074
5,154
5,233
5,312
5,389
5,472
5,556
5,640
5,723
5,808
5,891
5,976
6,060
6,144
6,229
6,314
6,399
6,485
6,570
6,656
6,742
7-45
-------
Draft Regulatory Impact Analysis
The following table summarizes the estimated volume of 500 ppm diesel fuel and 15 ppm
diesel fuel affected by the Two Step Nonroad Program with the 15 ppm sulfur standard being met
in 2009 (Option #2c) by each year from 2008 to 2036. The spillover of highway into the NRLM
diesel pools and the total volume of NRLM diesel fuel is also presented.
Table 7.1-37 - Future nonhighway Supply volumes for the U.S. outside of California
for the Nonroad Program which goes to 15 ppm in 2009 instead of 2010
(MMGallons/yr)
Totals (NR, Loc & Mar)
Year
2000
2001
2002
2003
2004
2005
2006
2007
2008
Jan-May '09
Jun-Dec '09
2010
2011
2012
Jan-May '13
Jun-Dec '13
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2031
2032
2033
2034
2035
2036
Total
Volume
12,145
12,523
12,719
13,009
13,289
13,571
13,839
14,745
15,006
15,274
15,274
15,543
15,844
16,107
16,366
16,366
16,625
16,890
17,171
17,438
17,705
17,973
18,235
18,514
18,795
19,076
19,358
19,641
19,923
20,205
20,489
20,773
21 ,058
21 ,344
21,631
21,918
22,205
22,494
22,783
Affected Volume for 500 ppm
2007-2009
5,449
9,504
3,731
2009-2013
2,859
4,959
5,036
5,089
2,141
2013+
2,354
4,063
4,095
4,137
4,172
4,205
4,240
4,270
4,313
4,356
4,399
4,443
4,487
4,532
4,578
4,624
4,670
4,717
4,765
4,813
4,861
4,910
4,960
5,010
Affected Volume for 15 ppm
2009-2013
3,418
5,423
5,543
5,663
2,409
2013+
4,370
7,647
7,803
7,957
8,111
8,265
8,419
8,573
8,727
8,881
9,035
9,190
9,344
9,497
9,650
9,803
9,956
10,110
10,263
10,416
10,569
10,723
10,876
1 1 ,029
Spillover Volume
3,739
3,858
3,918
4,009
4,096
4,184
4,268
4,352
4,431
5,229
5,296
5,161
5,264
5,355
5,444
5,444
4,914
4,993
5,076
5,156
5,235
5,314
5,391
5,474
5,558
5,642
5,726
5,810
5,894
5,978
6,062
6,147
6,231
6,316
6,402
6,487
6,572
6,658
6,744
7-46
-------
Estimated Costs of Low-Sulfur Fuels
The following table summarizes the estimated volume of 500 ppm diesel fuel and 15 ppm
diesel fuel affected by the Two Step Nonroad Program with locomotive and marine diesel fuel
complying with the 15 ppm sulfur standard along with nonroad in 2010 (Option #4) by each year
from 2008 to 2036. The spillover of highway into the NRLM diesel pools and the total volume
of NRLM diesel fuel is also presented.
Table 7.1 -38 Future nonhighway Supply volumes for the U.S. outside of California
for the Nonroad Program which has Locomotive and Marine going to 15 ppm in
2010 along with Nonroad (MMGallons/yr)
Totals (NR, Loc & Mar)
Year
2000
2001
2002
2003
2004
2005
2006
2007
2008
2009
Jan-May '10
Jun-Dec'10
2011
2012
2013
Jan-May '14
Jun-Dec'14
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2031
2032
2033
2034
2035
2036
Total
Volume
12,145
12,523
12,719
13,009
13,289
13,571
13,839
14,745
15,006
15,274
15,543
15,543
15,843
16,107
16,366
16,625
16,625
16,890
17,171
17,438
17,705
17,973
18,235
18,514
18,795
19,076
19,358
19,641
19,923
20,205
20,489
20,773
21,058
21,344
21,630
21,918
22,205
22,494
22,783
Affected Volume for 500 ppm
2007-2010
5,449
9,504
9,671
4,099
2010-2014
654
1,146
1,171
1,196
509
2014+
Affected Volume for 15 ppm
2010-2014
5,851
10,217
10,380
10,538
4,457
2014+
6,952
12,106
12,305
12,496
12,685
12,875
13,061
13,260
13,459
13,659
13,859
14,060
14,261
14,462
14,663
14,865
15,068
15,271
15,475
15,680
15,884
16,090
16,296
SpilloverVolume
3,739
3,858
3,918
4,009
4,096
4,184
4,268
4,352
4,431
4,513
4,595
4,392
4,480
4,557
4,632
4,707
4,707
4,784
4,865
4,943
5,020
5,098
5,173
5,254
5,336
5,417
5,499
5,580
5,662
5,744
5,826
5,908
5,990
6,073
6,155
6,238
6,321
6,404
6,488
7-47
-------
Draft Regulatory Impact Analysis
7.2 Refining Costs
The most significant cost involved in providing diesel fuel which meets 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 proposed 2007 Nonroad, Locomotive and Marine (NRLM) diesel fuel sulfur
standards and the 2010 nonroad diesel fuel standards,
comply with other NRLM diesel fuel sulfur options considered, and
• comply with the already promulgated 2006 highway diesel fuel sulfur standards (an update of
a previous cost analysis).
Finally, we compare our estimated costs to those developed by others who have evaluated the
refining costs of meeting tighter sulfur caps for non-highway diesel fuel.
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 three distinct desulfurization technologies:
conventional hydrotreating, the Linde Iso-Therming process and the Phillips SZorb adsorption
process. Conventional hydrotreating cost estimates were based on information from two
vendors, while the cost estimates for the other two more advanced processes was made from
information provided by the respective vendors. For all three 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 of
primarily advanced desulfurization technologies 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 cap standards, refiners are expected to have to desulfurize to 340 ppm and 7 ppm,
respectively.
Refining costs were developed for revamping existing hydrotreaters which 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 estimated the desulfurization
cost for each refinery in the U.S. producing distillate fuel in 2000. These refinery-specific costs
consider the volume of distillate fuel produced, the composition of this distillate fuel, 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,
7-48
-------
Estimated Costs of Low-Sulfur Fuels
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 judgement was
used to develop the needed specific cost estimate. The average desulfurization cost for each
refinery was based on the volume-weighted average of desulfurizing each of those blendstocks.
The production volumes used were those indicative of the year 2008, the first full year that the
proposed NRLM diesel fuel program would be applicable.
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,
Phillips, and Linde. 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.
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 2007 highway diesel fuel rulemaking. 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, the reader should
consult the Chapter 5 of the Regulatory Impact Analysis for the 2007 highway diesel fuel rule.
The few variations from the methodology described in that RIA are described below.
Next we present the methodology and resulting cost information used for developing the
refinery costs for the Phillips adsorption and Linde Iso-Therming processes. In this case, we
begin by presenting the estimates of the process design parameters provided by the developers of
these processes. These projections are then evaluated to produce sets of process design
parameters which 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 compete sets of such process design paramters from two out of the five or six
licensors of conventional desulfurization technologies.910 n 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.
7-49
-------
Draft Regulatory Impact Analysis
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
with the vendors.12 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 of the vendors indicated that operating pressures of no more than 900 pounds per square
inch (psi) would be sufficient to produce fuel meeting a 15 ppm sulfur cap. Most of the vendors
projected that 650 psi would be sufficient. Likewise, a number of refiners have indicated that
pressures well below 1000 psi would be sufficient. A contractor for API has indicated that they
believe that a 850 psi unit is all that is necessary meet a 15 ppm cap standard, although the
contractor also stated that lower pressure units would not be sufficient. Thus, we based our
estimate of capital cost on two different vendor submissions which were based on units operating
at 650 and 900 psi pressure.
Based on the information obtained from the two vendors of conventional hydrotreating
technologies, as well as that obtained from Phillips and Linde, we project that refiners would use
conventional hydrotreating to produce NRLM diesel fuel meeting the proposed 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 require a source of new hydrogen, an amine scrubber and a sulfur plant.
Most all 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. Thus, we project that any new hydrogen demand would likely have to be
produced from natural gas, either on-site or by a third party. Likewise, modest expansions of its
amine scrubber and sulfur plant would be required.
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 would follow the 500 ppm NRLM sulfur cap by only three
years, we project that refiners would have designed any new hydrotreaters built in 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 diesel fuel sulfur cap (average level of 7-
8 ppm) is summarized in Table 7.2-1.
7-50
-------
Estimated Costs of Low-Sulfur Fuels
Table 7.2-1
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
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 a number of hydrotreater
sub-units which are necessary for desulfurization and would save on both capital and operating
costs for a two stage revamp compared to 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 the feasibility section contained in Chapter 5.
In order to estimate the cost of meeting the proposed NRLM diesel fuel sulfur standards, it
was necessary to evaluate three situations which would be faced by refiners: 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 NRLM diesel fuel
meeting a 500 ppm cap from high sulfur distillate (i.e., grass roots 500 ppm hydrotreater). Sets
of process design parameters for the first two of these desulfurization configurations were
developed for the highway rule. 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 which 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.
One straightforward adjustment was made to all the capital costs developed for the 2007
highway diesel fuel rule. The capital costs developed for that rule were in terms of 1999 dollars.
These costs were increased by 2.5% to reflect construction costs in 2002 dollars.13
7.2.1.2.1.2 Revamping to Process 500 ppm Diesel Fuel to Meet a 15 ppm Cap
These process design projections developed in this section would 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 which apply to operating costs are
also relevant if a refiner would decide to replace their existing diesel fuel desulfurization unit
7-51
-------
Draft Regulatory Impact Analysis
with a new grassroots unit. In this case, the entire capital cost of the grass roots unit would be
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
RIA for the 2007 highway diesel fuel rule, with one adjustment. Diesel fuel complying with the
current 500 ppm sulfur standard typically contains 340 ppm sulfur. We expect refiners
complying with the proposed 500 ppm NRLM diesel fuel sulfur cap would also desulfurize down
to roughly 340 ppm sulfur. Thus, in revamping an existing 500 ppm hydrotreater to comply with
a 15 ppm cap, refiners would have to desulfurize from 340 ppm down to 7 ppm. This is
analogous to what we assumed in the analysis for the 2007 highway diesel fuel rule.
However, after the highway diesel fuel rule was finalized, 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 Draft RIA
to the 2007 highway diesel fuel rule, reducing the initial sulfur level from 500 ppm to 340 ppm
reduces hydrogen consumption by 3.5%. We assumed that all cost-related parameters (capital
cost, catalyst cost, yield losses, and utilities) would be reduced by the same 3.5%.
Table 7.2-6 presents the process design parameters for desulfurizing 500 ppm sulfur diesel
fuel to meet a 15 ppm cap standard.8
B There are no tables numbered 7.2-2 through 7.2-5.
7-52
-------
Estimated Costs of Low-Sulfur Fuels
Table 7.2-6
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 ($/BPSD)
Yield Loss (wt%)
Diesel
Naphtha
LPG
Fuel Gas
Straight Run
25,000
16
1.25
93
0.4
-
40
0.2
1.0
-0.7
-0.04
-0.04
Other Cracked Stocks
25,000
19
0.7
223
0.7
-
70
0.4
1.9
-1.3
-0.07
-0.11
Light Cycle Oil
25,000
22
0.6
362
0.8
-
80
0.5
2.1
-1.4
-0.08
-0.13
7.2.1.2.1.3 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 RIA for the 2007 highway diesel fuel rule. These costs would apply primarily to
refineries only producing, or predominantly producing, high sulfur diesel fuel today. In addition,
the capital cost portion of these costs would apply to a a refinery which replaced 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 which currently produce high sulfur
distillate fuel also produce some highway diesel fuel. In this case, we project costs which reflect
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 7.2-7 presents the process design parameters for desulfurizing high sulfur distillate fuel
to meet a 15 ppm cap standard in a grassroots unit.
7-53
-------
Draft Regulatory Impact Analysis
Table 7.2-7
Process Projections for Installing a New Grassroots Unit for Desulfurizing
Untreated Distillate Fuel Blendstocks to Meet a 15 ppm Cap 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
31
0.8
240
0.6
_
60
0.3
1.5
1.1
0.06
0.06
Other Cracked Stocks
25,000
37
0.5
850
1.1
_
105
0.6
2.9
2.0
0.11
0.17
Light Cycle Oil
25,000
42
0.4
1100
1.2
_
120
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 refiner to refiner and region to
region. A adjustment in hydrogen consumption was made for differing starting sulfur levels.
The basis for the amount of sulfur needed 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 Subsection 7.2.1.3, the average concentration
of sulfur in the overall distillate pool, and especially that part of the pool which is untreated,
varies by PADD. After estimating what this sulfur level is, 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.14) We did not adjust the
hydrogen consumption for the other qualities, polyaromatics and olefins, because we do not
believe that these would likely vary independently with the sulfur level. Since the removal of
sulfur consumes less than half the estimated hydrogen consumed as untreated 9000 ppm is
desulfurized to 15 ppm, 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 0.79 for PADD 5, which has an estimated untreated
distillate sulfur level of 2610 ppm, to 1.1 for PADD 3 which has an estimated untreated distillate
7-54
-------
Estimated Costs of Low-Sulfur Fuels
sulfur level of 11,320 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
which 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 5, which has
an estimated untreated sulfur level of 2540 ppm, to 0.87 for PADD 3 which has a starting sulfur
level of 5200 ppm, for these refineries without a distillate hydrotreater. The various hydrogen
consumption adjustment values are summarized in Table 7.2-8.c
Table 7.2-8
Hydrogen Consumption Adjustment Factors: Revamped Units
Refinery with Distillate HT
No Distillate HT
PADD1
0.90
0.81
PADD 2
0.88
0.80
PADD 3
1.1
0.87
PADD 4
0.84
0.79
PADD 5
0.79
0.79
7.2.1.2.1.4 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 cap 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 7.2-10.
c There is no table numbered 7.2-9.
7-55
-------
Draft Regulatory Impact Analysis
Table 7.2-10
Process Projections for Installing a New Unit for Desulfurizing
Untreated Diesel Fuel Blendstocks to Meet a 500 ppm Sulfur Cap 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
16
2.4
147
0.2
_
21
0.1
0.5
-0.4
-0.02
-0.02
Coker Distillate
25,000
19
1.9
628
0.4
_
37
0.2
1.1
-0.7
-0.04
-0.06
Light Cycle Oil
25,000
21
1.3
738
0.4
_
43
0.3
1.2
-0.85
-0.04
-0.07
Again, a hydrogen consumption adjustment was made for starting sulfur levels which 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. The adjustment is
applied as an adjustment ratio to each blendstock type and it ranged from 0.67 for PADD 5,
which has an estimated untreated distillate sulfur level of 2610 ppm, to 1.12 for PADD 3 which
has an estimated untreated distillate sulfur level of 11,320 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 which comprises the high sulfur pool. Thus, we estimate
a starting sulfur level which is somewhat lower. Our adjustment for these refineries ranged from
0.67 for PADD 5, which has an estimated untreated sulfur level of 2540 ppm, to 0.80 for PADD
3 which has a starting sulfur level of 5200 ppm. The various hydrogen consumption adjustment
values are summarized in Table 7.2-11.
7-56
-------
Estimated Costs of Low-Sulfur Fuels
Table 7.2-11
Hydrogen Consumption Adjustment Factors: Revamped Units
Refinery with Distillate HT
No Distillate HT
PADD 1
0.85
0.71
PADD 2
0.82
0.69
PADD 3
1.1
0.80
PADD 4
0.75
0.68
PADD 5
0.67
0.67
7.2.1.2.1.5 Hydrocrackate Processing and Tankage Costs
We believe that 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 cap
standard. The hydrocracker is a very severe hydrotreating unit capable of hydrotreating its
product from thousands of ppm sulfur to essentially zero ppm sulfur, however, hydrogen sulfide
recombination reactions which 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.1516 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 which can desulfurize the simple
sulfur compounds which are formed in the cracking stage of the hydrocracker.
Additionally, since the 15 ppm diesel sulfur standard is a very stringent cap standard, we are
taking into account tankage that would likely be needed. We believe that 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 in our cost model
of the installation of a tank that would store 10 days of 15 ppm sulfur diesel production sufficient
for a 10 day emergency turnaround which is typical for the industry, which would be about 3
million dollars for a 270,000 barrel storage tank.17 This amount of storage should be adequate
for most unanticipated turnarounds. We presumed that each refinery would need to add such
storage, (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 7.2-12.
7-57
-------
Draft Regulatory Impact Analysis
Table 7.2-12
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
—
—
—
—
nonea
Distillate Hydrocracker
Post Treat Reactor
25,000 (bbl/day)
5.718
0.98
4.2
18
5
—
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.19 The availability of a sulfur
analyzer at the refinery would provide essentially 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 cost for a diesel fuel sulfur analyzer would be about $50,000, and the installation cost would
be another $5000.20 Compared to 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 covered as a cost contingency described in
Subsection 7.2.1.4.1.
7.2.1.2.2. Sulfur Adsorption - Phillips SZorb
Phillips has developed a desulfurization technology applicable to either gasoline or diesel
fuel, as discussed in some detail in Chapter 5. At our request, Phillips provided process design
parameters for an SZorb diesel fuel desulfurization unit processing seven different feedstock
compositions. Table 7.2-13 summarizes the information provided.21'22
7-58
-------
Estimated Costs of Low-Sulfur Fuels
Table 7.2-13
SZorb Process Design Parameters
Diesel Fuel
Composition
Feed Sulfur (ppm)
Product Sulfur (ppm)
LHSV (hr1)
H2 Chemical
Consumption
Feed API Gravity
Feed 10%
Feed 50%
Feed 90%
Diesel A
HT: 80%
SR & 20%
LCOa
523
6
2
-5
33
440
513
604
Diesel B
83% SR &
17% HT
LCD
460
<1
2
-15
36
402
492
573
Diesel C
SR, CKR &
HT & nonHT
LCD
662
9
2
15
—
—
—
—
Diesel D
nonHT
SR
2000
<1
6
42
41
318
401
496
Diesel E
nonHT
LCD
2400
10
1
186
20
480
537
611
Diesel F
Diesel B
with some
non-hwy
1800
6
1.5
2
—
—
—
—
Diesel G
Diesel F
with HT
shutdown
3300
4
1
90
—
—
—
—
a HT = hydrotreated, nonHT = non-hydrotreated, SR = straight run, CKR = light coker gas oil
Diesel fuels A, B and C are somewhat representative of highway diesel fuel, although the
sulfur level is slightly higher than the average highway diesel fuel sulfur level found in the U.S.
Diesel fuel A is a hydrotreated blend of 80 percent straight run and 20 percent LCO with
distillation properties typical of today's highway diesel fuel. Diesel fuel B is a lighter blend than
diesel fuel A with 83 percent unhydrotreated straight run and 17 percent hydrotreated LCO.
Diesel fuel C is a rather typical diesel fuel composition for a refinery with an FCC unit and a
coker. The distillation qualities of this diesel fuel, and those of diesel fuels F and G are not
known.
Diesel fuels D, E, F and G have moderate sulfur levels more typical of non-highway distillate
fuels. However, these fuels' sulfur contents are not as high as most refiners' unhydrotreated
distillate, as discussed above. Diesel fuel D is comprised of unhydrotreated straight run with
distillation qualities lighter than the average diesel fuel. Diesel fuel E is 100% unhydrotreated
LCO although its relatively low sulfur level suggests that it is either from a sweet crude refinery
or from a refinery with a FCC feed hydrotreater. Its distillation curve is typical of the LCO
blended into the diesel fuel pool. Diesel fuel F consists of diesel fuel B plus some amount of
non-highway diesel fuel. The composition of the non-highway diesel fuel is unknown.
However, if we assume that the non-highway diesel fuel contains the national average sulfur
level of 3400 ppm, the sulfur level of this diesel fuel blend suggests that it may be about half
non-highway and half diesel fuel B. However, if the sulfur content is closer to the maximum
5000 ppm allowed under ASTM specifications, then diesel fuel F might only contain 25-30%
non-highway diesel fuel. Diesel fuel G consists of diesel fuel F with the highway hydrotreater
shutdown. The highway hydrotreater appears to have been only hydrotreating the LCO fraction
of diesel fuel B, which represents less than 17% of Diesel fuel G. Since the sulfur content of
7-59
-------
Draft Regulatory Impact Analysis
diesel fuel G exceeds that of diesel fuel F by 1500 ppm, the sulfur content of the unhydrotreated
LCO in diesel fuel B must be 9000 ppm or more, which is typical.
The design parameters provided by Phillips involve a stand-alone SZorb unit, sometimes
processing unhydrotreated feedstock, and sometimes processing partially hydrotreated feedstock
(i.e., following an existing conventional hydrotreater). In all cases, sulfur content is being
reduced to 10 ppm or less, very close to the 7-8 ppm target which we expect refiners to have to
achieve on average to comply with a 15 ppm cap. Below, we will use the process design
parameters shown in Table 7.2-13 to project the cost of an SZorb unit performing two different
tasks: 1) processing current high sulfur distillate to meet a 15 ppm cap, and 2) processing 500
ppm NRLM diesel fuel to meet a 15 ppm cap. The methodology used to develop the projected
costs for these two tasks are presented below. As was done for conventional hydrotreating, we
will develop cost estimates for processing three individual blendstocks: straight run, LCO and
light coker gas oil, in order to be able to project desulfurization costs for individual refineries
whose diesel fuel compositions vary dramatically.
7.2.1.2.2.1 Desulfurizing High Sulfur Distillate Fuel to Meet a 15 ppm Sulfur Cap
Phillips provided four sets of process design parameters for using SZorb to achieve a 15 ppm
cap from high sulfur distillate. Two of these designs treated a pure blendstock, straight run and
LCO. However, neither blendstock had properties typical of these blendstocks for the average
refinery. Thus, the four sets of process designs have to evaluated together to develop sets of
process design parameters for the three distillate blendstocks.
Also, the maximum initial sulfur level shown in Table 7.2-13 is 3300 ppm sulfur. Thus, we
believe that it is reasonable to limit the applicability of our projections to feedstocks containing
3300 ppm sulfur or less. Current high sulfur distillate averages 3400 ppm sulfur. Therefore, it is
reasonable to expect that just under half of all NRLM diesel fuel contains 3300 ppm sulfur or
less.
We have broken down the derivation of the cost of a stand-alone SZorb unit capable of
producing 15 ppm diesel fuel into four parts: hydrogen consumption, utilities and yield losses,
catalyst cost and capital cost.
Hydrogen Consumption: Phillips provided an estimate of hydrogen consumption for two
individual blendstocks, straight run and LCO. Diesel fuel D was an unhydrotreated straight run
stream hydrotreated to less than 1 ppm sulfur. Doing so consumed 42 scf/bbl of feed.
Comparing its sulfur content to those for typical high sulfur distillate from Chapter 5.1, it
contains 90% of the amount of sulfur in average unhydrotreated straight run. However, its
distillation properties show this stream to be lighter than the average straight run feedstock.
Thus, the hydrogen consumption for a more typical could be slightly higher, due to the greater
concentration of aromatics typical for heavier cuts of distillate.
7-60
-------
Estimated Costs of Low-Sulfur Fuels
Interestingly, the hydrogen consumption for a number of the diesel fuel feedstocks including
some LCO shown in Table 7.2-13 is less than 42 scf/bbl. A couple even show a net production
of hydrogen. One of the aspects of the SZorb process is that the temperature can be varied to
control the level of aromatics saturation. At low temperatures, aromatics content can actually be
increased, generating hydrogen. However, there is a practical limit to this, as higher aromatic
contents reduce cetane. This flexibility of the SZorb process makes it difficult to accurately
predict typical hydrogen consumption, as each refiner's ability to absorb a loss in cetane will
differ. Thus, to be conservative, we assumed that the 42 scf/bbl hydrogen consumption of diesel
fuel D was representative of treating typical straight run.
The LCO feed (diesel fuel E) contains 2400 ppm sulfur, which is about two-thirds of the
average amount of sulfur in unhydrotreated LCO, which contains 3500 ppm sulfur, as discussed
in Chapter 5.1. However, it was desulfurized to a lower sulfur level than what would be
expected for meeting the 500 ppm sulfur target (typically around 340 ppm). LCO usually
comprises about 25 percent of diesel fuel for the average refinery with an FCC unit. LCO will
likely contribute the most amount of sulfur after hydrotreating, because it generally contains
largest concentration of sterically hindered sulfur compounds. Thus, to meet a 7-8 ppm sulfur
target, LCO will likely need to be desulfurized to about 20 ppm and contribute about 5 ppm to
the diesel fuel pool. Coker distillate might be desulfurized to roughly 5 ppm and straight run to
1-2 ppm, resulting in an average diesel fuel sulfur level of about 7 ppm. Therefore, the hydrogen
consumption of 186 scf/bbl is lower than that for average LCO due to its low initial sulfur level,
but is high (for a 7-8 ppm target), due to its final sulfur level of 10 ppm. Lacking the ability to
compensate for either of these two factors, we assumed that the hydrogen consumption of 186
scf/bbl for diesel fuel E was representative of the hydrogen consumption for average LCO.
Phillips did not provide an estimate of the hydrogen consumption for treating 100% coker
distillate. Therefore, we assumed that the relationship for hydrogen consumption for straight run,
light coker distillate and LCO being treated by an SZorb unit would be the same as that for
conventional hydrotreating. As described in Table 7.2-7, the hydrogen consumptions for
conventionally hydrotreating straight run, coker distillate and LCO to 15 ppm are estimated to be
240, 850 and 1100 scf/bbl. Thus, for conventional hydrotreating, the hydrogen consumption for
coker distillate falls 70 percent of the way between straight run and LCO. Thus, if we apply this
same percentage to the straight run and LCO hydrogen consumption values for SZorb, we
estimate that coker distillate would consume 144 scf/bbl of hydrogen.
An adjustment to these hydrogen consumptions was developed to account for differences in
initial sulfur levels. The Phillips' feedstocks upon which the above hydrogen consumption
estimates were based contain about 2100 ppm sulfur. However, as described below in subsection
7.2.1.3, the average concentration of sulfur in the overall distillate pool, and especially that part
of the pool which is untreated, exceed 2100 ppm sulfur. To account for the additional hydrogen
which would be consumed when processing higher sulfur feeds, we increased hydrogen
consumption by 12.5 scf/bbl for each 1000 ppm of additional initial sulfur content. We did not
adjust hydrogen consumption for other feedstock qualities, such as polyaromatics and olefins,
because we do not believe that these would likely vary consistently with the sulfur level.
7-61
-------
Draft Regulatory Impact Analysis
Combination with sulfur represents a significant portion of total hydrogen consumption for
the Phillips process. The sulfur levels of the feeds reported by Phillips are significantly lower
than the average initial sulfur levels in each PADD. Thus, the adjustments are always greater
than 1.0.
For refineries with a distillate hydrotreater, the adjustment factor ranged from 1.1 for PADD
5, which has an estimated untreated distillate sulfur level of 2610 ppm, to 2.4 for PADD 3, which
has an estimated untreated distillate sulfur level of 11,320 ppm. No adjustment is necessary for
the already hydrotreated part of the distillate pool. This sub-pool is always assumed to contain
340 ppm sulfur, for which we use the hydrogen consumptions developed in the next section
which evaluates adding SZorb unit after a hydrotreater producing 500 ppm diesel fuel.
For refineries without a distillate hydrotreater, they have no hydrotreated blendstocks, but
still meet applicable sulfur limits. Thus, we estimate lower initial sulfur levels than those
mentioned above. Our adjustment factors for these refineries ranged from 1.1 for PADD 5,
which has an estimated initial sulfur level of 2540 ppm, to 1.5 for PADD 3 which has a initial
sulfur level of 5200 ppm. The various hydrogen adjustment factors for refineries with and
without a hydrotreater are summarized in the following table:
with Dist HT
no Dist HT
PADD1
1.6
1.2
PADD 2
1.5
1.1
PADD 3
2.4
1.5
PADD 4
1.3
1.1
PADD 5
1.1
1.1
Utilities and Yield Losses: Phillips did not provide specific estimates of the utility demands
for the seven designs shown in Table 7.2-13. Thus, we estimated utility demands based on a
comparison of SZorb to conventional hydrotreating, guided by some general estimates provided
by Phillips. The largest consumer of electricity in conventional hydrotreating is the hydrogen
compressor. The SZorb process differs, in that the absorption catalyst must be recycled between
the reactor and the regeneration reactor. While the SZorb process consumes much less hydrogen,
it operates like conventional hydrotreating in that it requires an excess amount of hydrogen to be
mixed with the diesel fuel. Thus, the total amount of hydrogen being compressed is roughly the
same. However, the SZorb process operates at 275-500 psi, which is about half of the pressure at
which conventional desulfurization operates. Although the SZorb process operates at about half
the pressure of conventional hydrotreating and demands less hydrogen, the need to recycle
catalyst likely offsets some of the savings related to hydrogen compression. Still, we estimated
that the electrical demand for SZorb would be one half that of conventional hydrotreating shown
in Table 7.2-7. Thus, we estimate SZorb's electricity demand to be 0.3, 0.55, and 0.6 kW-hr/bbl
for straight run, light coker gas oil, and LCO, respectively.
Concerning fuel gas, it is used to heat up the feed to enable the desulfurization reaction to
occur. The SZorb process operates at about the same temperature as conventional hydrotreating.
Thus, we assumed that fuel gas demand would be the same as conventional hydrotreating listed
7-62
-------
Estimated Costs of Low-Sulfur Fuels
in Table 7.2-7, or 60, 105, and 120 btu/bbl for straight run, light coker gas oil, and LCO,
respectively.
Due to its use of adsorption, instead of hydrogenation, SZorb essentially has no yield losses.
Catalyst Costs: Conversations with Phillips indicated that the catalyst is likely to be cheaper
than conventional hydrotreating catalysts. However a significant level of catalyst demand would
have to occur to realize economies of scale for this lower cost to be realized. Since this process
is just emerging and not many units have been licensed yet, we decided to assume the same
catalyst cost as for conventional hydrotreating (shown in Table 7.2-7), or 0.3, 0.6, and 0.8
$/BPSD for straight run, light coker gas oil, and LCO, respectively.
Capital Costs: In their literature, Phillips only provides capital cost estimates for an SZorb
unit processing 500 ppm sulfur diesel fuel to meet a 15 ppm sulfur cap. To be conservatived, we
assumed that the ratio of the capital costs of SZorb units treating high sulfur and 500 ppm diesel
fuel to meet a 15 ppm cap, would be the same as the ratio of the capital costs of conventional
hydrotreaters doing the same thing. This ratio is a factor of two for conventional hydrotreating.
Phillips estimates that an SZorb unit reducing sulfur content from 500 ppm to 15 ppm would be
$48 million for a 40,000 bbl/day unit, fully installed on the Gulf Coast. Thus, the cost of an
SZorb unit treating the same volume of high sulfur distillate would be $96 million.
A number of steps still needed to be performed before this cost could be converted to a
capital cost for processing the three individual diesel fuel blendstocks on a basis comparable to
those for conventional hydrotreating above. First, we assumed that this cost included some
provision for off-site costs, while the primary capital costs estimates, such as those shown in
Tables 7.2-6 and 7.2-7 for conventional hydrotreating, only include "inside battery limit" (ISBL)
costs. Thus, we divided the $96 million cost by a factor of 1.2 (from Table 7.2-23 below) to
remove off-site costs. This produced an ISBL cost of $80 million. We then scaled this ISBL
cost down to represent that for a 25,000 bbl/day unit using the "six-tenths rule with an exponent
of 0.65. The scaling factor representing this reduction in volumetric capacity is 0.74 ((25,000 /
40,000)°'65). Multiplying the $80 million cost by this factor produced a revised cost of $59
million for a 25,000 bbl/day unit.
The final step was to convert this cost for processing a mix of blendstocks to those for
processing individual blendstocks. We assumed that this unit was designed to process a typical
diesel fuel comprised of 69% straight run, 23% LCO and 8% other cracked stocks. We also
assumed that the relationship between the capital costs for specific blendstocks (straight run,
coker distillate and LCO) for SZorb was the same as those for conventional hydrotreating. Using
the capital costs for conventional hydrotreating of $31, $37 and $42 million for straight run,
coker distillate and LCO from Table 7.2-7, the capital cost for the above average feed
D The assumption is likely conservative because the SZorb desulfurization process is likely
able to handle higher sulfur levels partially or perhaps even primarily by a higher rate of catalyst
recycle as opposed to just increasing the unit size.
7-63
-------
Draft Regulatory Impact Analysis
composition would be $34 million. Thus, the capital cost for an SZorb unit would be 1.7 times
higher (59/34). Multiplying the capital costs for processing the individual blendstocks using
conventional hydrotreating by 1.73 produced SZorb capital costs of $53, $63 and $72 million for
straight run, coker distillate and LCO average.
Summary of Process Design Parameters: The process design parameters for a new, 25,000
bbl/day SZorb unit are summarized in Table 7.2-14.
Table 7.2-14
Process Design Parameters for a New SZorb Unit Desulfurizing High Sulfur
Distillate Fuel to Meet a 15 ppm Cap Standard
Capital Cost (Smillion)
Unit Size (BPSD)
Hydrogen Demand (scf/bbl)
Electricity Demand (kW-hr/bbl)
Fuel Gas Demand (btu/bbl)
Catalyst Cost ($/BPSD)
Yield Loss
Straight Run
52
25,000
42
0.3
60
0.3
0
Other Cracked Stocks
63
25,000
144
0.55
105
0.6
0
Light Cycle Oil (LCO)
71
25,000
186
0.6
120
0.8
0
7.2.1.2.2.2 Desulfurizing 500 ppm Diesel Fuel to Meet a 15 ppm Cap
We next estimated the process design parameters for an SZorb unit which treats distillate
which has already been hydrotreated to meet a 500 ppm cap down to 7-8 ppm. We assume that
this feed contains 340 ppm sulfur, the average sulfur content of highway diesel fuel today outside
of California.
Phillips provided three sets of process design parameters for using SZorb to achieve a 15 ppm
cap from distillate with sulfur contents just above 340 ppm (diesel fuels A, B, and C). None of
these designs treated a pure blendstock. Thus, the three sets of process designs have to evaluated
together to develop sets of process design parameters for the three distillate blendstocks.
As we did above, we have broken down the derivation of the cost of an SZorb unit capable of
producing 15 ppm diesel fuel from 500 ppm diesel fuel into four parts: hydrogen consumption,
utilities and yield losses, catalyst cost and capital cost.
Hydrogen Consumption: In order to estimate the hydrogen consumption for each
blendstock type, we focused on diesel fuel A, which is an 80/20 blend of straight run and LCO.
This fuel is the closest to the composition of average diesel fuel of diesel fuels A, B, and C
shown in Table 7.2-13. Processing diesel fuel A to 6 ppm sulfur actually produces 5 scf/bbl of
7-64
-------
Estimated Costs of Low-Sulfur Fuels
hydrogen. We assumed that processing straight run would produce 10 scf/bbl of hydrogen, while
processing LCO would consume 15 scf/bbl. An 80/20 weighting of these two figures produces a
net hydrogen production of 5 scf/bbl, precisely that for diesel fuel A. We again based the
hydrogen consumption for processing coker distillate on its relative hydrogen consumption when
using conventional hydrotreating. There, the hydrogen consumption to process coker distillate
falls 70% of the way between those for straight run and LCO. Here, 70% of the way between -5
and +10 is +5. Thus, we assumed that the hydrogen consumption for coker distillate would be 5
scf/bbl.
Utilities and Yield Losses: As we assumed for an SZorb unit processing high sulfur
distillate, we assumed that electricity demand for an SZorb unit processing 500 ppm diesel fuel
would be half that for conventional hydrotreating (shown in Table 7.2-6), or 0.19, 0.34, and 0.39
kW-hr/bbl for straight run, light coker gas oil, and LCO, respectively.
Concerning fuel gas, as we assumed for an SZorb unit processing high sulfur distillate, we
assumed that fuel gas demand for an SZorb unit processing 500 ppm diesel fuel would be the
same catalyst cost as for conventional hydrotreating (shown in Table 7.2-6), or 38, 68, and 77
btu/bbl for straight run, light coker gas oil, and LCO, respectively.
Due to its use of adsorption, instead of hydrogenation, SZorb essentially has no yield losses.
Catalyst Costs: As we assumed for an SZorb unit processing high sulfur distillate, we
assumed that the catalyst costs for an SZorb unit processing 500 ppm diesel fuel would be the
same catalyst cost as for conventional hydrotreating (shown in Table 7.2-6), or 0.1, 0.2, and 0.24
$/BPSD for straight run, light coker gas oil, and LCO, respectively.
Capital Costs: As mentioned above, Phillips estimates that an SZorb unit reducing sulfur
content from 500 ppm to 15 ppm would be $48 million for a 40,000 bbl/day unit installed on the
Gulf Coast, including off-site costs. We divided by a factor of 1.2 to remove off-sites for a new
unit (see Table 7.2-23), producing an ISBL cost of $40 million. We scaled this cost down to
represent that of a 25,000 bbl/day unit using the "six-tenths rule with an exponent of 0.65. This
produced a scaling factor of 0.74 and a revised ISBL cost of $29 million. As we did above, we
assumed that this unit was designed to process a typical diesel fuel comprised of 69% straight
run, 23% LCO and 8% other cracked stocks. We again assumed that the relationship between
the capital costs for specific blendstocks (straight run, coker distillate and LCO) for SZorb was
the same as those for conventional hydrotreating.
Using the capital costs for conventional hydrotreating of $15, $18 and $21 million for straight
run, coker distillate and LCO from Table 7.2-7, the capital cost for the above average feed
composition would be $17 million. Thus, the capital cost for an SZorb unit would be 1.7 times
higher, or $27, $32 and $37 million for straight run, coker distillate and LCO average.
Summary of Process Design Parameters: The process design parameters for a new, 25,000
bbl/day SZorb unit are summarized in Table 7.2-15.
7-65
-------
Draft Regulatory Impact Analysis
Table 7.2-15
Process Design Parameters for an SZorb Unit Desulfurizing 500 ppm
Distillate Fuel to Meet a 15 ppm Cap Standard
Blendstocks from 500 ppm to 15 ppm
Capital Cost ($MM)
Unit Size (bbl/stream Day)
Hydrogen Demand (scf/bbl)
Electricity Demand (kwh/bbl)
Fuel Gas Demand (btu/bbl)
Catalyst Cost ($/bpsd)
Yield Loss
Straight Run (SR)
27
25,000
-10
0.19
39
0.1
0
Other Cracked Stocks
32
25,000
8
0.34
68
0.19
0
Light Cycle Oil (LCD)
37
25,000
15
0.39
77
0.24
0
7.2.1.2.3 Linde Isotherming
Linde has licensed a technology called IsoTherming which is designed to desulfurize both
highway and non-highway distillate fuel. Upon our request, Linde provided basic design
parameters for their process which could be used to project the cost of using their process to meet
tighter sulfur caps.23 Specifically, Linde provided design parameters for a revamp of an existing
highway desulfurization unit to meet a 15 ppm cap standard. The revamp would put an
IsoTherming unit upstream of the existing highway diesel fuel hydrotreater.
Linde provided IsoTherming designs for three revamp situations. In the first 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 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
most refiners than 1500 psi. The higher pressure unit would reduce capital and catalyst costs, but
higher hydrogen consumption would offset much of the cost savings. The higher pressure
reactors and compressors would also have a longer delivery time and there would likely be fewer
fabricators to select from. Thus, given that the savings associated with the higher pressure unit
were small, we decided to focus on the 950 psi design.
7-66
-------
Estimated Costs of Low-Sulfur Fuels
The information provided by Linde for the 950 psi IsoTherming desulfurization unit is
summarized in the following table. The operation and product quality of the IsoTherming unit is
shown separate from those for the existing conventional hydrotreater. Again, prior to the
revamp, the conventional hydrotreater would have processed this feedstock down to roughly 300-
400 ppm sulfur.
Table 7.2-16
Linde 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 (hf1)
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
7.2.1.2.3.1 Hydrotreating High Sulfur Distillate Fuel to 15 ppm
The design parameters provided by Linde 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 Linde revamp design to project the costs of an IsoTherming unit which 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 proposed two-step fuel program. However, it is
7-67
-------
Draft Regulatory Impact Analysis
projected to be used under the one-step alternative fuel program, for which costs are also
estimated later in this chapter.
Also, as was done for conventional hydrotreating and the Phillips SZorb process, we will
develop cost estimates for applying the IsoTherming process to three individual blendstocks,
straight run, LCO and light coker gas oil, in order to be able to project desulfurization costs for
individual refineries whose diesel fuel compositions vary dramatically.
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. Linde
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, Linde provided information for a hybrid desulfurization unit which is comprised of
a Linde IsoTherming unit which is revamping a conventional highway hydrotreater. Upon our
request for additional information, Linde informed us that the highway hydrotreater in the above
described revamp is operating similar to an IsoTherming unit. Thus, we used the hydrogen
consumption estimates in Tables 7.2-16 as if they represented a stand-alone IsoTherming unit.
As a first step in estimating the IsoTherming hydrogen consumption for individual
blendstocks, we compared the hydrogen consumption of the Linde IsoTherming process with that
of conventional hydrotreating. Table 7.2-16 shows a total hydrogen consumption of consumes
420 scf/bbl to desulfurize untreated diesel fuel to 10 ppm. Using the projected hydrogen
consumption for conventional hydrotreating shown in Table 7.2-7 above, the total hydrogen
consumption to desulfurize this same feedstock to 7-8 ppm would be 559 scf/bbl. Based on this
example, the IsoTherming process appears to only consume 75 percent of that associated with
conventional hydrotreating. Thus, we assumed that the Linde IsoTherming process would only
use 71% of the hydrogen which we projected above for processing individual blendstocks using
conventional hydrotreating. The resulting hydrogen consumptions were 826 scf/bbl for LCO,
638 scf/bbl for other cracked stocks, and 180 scf/bbl for straight run.
As we did for conventional hydrotreating and the Phillips SZorb process, we developed
adjustments to these hydrogen consumptions 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 (see subsection
7.2.1.2.1.3). The adjustment is applied as a multiplicative factor to the above base hydrogen
consumption for each untreated blendstock type. For a refinery with a distillate hydrotreater, it
ranged from 0.79 for PADD 5, which has an estimated untreated distillate sulfur level of 2610
ppm, to 1.08 for PADD 3 which has an estimated untreated distillate sulfur level of 11,320 ppm.
No adjustment factors are applied to blendstocks which are already hydrotreated. These
7-68
-------
Estimated Costs of Low-Sulfur Fuels
blendstocks are assumed to contain 340 ppm sulfur. The hydrogen consumption for the
IsoTherming process when applied to diesel fuel with this initial sulfur level is described in the
next Section 7.2.1.2.3.2 below.
Refineries without a distillate hydrotreater do not have any any hydrotreated blendstocks to
blend into their high sulfur distillate. Thus, we estimate lower unhydrotreated sulfur levels for
these refineries. The adjustment factors for these refineries ranged from 0.79 for PADD 5, which
has an estimated untreated sulfur level of 2540 ppm, to 0.87 for PADD 3 which has a starting
sulfur level of 5200 ppm.
Utilities and Yield Losses: We next established the IsoTherming utility inputs for
individual blendstocks. The Linde IsoTherming process saves a substantial amount of heat input
by conserving the heat of reaction which occurs in the IsoTherming reactors. This conserved
energy is used to heat the feedstock to the unit. This differs from conventional hydrotreating
which normally must reject much of this energy to avoid coking the catalyst. According to
Linde, this allows the IsoTherming process to operate with essentially no external heat input. In
the highway hydrotreater revamp which is the source of the information provided by Linde, the
existing heater for the highway hydrotreater could essentially be turned off 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, this little amount of fuel used during start-up is negligible. Thus, we estimate
need for either fuel or steam with the IsoTherming process.
As shown in Table 7.2-16, Linde estimated electricity demand at 1525 kilowatts. The unit
was designed to process 20,000 bbl/day, so the unit electricity demand was 1.83 kilowatt-hour
per barrel (kw-hr/bbl). Because the electricity demand value was not provided separately for the
IsoTherming and the original conventional highway hydrotreater, we assumed that this demand
applied to a stand-alone IsoTherming unit, as well. Since most of the electrical demand is due to
the compression of hydrogen, and we are using the same hydrogen consumption as shown in
Table 7.2-16, it is consistent that the electrical demand would be the same.
We compared this electrical demand to that of conventionally hydrotreater treating the same
feedstock. Using the elecricity demands from Table 7.2-7 above, we project that the electrical
demand of conventional hydrotreating would be 0.83 kW-hr/bbl. Thus, IsoTherming appears to
consume 2.2 times as much electicity, probably due to increased liquid pumping associated with
liquid recycle to the reactors. We assumed that this 2.2 factor applied to each individual
blendstock. Thus, we estimate electricity demand at 1.3, 2.4, and 2.6 kW-hr/bbl for straight run,
light coker gas oil, and LCO, respectively.
Linde did not estimate the specific yield losses for the for the IsoTherming process. Upon
our request for further information, Linde indicated that their process causes slightly less than
half of the yield loss of conventional hydrotreating. Thus, the yield loss of the Linde 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 below:
7-69
-------
Draft Regulatory Impact Analysis
Diesel
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 Linde process was estimated based on the relative
catalyst volume compared to conventional hydrotreating. As shown in Table 7.2-16, Linde
indicated that the catalyst volume for the new IsoTherming reactors contained only 45% of the
volume of the new conventional hydrotreating reactors which Linde projects would have been
needed to revamp the existing hydrotreater to produce 10 ppm fuel. We assumed that this same
relationship would hold for a stand-alone IsoTherming unit. Thus, we multiplied the catalyst
costs for conventionally hydrotreating specific blendstocks (shown in Table 7.2-7) by 45%. The
resulting IsoTherming catalyst costs were 0.14, 0.27 and 0.36 $/BPSD for straight run, light
coker gas oil and LCO, respectively.
Capital Costs: The last aspect of the IsoTherming process to be determined on a per-
blendstock basis is its capital cost. Linde'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 to those involved in conventional hydrotreating. As indicated in
Table 7.2-16, 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 individul blenedstocks from Table 7.2-6). 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 only required 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 would be 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 overall capital cost for the
specific feed composition shown in Table 7.2-16 above would be $900 per BPSD for the
IsoTherming unit, versus $1400 per BPSD for a conventional hydrotreater. More recently, Linde
indicated that the capital cost would be roughly $800 per barrel for a 25,000 bbl per day unit.24
For this analysis, we retained the two-thirds factor relative to conventional hydrotreating ($900
per BPSD). We are considering reducing this cost by 11% to match that of the most recent Linde
estimate for our analyses following the proposed rule.
Summary of Process Design Parameters: Table 7.2-17 summarizes the design parameters
used for using the Linde IsoTherming process to desulfurize untreated distillate fuel to 10 ppm.
7-70
-------
Estimated Costs of Low-Sulfur Fuels
Table 7.2-17
Process Parameters for a Stand-Alone IsoTherming 25,000 BPSD Unit to Produce 10 ppm Sulfur
Fuel from Untreated Distillate Fuel
Capital Cost ($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
187
1.71
0
0.16
0.75
-0.55
-0.03
-0.03
Other Cracked Stocks
25
663
3.10
0
0.31
1.45
-1.00
-0.055
-0.085
Light Cycle Oil (LCD)
29
858
3.41
0
0.46
1.65
-1.10
-0.06
-0.10
7.2.1.2.3.2 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 Linde in Table 7.2-16 were for a revamp. As 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: Table 7.2-16 depicts the hydrogen consumption for an
IsoTherming revamp, but does not provide the hydrogen consumption for the original highway
hydrotreater. In estimating hydrogen consumption for a stand-alone IsoTherming unit above, we
estimated that the hydrogen consumption for a conventional hydrotreater processing that
feedstock to 10 ppm sulfur would consume 559 scf/bbl. The IsoTherming revamp is projected to
consumes only 420 scf/bbl, for a savings of 139 scf/bbl. Using the hydrogen consumptions
shown in Table 7.2-6, a conventional hydrotreating revamp is projected to consume 193 scf/bbl
of hydrogen over that being consumed in the original highway fuel hydrotreater. Thus, the
IsoTherming process appears to reduce this incremental consumption by 71% (139/193 * 100%).
Given that we had to project the hydrogen consumption of the original hydrotreater, we decided
to only project a 60% savings for an IsoTherming revamp, rather than 71%. We will review this
estimate for future analyses as additional data from the IsoTherming revamp being installed at a
Giant refinery becomes available. Reducing the hydrogen consumptions shown in Table 7.2-6 by
60%, the resulting hydrogen consumptions for an IsoTherming revamp were 150 scf/bbl for
LCO, 92 scf/bbl for other cracked stocks, and 38 scf/bbl for straight run.
Utilities and Yield Losses: We followed the same methodology for estimating electricity
demand as we did above for hydrogen consumption. We estimated above that the electricity
7-71
-------
Draft Regulatory Impact Analysis
demand for a conventional hydrotreater processing that feedstock to 10 ppm sulfur would
consume 0.83 kW-hr/bbl. The IsoTherming revamp is projected to consume 1.83 kW-hr/bbl, an
increase of 1.0 kW-hr/bbl. Using the electricity demand shown in Table 7.2-6, a conventional
hydrotreating revamp is projected to use 0.53 kW-hr/bbl of electricity over that being used in the
original highway fuel hydrotreater. Thus, the IsoTherming process appears to increase this
incremental usage by 190% ((1.53/0.53 - 1) * 100%). Given that we had to project the electricity
demand of the original hydrotreater, we decided to project a slight larger increase of 225% for an
IsoTherming revamp, rather than 190%. We will review this estimate for future analyses as
additional data from the IsoTherming revamp being installed at a Giant refinery becomes
available. Increasing the electricity demand shown in Table 7.2-6 by 225%, the resulting
electricity demand for an IsoTherming revamp were 2.6 kW-hr/bbl for LCO, 2.3 kW-hr/bbl for
other cracked stocks, and 1.3 kW-hr/bbl for straight run.
Regarding fuel gas consumption, the total fuel gas consumption for a stand-alone
IsoTherming unit was projected above to be zero, due the enhanced ability to conserve heat
generated in the aromatic saturation reactions which accompany desulfurization. The process
projections shown in Table 7.2-16 above show no consumption of natural gas by the unit. Thus,
it would seem reasonable to project that a Linde revamp would cause no increase in fuel gas
consumption. In fact, if the total use of natural gas was zero, one might expect that the
IsoTherming revamp actually reduced fuel gas consumption, as the original hydrotreater would
have been consuming some fuel gas. However, to be conservative, we projected that a
IsoTherming revamp would require the same fuel gas consumption as that of a conventional
hydrotreating revamp. We will review this estimate for future analyses as additional data from
the IsoTherming revamp being installed at a Giant refinery becomes available.
As mentioned above, Linde did not provide estimates of yield losses for the IsoTherming
process. We estimated that a stand-alone IsoTherming unit would reduce yield losses by 45%
compared to a stand-alone convention hydrotreater. Table 7.2-6 shows that the yield loss for
straight run feed is 1.0% for a conventional hydrotreating revamp and Table 7.2-7 shows a 1.5%
loss for a grass roots conventional hydrotreater. Thus, the yield losses for a conventional
hydrotreating revamp is two-thirds of the yield loss for a grass roots conventional hydrotreater.
Thus, the original highway fuel hydrotreater has a yield loss of 0.5% for straight run, consistent
with that shown in Table 7.2-11.
If the IsoTherming revamp reduces the yield loss by 45%, its yield loss for straight run is
55% of 1.5%, or 0.82%. Subtracting out the 0.5% loss of the original highway hydrotreater
means that the IsoTherming revamp had an incremental yield loss of 0.32%, or 32% of the 1.0%
yield loss projected for the conventional hydrotreating revamp. Thus, we projected that all of the
yield losses shown in Table 7.2-6 for a conventional hydrotreating revamp would be only 32% as
large for an IsoTherming revamp.
7-72
-------
Estimated Costs of Low-Sulfur Fuels
Diesel
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% 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% 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 the following table.
Table 7.2-18
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
38
1.30
0
0.09
0.25
-0.18
-0.01
-0.01
Other Cracked Stocks
12.5
25,000
92
2.28
0
0.17
0.48
-0.33
-0.02
-0.03
Light Cycle Oil (LCD)
14.5
25,000
150
2.60
0
0.22
0.55
-0.38
-0.02
-0.03
7.2.1.2.4 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.
7-73
-------
Draft Regulatory Impact Analysis
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 the straight run diesel fuel in the SZorb estimates, 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.25 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, it is appropriate to base our cost analysis for each refinery on
what we estimate would be typical or average qualities for 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
should be optimistic.
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.
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. Another option is for refiners to shift
some of their LCO to the locomotive and marine markets. While these markets would be
7-74
-------
Estimated Costs of Low-Sulfur Fuels
regulated to a 500 ppm sulfur standard, it is less stringent and does not require the aggressive
desulfurization of the sterically hindered compounds. Because of the low cetane value inherent
with LCO, refiners cannot simply dump a large amount into locomotive and marine diesel since
those two pools must meet an ASTM cetane specification. Thus, we believe that refiners could
distill its LCO into a light and heavy fraction and only shift the heavy fraction to off-highway,
locomotive, and marine diesel fuels. 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 non-highway pool to maintain current production
volumes of all fuels. In addition to the cetane limit which restricts blending of LCO into non-
highway diesel, the T-90 maximum established by ASTM limits would limit the amount of LCO,
and especially heavy LCO, which can be moved from nonroad diesel fuel into these other
distillate streams. The exception, 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 high sulfur pool from 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.
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.26 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 refineries 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 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
catalysts available. As catalysts continue to improve, the cost of desulfurizing diesel fuel will
continue to decrease.
7-75
-------
Draft Regulatory Impact Analysis
In summary, if the vendor cost estimates are optimistically low, there are a number of reasons
why the cost of desulfurizing highway diesel fuel to meet the 15 ppm cap standard are likely to
be low. 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. 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.
7.2.1.3 Composition of Distillate Fuel by Refinery
In the previous section, we established distinct desulfurization costs for the various
blendstocks comprising diesel fuel. To apply these costs to each refinery, we must estimate the
each refinery's diesel fuel composition. Refiners do not publish this information, so we
estimated these compositions from other publically available sources of information. The
fraction of LCO in distillate fuel is addressed first, then we estimate the fraction of other cracked
stocks and lastly, the fraction of hydrocrackated stocks. By estimating the fractions of each
refinery's number two distillate comprised by these various blendstocks, the remaining fraction is
comprised of straight run distillate. In addition to these primary sources of distillate blendstocks,
the fraction of distillate currently being hydrotreated also affects the cost of further sulfur control,
particularly the required consumption of hydrogen. Thus, the fraction of distillate fuel which is
currently hydrotreated is also estimated below for each refinery. Finally, how distillate
composition might be changing over time is discussed.
Light-Cycle Oil: First, we estimated the volume of LCO produced by each refinery based on
the capacity of its Fluidized Cat Cracker unit (FCC unit). The Oil and Gas Journal (OGJ)
publishes information on the capacity of major processing units for each refinery in the country,
including the FCC unit.27 Based on the results of API/NPRA's Refining Operations and Product
Quality survey, the FCC units typically operate at 90 percent of capacity.28 The API/NPRA
survey also shows that number two distillate produced nationally (outside of California) contains
21 percent LCO averaged over the entire pool. However before using this information for
estimating the amount of FCC feed which is produced as LCO for each refinery, we needed to
take two important steps to facilitate the calculation. First, it was necessary to account for the
LCO which is processed by the hydrocracker in the refinery. As discussed above, refineries with
hydrocrackers normally send their LCO to the hydrocracker and convert most of it to gasoline.
Thus, for refineries with distillate hydrocrackers, we reduced their estimated LCO production by
the operational capacity of their hydrocracker (again estimated to be 90% of rated capacity per
7-76
-------
Estimated Costs of Low-Sulfur Fuels
the above API/NPRA survey). Also, an FCC feed hydrotreater can significantly improve the
quality of LCO. However, we do not believe that refineries outside of California6 have FCC feed
hydrotreaters with sufficiently high hydrogen pressure to produce this quality improvement.
Also, we do not have a source of desulfurization costs for LCO produced from an FCC unit with
feed pretreatment versus more typical LCO. Thus, all LCO was assumed to have the same
quality.
Second, EIA regularly collects data from refiners, including their production volumes of high
and low sulfur distillate fuel. According to EIA, 138 U.S. refiners produced a total of 55 billion
gallons of distillate in 2000, with 17 billion gallons (about 31 percent) being high sulfur diesel
fuel produced by 105 refineries. The 1996 API/NPRA survey shows that the LCO fraction of
low sulfur and high sulfur distillate fuel are quite similar. Therefore, for the purpose of
estimating the fraction of each refinery's feed to the FCC unit which is produced as LCO, of we
assumed that the LCO fraction of each refinery's low and high sulfur distillate fuel were equal.
We then backcalculated from the aggregate figure that 21 percent of the nation's refineries'
number 2 distillate is LCO, and based on the premise that refineries with hydrocrackers
processed their LCO in that unit, that refineries with FCC units produce 25 percent LCO from the
feed to those units. We then categorized the 105 refineries producing high sulfur distillate based
on the LCO fraction of their distillate pool at 5 or 10 percent intervals from 0 to 60 percent. The
distribution of refineries by fraction of LCO is summarized in Table 7.2-19.
E This analysis does not model the Federal nonroad standards applying in California since it
is expected that California will be implementing its own program before the Federal program
takes effect.
7-77
-------
Draft Regulatory Impact Analysis
Table 7.2-19
Distribution of LCO in Distillate Fuel by Refinery
(High Sulfur Producing U.S. Refineries)
Number of Refineries
Cumulative Percentage
of US Nonroad Diesel
Volume
Percentage of LCO in the Distillate Pool
0%
44
26
<10%
45
27
<20%
51
34
<25%
61
52
<30%
81
74
<40%
96
87
<50%
101
97
<80%
104
99
As the table shows, we estimate that distillate fuel produced by refineries which produce high
sulfur distillate contains anywhere from zero LCO to over 80 percent LCO. The table also shows
that 44 U.S. refineries, which produce about 26 percent of the high sulfur distillate in the U.S.,
blend no LCO into their distillate. The above table also reveals that the distillate from the
remaining 61 refineries averages about 28 percent LCO by volume.
Other Cracked Stocks: In addition to LCO, nonroad diesel fuel is comprised of other
cracked stocks such as light coker gas oil, light visbreaker gas oil, and light thermally cracked
gas oil. These other cracked stocks are somewhat more difficult to treat than straight run
distillate, but less difficult to treat than LCO. Light coker gas oil dominates this intermediate
group of blendstocks and we have estimates available for its desulfurization costs. Therefore, we
estimated the fraction of all of these other cracked stocks in each refinery's high sulfur distillate
fuel and treat the volume as light coker gas oil for cost estimation purposes.
Similar to our approach for LCO, we based the volume of each of these cracked stocks on the
capacity of the refining units which produce them, namely delayed and fluid cokers, visbreakers,
and thermal crackers. Based on the above mentioned API/NPRA survey, we estimate that all of
these units operate at 90 percent of capacity. Based on confidential estimates from a refining
industry consultant, we estimate that 30 percent of delayed coker and 15 percent of the other
units' production is distillate blended into the distillate pool. As we did with our procedure for
LCO, refineries with hydrocrackers were assumed to send these other cracked stocks to the
hydrocracker for conversion to gasoline to the extent that capacity remained after any LCO was
processed by that unit. We also again spread the volume of these other cracked stocks
proportionately across each refinery's production of low and high sulfur distillate fuel based on
the information in the API/NPRA survey 29 which shows that the other cracked stocks is fairly
well equally distributed across these two pools.
Summing the volume of other cracked stocks in high distillate across all refineries, we found
that about 8 percent of the entire distillate fuel volume produced by high sulfur distillate
producing refineries is comprised of these other cracked stocks. This value agrees well with the
1996 API/NPRA survey of distillate fuel. The fractions which cracked stocks comprise of the
7-78
-------
Estimated Costs of Low-Sulfur Fuels
total distillate pool for refineries which produce high sulfur distillate fuel was characterized for
the industry as a whole and summarized below in Table 7.2-20.
Table 7.2-20
Distribution of Other Cracked Stocks in Distillate Fuel by Refinery
(High Sulfur Producing U.S. Refineries)
Number of Refineries
Cumulative Percentage of
US non-Highway Diesel
Volume
Percentage of Other Cracked Stocks in the Distillate Pool
0%
74
48
<10%
76
51
<15%
86
64
<20%
94
76
<25%
97
78
<30%
102
96
<40%
105
100
As shown, we estimate that almost half of distillate fuel in the U.S, which is produced by 74
refineries, does not contain other cracked stocks from cokers, visbreakers and thermal crackers.
Of the refineries which are projected to blend other cracked stocks into their distillate pool, the
analysis predicts that, on average, the distillate fuel from these refineries contains approximately
19 percent of other cracked stocks.
Hydrocracked: In the U.S., hydrocrackers are almost exclusively used to crack undesirable
distillate blendstocks, primarily LCO, into more desirable products, such as gasoline. However,
not all of this distillate material is converted to gasoline. The portion which remains distillate is
of high quality/ We again obtained the hydrocracker capacity by refinery from the OGJ. The
1996 API/NPRA survey of distillate fuel composition indicated that 5.8 percent of all distillate
was hydrocracked. Dividing 5.8% of total distillate production in 2000 by total hydrocracker
capacity, we found that about 20 percent of the hydrocracker capacity is hydrocracked distillate,
which is also termed hydrocrackate. This percentage was assumed to apply to all hydrocrackers.
Unlike the other blendstocks, the 1996 API/NPRA survey indicated that hydrocrackate
comprises a smaller percentage of low sulfur diesel fuel (4.4%) than high sulfur distillate fuel
(8.8%). Thus, for refineries which produced both low and high sulfur distillate fuel in 2000, we
allocated a higher percentage of hydrocrackate to their high sulfur pool than their low sulfur pool
until the overall percentage of hydrocrackate in low and high sulfur distillate fuel pools equaled
4.4 percent and 8.8 percent, respectively.
Hydrotreated Material in High Sulfur Distillate: The 1996 API/NPRA fuel quality survey
shows that a significant percentage of the blendstocks comprising high sulfur distillate are
F According to Mathpro, hydrocracked distillate is high in cetane (-46) and low in sulfur
(-100 ppm) relative to the feedstock which if it is LCO is low in cetane (-25) and high in sulfur
(-12,000 ppm)
7-79
-------
Draft Regulatory Impact Analysis
hydrotreated, 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 blendstocks which can exceed 10,000 ppm sulfur and have a cetane number
of less than 15 prior to hydrotreating. The fact that a portion of high sulfur distillate fuel is
currently hydrotreated is important, because this removes all of the olefins and saturates some of
the aromatics, reducing subsequent hydrotreating costs. While hydrotreating also removes some
of the sulfur from these streams, our estimates of the amounts of sulfur in high sulfur distillate
come from measurements of finished high sulfur distillate. Thus, while hydrotreated blendstocks
have reduced sulfur levels, this means that the unhydrotreated blendstocks have higher sulfur
levels, which combined, produce the final, surveyed sulfur level. The API/NPRA survey shows
that the fraction of the high sulfur distillate pool which is desulfurized varies significantly
between PADDs. Thus, the refinery model was calibrated against the 1996 API/NPRA Refinery
Survey at the PADD level instead of at the national level. Table 7.2-21 summarizes the fraction
of high sulfur distillate which is hydrotreated in each PADD.
Table 7.2-21
Hydrotreated Percentage of High Sulfur Distillate Blendstocks
PADD
1
2
3
4
5 (CA excluded)
AK
Percent Hydrotreated
27
31
44
17
2
0
As Table 7.2-3 shows, PADD 3 has the highest percentage of its high sulfur distillate pool
hydrotreated at 44 percent. None of Alaska's fuel is believed to be hydrotreated since none of
the refineries located in Alaska have distillate hydrotreaters.
The hydrotreated blendstocks of the high sulfur distillate pool are assumed to be treated to
meet the current highway sulfur standard average of 340 ppm. We believe that this is reasonable
because many refiners who are blending their nonroad diesel fuel using both hydrotreated and
unhydrotreated streams likely only have a single hydrotreater and they are simply blending some
of their highway diesel fuel with high sulfur distillate to produce a product which meets either
nonroad or heating oil standards. There could be refiners who have dedicated hydrotreaters in
their refineries for treating high sulfur distillate for producing nonroad or heating oil directly, or
for blending with other high sulfur distillate for producing nonroad or heating oil. Thus the
hydrotreated product could be higher or lower than the current average of 340 ppm. However, as
seen below, this would simply result in a lower or higher starting sulfur level for the balance of
the pool which is not desulfurized and the net desulfurization cost would be about the same.
Also, one cannot tell by looking at the U.S. refinery unit capacities in the Oil and Gas Journal if
7-80
-------
Estimated Costs of Low-Sulfur Fuels
there is dedicated nonroad distillate hydrotreating capacity or not. Assigning the hydrotreated
stocks a sulfur level of 340 ppm simplifies the analysis.
As seen in Section 7.1, average sulfur levels were calculated for each PADD. Using these as
a starting point, PADD-specific starting sulfur levels were estimated for each PADD depending
on whether the refinery had a distillate hydrotreater or not. If a refinery did not have a distillate
hydrotreater, then its starting sulfur level is the same as that reported in Section 7.1. However, if
the refinery did have a diesel hydrotreater, then the sulfur level for the unhydrotreated portion of
the nonroad pool was calculated. One adjustment was made to the average starting sulfur levels
of PADDs 1 and 3. Because the high sulfur producing refineries in those two PADDs have
hydrocrackers, the sulfur levels were adjusted to estimate the sulfur level of the nonhydrocracked
blendstocks. Excluding hydrocracked distillate from the average sulfur level is important as
hydrocracked distillate is not expected to be treated in new distillate hydrotreater equipment
added to comply with the 500 ppm and 15 ppm sulfur cap standards. The various sulfur levels
are summarized in Table 7.2-22 and these are used to estimate the cost of desulfurizing nonroad
diesel fuel.
Table 7.2-22
High Sulfur Distillate Fuel Sulfur Levels3
(Excludes Hydrocracked Blendstocks)
PADD
1
2
3
4
5 (Excluding CA)
Alaska
Sulfur Level of High Sulfur
Distillate in Refineries
without Hydrotreaters
3420
3000
5200
2700
2540
2540
Sulfur Level of
Hydroteated Blendstocks
in Refineries with
Distillate Hydrotreaters
340
340
340
340
340
—
Sulfur Level of non-
Hydrotreated High Sulfur
Distillate in Refineries
with Distillate
Hydrotreaters
6130
5400
11,320
4200
2600
2540
a The values in the third column are calculated from the sulfur levels of the first column, the sulfur levels of the second
column and the percentages in Table 7.3-3
Trends in Distillate Fuel Composition: It is likely that refiners will want to shift their
blendstocks in an effort to reduce their costs for complying with the 15 ppm highway diesel fuel
standard in 2006. Directionally, refiners would likely shift their more difficult to treat
blendstocks (LCO and other cracked stocks) to high sulfur distillate and their easier to treat
blendstocks (straight run and hydrocrackate) to highway fuel. Heating oil must meet at least the
5000 ppm sulfur specification, as well as more stringent state specifications, as low as 2000 ppm.
Most high sulfur diesel fuel and heating oil is shipped as a single fuel, so high sulfur diesel fuel
often must meet state sulfur standards for heating oil and heating oil must have a cetane number
of at least 40. However, nonroad diesel fuel must continue to meet a 40 cetane specification and
7-81
-------
Draft Regulatory Impact Analysis
a 500 ppm sulfur specification. Since straight run and hydrocrackate generally have higher
cetane levels and lower sulfur levels than the other blendstocks, only a limited amount of such
shifting is feasible.
The sulfur content and cetane number of each type of blendstock can vary widely depending
on the specific crude oil processed and the design of the refining equipment employed.
Therefore, estimating the degree to which each refinery could shift its blendstocks around to
minimize its desulfurization costs under the highway diesel fuel rule was not possible. However,
in an attempt to partially reflect shifting that could occur, we shifted some hydrocrackate from
selected refineries' high sulfur distillate pool to their low sulfur pool. This was only done when a
blendstock with relatively high cetane and low sulfur was available to replace the hydrocrackate.
The only blendstock considered to be of sufficiently high quality to replace hydrocrackate was
distillate which had been processed through a distillate hydrotreater. Once hydrotreated,
typically blended distillate tends to have reasonable cetane and sulfur levels, so the specific
composition of the hydrotreated distillate is not critical. Thus, only those refineries with both
hydrocrackers and distillate hydrotreaters were assumed to move the hydrocracked distillate over
to the highway pool. The composition of the hydrotreated distillate shifted to the high sulfur
distillate fuel pool was assumed to match the composition of all the non-hydrocracked distillate
material produced by the refinery. Since the majority of current low sulfur diesel fuel is
hydrotreated to meet the 500 ppm cap, plenty of hydrotreated material was usually available for
swapping for the hydrocracked distillate available from the high sulfur distillate pool. Thus,
most of the hydrocrackate from refineries producing both low and high sulfur distillate in 2000
was shifted to the low sulfur fuel pool. After this shift, hydrocracked distillate comprised 2.6
percent of the high sulfur distillate pool and 7.2 percent of the highway diesel pool.
The distillate fuel compositions estimated above are based primarily on data from the 1996
API/NPRA survey and current refinery unit capacities. We assumed that these compositions
would remain constant throughout the timeframe of the analysis: 2007-2030. A recent
presentation by EIA indicates that this is not likely to be the case. While the volumes of light and
medium crude oils processed by U.S. refineries has been relatively constant over the past 15
years, the volume of heavy crude oils has increased significantly. This has lead to an increase in
the fraction of crude oil volume processed through conversion units, particularly cokers and
hydrocrackers. FCC unit capacity also increased slightly as a fraction of total crude oil
distillation capacity Thus, the 1996 API/NPRA distillate fuel quality survey likely
underestimates the amount of other cracked stocks and hydrocracked material in distillate fuel.
While this analysis does not reflect this trend, we plan to incorporate this trend into future
estimates of the cost of desulfurizing diesel fuel.
7.2.1.4 Summary of Cost Estimation Factors
This section presents a number of costs, such as those for electricity and natural gas, as well
as cost adjustment factors which are applicable to all three of the above desulfurization
technologies.
7-82
-------
Estimated Costs of Low-Sulfur Fuels
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."g 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 which they are desulfurizing.
Stream Day Basis: The EIA data for the production of distillate by various refineries is 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 during which the unit is actually or 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 three
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%
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 have assumed that a desulfurization unit
would be designed to meet its annual production target while operating only 80% 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.
The off-site cost factors and construction location cost factors used in this analysis were
taken from Gary and Handewerk.30 The offsite 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%.31
G 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).
7-83
-------
Draft Regulatory Impact Analysis
The off-site factors vary by refinery capacity, while the construction location factors vary
between regions of the country.32 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 7.2-23.
Table 7.2-23
Offsite and Construction Location Factors
Offsite Factor
- New Unit
- Revamped Unit
Construction Location Factor
PADD1
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
would be expansions of the amine and sulfur plants to address the additional sulfur removed, a
new sulfur analyzer. To account for these other capital costs, and for other contingencies, capital
costs (including off-sites) were increased by 15 percent, typical for this type of analysis.33 In
addition, we increased this factor to 18% to include the costs of starting up a new unit.34
Capital Amortization: The economic assumptions used to amortize capital costs over
production volume and the resultant capital amortization factors are summarized below in Table
7.2-24.35 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.h
Table 7.2-24
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
H The capital amortization factor is applied to a one time capital cost to create an amortized
annual capital cost which occurs each and every year for the 15 years of the economic and project
life of the unit. This implicitly assumes that refiners would reinvest in desulfurization capacity
after 15 years at the same capital cost and amortized annual cost, and amortized cost per gallon.
7-84
-------
Estimated Costs of Low-Sulfur Fuels
The capital amortization scheme labeled Societal Cost is used most often in our estimates of
cost made below. It excludes the consideration of taxes, since taxes are considered to be transfer
payments between various sectors of the economy and not true economic costs. The other two
cost amortization schemes include corporate taxes, to represent the cost as the regulated industry
might view it. The lower, 6%, rate of return represents the rate of return for the refining industry
over the past 10-15 years. The higher, 10%, rate of return represents the rate of return which
would be expected for an industry having the general aspects of the refining industry.
7.2.1.4.2 Fixed Operating Costs
Operating costs which are 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% 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.36
The breakdown of the base fixed operating cost percentage is as follows:
Maintenance costs: 3%
Buildings: 1.5%
Land: 0.2%
Supplies: 1%
Insurance: 1%.
Annual labor costs were taken from the refinery model developed by the Oak Ridge National
Laboratory (ORNL).37 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
Variable operating costs only accrue as the unit is operating. Thus, they are usually based on
unit throughput. When the unit is not operating, variable operating costs are zero. Thus, variable
operating costs are based on calendar day throughput, not stream day capacity, to avoid over-
counting these costs.
We obtained utility costs from EIA's 1999 Petroleum Marketing Annual report, which
provides these costs by PADD.38 We considered updating these costs. However, a review of
more recent electricity showed little change from 1999. The price of liquid fuels changed
significantly from 1999, but did not appear to represent long term trends. Thus, 1999 liquid fuel
prices were also retained, with one change. We did add 5 c/gal to the price of high sulfur
distillate fuel to represent the added cost of meeting the 15 ppm sulfur cap to.
7-85
-------
Draft Regulatory Impact Analysis
Natural gas prices have been particularly volatile over the past three years, as natural gas use
in electricity generation is increasing rapidly. Therefore, the price over any short period of time
is unlikely to represent long term prices well. Thus, natural gas prices were averaged over 5
years starting at the end of 1996 and this average price was used here.
Steam demand is presented above in terms of pounds per hour. This was converted to BTUs
per hour, assuming that the steam was provided at 300 psi (809 BTU per pound). We assume
that the steam is generated using natural gas as fuel, at an efficiency of 50%, which was taken
from Perry's Handbook.39
These utility and fuel costs are summarized in Table 7.2-25. For future analyses, we are
considering using projections of future utility and fuel prices from EIA's most recent Annual
Energy Outlook.
Table 7.2-25
Summary of Costs From EIA Information Tables for 1999, and Other Cost Factors
Electricity (cents per kilowatt -hour)
LPG (dollars per barrel)
Highway Diesel (cents per gallon)
Nonhighway Diesel (cents per gallon)
Gasoline (dollars per barrel)
Natural Gas ($/MMbtu)
PADD 1
8.35
17.09
53.1
49.3
27.0
4.15
PADD 2
6.40
14.11
55.9
55.7
25.9
4.24
PADD 3
6.66
14.49
51.5
48.6
24.9
2.98
PADD 4
5.4
14.53
62.4
60.4
28.9
3.15
PADD 5
7.18
17.05
64.0
58.9
30.0
3.91
7.2.1.4.4 Hydrogen Costs
Hydrogen costs are 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
costs is a technical paper published by Air Products, which is a large provider of hydrogen to
refineries and petrochemical plants.40 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% factor for offsites. The process design parameters from this paper are summarized in the
Table 7.2-26.
7-86
-------
Estimated Costs of Low-Sulfur Fuels
Table 7.2-26
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 ($/MSCF)
1.18
0.03
0.03
-0.07
0.83
2.00
Notes: Natural Gas @ S2.75/MMBTU; Steam @ S4.00/M Ibs; Electricity @ $0.045 KWH
The estimates shown in Table 7.2-26 were adjusted to reflect natural gas and utility costs in
each PADD (shown in Table 7.2-25). The steam costs were adjusted based on the cost of natural
gas. The capital cost and fixed operating costs were increased by 8% 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.41 We
incorporated PADD-specific offsite and construction location factors from Table 7.2-23, again
assuming a 50-50 mix of new and revamped units. Table 7.2-27 summarizes the average plant
size and the offsite and location factors for the installation of hydrogen plant capital for each
PADD.
Table 7.2-27
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
Offsite 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
7-87
-------
Draft Regulatory Impact Analysis
The adjusted hydrogen costs in each PADD are summarized in Table 7.2-28.
Table 7.2-28
Estimated Hydrogen Costs by PADD
PADD
1
2
3
4
5 Excluding CA and AK
AK
Cost ($71000 scf)
3.26
2.80
1.89
2.82
2.91
3.69
7.2.1.4.5 Other Operating Cost Factors
Similar to the 15% contingency factor for capital costs, we included a 10% contingency
factor to account for operating costs which are beyond the those directly related to operating the
desulfurization unit.42 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 which
may be incurred but not explicitly accounted for in our cost analysis. We then increased this
factor by 2% to account for reprocessing of off-specification material. Above, we estimated that
5% of all batches could require re-processing. However, since this material would 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.
We also believe that 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 proposed 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 would be important for effective desulfurization. Finally, the main fractionator of
the FCC unit would have to be carefully controlled to avoid significant increases in the
distillation endpoint, as this could increase the amount of sterically hindered compounds sent to
the diesel hydrotreater.
Improved control each of these units could involve enhancements to computer control
systems, as well as improved maintenance practices.43 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% of those estimated below for diesel fuel desulfurization.44 45 No
costs were included in the cost analysis for these potential issues.
-------
Estimated Costs of Low-Sulfur Fuels
7.2.1.5 How Refiners are Expected to Meet the Nonroad Sulfur Requirements
This section presents the methodology used to determine which refiners produce 15 and 500
ppm highway diesel fuel, 500 ppm NRLM diesel fuel, 15 ppm nonroad diesel fuel and heating oil
during the four phases of the highway and NRLM diesel fuel programs. These four phases are:
1) June 1, 2006 - May 31, 2007: 15 ppm highway cap with temporary compliance option and
small refiner provisions; no NRLM caps
2) June 1, 2007- May 31, 2010: 15 ppm highway cap with temporary compliance option and
small refiner provisions; 500 ppm NRLM cap with small refiner provisions
3) June 1, 2010 - May 31, 2014: 15 ppm highway cap; 15 ppm nonroad cap with small refiner
provisions; 500 ppm locomotive/marine cap
4) June 1, 2014 and beyond: 15 ppm highway cap; 15 ppm nonroad cap; 500 ppm
locomotive/marine cap
As can be seen from these phases, there is significant overlap between the highway and
NRLM diesel fuel sulfur programs. Thus, we begin our analysis below with a projection of
which refiners would likely produce 15 ppm highway diesel fuel, first in 2006 and then in 2010.
Then, we project which refiners would invest to produce 500 ppm NRLM diesel fuel and then 15
ppm nonroad diesel fuel.
In order to make these projections, we estimated how much highway and NRLM diesel fuel
each refiner could produce. We obtained each U.S. refinery's actual production volumes of low-
sulfur (highway) and high sulfur distillate during 2000 from the Energy Information
Administration (EIA). Since the highway and NRLM diesel fuel programs phase in from 2006-
2010, we projected these 2000 production volumes out to 2008, the mid-point of this time period.
All the costs developed below presume economies of scale projected to exist in 2008.
Over the past 20 years, the production capacity of refineries which have remained in
operation has steadily increased. EIA projects that this is likely to continue. Ideally, we would
project each refinery's individual growth in production between 2000 and 2008. However, this
information in not available. Therefore, we projected national average growth rates for highway
and high sulfur distillate fuel, respectively. This appears to be quite reasonable. Not every
refinery has increased capacity, nor has the increase been the same for every refinery showing an
increase. However, a comparison of the crude oil distillation capacities of refineries in 1990 and
2002 indicate that a large majority of refineries have increased capacity. Thus, projecting the
same growth rate for all refiners is reasonably consistent with past growth.
We projected growth in domestic refineries' production of diesel fuel based on growth in
diesel fuel consumption. Based on the demand for low and high sulfur distillate fuel in 2000 and
2008 (discussed in Section 7.1 above), we determined that, absent this rule, the demand for
highway diesel fuel and high sulfur distillate would increased by 24% and 8% between 2000 and
2008. Thus, the production volume of highway diesel fuel by each domestic refinery in 2000,
from EIA, was increased by 24%, and that for high sulfur distillate was increased by 8%. This
implicitly assumes that imports of both fuels will remain a constant percentage of total demand.
7-89
-------
Draft Regulatory Impact Analysis
We made no changes in the production volumes of distillate fuel to account for any reduction
in wintertime blending of kerosene that might occur with the implementation of 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 could mean adding kerosene to more diesel
fuel than actual requires it. The latter approach would require 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 refinery could simply
be added to the distillate being fed to the hydrotreater and desulfurized along with the rest of the
15 ppm diesel fuel pool.
The current amount of terminal blending of kerosene is difficult to estimate. Therefore, we
have not attempted to estimate its current or future level and account for any change in this
practice. In either case (terminal kerosene blending or the use of additives), the volume of
distillate provided to terminals is roughly the same. Thus, we have simply based our projected
costs of today's proposal on current diesel fuel demand. This way, we are assured of including
the cost of desulfurizing the total volume of diesel fuel consumed in NRLM diesel engines. The
most important assumption here is that we have assumed that a separate, 15 ppm grade of
kerosene will not be produced and distributed for downstream blending. This would entail
additional production and distribution costs, which we believe will discourage this practice.
Thus, we have not included these costs here.
The remainder of this section provides an overview of how we projected which refineries
would likely produce highway and NRLM diesel fuel during the various phases of the program
and how they would likely try to optimize the construction of their desulfurization equipment in
order to comply with both programs.
7.2.1.5.1 Complying with the Highway Diesel Fuel Sulfur Program
The 15 ppm cap on highway diesel fuel takes effect June 1, 2006, when 80% of the highway
diesel fuel produced by non-small refiners must meet this standard. Twenty percent of highway
diesel fuel can remain at 500 ppm sulfur. Small refiners, which produce roughly 5 percent of
highway diesel fuel are allowed to continue producing 500 ppm highway diesel fuel until January
1, 2010. Credits for over-production of 15 ppm diesel fuel can be traded within the PADD they
are generated. These credits can be used through May 31, 2010. Thus, roughly 25% of highway
diesel fuel can remain at 500 ppm sulfur through May 31, 2010.
The implementation date of the 15 ppm highway diesel fuel, June 1, 2006, occurs only 3
years from now. By the time that this proposal is finalized, only slightly more than two years
will remain before June 1, 2006. This leadtime is not likely to be sufficient for refiners planning
7-90
-------
Estimated Costs of Low-Sulfur Fuels
on producing 15 ppm highway diesel fuel in 2006 to fully coordinate these plans with this NRLM
rule. Thus, as described below, we generally made the conservative assumption that refiners
would make their plans for 2006 independent of the proposed NRLM diesel fuel program.
However, as indicated above, many refineries could delay the production of 15 ppm highway
diesel fuel until 2010 by purchasing credits. This would give them four additional years of
leadtime and allow them to fully coordinate their plans for desulfurizing both highway and
NRLM diesel fuel. Therefore, we have incorporated such coordination in our projections below.
As mentioned above, small refiners are allowed to continue producing 500 ppm highway
diesel fuel until 2010 without the need to purchase credits. In addition, if a small refiner chooses
to meet the 15 ppm cap with their highway diesel fuel, they are allowed to produce gasoline
under their interim Tier 2 sulfur standards before the final 30 ppm Tier 2 standard applies. Other
refineries located in the Geographic Phase-in Area (GPA) also have this option under the 2007
highway diesel fuel program.
Small and GPA refiners have already indicated to EPA whether they plan to take this option.
This information was incorporated into our analysis by projecting that these refiners would begin
producing 15 ppm highway diesel fuel in 2006, as opposed to 2010.
In order to produce 15 ppm highway diesel fuel, refiners have the choice of revamping their
existing distillate hydrotreater or construct a new, grassroots hydrotreater. In the 2007 highway
diesel fuel rule, we projected that 80% of the volume of 15 ppm fuel would be produced using
revamped hydrotreaters and 20% would be produced using new, grassroots hydrotreaters. We
have retained this projection in this analysis. As described in Chapter 5, refiners are still in the
process of determining how they will produce 15 ppm highway diesel fuel and revamping their
existing hydrotreater still appears likely to be feasible for most refiners.
A refiner's decision to revamp or construct a new unit will depend on many factors specific
to that refinery. We lack the information necessary to project which decision individual
refineries will make. Thus, we projected the cost of producing 15 ppm highway diesel fuel at
each refinery using both a revamped hydrotreater and using a new, grass roots unit. An average
cost was determined by weighting the revamp cost by 80% and the grass roots cost by 20%. As
described in more detail in Section 7.2.2 below, we then used these average highway diesel fuel
costs to determine which refineries were most likely to produce 15 ppm diesel fuel in 2006 and
2010.
The use of advanced desulfurization technologies was estimated in the same way. We made
no attempt to determine which specific refineries would use each technology. We estimated the
cost of producing 15 ppm diesel fuel using each technology at each refinery and then weighted
these costs by the projected mix of desulfurization technologies applicable in that year.
For 2006, we assumed that refiners would only process their current highway diesel fuel
volume (grown to 2008 production levels) to 15 ppm. In other words, no 15 ppm highway diesel
fuel would be produced from current high sulfur distillate. However, for 2010, we evaluated the
7-91
-------
Draft Regulatory Impact Analysis
production of allowed 15 ppm highway diesel fuel from current high sulfur distillate. This could
be at a refinery currently producing a mix of highway and high sulfur distillate, or just high sulfur
distillate. This was done because some refineries currently produce very small quantities of
highway diesel fuel, likely from naturally, low sulfur blendstocks. These refiners are unlikely to
produce 15 ppm diesel fuel at such low volumes. Conversely, some refineries produce very large
volumes of high sulfur distillate which could be controlled to 15 ppm with good economies of
scale.
Once we had estimated each refinery's cost to produce 15 ppm highway diesel fuel, we
assumed that those with the lowest cost in each PADD would be the most likely to produce this
fuel in 2006. Thus, after considering the production of 15 ppm fuel by small and GPA refiners
choosing to delay compliance with the Tier 2 gasoline sulfur standards, we fulfilled the
remainder of each PADD's highway diesel fuel demand with 15 ppm highway diesel fuel from
the other refineries, starting with those with the lowest cost per gallon and moving up until
demand was met. Then in 2010, the remainder of highway diesel fuel demand was met by the
next lowest cost production, either from current highway diesel fuel or high sulfur distillate fuel.
7.2.1.5.2 Complying with the 500 ppm NRLMDiesel Fuel Sulfur Standard in 2007
We used two basic criteria to project which refineries would likely produce 500 ppm NRLM
diesel fuel in 2007. The first criterion was the refinery's ability to continue to sell high sulfur
distillate. The Northeast has a large heating oil market. Thus, PADD 1 refineries were assumed
to be able to continue to sell high sulfur distillate to this market if they desired. The same
flexibility was assumed to apply to PADD 3 refineries which are either connected to one of the
two large pipelines running from the Gulf Coast to the Northeast (Plantation and Colonial) or
have access to ocean transport. 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. Besides these refineries,
however, all refineries in PADDs 2 and 4 and those in PADDs 3 and 5 not meeting the above
criteria were assumed to have to produce 500 ppm NRLM diesel fuel starting June 1, 2007.
While the proposed rule would not directly require this, we believe that for cost estimation
purposes, this is a reasonable assumption.
Under the proposed NRLM diesel fuel program small refiners could continue selling high
sulfur NRLM diesel fuel until June 1, 2010. Thus, these small refiners have more flexibility in
selling high sulfur distillate fuel, as they can sell this fuel to either the heating oil or NRLM
diesel fuel markets. We evaluated small refiners' ability to distribute high sulfur NRLM diesel
fuel, as it is unlikely that common carrier pipelines would carry this fuel. Starting with the
demand for NRLM diesel fuel in each PADD in 2008 from Section 7.1 above, we divided this
demand by the square milage of each PADD to estimate an 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 in order to maintain its current 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, it was reasonable to assume that all small refiners could continue selling all of
7-92
-------
Estimated Costs of Low-Sulfur Fuels
their high sulfur distillate fuel as either high sulfur distillate fuel or heating oil and delay
producing any 500 ppm NRLM diesel fuel until 2010 at the earliest.
Table 7.2-29 compares the the number of refineries projected to have no choice but to
produce 500 ppm NRLM diesel fuel to the number of refineries currently producing high sulfue
distillate fuel.
Table 7.2-29
Number of Refineries Assumed to Have to Produce 500 ppm NRLM Diesel Fuel in 2007
Having to Participate
Total Producing High Sulfur
Distillate Fuel Today
PADD1
1
13
PADD2
23
23
PADD3
11
41
PADD4
8
8
PADD5
10
20
For each PADD, we then added the production volumes of those refineries projected to have
no choice but to produce 500 ppm NRLM diesel fuel to the volume of high sulfur NRLM diesel
fuel which could be produced small refiners. We then compared these initial volumes of NRLM
diesel fuel to the projected demand for NRLM diesel fuel in each PADD, as estimated in Section
7.1 above. We found that the demand for NRLM diesel fuel in PADDs 2 and 4 were already
fulfilled by these refineries. This is not surprising given the assumptions described above.
However, greater production of NRLM diesel fuel was required in PADDs 1, 3 and 5. This
NRLM fuel would have to meet the proposed 500 ppm cap. We projected the refineries most
likely to produce this fuel would be those facing the lowest per gallon desulfurization costs in
each PADD.
All 500 ppm NRLM diesel fuel was assumed to be produced using conventional
hydrotreating technology. The operating cost of this desulfurization is simply a function of the
composition of each refinery's high sulfur distillate fuel, as well as some costs which vary by
PADD, such as hydrogen, utilities, etc. However, a number of ways existed to estimate the
capital cost, depending on how the potential production of 15 ppm diesel fuel in the future was
considered and whether the refinery was already producing some of its distillate fuel as 500 ppm
highway diesel fuel. The methodology used to estimate capital costs is summarized below.
As mentioned above, we generally presume that refiners projected to produce 15 ppm
highway diesel fuel in 2006 cannot incorporate the production of 500 ppm or 15 ppm NRLM
diesel fuel into their 2006 plans. Thus, with two exceptions, these refiners would have to
construct a new, grass roots hydrotreater to produce 500 ppm NRLM in 2007.
One exception applied to refineries which produce only a very small amount of high sulfur
distillate fuel compared to their volume of highway diesel fuel. This small volume is likely
either off-specification diesel fuel or opportunistic sales to the non-highway diesel fuel market
because of advantageous prices. Thus, in the cases where high sulfur distillate production
7-93
-------
Draft Regulatory Impact Analysis
represented 5% or less of total distillate fuel production, we assumed that the refinery could
incorporate the 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 hydrotreater sized to process all the refinery's distillate fuel and
that for a hydrotreater sized to treat just the highway diesel fuel volume. In this case, both
hydrotreaters were assumed to be grass roots hydrotreaters. In other words, even if the high
sulfur distillate fuel was being incorporated into a revamp of an existing highway diesel fuel
hydrotreater, the incremental cost of increasing capacity was assumed to occur at a grass roots
cost. As mentioned above, the operating cost was simply estimated based on the particular mix
of blendstocks for that refinery and its location (i.e., PADD). As described above in subsection
7.2.1.2, this operating cost depends on how much of that refinery's high sulfur distillate is
already being processed by a hydrotreater. Based on API/NPRA survey findings, refineries
which currently have diesel fuel hydrotreaters were projected to blend a certain amount of
hydrotreated material into their nonroad pool. This reduces the net hydrotreating cost, as the
olefins and some polynuclear aromatics in the high sulfur distillate are already being saturated
today.
The other exception is a refinery which is projected to construct a new, grass roots
hydrotreater in 2006 to produce 15 ppm highway diesel fuel. This refinery would be able to
produce 500 ppm NRLM fuel in 2007 with its existing highway unit. As mentioned above, we
do not identify which individual refineries would likely construct a new grassroots unit in 2006.
Thus, we simply assumed that 20% of the high sulfur distillate volume being produced from
refineries projected to produce 15 ppm highway diesel fuel in 2006 could be desulfurized to 500
ppm with no capital costs.
The next set of refineries to be discussed are those which currently produce both highway and
high sulfur distillate fuel and are not projected to produce 15 ppm highway diesel fuel until 2010.
We presume that these refineries would have to build a new hydrotreater in 2007 in order to
desulfurize their current high sulfur distillate to 500 ppm. However, due to the significant
amount of leadtime available, we project that these refiners could design a revamp that would
desulfurize all of their distillate fuel to 15 ppm in 2010 if they so desire.
Of course, refineries which only produce high sulfur distillate fuel today would have to install
a new hydrotreater to produce 500 ppm NRLM fuel in 2007. We presume that this unit could be
revamped in 2010 to produce 15 ppm nonroad diesel fuel in 2010, if so desired.
Table 7.2-30 presents the percentages of high sulfur distillate fuel production which falls in
the categories described above.
7-94
-------
Estimated Costs of Low-Sulfur Fuels
Table 7.2-30
Production of High Sulfur Distillate: Interaction with Highway Diesel Fuel Program f0'
Number of
Refineries
Percent of
Nonroad Fuel
High Sulfur
Refineries
W/Dist
HT
5
22
No Dist
HT
20
18
Mixed Refineries
Producing 15 ppm
Highway Fuel in 2006
W/Dist
HT
27
32
No Dist
HT
8
6
Mixed Refineries
Producing 15 ppm
Highway in 20 10
W/Dist
HT
15
13
No Dist
HT
13
7
Highway Refineries
W/Dist
HT
15
2
No Dist
HT
2
0
a "Highway" refinery: high sulfur distillate fuel production <5% of total distillate fuel production
"High sulfur" refinery: high sulfur distillate fuel production > 90% of total distillate fuel production
"Mixed refinery: refineries which are neither highway or high sulfur refineries
" W/Dist HT" means refineries currently having a distillate hydrotreater
"No Dist HT means refineries which do not currently have a distillate hydrotreater
Table 7.2-31 presents the estimation of the volume of NRLM diesel fuel which must be
desulfurized to 500 ppm in 2007 in each PADD. PADDs 1 and 3 are shown combined since we
assume that PADD 3 refineries can produce and ship 500 ppm NRLM fuel to PADD 1. The first
line shows total volume of NRLM diesel fuel demand from Section 7.1. The next line shows the
projected volume of highway fuel spillover to the NRLM fuel pool. This volume is subtracted
from NRLM demand, as the spillover already meets the proposed 500 ppm cap. The difference is
total demand for high sulfur NRLM diesel fuel, which is shown on the third line. The fourth line
shows total small refiner volume, which does not need to be desulfurized in 2007. Then, current
production volumes of high sulfur distillate from refineries which we project would not be able
to continue marketing high sulfur distillate are shown. The difference, if any, is the final volume
of 500 ppm NRLM diesel fuel which must be desulfurized. We presume that this final volume
would be produced by refiners facing the lowest desulfurization costs in each PADD. In PADDs
2 and 4, this last volume is zero, because we project that all refineries in these PADDs would
likely have to desulfurize their high sulfur distillate to 500 ppm in order to market it. The total
volume of the last two rows of the table (highlighted in bold) yields the estimated total amount of
high sulfur distillate which is expected to be hydrotreated to meet the 500 ppm NRLM diesel fuel
in 2007. In PADD 2 and 4, this value is larger than the required volume. Thus, some volume of
heating oil is being desulfurized in these PADDs to 500 ppm.
7-95
-------
Draft Regulatory Impact Analysis
Table 7.2-31
NRLM Diesel Fuel Volume Needing Desulfurization: 2007-2010" (million gallons per year)
NRLM Diesel Fuel Demand
Highway Spillover
Base High Sulfur NRLM Demand
Small Refiner Volume
Non- Small High Sulfur Demand
Non-Small Volume Required to
Produce 500 ppm NRLM Fuel
Remaining Demand for 500 ppm
NRLM Diesel Fuel
PADDs 1 & 3
7143
1709
5434
490
4944
914
4030
PADD2
5111
1728
3382
369
3013
3385
0
PADD4
1173
779
394
10
384
488
0
PADDS
937
215
722
202
519
84
435
a Based on projected volumes for 2008
Table 7.2-32 presents an analogous set of volumes for 2010 assuming that no 15 ppm
nonroad diesel 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 is the absence of the small refiner volume.
Table 7.2-32
NRLM Diesel Fuel Volume Needing Desulfurization in the Absence of a 15 ppm Nonroad
Diesel Fuel Cap: 2010 and beyond" (million gallons per year)
NRLM Diesel Fuel Demand
Highway Spillover
Base High Sulfur NRLM Demand
Small Refiner Volume
Net High Sulfur Volume
Non-Small Volume Required to
Produce 500 ppm NRLM Fuel
Remaining Demand for 500 ppm
NRLM Diesel Fuel
PADDs 1 & 3
7,143
1,709
5,434
0
5,434
1,344
4,090
PADD2
5,111
1,728
3,382
0
3,382
3,755
0
PADD4
1,173
779
394
0
394
498
0
PADDS
937
215
722
0
722
286
435
a Based on projected volumes for 2008
In Table 7.2-33, the refineries which are projected to produce 500 ppm NRLM diesel fuel
after the program is fully phased in are characterized by whether they produce predominantly
high sulfur distillate, a mix of highway and high sulfur distillate or predominantly highway diesel
fuel. Like Table 7.2-32, Table 7.2-33 is provided to enable the long-term evaluation of the 500
ppm NRLM standard in the absence of the 15 ppm nonroad diesel fuel cap.
7-96
-------
Estimated Costs of Low-Sulfur Fuels
Table 7.2-33
Characterization of the Refineries Projected to Produce 500 ppm NRLM Fuel for 2007
Number of
Refineries
Nonroad Only
Refineries
W/Dist
HT
0
No Dist
Ht
14
Mixed Refineries
Complying with
Highway in 2006
W/Dist
HT
15
No Dist
Ht
4
Mixed Refineries
Complying with
Highway in 20 10
W/Dist
HT
10
No Dist
Ht
9
Highway Only
Refineries
W/Dist
HT
10
No Dist
Ht
0
7.2.1.5.3 Complying with the 15 ppm Nonroad Sulfur Standard for 2010
We followed the same basic methodology for projecting the cost of 15 ppm nonroad diesel
fuel in 2010, as was described in the previous section for the production of 500 ppm NRLM
diesel fuel in 2007. We first considered whether refineries projected to produce 500 ppm NRLM
diesel fuel in 2007 could continue to do so in 2010 if they so desired. A few refineries were
found to have a sufficiently large volume of 500 ppm NRLM diesel fuel in 2007 and were
located distant from a pipeline or a navigable waterway that it was deemed unlikely that they
could sell all of this fuel to the locomotive and marine diesel fuel markets. These refineries were
assumed to have to process this fuel further to 15 ppm.
All other refineries which produced 500 ppm NRLM diesel fuel in 2007 were assumed to
have the option of producing 15 ppm nonroad diesel fuel in 2010 or continuing their production
of 500 ppm fuel for the locomotive and marine diesel fuel markets. Refineries which did not
produce 500 ppm NRLM diesel fuel in 2007 (or 2010 after the expiration of small refiner
provisions) were not considered likely to produce 15 ppm nonroad fuel in 2010 (or 2014 after the
expiration of small refiner provisions). Since the locomotive and marine diesel fuel markets
exist essentially everywhere in the country, far fewer refineries were projected to have to produce
15 ppm nonroad fuel in 2010 compared to 500 ppm NRLM fuel in 2007.
Again, we evaluated small refiners' ability to market 500 ppm NRLM fuel in 2010 and found
that they could do so by truck. Thus, we assumed that they would either do so or would produce
15 ppm fuel and sell credits to other refiners. In either case, their current high sulfur distillate
production volume would only have to meet a 500 ppm cap in 2010. Table 7.2-34 shows the
number of refineries projected to have little flexibility to avoid producing 15 ppm nonroad diesel
fuel in 2010.
7-97
-------
Draft Regulatory Impact Analysis
Table 7.2-34
Number of Refineries Assumed to Have to Produce 15 ppm Nonroad Diesel Fuel in 2010
Having to Participate
Total Producing High Sulfur
Distillate Today
PADD1
1
13
PADD2
8
23
PADD3
0
41
PADD4
4
8
PADD5
5
17
As already described in the previous section, the capital cost for producing 15 ppm nonroad
diesel fuel depends on whether the refinery currently produces highway diesel fuel and when it is
projected to first produce 15 ppm diesel fuel. Refineries producing less than 5% of their
distillate as high sulfur were assumed to process all their distillate in one desulfurization unit,
regardless of whether this was a new unit or a revamp, whether it was first produced in 2006 or
2010.
As mentioned above, 20% of the refineries producing 15 ppm highway diesel fuel were
presumed to install a new, grass roots unit to do so. These new units would desulfurize high
sulfur distillate now being desulfurized to 500 ppm down to 15 ppm. These refineries could
therefore produce 500 ppm NRLM diesel fuel using their existing highway diesel fuel
hydrotreater. In order to produce 15 ppm nonroad diesel fuel in 2010, we assumed that these
refineries would need to construct a grass roots desulfurization unit. Since the highway
hydrotreater could not be revamped in 2006 in order to produce 15 ppm highway diesel fuel, it
could also not be revamped to produce 15 ppm nonroad diesel fuel. We also assume that any
new hydrotreater constructed in 2007 could be revamped to produce 15 ppm nonroad diesel fuel
in 2010, due to its recent construction and the presumption that refiners would consider the 15
ppm cap when designing their 2007 unit.
Otherwise, the projection of the types of units that could be installed to produce 15 ppm
nonroad units was consistent with the description presented in the previous section. Refineries
only producing high sulfur distillate today could revamp the hydrotreater added in 2007 to
produce 15 ppm nonroad diesel fuel. Refineries producing less than 5% of their distillate fuel as
high sulfur fuel and projected to produce 15 ppm highway fuel in 2010 were assumed to be able
to revamp their highway unit, allowing them to process the small amount of high sulfur distillate
fuel in the same unit. Refineries producing both highway and high sulfur distillate today and
projected to produce 15 ppm highway diesel fuel in 2010 were assumed to be able to process all
their distillate to 15 ppm in a single unit (80% revamped, 20% grass roots).
The methodology for estimating the capital costs for the mixed refineries is somewhat
complex. Table 7.2-34b shows a description of the different new and revamp unit options to
enable refiners to meet the 15 ppm highway and nonroad standards.
7-98
-------
Estimated Costs of Low-Sulfur Fuels
Table 7.2-34b
Summary of New and Revamp Options by Refinery Situation
New vs Revamp
Revamped Highway
Hydrotreaterin2010
New Highway
Hydrotreater added
in 2010
Refinery Configuration
Refinery with Distillate
Hydrotreater
Refinery w/o Distillate
Hydrotreater
Refinery with Distillate
Hydrotreater
Refinery w/o Distillate
Hydrotreater
Fuel Category
Highway
Non-Highway treated in
Hwy Hydrotreater
Non-Highway
Highway
Non-Highway
Highway
Non-Highway treated in
Hwy Hydrotreater
Non-Highway
Highway
Non-Highway
Type of Added Unit
Revamp of Existing Highway
Hydrotreater
Revamped Treater
Revamp of Hydrotreater
installed in 2007
Revamp of Existing Highway
Hydrotreater
Revamp of Hydrotreater
installed in 2007
New Hydrotreater
New Hydrotreater
Revamp of Hydrotreater
installed in 2007
New Treater
Revamp of Hydrotreater
installed in 2007
An example is provided here to better explain the capital cost calculation methodology. This
example is made for a refinery on the Gulf Coast with a distillate hydrotreater and this refinery
will comply with the highway diesel sulfur program in 2010 and also comply with the nonroad
diesel fuel sulfur program in 2010. This refinery also has an FCC unit and a hydrocracker which
is large enough to process all the LCO from the FCC unit. Thus, the highway and nonhighway
pools would be composed of straight run diesel fuel only. The refinery produces 40,000 bbl/day
of highway diesel fuel and 20,000 bbl/day of nonhighway distillate, and the hydrocracker
produces 15,000 bbl/day of hydrocracked distillate, 10,000 bbl of which goes into the highway
pool and 5,000 bbl of which goes into the nonhighway pool. This refinery is presumed to use the
Linde hydrotreating technology to comply with the 2010 standards. As shown in Table 7.2-21,
refineries with distillate hydrotreaters on the Gulf Coast are presumed to hydrotreat 44 percent of
their nonhighway distillate. Thus, 44 percent of the 15,000 bbl per day nonhighway pool (20,000
bbl/day total nonhighway volume minus the 5000 bbl/day which is hydrocracked), or 6600
bbl/day is already hydrotreated, while 8400 bbl/day is not hydrotreated. Thus to calculate the
capital cost for the highway and nonroad programs the following apply:
The capital cost of the highway hydrotreater is estimated using an exponential equation termed
the "six-tenths rule" (from Subsection 7.2.1.4.1 above). The equation is as follows:
7-99
-------
Draft Regulatory Impact Analysis
(Sb/Sa)exCa=Cb,
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, which is 0.65 for desulfurization units
Ca is the cost of the unit quoted by the vendor (which is 25000 bbl/day), and
Cb is the desired cost for the different sized unit.
The cost of the highway hydrotreater would therefore be calculated to be
(30,000/25,000)a65x $21 million = $23.6 million
This value needs to be increased by a factor of 1.2 to account for the offsites, and for the location,
which is 1.0 for the Gulf Coast (from Subsection 7.2.1.4.1 above).
This increases the highway diesel fuel hydrotreater cost to $28.4 million.
If the highway hydrotreater were to be revamped, it would cost $13.1 million using the same
methodology but substituting a $10.6 million figure for the base unit with a 25,000 bbl/day
capacity for the $21 million figure and using a 1.1 factor for offsites instead of 1.2. Although
calculating the revamped cost seems irrelevant for this example, this value is actually used as
described below.
The cost for complying with the Nonroad Program is calculating by calculating the combined
Nonroad/Highway capital cost and subtracting the highway program capital cost from it. Thus,
using the same equation, the cost for the combined new Nonroad/Highway hydrotreater is
calcualated as follows:
(45,000/25,000)a65x $21 million = Cb which is $30.8 million which increases to $36.9 using the
offsite factor (1.2).
To calculate the capital cost of a new Nonroad unit, the new Highway unit capital cost is
subtracted from the combined, new Highway/Nonroad capital cost ($36.9 - $23.6 to yield the
Nonroad new unit capital cost which is $13.3. (The economy of scale benefit is apparent by
calculating the capital cost of a dedicated new, nonroad only unit which is $18.1 million and
comparing it to the $13.3 million figure.)
However, a portion of nonhighway distillate which is being desulfurized down to 500 ppm in
2007 only needs to be revamped. A credit is claimed for this fraction by calculating the economy
of scale capital cost for a revamped unit and ratio the two costs.
A combined Highway/Nonroad revamped unit is calculated as follows
(45,000/25,000)a65x $10.6 million x 1.1 = $17.1 million, then the Nonroad portion is calculated
by subtracting the revamped highway hydrotreater capital cost ($17.1 million - $13.1 million)
which is $4.0 million.
7-100
-------
Estimated Costs of Low-Sulfur Fuels
The Nonroad capital cost is calculating by apportioning the capital cost estimates for to the
respective portions of the Nonroad pool. As stated above, 44 percent of the Nonroad pool is
hydrotreated and this portion requires a new unit cost while the 56 percent balance only requires
the cost of a revamp. Thus, 44 percent of the capital cost is $13.3 million and 56 percent of the
capital cost is $4.0 million, yielding a volume weighted cost of $8.1 million.
Table 7.2-35 presents the estimated volume of nonroad diesel fuel which must be
desulfurized in 2010 to 15 ppm by PADD. The methodology used to develop these figures is the
same as that described above for the required volume of 500 NRLM diesel fuel (Table 7.2-31).
Table 7.2-35
Nonroad Diesel Fuel Needing Desulfurization: 2010-2014a
(million gallons per year)
Nonroad, Diesel Fuel Demand
Highway Spillover
Base 500 ppm Nonroad Volume
Small Refiner Volume
Net Volume of 1 5 ppm Nonroad Fuel
Non-Small Volume Having to Produce 15
ppm Nonroad Diesel Fuel
Remaining Demand for 1 5 ppm Nonroad
Diesel Fuel
PADDs 1 & 3
4440
1018
3422
407
3015
0
3015
PADD 2
3559
1168
2391
369
2022
1032
989
PADD 4
822
524
298
10
288
370
0
PADDS
665
157
508
202
306
84
222
a Based on projected volumes for 2008
Table 7.2-36 presents an analogous set of volumes for 2014. The difference is the absence of
the small refiner volume.
7-101
-------
Draft Regulatory Impact Analysis
Table 7.2-36
Nonroad Diesel Fuel Volume Needing Desulfurization: 2014 and Thereaftera
(million gallons per year)
Required Supply of Nonroad, Diesel
Fuel
Highway Spillover
Net 500 ppm volume to be treated
Small Refiner Volume
Net Volume of 1 5 ppm Nonroad Fuel
Non-Small Volume Required to Produce
15 ppm Nonroad Diesel Fuel
Remaining Demand for 1 5 ppm Nonroad
Diesel Fuel
PADDs 1 & 3
4,440
1,018
3,422
0
3,422
23
3,399
PADD2
3,559
1,168
2,391
0
2,391
1,157
1,234
PADD4
822
524
298
0
298
370
0
PADDS
665
157
508
0
508
108
401
a Based on projected volumes for 2008
In Table 7.2-37, the refineries which are projected to produce 15 ppm nonroad diesel fuel
after the program is fully phased in are characterized by whether they produce predominantly
high sulfur distillate, a mix of highway and high sulfur distillate or predominantly highway diesel
fuel.
Table 7.2-37
Characterization of the Refineries Projected to Produce 15 ppm Nonroad Diesel Fuel
Number of
Refineries
Nonroad Only
Refineries
W/Dist
HT
0
No Dist
Ht
4
Mixed Refineries
Complying with
Highway in 2006
W/Dist
HT
9
No Dist
Ht
4
Mixed Refineries
Complying with
Highway in 2010
W/Dist
HT
7
No Dist
Ht
5
Highway Only
Refineries
W/Dist
HT
0
No Dist
Ht
8
m Refineries listed in No dist Ht column do not currently have a highway diesel hydrotreater, ie make highway fuel from
straight run, hydrocrackate and other low sulfur blendstocks. These refineries would install a new hydrotreater to make
500 ppm diesel for the two step program which is revamped to make 15 ppm nonroad.
7.2.1.5.4 Projected Use of Advanced Desulfurization Technologies
In Chapter 5, we projected of the mix of technologies used to comply with a program being
implemented in any year. This projection took into account the factors which affect the decisions
by refiners in choosing a new technology. The projected mix of technologies for certain
important years is summarized here for the reader's benefit.
7-102
-------
Estimated Costs of Low-Sulfur Fuels
Table 7.2-38
Projected Use of Advanced Desulfurization Technologies for Future Years
Conventional Technology
Linde Isotherming
Phillips SZorb
2008
60
20
20
2009
40
30
30
2010
20
40
40
2012+
0
50
50
7.2.2 Refining Costs
In this section, we present the refining costs for the proposed NRLM diesel fuel program, as
well as for several alternative fuel programs evaluated in the process of developing this proposal.
The first step in developing the refining costs for the proposal was to estimate the cost of
producing 500 and 15 ppm diesel fuel for each of the 143 refineries currently producing either
highway diesel fuel or high sulfur diesel fuel, or both fuels. These costs were estimated for both
conventional and advanced desulfurization technologies using the methodology developed in
Section 7.2 above. The capital and operating cost factors for each desulfurization technology, are
the same for each refinery. However, each refinery's projected 2008 production of highway
diesel fuel and high sulfur distillate, its LCO fraction and other cracked stocks fraction and its
location (i.e., PADD) were also used, which led to different projected costs to produce 500 and
15 ppm diesel fuel for each refinery. As the mix of desulfurization technologies is projected to
vary with the implementation year for the 15 ppm standard, the cost of producing 15 ppm fuel
varies with year of implementation for each refinery in the U.S.
The remainder of this section presents the refining costs for the various fuel programs.
Refining costs to meet the 2007 highway diesel fuel program are presented first, as this provides
the basis for evaluating the additional costs for NRLM diesel fuel sulfur control. Refining costs
for the proposed two-step NRLM fuel program are presented next, followed by the refining costs
for the alternative NRLM fuel programs evaluated in the developing the proposal. 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, namely the
Tier 2 gasoline sulfur program and the 2007 highway diesel fuel program. All per gallon costs
presented in this section would apply to the volume of NRLM disel fuel actually being
desulfurized under the proposed fuel program. These costs would not apply to NRLM diesel fuel
already meeting highway diesel fuel sulfur standards (i.e., spillover fuel).
7.2.2.1. 15 ppm Highway Diesel Fuel Program
Highway diesel desulfurization cost to 15 ppm were estimated in 2006 and 2010 to provide a
basis from which to estimate the costs of the NRLM program. The methodology used here is
nearly identical to that used to develop the costs presented in the 2007 highway diesel fuel
rulemaking. The two differences are: 1) we used more recent estimates of each refinery's current
production of highway diesel fuel and high sulfur distillate, and 2) we modified the methodology
7-103
-------
Draft Regulatory Impact Analysis
used to estimate the cost of expanding the production volume of highway fuel by desulfurizing
current high sulfur distillate. Both of these changes were described in Section 7.2.1 above.
We projected the specific refineries which will produce 15 ppm highway diesel fuel in 2006
based on their projected cost per gallon. We did not consider the potential for refineries to
desulfurize their current high sulfur distillate fuel. The lowest cost refiners were assumed to
produce 15 ppm highway diesel fuel until at least 80% of the required supply of highway diesel
fuel was fulfilled. The exception to this was that several refineries with potentially higher
desulfurization costs were also assumed to produce 15 ppm fuel in 2006. These refineries are
eligible to select a delay in their applicable Tier 2 gasoline sulfur standards if they produce 15
ppm highway diesel fuel in 2006. Several refiners have informed EPA that they are planning to
select this compliance option. Therefore, we projected that these refineries would produce 15
ppm highway diesel fuel in 2006.
We projected specific refineries to produce additional 15 ppm highway diesel fuel in 2010
again based on their projected cost per gallon. The lowest cost refiners were assumed to produce
15 ppm highway diesel fuel until at least 100% of the required supply of highway diesel fuel was
fulfilled. Initially, only distillate volume which is currently highway diesel fuel was considered.
After doing so, we determined that 13 refineries faced very high costs of producing 15 ppm
highway diesel fuel, due solely to their extremely low production volumes and resulting poor
economies of scale for a new or revamped hydrotreater. It is very likely that these refineries
produce highway diesel fuel today from blendstocks which are naturally low in sulfur. It is very
unlikely that they currently have a hydrotreater of such low capacity. Therefore, we do not
believe that it is likely that these refineries would construct a new hydrotreater to produce such a
low volume of highway fuel. Thus, we assumed that they would not produce 15 ppm highway
diesel fuel in 2010. We replaced their production volume with 15 ppm diesel fuel produced from
high sulfur distillate currently being produced by five refiners currently producing both highway
diesel fuel and high sulfur distillate fuel.
The projected costs for producing 15 ppm highway diesel fuel are summarized in Table 7.2-
39.
7-104
-------
Estimated Costs of Low-Sulfur Fuels
Table 7.2-39
Highway Diesel Desulfurization Costs to Meet a 15 ppm Cap Standard
($2002, 7% ROI before taxes)
Number of Refineries
Total Capital Cost (SMillion)
Average Capital Cost per Refinery
(SMillion)
Average Operating Cost per Refinery
($Million/yr)
Total Cost (c/gal)
Refineries
Producing 15 ppm
in 2006
74
4,210
56.9
13.6
3.5
Refineries First
Producing 1 5 ppm in
2010
40
1,240
31.1
4.7
3.8
All Refineries
114
5,450
47.8
9.0
3.6
As can be seen, we project that 74 refineries will invest to produce 15 ppm highway fuel in
2006, with a total capital cost of $4.21 billion ($57 million per refinery). All of the fuel
desulfurized to 15 ppm is produced from current highway diesel fuel. The average cost to
produce 15 ppm highway diesel fuel is 3.5 cents per gallon. These costs assumed that all this 15
ppm fuel is being produced using conventional hydrotreating.
We project that 40 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 $1.24 billion
($31 million per refinery). The average cost for 15 ppm fuel newly produced in 2010 is 3.8 cents
per gallon, which is 0.3 cents higher than 15 ppm fuel first produced in 2006. Five refineries
invest to desulfurize both their current highway and high sulfur distillate fuels to make 15 ppm
fuel, while 13 refineries cease production of highway diesel fuel.
Overall, 114 refineries produce the 15 ppm diesel fuel under the 2007 highway diesel fuel
program, with a total capital cost of $5.45 billion ($47.8 million per refinery). The average
refining cost in 2010 will be 3.6 cents per gallon of fuel.
7.2.2.2 Costs for Proposed Two Step Nonroad Program
The proposed two step program specifies that nonroad, locomotive and marine volumes have
sulfur caps of 500 ppm in year 2007 with nonroad sulfur further reduced to 15 ppm in year 2010.
Small refineries have three and four year delay provisions for complying with the 500 ppm and
15 ppm, respectively. Small refiner's can sell high sulfur diesel fuel in the NRLM market in
years 2007-2010, while small refiners can sell 500 ppm fuel in the nonroad market in years 2010-
2014. In lieu of physically selling these higher sulfur fuels to the NRLM and nonroad markets,
small refiners can sell their credits to other refiners, who can then do the same. From the point of
view of this cost analysis, because these small refiner credits can be sold to others, the small
7-105
-------
Draft Regulatory Impact Analysis
provisions have the net result of reducing the volume of NRLM diesel fuel which would have to
meet the 500 ppm cap in 2007 and the volume of nonroad diesel fuel which would have to meet
the 15 ppm cap in 2010. Small refiners need not be the refiners producing the high sulfur NRLM
diesel fuel in 2007-2010, nor the 500 ppm nonroad diesel fuel in 2010-2014.
Below, we first present an overall summary of the costs of the entire proposed NRLM fuel
program. Then we present in greater detail the refining costs for the three steps of the proposed
NRLM fuel program: 1) the 500 ppm NRLM diesel fuel cap in 2007, 2) the 15 ppm nonroad
diesel fuel cap and 500 ppm locomotive and marine diesel fuel cap in 2010, and 3) the 15 ppm
nonroad diesel fuel cap and 500 ppm locomotive and marine diesel fuel cap in 2014 after the
expiration of small refiner provisions.
Overall, by 2014, we project that 62 refineries would invest to make either 15 or 500 ppm
NRLM diesel fuel. We project that 37 of these refineries would produce 15 ppm nonroad diesel
fuel, with the remaining 25 producing 500 ppm locomotive and marine diesel fuel. The
projected costs to meet these standards are summarized in the two tables below. Table 7.2-40
presents the total refining costs per gallon for the various steps in and fuels of the proposed
program. Table 7.2-41 presents the costs for average and small refineries.
Table 7.2-40
Number of Refineries and Refining Costs for the Proposed Two Step Program
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
All
Refineries3
42
37
25
2.1
2.3
2.2
Small
Refineries
0
19
12
0
3.3
3.3
1 5 ppm Fuel
All
Refineries3
0
25
37
0
4.2
4.4
Small
Refineries
0
0
7
0
0
8.2
' Includes small refiners.
7-106
-------
Estimated Costs of Low-Sulfur Fuels
Table 7.2-41
Refining Costs for Fully Implemented (2014 and Beyond)
Proposed Two Step Program ($2002, 7% ROI before taxes)
Number of Refineries
Total Refinery Capital Cost (SMillion)
2007
2010
2014
Average Refinery Capital Cost (SMillion)
Average Refinery Operating Cost ($Million/yr)
All Refineries
62
1240
449.0
627.0
163.0
20.0
4.1
Small Refineries
19
215
0
131.0
86.0
11.3
1.3
As can be seen, total capital costs would be $1,240 million for the entire proposed NRLM
fuel program (average of $20.0 million per refinery). The per gallon cost of both 500 ppm and
15 ppm diesel fuels would be 2.2-2.3 and 4.2-4.4 cents, respectively. Small refiners projected to
produce either 500 or 15 ppm NRLM diesel fuel would face higher costs on a per gallon basis.
At the 500 ppm level, small refiner costs would be about 50% greater, at 3.3 cents per gallon. At
the 15 ppm level, small refiner costs would be over 80% greater, at 8.2 cents per gallon. Total
capital costs for the 19 small refineries would be $215 million (average of $11.3 million per
refinery).
7.2.2.2.1 Refining Costs in Year 2007
We projected the specific refineries which would produce 500 ppm NRLM fuel beginning in
2007 in two steps. First, we identified specific refineries which would have difficulty marketing
high sulfur distillate fuel in 2007 because of the small volume of heating oil sales in their PADD.
These refineries were projected to hydrotreat all their high sulfur distillate fuel to 500 ppm
regardless of the cost per gallon. However, we excluded small refiners in this step, as they could
sell their high sulfur diesel fuel to either the NRLM diesel fuel market or the heating oil market.
Second, if these refineries did not produce the required volume of 500 ppm NRLM fuel in a
specific PADD, the refineries with the lowest cost of producing additional volume of 500 ppm
fuel were projected to do so until sufficient 500 ppm NRLM fuel was produced in each PADD.
We project that 42 refiners would produce 500 ppm NRLM fuel in 2007. Of these 42
refineries, we project that 32 would install new hydrotreaters, seven "highway" refiners would
perform a relatively minor revamp to their highway distillate hydrotreaters and three refineries
could produce 500 ppm NRLM diesel fuel with an idled highway hydrotreater.1 These last three
1 "Highway" refineries' high sulfur diesel fuel production is no more than 5 percent of their
total no. 2 distillate production. High sulfur refineries high sulfur diesel fuel production is no
7-107
-------
Draft Regulatory Impact Analysis
refineries were projected to build a new hydrotreater to comply with the 15 ppm highway diesel
fuel standard. Therefore, their current highway hydrotreater would be available to produce 500
ppm NRLM fuel.
Small refiners were assumed to exercise small refiner delay provisions and not produce 500
ppm fuel in 2007 unless their desulfurization costs were competitive with other refiners who
invested to make 500 ppm diesel fuel. However, none of the small refiners costs for producing a
15 ppm fuel were competitive with the other refineries which produced sufficient volumes of 500
ppm NRLM fuel to satisfy market demand. Thus, small refiners have no cost associated with
implementing the 500 ppm standard in 2007. Small refiners would sell their high sulfur diesel
fuel to the NRLM market with no attendant refining cost.
The cost of the 500 ppm NRLM cap in 2007 is summarized in Table 7.2-42 below.
Table 7.2-42
Refining Costs for 500 ppm
NRLM Diesel Fuel in 2007 ($2002, 7% ROI before taxes)3
Number of Refineries
Total Refinery Capital Cost (SMillion)
Average Refinery Capital Cost (SMillion)
Average Refinery Operating Cost ($Million/yr)
Amortized Capital Cost (c/gal)
Operating Cost (c/gal)
Cost Per Affected Gallon (c/gal)
All Refineries
42
449
10.7
3.3
0.6
1.5
2.1
' With consideration of small refiner provisions.
We project that the 42 refiners would incur a total capital cost of $499 million (average of
$11 million per refinery). The total refining cost for the 500 ppm NRLM diesel fuel sulfur cap is
2.1 cents per gallon of affected fuel volume, including both operating and amortized capital
costs.
We repeated this 2007 analysis without the small refiner provisions (i.e., for a higher volume
of 500 ppm NRLM diesel fuel. (This situation is equivalent to the proposed 500 ppm NRLM
standard in 2010 without the addition of the 15 ppm nonroad diesel fuel standard). The
less than 95 percent of their total no. 2 distillate production. All other refiners are termed mix
refineries. Mix refineries projected to produce 15 ppm highway diesel fuel in 2006 and 2010 are
termed 2006 mix refiners and 2010 mix refiners, respectively.
7-108
-------
Estimated Costs of Low-Sulfur Fuels
availability of this long term cost is useful in the legal justification of the 500 ppm standard.
With the expiration of the small refiner provisions regarding the 500 ppm NRLM marine
diesel fuel sulfur standard, an additional 20 refiners would invest to produce 500 ppm NRLM
diesel fuel, for a total of 62 refineries producing 500 ppm NRLM diesel fuel. The overall
refining cost would increase very slightly to 2.2 cents per gallon. Of the 20 new refineries, 19
would be small refineries. The reason for the predominance of small refiners in this step is that
most of these 19 small refiners would have difficulty marketing high sulfur distillate fuel in 2010
because of the small volume of heating oil sales in their PADD. On average, the refining cost for
small refiners would be more than 60% higher than that of the non-small refiner, 3.3 cents per
gallon. Various costs of the 500 ppm NRLM diesel fuel cap without the small refiner provisions
are summarized in Table 7.2-43.
Table 7.2-43
Refining Costs for 500 ppm NRLM Diesel Fuel
in 2007 without Small Refiner Provisions ($2002, 7% ROI before taxes)3
Number of Refineries
Total Refinery Capital Cost (SMillion)
Average Refinery Capital Cost (SMillion)
Average Refinery Operating Cost ($Million/yr)
Capital Cost (c/gal)
Operating Cost (c/gal)
Cost Per Affected Gallon (c/gal)
All
Refineries
62
600
9.7
2.8
0.6
1.6
2.2
Nonsmall
Refineries
43
468
10.9
3.6
0.5
1.5
2.0
Small
Refineries
19
131
6.9
0.9
1.5
1.8
3.3
1 Equivalent to the costs of the 500 ppm NRLM cap in 2010 without the 15 ppm nonroad cap.
7.2.2.2.2 Refining Costs in Year 2010
In 2010 under the proposal, all nonroad diesel fuel except that represented by small refiners
must meet a 15 ppm cap. The specific refineries producing this 15 ppm nonroad diesel fuel were
identified in a two step process, analogous to the procedure followed for 2007. First, of those
refineries producing 500 ppm NRLM fuel in 2007, we identified specific refineries which would
have difficulty marketing 500 ppm locomotive and marine diesel fuel in 2010 because of
difficulty economically transporting this fuel in large quantities. Second, the refineries with the
lowest cost of producing 15 ppm fuel were projected to do so until sufficient 15 ppm nonroad
fuel was produced in each PADD.
After the refineries projected to produce 15 ppm nonroad diesel fuel in 2010 were identified,
this left a few refineries still producing 500 ppm diesel fuel from those first producing 500 ppm
7-109
-------
Draft Regulatory Impact Analysis
NRLM diesel fuel in 2007. Additional refineries were then identified in each PADD until the
total production volume of 500 ppm diesel fuel reached the required volume of locomotive and
marine and small refiner nonroad diesel fuel.
We project that 25 refineries would produce 15 ppm nonroad diesel fuel in 2010. Two of
these refineries would install new hydrotreaters, as they were using their existing highway diesel
hydrotreater to produce 500 ppm NRLM diesel fuel in 2007. Five "highway" refineries would
incorporate their current high sulfur distillate fuel with their highway diesel fuel when they
revamp their highway hydrotreater to produce 15 ppm highway diesel fuel in 2010. The
remaining 18 refineries are projected to revamp their new nonroad hydrotreater built in 2007 to
produce 500 ppm NRLM diesel fuel.
The refining costs to produce 15 ppm nonroad fuel in 2010 are presented in Table 7.2-44.
The first column of costs shows the total refining cost relative to today's uncontrolled sulfur
levels. The last column shows the incremental costs relative to the cost of producing 500 ppm
fuel in 2007. Small refiners were assumed to exercise small refiner delay provisions and not
produce 15 ppm fuel in 2010 unless their desulfurization costs were competitive with other
refiners in whom invested to make 15 ppm diesel fuel. However, none of the small refiners costs
for producing a 15 ppm fuel were competitive with the other refineries which produced sufficient
volumes of 15 ppm nonroad fuel to satisfy market demand. Thus, small refiners are projected to
have no cost associated with the 15 ppm nonroad diesel fuel standard in 2010.
Table 7.2-44
Refining Costs to Produce 15 ppm Nonroad Diesel Fuel in 2010
($2002, 7% ROI before taxes)
Number of Refineries
Total Refinery Capital Cost (SMillion)
Average Refinery Capital Cost (SMillion)
Average Refinery Operating Cost ($Million/yr)
Capital Cost (c/gal)
Operating Cost (c/gal)
Cost Per Affected Gallon (c/gal)
All
Refineries
25
720
28.8
6.0
1.5
2.7
4.2
Incremental Desulfurization Cost
SOOppm to 15 ppm All Refineries
25
477
19.1
2.6
0.9
1.2
2.1
The desulfurization equipment used to meet the 500 ppm standard would have been built
three years prior, and we expect it would have been designed to facilitate further processing to 15
ppm sulfur through a revamp. However, a few refiners which were expected to use their existing
7-110
-------
Estimated Costs of Low-Sulfur Fuels
highway diesel hydro treaters to meet the proposed 500 ppm cap in 2007 would likely have to
construct new equipment in 2010 to meet the 15 ppm cap on nonroad diesel fuel.
We project that 25 refineries would invest to produce 15 ppm nonroad in 2010 at an
incremental capital cost of $477 million. Including the cost of meeting the 500 ppm NRLM cap
in 2007, these 25 refineries' total capital costs would be $720 million. The incremental cost of
producing 15 ppm nonroad diesel fuel is 2.1 cents per gallon, for a total cost of 4.2 cents per
gallon. The incremental cost of 2.1 cents per gallon to desulfurize 500 ppm diesel fuel to a 15
ppm cap is 1.5 cents per gallon less than the 3.6 cent per gallon cost estimated above for the 15
ppm highway diesel fuel cap. There are three reasons for this. One, most 15 ppm highway fuel
is being initially produced in 2006, when we project little or no use of advanced desulfurization
technologies. Two, current highway hydrotreaters are at least 10 years old. We project that only
80% of them can be revamped to produce 15 ppm diesel fuel. Thus, the cost for 15 ppm highway
diesel fuel includes new hydrotreaters for 20% of the volume. However, over 90% of 500 ppm
nonroad diesel fuel would be produced in 2007 using new hydrotreaters. All of these new units
could be designed to be revamped in 2010. Three, we focused the production of 15 ppm
highway diesel fuel to a large degree on those refiners already producing 500 ppm highway diesel
fuel. This included some refiners with relatively high costs of producing 15 ppm fuel. As
described above, we did exclude 13 refineries with very high costs of producing 15 ppm fuel, and
replaced their highway fuel with 15 ppm fuel produced from current high sulfur distillate fuel by
four selected refineries. However, we did not include a few current "high sulfur" refineries
which are projected to have the lowest cost of producing 15 ppm diesel fuel from high sulfur
distillate fuel. We will reconsider this decision in the analysis for the final rule, as it may have
increased the projected cost of 15 ppm highway diesel fuel and lowered the cost of producing 15
ppm nonroad diesel fuel to too great a degree. However, we do not expect this change to
substantially change the average costs per gallon, as the total fuel volume affected by this
decision is small.
With respect to the 500 ppm locomotive and marine diesel fuel cap in 2010, we project that
20 refiners would invest in new hydrotreaters to produce 500 ppm fuel, with 19 of these being
small refiners. The reason most of these additional refiners would be small refiners is due to the
expiration of the small refiner provisions related to 500 ppm NRLM diesel fuel. An additional 17
refineries would continue producing 500 ppm diesel fuel (which they started doing in 2007).
The costs of producing 500 ppm diesel fuel in 2010 are presented in Table 7.2-45. This fuel
includes locomotive and marine diesel fuel, as well as 500 ppm nonroad diesel fuel produced by
small refiners (or by other refiners purchasing small refiner credits). Of the 20 refineries which
initially comply with the 500 ppm standard in year 2010, 17 refiners would install new hydro
treaters and three "highway" refiners would modify their existing highway hydrotreater to
process their high sulfur distillate fuel.
7-111
-------
Draft Regulatory Impact Analysis
Table 7.2-45
Refining Costs for 500 ppm Diesel Fuel in 2010 ($2002, 7% ROI before taxes)
Number of Refineries
Total Refinery Capital Cost (SMillion)
Average Refinery Capital Cost (SMillion)
Average Refinery Operating Cost ($Million/yr)
Capital Cost (c/gal)
Operating Cost (c/gal)
Cost Per Affected Gallon (c/gal)
All Refineries in
2010
37
357
9.7
5.3
0.7
1.6
2.3
New Refineries in
2010
20
150
7.5
1.6
0.8
1.6
2.4
Small Refineries
19
131
6.9
0.9
1.5
1.8
3.3
The average cost per gallon of producing 500 ppm fuel for the 20 new 500 ppm refineries is
almost identical to that for the 17 refineries already producing 500 ppm fuel. However, small
refiners would face costs roughly 40% higher than those of the average refiner producing 500
ppm fuel.
7.2.2.2.3 Refining Costs in Year 2014
15 ppm Nonroad Diesel Fuel: In 2014, small refiner provisions related to the 15 ppm
nonroad diesel fuel cap expire, increasing the total required volume of 15 ppm nonroad diesel
fuel. The total production volume of 500 ppm NRLM diesel fuel decreases to just that used in
locomotives and marine vessels. The specific refineries producing the additional volume of 15
ppm nonroad diesel fuel were those facing the lowest projected costs per gallon in each PADD,
plus some refineries which we projected would have difficulty distributing large volumes of 500
ppm locomotive and marine diesel fuel. The volume of combined 15 ppm and 500 ppm NRLM
fuel in 2014 is the same as that in 2010.
We project that 12 additional refineries would produce 15 ppm nonroad diesel fuel in 2014,
with 7 of these being refineries owned by small refiners. None of these refineries would install
new hydrotreaters, as none were using their existing highway diesel hydrotreater to produce 500
ppm NRLM diesel fuel in 2007. Three "highway" refineries would incorporate their current high
sulfur distillate fuel with their highway diesel fuel when they revamp their highway hydrotreater
to produce 15 ppm highway diesel fuel in 2010. The remaining 9 refineries are projected to
revamp their new nonroad hydrotreater built in 2007 to produce 500 ppm NRLM diesel fuel.
The refining costs to produce 15 ppm nonroad fuel in 2014 are presented in Table 7.2-46.
The first two columns containing costs show the total and incremental refining costs for all
7-112
-------
Estimated Costs of Low-Sulfur Fuels
refineries. The last two columns containing costs show the total and incremental refining costs
for small refiners. Total refining costs are those relative to today's uncontrolled sulfur levels.
Incremental costs are those relative to the cost of producing 500 ppm fuel in 2007 or 2010.
Table 7.2-46
Refining Costs for 15 ppm Nonroad 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
12
256
21.3
2.9
2.4
3.0
5.3
Incremental to
500 ppm
12
163
13.6
1.5
1.5
1.5
3.0
Small Refineries
Total
7
161
23.0
2.2
4.4
3.8
8.2
Incremental to
500 ppm
7
86
12.23
0.9
2.3
1.1
3.9
The total refining cost to produce 15 ppm fuel is 5.3 cents per gallon, or 1.1 cent per gallon
more than in 2010. The average incremental cost to desulfurize from 500 ppm to 15 ppm is 3.0
cents per gallon, or 0.9 cents per gallon higher than in 2010. Small refiners' average cost to
produce 15 ppm nonroad diesel fuel 8.2 cents per gallon, or more than 50% higher than that of
the average refiner. The average refinery first producing 15 ppm nonroad diesel fuel in 2014
faces a capital investment of $14 million, while the investment for the average small refiner
would be only slightly smaller, $12 million.
The following two tables present the total and incremental refining costs for all 15 ppm
nonroad diesel fuel being produced in 2014, after the expiration small refiner provisions. These
costs include those for refiners first producing 15 ppm nonroad diesel fuel in 2010 and 2014.
7-113
-------
Draft Regulatory Impact Analysis
Table 7.2-47
Total Refinery Costs for 15 ppm Nonroad Diesel Fuel 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
37
976
26.4
5.0
1.6
2.8
4.4
Non-small
Refineries
30
813
27.1
5.7
1.4
2.7
4.1
Small
Refineries
7
161
23.0
2.2
4.4
3.8
8.2
Table 7.2-48
Incremental Refinery Costs for All 15 ppm Nonroad Fuel 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
37
640
17.3
2.0
1.1
1.1
2.3
Non-small
Refineries
30
556
18.5
2.6
1.0
1.2
2.2
Small
Refineries
7
84
11.9
1.0
2.3
1.7
4.0
With full implementation of the 15 ppm nonroad diesel fuel cap, we project that 37 refineries
would produce this fuel. The total refining cost measured from today's high sulfur level would
be 4.4 cents per gallon, or 2.3 cents per gallon cost over that to produce 500 ppm fuel. Small
refineries would have an average cost of 8.2 cents per gallon, or twice as high as the average non-
small refineries. The average capital cost to produce 15 ppm nonroad fuel would be $23.0
million for the average small refiner, or $4 million less than the average non-small refinery.
However, the amortized capital cost per gallon would be much higher for the average small
refinery due to their lower production volumes.
500 ppm Locomotive and Marine Diesel Fuel: In 2014, the number of refineries producing
500 ppm fuel drops from 37 to 25, as no new volume of 500 ppm diesel fuel would be required
and 12 refineries producing 500 ppm diesel fuel in 2010 shift to 15 ppm nonroad diesel fuel.
7-114
-------
Estimated Costs of Low-Sulfur Fuels
There is no new investment to produce 500 ppm diesel fuel, as all of the 500 ppm locomotive
and marine diesel fuel being produced in 2014 was already being produced in 2010. The costs of
the remaining 500 ppm diesel fuel being produced in 2014 are shown in Table 7.2-49.
Table 7.2-49
Refining Costs to Produce 500 ppm Locomotive and Marine Diesel Fuel in 2014
($2002, 7% ROI before taxes)
Number of Refineries
Average Refinery Capital Cost (SMillion)
Average Refinery Operating Cost ($Million/yr)
Capital Cost (c/gal)
Operating Cost (c/gal)
Cost Per Affected Gallon (c/gal)
All Refineries
25
10.6
2.8
0.6
1.6
2.2
Small Refineries
12
4.5
0.8
1.0
1.7
2.7
The cost to produce 500 ppm diesel fuel in 2014 is 2.2 cents per gallon. This is a slight
decrease from the cost in 2010 due a number of higher cost refineries (mostly owned by small
refiners) exiting the 500 ppm market to make 15 ppm fuel. The average cost for small refiners
still producing 500 ppm diesel fuel is only slightly greater than that for the average refinery, 2.7
cents per gallon.
7.2.2.2.4 Total Refining Costs at Different Rates of Return on Investment
We also estimated the total refining cost of the proposed NRLM fuel program using two
alternative rates of return on investment: 1) 6% per year after taxes, and 2) 10% per year after
taxes. The 6% rate is indicative of the economic performance of the refining industry over the
past 10-15 years. The 10% rate is indicative of economic performance of an industry like
refining which would attract additional capital investment. The total refining costs for both 500
and 15 ppm NRLM fuels once the proposed program is fully implemented in 2014 are shown
below for our standard 7% before tax rate, and the two alternative rates. As can be seen, the
difference in the rates of return on investment range from 0.1-0.8 cents per gallon.
7-115
-------
Draft Regulatory Impact Analysis
Table 7.2-50
Total Refining Costs for the Fully Implemented Proposed NRLM Program with Different Capital
Amortization Rates (cents per gallon, $2002)
Societal Cost
7% ROI before Taxes
Capital Payback
(6% ROI, after Taxes)
Capital Payback
(10% ROI, after Taxes)
500 ppm Locomotive and Marine
Diesel Fuel
2.2
2.3
2.6
15 ppm Nonroad Diesel Fuel
4.4
4.5
5.2
7.2.2.3 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 costs estimates
to the level of advanced technology projected, we developed costs for producing 15 ppm nonroad
diesel fuel with only the use of conventional hydrotreating.
Total refining costs to produce 15 ppm nonroad diesel fuel in 2014 are shown in Table 7.2-
51. The number of refiners required to invest (37 refiners) and types of hydrotreating
modifications are the same with conventional technology as described above for a mix of
advanced and conventional technology. Total capital costs would be $983 million with
conventional technology, essentially identical to the $976 million investment with advanced
technology (see Table 7.2-40). However, operating costs would be nearly 40% higher with
conventional technology, $6.9 million as compared to $5.0 million with advanced technology.
The same comparison applies to the impact of advanced technology on the capital costs faced by
small refiners. While the use of conventional technology increases operating costs for small
refiners ($2.6 million per year versus $2.2 million per year with advanced technology), the
reduction is smaller at just over 15%. This smaller benefit is due to their lower production
volumes and lower fractions of LCO and other cracked stocks. The total cost to produce 15 ppm
nonroad diesel fuel in 2014 with conventional technology would be 5.4 cents per gallon, versus
4.4 cents with a mix of conventional and advanced technology.
7-116
-------
Estimated Costs of Low-Sulfur Fuels
Table 7.2-51
Total Refining Costs to Produce 15 ppm Nonroad Diesel Fuel
with Conventional Technology 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 Cost (c/gal)
All Refineries
37
983
26.6
6.9
1.6
3.8
5.4
Small Refineries
7
150
21.5
2.6
4.1
4.5
8.5
The previous comparisons involved the total cost of producing 15 ppm diesel fuel from high
sulfur diesel fuel. However, we are only projecting that the advanced technology would be
applied to the step from 500 ppm to 15 ppm sulfur. Table 7.2-52 compares the refining costs of
producing 15 ppm nonroad diesel fuel from 500 ppm diesel fuel using 100 percent conventional
hydrotreating and with a mix of advanced and conventional technology.
Table 7.2-52
Impact of Advanced Technology on the Incremental Refining Costs
to Produce 15 ppm Nonroad Diesel Fuel ($2002, 7% ROI before taxes)
Average Capital Cost (SMillion)
Operating Cost ($Million/yr)
Capital Cost (c/gal)
Operating Cost (c/gal)
Cost Per Gallon (c/gal)
Refineries Producing
15 ppm Fuel First in 2010
Advanced and
Conventional
19.1
2.6
0.9
1.2
2.1
Conventional
Technology
Only
19.4
5.4
0.9
2.5
3.4
Refineries Producing
15 ppm Fuel First in 2014
Advanced and
Conventional
13.6
1.5
1.5
1.5
3.0
Conventional
Technology
Only
12.7
2.11
1.4
2.1
3.5
For refiners that first produce 15 ppm nonroad diesel fuel in 2010, the projection that 80
percent would use advanced technology versus conventional technology decreases incremental
refining costs relative to 500 ppm fuel by 1.4 cents per gallon, or more than 25%. Capital costs
decrease only slightly, while operating costs decrease by more than 50%.
7-117
-------
Draft Regulatory Impact Analysis
For refiners that first produce 15 ppm nonroad diesel fuel in 2014, the projection that 100
percent would use advanced technology versus conventional technology decreases incremental
refining costs relative to 500 ppm fuel by 0.5 cents per gallon, or roughly 15%. Capital costs
actually increase, while operating costs decrease by roughly 30%. The lower savings occurring
in 2014 relative to 2010 are due to the relative volumes of distillate being desulfurized at each
refinery and the percentages of LCO and other cracked stocks at the refineries producing 15 ppm
fuel in the two timeframes.
7.2.2.4 Refining Costs for Alternative NRLM Fuel Programs
7.2.2.4.1 One Step NRLM Fuel Program in Year 2008
This one step program specifies that nonroad diesel fuel would have to meet a sulfur cap of
15 ppm starting on June 1, 2008, while locomotive and marine diesel fuel would have to meet a
500 ppm cap at the same time. Small refiners would have four more years before having to meet
these standards. In the meantime, small refiners could sell high sulfur diesel fuel to the NRLM
fuel markets.
Once fully implemented, the same refineries would produce the same 15 and 500 ppm
NRLM diesel fuels as those projected under the proposed NRLM fuel program. Still, moving up
the 15 ppm nonroad diesel fuel cap by two years would increase costs in two ways. One, the cost
of 15 ppm nonroad fuel would be incurred two years earlier. That effect is addressed in Chapter
12, where aggregate costs are estimated for the various alternatives. Two, the cost of producing
15 ppm nonroad fuel would increase, as the earlier implementation date is projected to reduce the
penetration of advanced desulfurization technology. As described in section 7.2.1 above, we
project that refiners would use a mix of 60 percent conventional and 40 percent advanced
technology to produce 15 ppm diesel fuel in 2008, as compared to a 20/80 mix in 2010. Fifteen
ppm diesel fuel initially produced in 2012 would be desulfurized using 100 percent advanced
technology, as was projected for 2014. Cost are only presented for the fully implemented
program in 2012.
We project that 62 refineries would produce 500 ppm locomotive and marine diesel fuel or
15 ppm nonroad diesel fuel in 2012, when the program would be fully implemented. The total
refining costs for the one step fuel program are shown in Table 7.2-53. The total refining cost for
the one step fuel program for 15 ppm nonroad fuel would be 4.8 cents per gallon, or 0.4 cents per
gallon more than that for nonroad fuel cost in the proposed two step program. The total capital
cost of the one step program would also exceed those of the proposed two step program by $55
million.
7-118
-------
Estimated Costs of Low-Sulfur Fuels
Table 7.2-53
Total Refining Cost to Produce 15 ppm Nonroad Diesel Fuel
Under One Step Program ($2002, 7% ROI before taxes)
Number of Refineries
Total Refinery Capital Cost (SMillion)
Average Refinery Capital Cost (SMillion)
Average Refinery Operating Cost ($Million/yr)
Cost Per Affected Gallon Cost (c/gal)
One Step Program
37
1,031
27.9
5.6
4.8
Proposed Program
37
976
26.4
5.0
4.4
The cost of the 500 ppm locomotive and marine diesel fuel under the one step pro
gram would not differ from that under the proposed two step program, as the same refineries
using the same conventional hydrotreating are projected to be used in both cases. The difference
in total costs of the two programs lies in the production of 15 ppm nonroad diesel fuel.
7.2.2.4.2 Proposed Two Step NRLMFuel Program with Nonroad 15ppm Cap in 2009
This program would be identical to the proposed NRLM fuel program except for one
difference: the 15 ppm nonroad sulfur cap would be implemented one year earlier. The 500 ppm
sulfur standard for nonroad, locomotive and marine would still begin in mid-2007. Small
refiners would be able to sell high sulfur diesel fuel to the NRLM markets until mid-2009, and
would be able to sell 500 ppm diesel fuel to the nonroad market until mid-2013.
Moving up the 15 ppm nonroad diesel fuel cap by one year would increase costs in two ways.
One, 15 ppm nonroad fuel would be incurred one year earlier. That effect is addressed in
Chapter 12, where aggregate costs are estimated for the various alternatives. Two, the cost of
producing 15 ppm nonroad fuel would increase due to the earlier implementation date. The 37
refineries planning to produce 15 ppm nonroad diesel fuel in 2009 would only be producing 500
ppm NRLM fuel for two years. Thus, we projected that they would fully construct their 15 ppm
desulfurization equipment in 2007. This moved up the capital needed to meet the 15 ppm cap by
one year, increasing amortized costs per gallon of 15 ppm fuel produced. It also reduced the
projected penetration of advanced desulfurization technology. Specifically, we project that 60%
of the volume of 15 ppm fuel would be produced using advanced technology with a 2007
construction date, compared to the 80% level a year later. Small and other refiners first
producing 15 ppm fuel in 2013 would be all projected to use 100 percent advanced technology.
The cost of the 500 ppm locomotive and marine diesel fuel cap would not be affected by
moving up the 15 ppm cap one year, as the same refineries using conventional hydrotreating are
projected to be used in both programs. The difference in total costs of the two programs lies in
the production of 15 ppm nonroad diesel fuel. Thus, we have summarized the costs of producing
15 ppm nonroad diesel fuel in Table 7.2-54. Overall, the same refineries are projected to produce
7-119
-------
Draft Regulatory Impact Analysis
15 ppm nonroad diesel fuel. While total capital costs are essentially identical, operating costs
increase relative to the proposed two step program by 10% and per gallon costs increase by 5%.
Table 7.2-54
Total Refining Costs for 15 ppm Nonroad Fuel Under a
Two Step Program with the 15 ppm Standard in 2009 ($2002, 7% ROI before taxes)
Number of Refineries
Total Refinery Capital Cost (SMillion)
Average Refinery Capital Cost (SMillion)
Average Refinery Operating Cost ($Million/yr)
Cost Per Affected Gallon Cost (c/gal)
Two Step Program
with 15 ppm in 2009
37
977
26.4
5.5
4.6
Proposed
Two Step Program
37
976
26.4
5
4.4
7.2.2.4.3 Proposed Two Step Program with a 15 ppm Cap for Locomotive and Marine Fuel
in 2010
This program would be identical to the proposed NRLM fuel program except for one
difference: the 15 ppm nonroad sulfur cap would be extended to locomotive and marine diesel
fuel. The 500 ppm sulfur standard for nonroad, locomotive and marine would still begin in mid-
2007. Small refiners would be able to sell high sulfur diesel fuel to the NRLM markets until
mid-2010, and would be able to sell 500 ppm diesel fuel to the NRLM market until mid-2014.
The cost of the 500 ppm locomotive and marine diesel fuel cap in 2007 would not be affected
by moving up the 15 ppm cap one year, as the same refineries using conventional hydrotreating
are projected to be used in both programs. The difference in total costs of the two programs lies
in the increased production of 15 ppm nonroad diesel fuel in 2010 and 2014. The total costs of
producing NRLM diesel fuel for both the proposed program and that with the 15 ppm cap for
locomotive and marine diesel fuel are shown in Table 7.2-55.
7-120
-------
Estimated Costs of Low-Sulfur Fuels
Table 7.2-55
Total Refining Costs for Two Step Program:
All NRLM Fuel to 15 ppm in 2010 ($2002, 7% ROI before taxes)3
Number of Refineries
Total Capital Cost (SMillion)
Average Capital Cost per Refinery (SMillion)
Average Refinery Operating Cost ($Million/yr)
Cost Per Affected Gallon Cost (c/gal)
One Step Program
62
1,720
27.7
4.9
4.6
Proposed Program
62
1,240
20.0
4.1
4.1
1 Fully implemented program in 2014.
Overall, the same refineries are projected to be affected. The difference is that refineries
producing 500 ppm locomotive and marine diesel fuel in 2014 under the proposed program now
produce 15 ppm diesel fuel. Extending the 15 ppm cap to locomotive and marine diesel fuel
increases total capital cost by $480 million. The total cost per gallon of fuel affected would
increase by 0.5 cent per gallon, or just over 10%. The cost of 15 ppm diesel fuel would increase
from to 4.6 from 4.4 cents per gallon, or just 5%. However, this approach spreads out the
increased costs of extending the 15 ppm cap to greater fuel volume of all NRLM diesel fuel
volume. The cost of the 15 ppm locomotive and marine diesel fuel would be 4.8 cents per
gallon, or about 10% greater than the 15 ppm nonroad diesel fuel.
7.2.2.5 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 proposed NRLM fuel program.
The previous sections estimated the total capital cost associated with the proposal and two
alternative programs. 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 proposal and two 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 proposed
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.j The
J 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,
7-121
-------
Draft Regulatory Impact Analysis
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 total capital costs
associated with the 2007 highway diesel fuel program, as 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 7.2-56.
Table 7.2-56
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
1.33
2.15
0.82
0.41
0.62
0.21
Total
1.76
1.15
2.21
2.76
0.98
0.06
0.47
0.64
0.21
a2002 dollars obtained by
2 gasoline program (1997
use of Chemical Engineering Plant Annual Cost Index to adjust capital costs for Tier
dollars) and highway 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% of the capital cost was
expended in the calendar year two years prior to start up, 50% was expended in the year prior to
start up and the remaining 20% was expended in the year of start up. We repeated this analysis
for the one step NRLM program and the proposed NRLM program with 15 ppm cap for
locomotive and marine diesel fuel. The results are summarized in Table 7.2-57 below.
December 1999, EPA 420-R-99-023. Adjusted to 2002 dollars using Chemical Engineering
Plant Cost Index.
7-122
-------
Estimated Costs of Low-Sulfur Fuels
Table 7.2-57
Capital Expenditures for NRLM Fuel Programs 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
Proposed Two Step
NRLM Fuel Program
Increment
0.14
0.23
0.09
0.19
0.31
0.13
0.05
0.08
0.03
Total3
1.76
1.15
2.21
2.90
1.21
0.15
0.66
0.95
0.34
0.05
0.08
0.03
Proposed Program +
Locomotive and Marine
to 15ppmin2010
Increment
0.14
0.23
0.09
0.32
0.54
0.22
0.54
0.91
0.36
Total3
1.76
1.15
2.21
2.90
1.21
0.15
0.79
1.18
0.43
0.54
0.91
0.39
One Step
NRLM Program in 2008
Increment
0.31
0.52
0.21
0.64
0.08
0.13
0.05
Total3
1.76
1.15
2.21
2.76
1.29
0.58
0.68
0.64
0.29
0.13
0.05
32002 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 all NRLM programs. The proposed two
step NRLM program increases this peak by $140 million, or about 5%. Thereafter, capital
requirements drop dramatically. The proposed two step NRLM program with a 15 ppm cap on
locomotive and marine diesel fuel would require the same capital investments increases through
2007. Thereafter, it causes increased capital requirements, but this is well after the peak
investment requirements have occurred. The one step NRLM fuel program avoids increasing
capital investment in 2005, but more than makes up for this in 2006, though at a lower total
investment for all three programs. In all cases, the vast majority of capital investment in the
2002-2006 timeframe, when capital investment requirements are the highest, are causes by the
Tier 2 gasoline sulfur and 2007 highway diesel fuel programs. In comparison, the capital
investment requirements for the proposed NRLM fuel program are much smaller and more
spread out.
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
7-123
-------
Draft Regulatory Impact Analysis
(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
1990's.k 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 proposed
NRLM fuel program is about two-thirds of the historic peak level of investment for meeting
environmental programs experienced during 1992-1994.46 Given that the capital required by the
proposed NRLM fuel program contributes only 5% 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.
7.2.2.6 Other Cost Estimates for Desulfurizing Highway Diesel Fuel
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 the implementation of the 15 ppm cap standard for
highway diesel fuel. The cost estimate from this study is presented here and compared to our
costs.
In a study conducted for the EMA,4748 MathPro, Inc. first estimated the cost of desulfurizing
diesel fuel to meet a 15 ppm highway diesel fuel sulfur cap standard followed by a two step
nonroad standard of caps of 500 ppm and 15 ppm. MathPro assumed that desulfurization would
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.
Mathpro made a number of 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
K Rasmussen, Jon A., "The Impact of Environmental Compliance Costs on U.S. Refining
profitability," EIA, October 29, 1997.
L API Reported Refining and Marketing Capital Investment 1990-1998.
7-124
-------
Estimated Costs of Low-Sulfur Fuels
variance in seasonal demand. In addition, Mathpro applied a factor which falls somewhere in the
range of 1 - 8 percent for sizing the desulfurization unit larger for reprocessing off-spec material
to meet a number of different sulfur targets. Since meeting a 500 ppm cap standard is not very
stringent, Mathpro likely assumed that a desulfurization unit would need to be sized larger by 1 -
4 percent. For meeting the 15 ppm cap standard which is a relatively stringent sulfur standard
compared to the 500 ppm sulfur level studied, Mathpro likely assumed the desulfurization unit
would be sized larger by 5 - 8 percent. Onsite 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 that the existing highway diesel
hydrotreater must be removed from service and an new grassroots unit takes its place which
desulfurizes untreated distillate down to under 15 ppm. The difference between scenarios 1 and
2 is that scenario 1 does not allow some flexibilities which 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 cap 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 that it can achieve 15 by itself and then another unit is added in
parallel which is also being fed by a low throughput which allows it to meet the 15 ppm cap
standard. Scenario 4 installs the new capital in series with the existing hydrotreater with both
units handling the entire feed rate.
7-125
-------
Draft Regulatory Impact Analysis
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 understand how different
investment scenarios would affect the cost for the highway rule, they have implications for the
nonroad rule as well. For meeting a 500 ppm cap nonroad diesel fuel standard, the used highway
units which are freed up in Scenarios 1 and 2 can thus be converted over to nonroad service
which dramatically reduces the capital cost of compliance, and this supplements the existing
nonroad capacity which is already in place. However, for Scenario 2, the installed grassroots
capacity installed for the highway 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 cap
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 which 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 which the
refinery model was granted. For Scenario 3, the refinery model added some new grassroots unit
capacity compared to 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
7.2-58.
7-126
-------
Estimated Costs of Low-Sulfur Fuels
Table 7.2-58
Mathpro Capital Costs for Desulfurizing Highway and Nonroad Diesel Fuel
Reference Case
Highway 15 ppm
Cap Std
Nonroad Meeting
a 500 ppm Cap
Standard
Nonroad Meeting
a 1 5 ppm Cap
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 needed to determine which of the Mathpro cases which would best approximate the
investment scenarios which we are using in our 500 ppm cost analysis. As described above in
this section, the refineries which comply with the highway rule in 2006 by putting in a new
hydrotreater (20 percent of the mixed highway and nonroad refineries which comply with the
highway requirements in 2006 and which have a distillate hydrotreater), thus idling the existing
hydrotreater, is projected to use the existing hydrotreater to produce 500 ppm sulfur nonroad
diesel fuel in 2007. Those refineries comprise 7 percent of the nonroad pool. The rest of the
refineries are expected to install a new unit in 2007 to comply with the 500 ppm sulfur standard.
Next, we examined the Mathpro investment cases to match them with the scenarios in our cost
analysis. There were no cases which matched our scenario exactly, but we found two Mathpro
cases which, together, matched our investment scenario. The first is the No Retrofit Inflexible
case which met the nonroad requirements exclusively through using existing capacity. The
second case is the retrofitting derating case which met the nonroad requirements through new
capital investment. Since our analysis had only 9 percent of the nonroad volume as being
produced by refineries which would use the existing hydrotreater to produce 500 ppm sulfur
7-127
-------
Draft Regulatory Impact Analysis
nonroad diesel fuel, the Mathpro costs were weighted 7 percent No Retrofit Inflexible costs and
93 percent and Retrofit DeRate costs.
We then examined the Mathpro 15 ppm cases to determine which of them best matched our
15 ppm scenario. Since we already have identified 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 to consider how to adjust the costs to remove any costs associated with
going from untreated to 500 ppm. As discussed above in this section, our 15 ppm scenario has
new nonroad diesel fuel hydrotreating units being installed in 2010, although those which are
mixed highway and nonroad refineries are expected to install their highway and nonroad units
together taking advantage of the economies of scale for doing so. 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. Since the
second Mathpro case apparently allowed backsliding in the highway grassroots units needed for
complying with the highway 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, we decided to use Mathpro's case one.
Case one, however, needed to be adjusted to develop a scenario which we believe is more
realistic based on how refiners are likely to comply with the highway and nonroad programs.
Mathpro's case one was associated with the replacement of the existing hydrotreating capacity
which was likely used for desulfurizing nonroad down to 500 ppm. However, we believe that 80
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 the extra grassroots
hydrotreating capacity. We accomplished this by estimating what percent of the capital costs is
necessary for complying with 15 ppm standard and 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 bpd, to yield a total of 8,500 barrels per day of
replaced capital capacity which we assumed would be untreated to 500 ppm nonroad
hydrotreated capacity. Since we believe that it is reasonable that 20 percent of this existing
capacity would be replaced, we maintained 20 percent of 8,500 bpd, or an additional 1,700
barrels of the new nonroad hydrotreating capacity. Therefore, we maintained 13,500 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 13,500/20,300. 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 which deserve
7-128
-------
Estimated Costs of Low-Sulfur Fuels
mentioning or which were adjusted. 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 which 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, it appears that some of the reblending
costs in the MathPro study appear to be transfer payments,"1 not costs. 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 which can be produced
cheaper in many cases. As a result, their hydrogen cost may be exaggerated, which would tend to
increase costs. Finally, it should be noted that the MathPro study did take into consideration the
need for lubricity additives, but did not address costs that might be incurred in the distribution
system. Thus, in a comparison of our costs with Mathpro's, we will include our cost estimate for
adding the appropriate amount of lubricity additive, but not add on the distribution costs. For
comparing the aggregate capital costs, the Mathpro aggregate capital costs for the cases which
were chosen 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 which we used for making the adjustments are presented in Section 7.1 of the draft RIA.
A comparison of Mathpro's costs and our costs to desulfurize highway diesel fuel to meet a 500
ppm sulfur cap standard and then a 15 ppm sulfur cap standard is shown below in Table 7.2-59.
Table 7.2-59
Comparison of Mathpro's and EPA's
Refining Costs for Meeting a 500 ppm and a 15 ppm Nonroad Diesel Fuel Sulfur Cap
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
2.5
925
3.3
836
5.8
1761
EPA' s Costs
Advanced Tech
in 2010
2.2
612
2.2
649
4.4
1261
No Advanced
Tech
2.2
612
3.3
606
5.5
1218
M A transfer payment is when money changes hands, but no real resources (labor, natural
resources, manufacturing etc.) are consumed.
7-129
-------
Draft Regulatory Impact Analysis
7.3 Cost of Distributing Non-Highway Diesel Fuel
7.3.1 Distribution Costs Under the 500 ppm Sulfur Non-Highway Diesel Fuel Program
7.3.1.1 Fuel Distribution-Related Capital Costs Under the 500 ppm Sulfur Non-
Highway Diesel Fuel Program
The potential capital costs associated with distributing 500 ppm sulfur non-highway diesel
fuel pertain to the need for additional product segregation which might result. Section 5.4.2 of
this draft RIA evaluates the potential for additional product segregation in each segment of the
distribution system. The projected capital costs associated with distribution non-highway diesel
fuel meeting the proposed 500 ppm standard are limited to the need for approximately 1,000 bulk
plants to add an additional storage tank and demanifold their delivery truck to handle an
additional diesel product.
In its comments to the government/industry panel convened in accordance with the Small
business Regulatory enforcement Act (SERFA), 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 would range 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 diesel 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 would be
installed on the tank truck(i.e. so that there would be a separate delivery system for each fuel
carried by the delivery truck). The cost of demanifolding a tank truck and installing a an
additional fuel delivery system is estimated at $10,000, of which $6,000 is the cost of installing a
new fuel delivery system.49 Thus, the cost to each of the affected bulk plants would be $100,000
for a total cost of $100,000,000.
Amortizing the capital costs over 20 years, results in a estimated cost for tankage at such bulk
plants of 0.1 cent per gallon of affected non-highway diesel engine fuel supplied. Twenty years
was chosen due to the very long life of fuel storage tanks, and their lack of obsolescence.
Although the impact on the overall cost of the proposed 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 could make other arrangements to continue servicing
both heating oil and NRLM markets. In some cases, two or more bulk plants within a given
service area may a have a single owner. In these cases, the bulk plant operator could continue to
serve both markets by storing heating oil at one facility and nonroad fuel at the other. However,
it would be more likely that multiple bulk plants serving a given geographic area would have
different owners. In such cases, exchange agreements could be worked out between the two bulk
plant operators so that they could continue to serve both markets.
7-130
-------
Estimated Costs of Low-Sulfur Fuels
7.3.1.2 Distribution Costs Due to the Reduction in Fuel Volumetric Energy Content
Under the Proposed 500 ppm Sulfur Diesel Fuel Program
We estimate that desulfurization of non-highway diesel fuel to meet the proposed 500 ppm
sulfur standard would result in a 0.7 percent reduction in the volumetric energy content (VEC) of
the affected fuel (see section 5.9.2 of this draft RIA). This increases the cost to distribute diesel
fuel due to the increased volume.
We believe that 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 #2 diesel fuel to end-users versus the price to resellers. We used
the five year average of the difference between these two prices to arrive at an estimated typical
cost of distributing non-highway diesel fuel to the end-user of 10 cents per gallon. The following
table (7.3-1) presents the EIA data that we used in estimating the cost of distributing non-
highway diesel fuel.
Table 7.3-1
Cost of Distributing High-Sulfur #2 Diesel Fuel3
(cents per gallon, excluding taxes)
Year
1995
1996
1997
1998
1999
5 Year Average
Sales to End Users
52.4
63.9
60.2
43.7
51.9
54.4
Sales to Resellers
61.4
73.2
69.8
55.5
62.0
64.4
Difference Between Sales
to End Users &
Sales to Resellers
9.0
9.3
9.6
11.8
10.1
10.0
' Energy Information Administration, Annual Energy Review 2001
We assumed that the current 10 cent per gallon cost of distributing diesel fuel would stay
constant. For example, a one percent increase in the amount of fuel distributed would increase
total distribution costs by one percent. Thus, the 0.7 percent reduction in VEC is estimated to
result in a 0.07 cents per gallon increase in the cost to distribute non-highway diesel fuel. This
cost was applied to the gallons of non-highway diesel fuel that would need to be desulfurized to
meet the proposed 500 ppm sulfur standard. This cost was applied to NRLM from June 2007
through June 2010 when the proposed 15 ppm sulfur standard for nonroad diesel fuel would be
implemented. After June 2010, this cost applies to LM fuel only. The additional costs
associated with the further reduction in nonroad diesel fuel VEC associated with desulfurization
to meet a 15 ppm sulfur specification are discussed in section 7.3.2.2 of this draft RIA.
7-131
-------
Draft Regulatory Impact Analysis
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 would require slightly less
energy to be distributed also indicates that this estimate is conservative.
7.3.1.3 Other Potential Distribution Costs Under the Proposed 500 ppm Sulfur
Diesel Fuel Program
We anticipate that there would be no other significant distribution costs associated with the
adoption of the proposed 500 ppm non-highway diesel sulfur standard beyond those described in
sections 7.3.1.1 & 2 above. As discussed in section 5.4 of this draft RIA, we do not expect the
need for additional storage tanks beyond that discussed in section 7.3.1.2 above, an increase in
pipeline downgrade or transmix volumes, or the need for additional facilities at the refinery to
comply with the proposed fuel marker requirements.
Bulk plant and tank truck who previously only handled high-sulfur diesel fuel would need to
begin observing practices to limit sulfur contamination during the distribution of 500 ppm diesel
fuel. However, these practices are well established and are primarily associated with purging
storage tanks and fuel delivery systems of high-sulfur diesel fuel prior to the use in handling 500
ppm diesel fuel. Such tasks can be readily accomplished. Training of employees would be
necessary to impress the importance of consistently and carefully observing the practices to limit
sulfur contamination. However, we estimate the costs associated would be minimal. In addition,
we are estimating that most of the affected bulk plant operators would install dedicated storage
tanks and truck delivery systems. This would obviate 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.
7.3.2 Distribution Costs Under the 15 ppm Sulfur Nonroad Diesel Fuel Program
7.3.2.1 Fuel Distribution-Related Capital Costs Under the 15 ppm Sulfur Nonroad
Diesel Fuel Program
As discussed in section 5.6 of this draft RIA, we do not anticipate that the implementation of
the proposed 15 ppm sulfur standard would result in the need for fuel distribution industry to
make changes that would require investment capital. Specifically, we project that there would be
no substantial need for additional storage tanks or other facility changes to ensure product
segregation.
7.3.2.2 Distribution Costs Due to the Reduction in Fuel Volumetric Energy Content
Under the 15 ppm Sulfur Nonroad Diesel Fuel Program
We project that desulfurizing diesel fuel to 15 ppm would reduce volumetric energy content
of the affected fuel by an additional 0.35 percent in addition the 0.7 percent reduction in VEC
which accompanied desulfurization to meet the proposed 500 ppm standard. Thus, the total
7-132
-------
Estimated Costs of Low-Sulfur Fuels
reduction in the VEC of nonroad diesel fuel which would need to be desulfurized to meet the
proposed 15 ppm standard would be 1.1 percent (see section 5.9.2).
The methodology described in 7.3.1.2. for the calculation of the increase in distribution costs
due to the reduction in VEC associated with meeting the proposed 500 ppm sulfur standard is
also applicable in calculating the increase in distribution costs associated with meeting the
proposed 15 ppm nonroad standard. Using this methodology, we estimate that the additional
0.35 percent reduction in the VEC of nonroad diesel fuel would increase the cost of distributing
the affected gallons of 15 ppm nonroad diesel fuel by an additional 0.04 cent per gallon. Thus,
the total increase in distribution costs associated with the decrease in VEC of 15 ppm nonroad
diesel fuel would be 0.11 cent per gallon of affected nonroad diesel pool. This cost was applied
to the volume of nonroad diesel fuel that would need to be desulfurized to meet the proposed 15
ppm standard beginning in June 2010.
7.3.2.3 Other Potential Distribution Costs Under the 15 ppm Sulfur Nonroad Diesel
Fuel Program
We anticipate that there would be no other significant distribution costs associated with the
adoption of the proposed 500 ppm non-highway diesel sulfur standard beyond those described in
sections 7.3.1.1 & 2 above. As discussed in section 5.4 of this draft RIA, we do not expect the
need for additional storage tanks beyond that discussed in section 7.3.1.2 above, an increase in
pipeline downgrade or transmix volumes, or the need for additional facilities at the refinery to
comply with the proposed fuel marker requirements.
Bulk plant operators who previously only handled high-sulfur heating oil would need to
begin observing practices to limit sulfur contamination during the distribution of 1 ppm diesel
fuel. However, these practices will be established well in advance as entities comply with the 15
ppm highway standard in 2006. These practices include purging storage tanks and fuel delivery
systems of high-sulfur diesel fuel prior to the use of the equipment in handling 1500 ppm diesel
fuel.. Training of employees would be necessary to impress the importance of consistently and
carefully observing the practices to limit sulfur contamination. However, we estimate the costs
associated would be minimal. In addition, we are estimating that most of the subject bulk plant
operators would install dedicated storage tanks and truck delivery systems. This would obviate
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 4.6 in this draft RIA, 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 during the distribution of 15 ppm diesel fuel sulfur
contamination 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 highway diesel
program's RIA.50 Highway diesel fuel and nonroad diesel fuel meeting a 15 ppm sulfur
specification would share the same distribution system until nonroad diesel fuel would be dyed
as it leaves the terminal to meet IRS requirements. Therefore, we do not expect there would be
7-133
-------
Draft Regulatory Impact Analysis
any additional actions and associated costs needed to optimize the distribution system to limit
sulfur contamination during the distribution of 15 ppm nonroad diesel fuel.
A small fraction of bulk plant and tank truck operators who do not handle highway diesel
may have had no prior experience in limiting sulfur contamination during the distribution of 15
ppm diesel fuel prior to the implementation of the proposed 15 ppm nonroad diesel sulfur
standard in 2010. These would be the same entities that may have had no prior experience in
distributing 500 ppm diesel fuel prior to the implementation of the 500 ppm NRLM sulfur
standard in 2008. Consistent with the projections developed in the final highway diesel fuel rule
regarding the handling practices for 15 ppm diesel fuel we believe that such entities would only
need to more carefully and consistently observe standard industry practices regarding purging
tanks and delivery lines of higher-sulfur product prior to the use in delivering 15 ppm nonroad
diesel fuel.51 Additional training may be needed of there operators to emphasize the criticality
following such procedures. However, we believe that such training and the associated costs
would be minimal.
7.3.3 Cost of Lubricity Additives
Our evaluation of the potential impact of the proposed non-highway diesel sulfur standards
on fuel lubricity is contained in section 5.9 of this draft RIA. We concluded that the increased
need for lubricity additives that would result from the adoption of these proposed sulfur
standards would be similar to that for highway diesel fuel meeting the same sulfur standard.
The highway diesel final rule estimated that all diesel fuel meeting a 15 ppm sulfur standard
would require the use of lubricity additives at a cost would be 0.2 cents per gallon.52 As noted
above, we concluded that the impact on fuel lubricity of meeting a 15 ppm sulfur standard for
non-highway diesel fuel would be similar to that experienced in meetinglS ppm highway diesel
sulfur standard. Therefore, consistent with the estimated cost due 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 associated with today's action to account for cost for the increased use of
lubricity additives in 15 ppm nonroad diesel fuel. This lubricity additive cost would be
applicable to the affected nonroad diesel fuel pool beginning in 2010.
In estimating lubricity additive costs for 500 ppm diesel fuel we assumed that the same
additive concentration needed in 15 ppm diesel fuel would also be needed in 500 ppm diesel fuel
that needs such an additive, and that 5 percent of all 500 ppm diesel fuel would require a lubricity
additive. Based on these assumptions, we estimate that the cost of additional lubricity additives
for the affected 500 ppm NRLM diesel fuel would be 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 of this Draft RIA) it is likely that substantially
less that 5 percent of 500 ppm diesel fuel outside of California requires a lubricity additive.
Therefore, we believe that 0.01 cent 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 would likely be considerably less, we have no information with
7-134
-------
Estimated Costs of Low-Sulfur Fuels
which to better quantify the percentage of 500 ppm diesel fuel that is currently treated with a
lubricity additive or the appropriate additive treatment rate. Hence, it seems most appropriate to
use an estimate that is conservatively high. The 0.01 cents per gallon lubricity additive cost
would be applicable to the affected non-highway diesel pool (NRLM) until the proposed
reduction in the sulfur standard for nonroad diesel fuel to 15 ppm would be implemented in
2010. After 2010, the 0.01 cents per gallon lubricity additive cost would be applicable to the
affected locomotive and marine pool.
7.3.4 Fuel Marker Costs
Under the proposed requirement, high sulfur heating oil would be marked between 2007 and
2010 and locomotive and marine diesel fuel would be marked from 2010 until 2014. After 2014
the proposed marker requirement would expire.
Our conversations with marker manufactures indicate that the cost to treat fuel with either of
the markers considered in the proposed rule would be lower than the costs to treat non-highway
diesel fuel with red dye to meet IRS requirements. A major pipeline charges 0.2 cents per gallon
to inject red dye to IRS specifications. We believe that this represents a conservatively high
estimate of treatment costs for the markers under consideration. For the purposes of our cost
calculations, we applied the annual cost of treating heating oil volumes with marker to the
affected NRLM pool from June 2007 through June 2010. This results in a charge for the heating
oil marker used during this time period of 0.16 cents per gallon of affected NRLM fuel. For the
time period from June 2010 though June 2014 the cost of marking locomotive and marine diesel
fuel was applied to the locomotive and marine pool itself. Thus, the marker costs during this
time period are estimated at 0.2 cents per gallon of affected locomotive and marine diesel fuel.
Please refer to section 7.1 of this draft RIA regarding the volume of 15 ppm diesel fuel we
estimate would be used in locomotive and marine diesel fuel.
7.3.5 Distribution, Lubricity, and Marker Costs Under Alternative Sulfur Control Options
Distribution costs vary from 0.2 to 0.4 cents per gallon of affected diesel fuel under the
alternative options considered. The variation in distribution cost is relatively insignificant
compared to the variation in refining costs (see section 7.2 of this draft RIA).
Distribution costs vary due to differences in the volumetric energy density (VEC), marker and
lubricity additive cost components. Under all of the alternative options considered the cost of
additional storage tanks remains constant (0.1 cents per gallon).
Since the reduction in VEC is a side-effect of the desulfurization process, the increase in
distribution costs associated it varies directly with the timing and applicability of the sulfur
standards for NRLM. The earlier NRLM would be desulfurized, the earlier the charge for VEC
must be applied. Since the reduction in VEC is higher in meeting a 15 ppm standard (0.07 cents
per gallon) versus a 500 ppm standard (0.11 cents per gallon), costs related to reduced VEC
7-135
-------
Draft Regulatory Impact Analysis
increases if the 15 ppm sulfur standard would be more broadly applicable and/or the
implementation of the standard would be earlier.
There is relatively little lubricity additive cost associated with desulfurization to meet a 500
ppm standard (0.01 cents per gallon) compared to that associated with desulfurization to meet a
15 ppm standard (0.2 cents per gallon). Consequently, distribution costs related to the need for
additional lubricity additive increases if the 15 ppm sulfur standard would be more broadly
applicable and/or the implementation of the standard would be earlier.
Marker related costs also vary based on the timing and applicability of the sulfur standards
under consideration. Under some alternative options, marker costs apply for a longer or shorter
duration and/or to a larger or smaller diesel pool.
A summary of the distribution costs under the various alternative options is presented in
following table 7.3-2. A more complete discussion of the alternative options considered can be
found in chapter 12 of this draft RIA. Please refer to section 7.1. regarding the volumes of fuel
that these costs apply to. The net fuel related costs under the various sulfur control options under
consideration is contained in section 7.5 of this draft RIA.
7-136
-------
Estimated Costs of Low-Sulfur Fuels
Table 7.3-2
Distribution and Additive Costs under Non-Highway Diesel Control Options (c/gal)a
Control Option
Proposal
Proposal with
15ppmNRin2009
Proposal with
NR,L&M
to 15 ppmin2010
Sulfur Specifications
500 ppm NR, L & M
500 ppm L & M
ISppmNR
500 ppm L & M
500 ppm NR, L & M
500 ppm L & M
ISppmNR
(total incl 2007)
500 ppm L & M
500 ppm NR, L & M
ISppmNR, L&M
ftotal incl 2007s)
Year
2007-2010
2010-2014
2010 +
2014 +
2007 - 2009
2009-2013
2009 +
2013 +
2007-2010
2010 +
Reduction in
VEC
0.07
0.07
0.11
0.07
0.07
0.07
0.11
0.07
0.07
0.11
Additional
Storage Tanksb
0.1
0.1
0.1
0.1
0.1
0.1
0.1
0.1
0.1
0.1
Additional
Lubricity
Additive
0.01
0.01
0.2
0.01
0.01
0.01
0.2
0.01
0.01
0.2
Marker0
0.16
0.2
NA
NA
0.16
0.2
NA
NA
0.16
NA
Total
Distribution &
Additive Costs
0.3
0.4
0.4
0.2
0.3
0.4
0.4
0.2
0.3
0.4
a Legend: NR= Nonroad diesel, L = Locomotive diesel, M = Marine diesel, VEC = Volumetric energy content
b Costs applied to "affected" gallons, i.e., gallons of fuel destined for a given end-use that would be desulfurized under given control option.
0 When marker would be required in heating oil, costs are applied to affected NR, L, & M volume. When marker would be required in L &M, costs are applied to
affected L, & M volume.
7-137
-------
Draft Regulatory Impact Analysis
7.4 Net Cost of the Two-Step Nonroad Diesel Fuel Program
The estimated refining costs from Subsection 7.2.2 and distribution and additive costs from Sections 7.3 and 7.4 for the
Nonroad Program and the other fuel options considered are summarized together in the following table. Both the 2007 and
costs are presented in the table. Note that these fuel costs do not include the impacts of the small refiner exemptions.
Table 7.4-1
Table of Fuel Costs for Nonroad Program Control Options (cents per gallon and $2002)
Proposed
the 2010
Option
Proposal - Locomotive and Marine
to 500 ppm and NR to 1 5 ppm
One Step Locomotive & Marine to
500 ppm and NR to 1 5 ppm
Nonroad goes to 15 ppm in 2009
Nonroad, Locomotive and Marine
go to 15 ppm
Specification
500ppmNR,L&M
500 ppm L & M
15 ppm NR (total incl 2007)
500 ppm L & M
500 ppm L & M
ISppmNR
500 ppm L & M
500 ppm NR, L & M
500 ppm L & M
15 ppm NR (total incl 2007)
500 ppm L & M
500 ppm NR, L & M
ISppmNR, L&M
(total incl 2007)
Year
2007
2010
2010+
2014+
2008
2008+
2012+
2007
2009
2009+
2013+
2007+
2010+
Refining Costs
(c/gal)
2.2
2.2
4.4
2.2
2.2
4.8
2.2
2.2
2.2
4.6
2.2
2.2
4.6
Distribution & Additive
Costs (c/gal)
0.3
0.4
0.4
0.2
0.4
0.4
0.2
0.3
0.4
0.4
0.2
0.3
0.4
Total Costs
(c/gal)
2.5
2.6
4.8
2.4
2.5
5.2
2.4
2.5
2.5
5.0
2.4
2.5
5.0
7-138
-------
Estimated Costs of Low-Sulfur Fuels
Our projected total cost for producing 500 ppm fuel is essentially identical to the historical
price differential between 500 ppm highway diesel and uncontrolled high sulfur diesel. This
differential has averaged about 2.5 cents per gallon for the five year period from 1995 to 1999.
Arguably, this differential would minimally account for refiners costs to hydrotreat and distribute
a 500 ppm diesel fuel. While cost and prices are not always directly related, the fact that the two
numbers are so closely aligned provides added assurance that our cost estimates are reasonable.
7.5 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, the cost of alternative
energy sources (e.g., coal, natural gas), etc. Most of these factors would be unaffected by the
proposed NRLM diesel fuel standards. However, a few, namely fuel processing costs and
refinery capacity utilization could or would be affected by the proposed NRLM fuel program.
Fuel processing and distribution costs would clearly be affected due to the cost of
desulfurizing NRLM diesel fuel to either the 500 or 15 ppm sulfur cap. Refinery utilization
levels could be affected as the capacity to produce 500 ppm or 15 ppm NRLM diesel fuel would
depend on refiners' investment in desulfurization capacity. The potential impact of increased
fuel processing and distribution costs on the prices will be assessed below. The impact of the
proposed rule on refinery utilization levels is beyond the scope of this analysis. In the long run,
refiners would 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 was already addressed in Chapter 5.9 above.
Two approaches to projecting future price impacts are evaluated here. The most direct
approach to estimating the impact of the proposed 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
essentially the same as 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
marketed today in the U.S. This fuel is designed to be used in vehicle fleets which have been
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 U.S. are not at all representative of what might be expected in 2010 under
the proposed standard.
7-139
-------
Draft Regulatory Impact Analysis
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 essentially 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 10-15 ppm sulfur fuel.
Thus, as in the U.S., the prices paid for this fuel are not representative of what would occur in the
U.S. in 2010. Therefore, we did not attempt to use fuels sold today which have sulfur levels
similar to the standards we are proposing to evaluate our cost estimate for complying with the 15
ppm cap 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 would affect fuel processing and
distribution costs across the nation. (The exception would be California, where we presume that
sulfur caps at least as stringent as those being proposed federally 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 U.S. 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 essentially 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 which are 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 U.S., 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 U.S.
7-140
-------
Estimated Costs of Low-Sulfur Fuels
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 $0.49 per barrel in 1992 to a high of 2.23 per barrel in 2000." 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
proposed NRLM rule would be very unlikely to have a major impact on factors such as these.
Thus, projecting the likely price impact of the proposed rule 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 which have existed in the past.
To do this, we developed three projections for the potential impact of the proposed fuel
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 only reflect incremental operating costs. Consistent with this, under this assumption, we
project that the price of NRLM diesel fuel within a PADD would 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 prior to this rule. This may or may not be the case. If not, the price increase could be
even lower than that projected below. Under this "low cost" set of assumptions, the refiner with
the highest operating cost would not recover any of his invested capital related to desulfurizing
NRLM diesel fuel, but all other refiners would recover some of their investment.0
The mid-range estimate of price impacts could 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% per annum before
taxes). Unlike the low and high price scenarios, the mid-range, full cost price scenario does not
have a direct connection with economic pricing theory. It simply represents a convenient price
estimate which falls between the low and high price estimates.
Under this full cost price scenario, lower cost refiners would recover their capital investment
plus economic profit, while those with higher than average costs would recover some of their
invested capital, but not all of it (i.e., at a lower rate of return than 7% per annum).
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 brought on barely fulfills demand and does
N Inflation adjusted dollars. EIA, Form EIA-28, Financial Reporting System, updated June
27, 2002.
0 Theoretically, some refiners might recover all of 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.
7-141
-------
Draft Regulatory Impact Analysis
not create an excess in capacity which 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% per annum before taxes). All other refiners would recover more than
their capital investment.
The range of potential price increases resulting from these three sets of assumptions are
shown in Table 7.5-1. The wholesale price of high sulfur distillate fuel has varied widely even
over the past 12 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 would be equivalent to the price increase in terms of cents per
gallon shown below.
Table 7.5-1
Range of Possible Total Diesel Fuel Price Increases (cents per gallon)"
Low Price
Mid-Point
High Price
2007 500 ppm Sulfur Cap: Nonroad, Locomotive and Marine Diesel Fuel
PADDs 1 and 3
PADD 2
PADD 4
PADDS
0.9
2.3
1.7
1.0
1.5
3.0
4.1
2.8
3.4
4.8
5.8
4.3
2010 15 ppm Sulfur Cap: Nonroad Diesel Fuel
PADDs 1 and 3
PADD 2
PADD 4
PADDS
1.8
2.9
3.0
1.7
3.0
6.1
8.9
5.9
5.4
7.4
9.3
8.4
' At a wholesale price of approximately $ 1.00 per gallon, these values also represent the percentage increase in diesel fuel
price.
Under the low price scenario, the price of nonroad, locomotive and marine diesel fuel would
increase in 2007 by 1-2 cents per gallon, depending on the area of the country. In 2010, the price
of nonroad diesel fuel would increase a total of 2-3 cents per gallon. Locomotive and marine
diesel fuel prices would continue to increase by 1-2 cents per gallon.
Under the mid-point price scenario, the price of nonroad, locomotive and marine diesel fuel
would increase in 2007 by 2-4 cents per gallon, depending on the area of the country. In 2010,
7-142
-------
Estimated Costs of Low-Sulfur Fuels
the price of nonroad diesel fuel would increase a total of 3-9 cents per gallon. Locomotive and
marine diesel fuel prices would continue to increase by 2-4 cents per gallon.
Under the high price scenario, the price of nonroad, locomotive and marine diesel fuel would
increase in 2007 by 3-6 cents per gallon, depending on the area of the country. In 2010, the price
of nonroad diesel fuel would increase a total of 5-9 cents per gallon. Locomotive and marine
diesel fuel prices would continue to increase by 3-6 cents per gallon.
7-143
-------
Draft Regulatory Impact Analysis
Appendix 7A: Estimated Total Off-Highway Diesel Fuel Demand and Diesel Sulfur
Levels
Table 7A-1 was used to derive Table 7.1-3.
Table 7A-1
Estimated Total Non-Highway Diesel Fuel Demand"
Category
Total
Residential
Commercial
Industrial
Area
US
California
Alaska
Hawaii
49 State
PADDI
PADDII
PADD III
PADDIV
PADDV
US
California
Alaska
Hawaii
49 State
PADDI
PADDII
PADD III
PADDIV
PADDV
US
California
Alaska
Hawaii
49 State
PADDI
PADDII
PADD III
PADDIV
PADDV
US
California
Alaska
Hawaii
49 State
PADDI
PADDII
PADD III
PADDIV
PADDV
No. 1
0.236
0.001
0.043
0.000
0.235
0.012
0.134
0.004
0.033
0.053
0.118
0.000
0.023
0.118
0.008
0.072
0.000
0.009
0.029
0.064
0.000
0.01
0.064
0.003
0.036
0.001
0.011
0.013
0.054
0.000
0.01
0.054
0.001
0.026
0.003
0.013
0.011
L.S. Diesel
1.871
0.122
0.004
0.004
1.749
0.534
0.452
0.339
0.271
0.274
0.000
1.061
0.079
0.004
0.003
0.982
0.418
0.276
0.146
0.069
0.151
0.81
0.043
0.00002
0.001
0.767
0.116
0.176
0.193
0.202
0.123
Diesel
3.260
0.261
0.006
0.019
2.999
0.430
1.587
0.558
0.221
0.462
0.000
0.000
0.000
No. 2 P.O.
8.019
0.010
0.041
0.000
8.009
6.914
0.770
0.147
0.045
0.144
6.086
0.007
0.025
0.000
6.079
5.391
0.557
0.001
0.030
0.108
1.576
0.003
0.012
0.000
1.573
1.304
0.102
0.142
0.007
0.021
0.357
0.000
0.004
0.000
0.357
0.219
0.111
0.004
0.008
0.015
Distillate
P.O.
8.961
0.520
0.160
0.106
8.441
1.984
2.318
2.992
0.563
1.103
0.000
0.000
0.000
Total
Distillate
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
Other
Distillate
0.143
0.000
0.000
0.000
0.143
0.073
0.061
0.003
0.004
0.002
0.000
0.000
0.000
H.S.
Diesel
1.363
0.007
0.000
0.000
1.248
0.500
0.438
0.276
0.034
0.115
0.000
0.474
0.005
0.446
0.219
0.153
0.058
0.016
0.028
0.889
0.002
0.802
0.281
0.285
0.218
0.018
0.087
Total*
23.853
0.921
0.254
0.129
22.824
10.447
5.760
4.319
1.171
2.153
6.204
0.007
0.048
0.000
6.197
5.399
0.629
0.001
0.039
0.137
3.175
0.087
0.026
0.003
3.065
1.944
0.567
0.347
0.103
0.213
2.110
0.045
0.014
0.001
1.980
0.617
0.598
0.418
0.241
0.236
7-144
-------
Estimated Costs of Low-Sulfur Fuels
Category
Oil Company
Farm
Electric Utility
Railroad
Vessel*
Area
US
California
Alaska
Hawaii
49 State
PADDI
PADDII
PADD III
PADDIV
PADDV
US
California
Alaska
Hawaii
49 State
PADDI
PADDII
PADD III
PADDIV
PADDV
US
California
Alaska
Hawaii
49 State
PADDI
PADDII
PADD III
PADDIV
PADDV
US
California
Alaska
Hawaii
49 State
PADDI
PADDII
PADD III
PADDIV
PADDV
US
California
Alaska
Hawaii
49 State
PADDI
PADDII
No. 1
0.000
0.000
0.000
0.000
0.000
L.S. Diesel
0.000
0.000
0.000
0.000
0.000
Diesel
0.000
3.08
0.254
0.0000
3
0.008
2.826
0.389
1.572
0.549
0.219
0.351
0.000
0.000
0.000
No. 2 P.O.
0.000
0.000
0.000
0.000
0.000
Distillate
P.O.
0.685
0.006
0.026
0.000
0.679
0.019
0.042
0.561
0.029
0.034
0.000
0.793
0.008
0.036
0.09
0.785
0.305
0.134
0.195
0.009
0.151
3.071
0.189
0.004
2.882
0.5
1.233
0.686
0.345
0.307
2.081
0.101
0.08
0.013
1.980
0.49
0.301
Total
Distillate
0.000
0.000
0.000
0.000
0.000
Other
Distillate
0.000
0.089
0.000
0.089
0.044
0.040
0.003
0.002
0.000
0.000
0.000
0.000
H.S.
Diesel
0.000
0.000
0.000
0.000
0.000
Total*
0.685
0.006
0.026
0.000
0.679
0.019
0.042
0.561
0.029
0.034
3.169
0.254
0.000
0.008
2.915
0.433
1.612
0.552
0.221
0.351
0.793
0.008
0.036
0.090
0.785
0.305
0.134
0.195
0.009
0.151
3.071
0.189
0.004
0.000
2.882
0.500
1.233
0.686
0.345
0.307
2.081
0.101
0.080
0.013
1.980
0.490
0.301
7-145
-------
Draft Regulatory Impact Analysis
Category
Military
Construction
Other Off High
Area
PADD III
PADDIV
PADDV
US
California
Alaska
Hawaii
49 State
PADD I
PADD II
PADD III
PADDIV
PADDV
US
California
Alaska
Hawaii
49 State
PADD I
PADD II
PADD III
PADDIV
PADDV
US
California
Alaska
Hawaii
49 State
PADD I
PADD II
PADD III
PADDIV
PADDV
No. 1
0.000
0.000
0.000
0.000
L.S. Diesel
0.000
0.000
0.000
Diesel
0.18
0.007
0.006
0.011
0.173
0.041
0.015
0.009
0.002
0.111
0.000
0.000
No. 2 P.O.
0.000
0.000
0.000
Distillate
P.O.
1.033
0.0002
0.256
0.000
1.9
0.194
0.007
0.003
1.706
0.511
0.549
0.394
0.15
0.295
0.431
0.022
0.007
0.000
0.409
0.159
0.059
0.123
0.03
0.06
Total
Distillate
0.000
0.000
0.000
Other
Distillate
0.054
0.000
0.00005
0.054
0.029
0.021
0.000
0.002
0.002
0.000
0.000
H.S.
Diesel
0.000
0.000
0.000
Total*
1.033
0.000
0.256
0.234
0.007
0.006
0.011
0.227
0.070
0.036
0.009
0.004
0.113
1.900
0.194
0.007
0.003
1.706
0.511
0.549
0.394
0.150
0.295
0.431
0.022
0.007
0.000
0.409
0.159
0.059
0.123
0.030
0.060
a Energy Information Administration. Fuel Oil and Kerosene Sales 2000. DOE/EIA-0535(00). Office of Oil and Gas, U.S. Department of
Energy. Washington, D. C. September, 2001.
7-146
-------
Estimated Costs of Low-Sulfur Fuels
The information in Tables 7A-2 through 7 A-7 was used to derive Table 7.1-21
Table 7A-2
1996 Off-highway Diesel Sulfur Levels from TRW (Niper)
Sample
39
40
41
42
43
45
46
47
48
41
42
43
48
49
50
51
40
45
46
47
48
52
52
53
Region
Eastern
Eastern
Eastern
Eastern
Eastern
Southern
Southern
Southern
Eastern
Eastern
Central
Central
Eastern
Central
Central
Central
Southern
Southern
Southern
Southern
Southern
Rocky Mtn
Western
Western
District
Bl
B1,D1
C2
C1,E3
C2,E3
Dl
Dl
D3
D, B, C, A
C2
C1,E3
C2,E3
D, B, C, A
F1,E3
G
G
B1,D1
Dl
Dl
D3
D, B, C, A
K3, L3, M3
K3, L3, M3
Ml
PADD
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
3
3
3
3
3
4
5
5
National
Use Category
SM
RR, SM
TT
RR, SM
RR, SM
TT
TT
SM
SM,RR
TT
RR, SM
RR, SM
SM,RR
RR, SM
SM
TT
TT
SM
RR, SM
TT
TT
TT
Presumed Volume
1,500,000
1,500,000
500,000
1,500,000
1,000,000
750,000
750,000
137,500
7,637,500
500,000
275,000
275,000
1,775,000
2,825,000
1,500,000
750,000
750,000
137,500
3,137,500
412,500
412,500
412,500
1,500,000
1,912,500
15.925.000
Sulfur, ppm
1700
4000
1800
3700
2900
4330
4900
9600
1600
[
1800
3700
2900
1600
4200
2050
1640
[
4000
4330
4900
9600
1600
[
4100
L
4100
2700
[
1
Sulfur * Volume
2,550,000,000
6,000,000,000
900,000,000
5,550,000,000
2,900,000,000
3,247,500,000
3,675,000,000
1,320,000,000
3,423
900,000,000
1,017,500,000
797,500,000
7,455,000,000
3,600
6,000,000,000
3,247,500,000
3,675,000,000
1,320,000,000
4,539
1,691,250,000
4,100
1,691,250,000
4,050,000,000
3,002
3.641
7-147
-------
Draft Regulatory Impact Analysis
Table 7A-3
1997 Off-highway Diesel Sulfur Levels from TRW (Niper)
Sample
47
48
49
50
51
52
53
51
52
53
57
60
61
48
50
55
56
57
58
59
63
63
64
Region
Eastern
Eastern
Eastern
Eastern
Eastern
Eastern
Eastern
Central
Central
Central
Southern
Central
Central
Southern
Southern
Southern
Southern
Southern
Southern
Southern
Rocky Mtn
Western
Western
District
B
B,D
Bl
B1,D1
C1,E3
C2
C2,E3
C1,E3
C2
C2,E3
D1,G2
F1,E3
G
B,D
B1,D1
Dl
Dl
D1,G2
D2
D3
K3, L3, M3
K3, L3, M3
M1,N1
PADD
1
1
1
1
1
1
1
2
2
2
2
2
2
3
3
3
3
3
3
3
4
5
5
National
Use Category
SM
SM
RR, SM
RR, SM
TT
RR, SM
RR, SM
TT
RR, SM
TT
RR, SM
RR
SM
RR, SM
TT
TT
TT
RR, SM
SM
TT
TT
TT
Presumed Volume
1,500,000
2,250,000
750,000
750,000
750,000
6,000,000
1,025,000
500,000
775,000
1,000,000
1,775,000
2,775,000
750,000
750,000
750,000
750,000
500,000
137,500
3,637,500
275,000
275,000
550,000
3,000,000
3,550,000
16.237.500
Sulfur, ppm
1900
1200
1600
4000
4100
2000
3220
L
4100
2000
3220
1640
3360
2160
[
1200
4000
5000
3460
1640
4800
10000
[
1000
[
1000
2500
[
L
Sulfur * Volume
2,400,000,000
9,000,000,000
3,075,000,000
1,500,000,000
2,663
4,202,500,000
1,000,000,000
2,495,500,000
1,640,000,000
5,964,000,000
2,740
3,000,000,000
3,750,000,000
2,595,000,000
1,230,000,000
2,400,000,000
1,375,000,000
3,945
275,000,000
1,000
550,000,000
7,500,000,000
2,268
2.849
7-148
-------
Estimated Costs of Low-Sulfur Fuels
Table 7A-4
1998 Off-highway Diesel Sulfur Levels from TRW (Niper)
Sample
43
44
45
48
49
50
51
52
45
50
53
48
49
50
51
52
73
73
74
Region
Eastern
Eastern
Eastern
Southern
Southern
Southern
Southern
Southern
Central
Central
Central
Southern
Southern
Southern
Southern
Southern
Rocky Mtn
Western
Western
District
B
B2
C2,E2
Dl
Dl
Dl
D1,G2
D3
C2,E2
D1,G2
F3
Dl
Dl
Dl
D1,G2
D3
K3, L3, M3
K3, L3, M3
M2
PADD
1
1
1
1
1
1
1
1
2
2
2
3
3
3
3
3
4
5
5
National
Use Category
TT
SM
TT
RR, SM
TT
RR, SM
TT
TT
RR, SM
TT
RR, SM
TT
TT
TT
Presumed Volume
1,000,000
500,000
750,000
750,000
750,000
750,000
137,500
4,637,500
1,500,000
1,000,000
275,000
1,275,000
750,000
750,000
750,000
750,000
137,500
3,137,500
275,000
275,000
550,000
1,000,000
1,550,000
10.875.000
Sulfur, ppm
1600
1540
2600
4900
3870
1300
10000
4700
2600
1300
3700
4900
3870
1300
10000
4700
3400
3400
2900
Sulfur * Volume
1,540,000,000
1,300,000,000
3,675,000,000
2,902,500,000
975,000,000
7,500,000,000
646,250,000
| 3,998
3,900,000,000
1,300,000,000
1,017,500,000
| 1,818
3,675,000,000
2,902,500,000
975,000,000
7,500,000,000
646,250,000
| 5,004
935,000,000
| 3,400
1,870,000,000
2,900,000,000
| 3,077
1 3 886
7-149
-------
Draft Regulatory Impact Analysis
Table 7A-5
1999 Off-highway Diesel Sulfur Levels from TRW (Niper)
Sample
65
66
67
68
69
70
71
66
70
72
44
45
67
68
69
70
71
73
73
74
Region
Eastern
Eastern
Southern
Southern
Southern
Southern
Southern
Eastern
Central
Central
Southern
Southern
Southern
Southern
Southern
Southern
Southern
Rocky Mtn
Western
Western
District
B2
C3
Dl
Dl
Dl
D1,E1,G2
D3
C3
D1,E1,G2
F3
Dl
Dl
Dl
Dl
Dl
D1,E1,G2
D3
K3, L3, M3
K3, L3, M3
M2
PADD
1
1
1
1
1
1
1
2
2
2
3
3
3
3
3
3
3
4
5
5
National
Use Category
SM
SM,RR
TT
TT
TT
RR, SM
TT
TT
SM,RR
SM,RR
TT
TT
TT
RR, SM
TT
TT
TT
Presumed Volume
1,000,000
137,500
750,000
750,000
750,000
750,000
137,500
4,275,000
137,500
2,500,000
275,000
2,912,500
750,000
750,000
750,000
750,000
750,000
750,000
137,500
4,637,500
275,000
275,000
550,000
1,000,000
1,550,000
13.650.000
Sulfur, ppm
1560
4950
4290
4590
4900
1200
10000
L
4950
1200
4800
I
4900
4113
4290
4590
4900
1200
10000
L
2000
L
2000
2100
L
1
Sulfur * Volume
1,560,000,000
680,625,000
3,217,500,000
3,442,500,000
3,675,000,000
900,000,000
1,375,000,000
3,474
680,625,000
3,000,000,000
1,320,000,000
1,717
3,675,000,000
3,084,750,000
3,217,500,000
3,442,500,000
3,675,000,000
900,000,000
1,375,000,000
4,177
550,000,000
2,000
1,100,000,000
2,100,000,000
2,065
3.148
7-150
-------
Estimated Costs of Low-Sulfur Fuels
Table 7A-6
2000 Off-highway Diesel Sulfur Levels from TRW (Niper)
Sample
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
41
42
43
46
51
52
53
44
45
46
47
48
49
50
60
Region
Eastern
Eastern
Eastern
Eastern
Eastern
Eastern
Eastern
Southern
Southern
Southern
Southern
Southern
Southern
Southern
Eastern
Eastern
Central
Eastern
Central
Central
Central
Central
Southern
Southern
Southern
Southern
Southern
Southern
Southern
Rocky Mtn
District
B
B2
B2
B2
Cl
C2,E2
C3
Dl
Dl
D1,E1,G2
D2
D2
D2
D3
E1,C1
Cl
C2,E2
C3
D1,E1,G2
E1,C1
F3
G1,E1
Dl
Dl
D1,E1,G2
D2
D2
D2
D3
K3
PADD
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
3
3
3
3
3
3
3
4
1996 Off-highway Diesel Sulfur Levels from TRW
52
53
Western
Western
K3, L3, M3
Ml
5
5
National
Use Category
TT
TT
SM
TT,RR
TT
SM,RR
TT
TT
RR, SM
TT,RR
TT,RR
TT,RR
TT
TT,RR
TT,RR
TT
SM,RR
TT
TT
RR, SM
TT,RR
TT
(Niper)
TT
TT
Presumed Volume
1,000,000
1,000,000
1,000,000
750,000
500,000
137,500
750,000
750,000
750,000
500,000
500,000
500,000
137,500
750,000
9,025,000
750,000
1,500,000
137,500
2,500,000
2,250,000
275,000
3,000,000
10,412,500
750,000
750,000
750,000
500,000
500,000
500,000
137,500
3,887,500
275,000
275,000
412,500
1,500,000
1,912,500
25.512.500
Sulfur, ppm
1370
3000
2900
1280
3600
7200
4240
4900
4113
1150
3100
9800
4440
4800
2600
[
3600
2200
4240
1150
2600
4120
4720
r
4900
4113
1150
3100
9800
4440
4800
r
2600
1
4100
2700
r
i
Sulfur * Volume
3,000,000,000
2,900,000,000
1,280,000,000
2,700,000,000
3,600,000,000
583,000,000
3,675,000,000
3,084,750,000
862,500,000
1,550,000,000
4,900,000,000
2,220,000,000
660,000,000
1,950,000,000
3,653
2,700,000,000
3,300,000,000
583,000,000
2,875,000,000
5,850,000,000
1,133,000,000
14,160,000,000
2,939
3,675,000,000
3,084,750,000
862,500,000
1,550,000,000
4,900,000,000
2,220,000,000
660,000,000
4,361
715,000,000
2,600
1,691,250,000
4,050,000,000
3,002
3.409
7-151
-------
Draft Regulatory Impact Analysis
Table 7A-7
2001 Off-highway Diesel Sulfur Levels from TRW (Niper) 14 samples total
Sample
48
49
50
51
52
53
54
55
56
50
51
52
57
58
59
53
54
55
56
60
Region
Eastern
Eastern
Eastern
Eastern
Eastern
Southern
Southern
Southern
Southern
Eastern
Central
Central
Central
Central
Central
Southern
Southern
Southern
Southern
Rocky Mtn
District
B2
B3
Cl
C2,E2
C3
Dl
Dl
D3
D3
Cl
C2,E2
C3
E4
F3
G1,E1
Dl
Dl
D3
D3
K3
1996 Off-highway Diesel Sulfur
52
53
Western
Western
K3,L3,
M3
Ml
PADD Use Category
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
3
3
3
3
4
Levels from TRW
5
5
National
SM
TT
TT
SM,RR
TT
TT
TT
TT
TT,RR
SM,RR
TT
TT
(Niper)
TT
TT
Presumed Volume
1,000,000
275,000
750,000
1,000,000
137,500
750,000
750,000
137,500
137,500
4,937,500
750,000
1,000,000
137,500
50,000
275,000
3,000,000
5,212,500
750,000
750,000
137,500
137,500
1,775,000
275,000
275,000
412,500
1,500,000
1,912,500
14.112.500
Sulfur, ppm
1560
1600
3020
3000
1360
4330
4600
4980
1800
[
3020
3000
1360
3100
4150
4590
[
4330
4600
4980
1800
[
2340
L
4100
2700
[
1
Sulfur * Volume
1,560,000,000
440,000,000
2,265,000,000
3,000,000,000
187,000,000
3,247,500,000
3,450,000,000
684,750,000
247,500,000
3,055
2,265,000,000
3,000,000,000
187,000,000
155,000,000
1,141,250,000
13,770,000,000
3,936
3,247,500,000
3,450,000,000
684,750,000
247,500,000
4,298
643,500,000
2,340
1,691,250,000
4,050,000,000
3,002
3.516
7-152
-------
Estimated Costs of Low-Sulfur Fuels
Appendix 7B: Land-Based Nonroad Engine Growth Rate Based on Annual Energy Outlook 2002
Table 7.B-1
2000-2008 Composite Growth Factor for Land-Based Nonroad Engines
Based on Annual Energy Outlook 2002 (AEO2002)
End Use
Commercial
Industrial
Farm
Construction
Railroad
Military
Other No n-
Highway
Oil Company
On-Highway
Composite
Average
2000 Land-
Based Nonroad
Diesel Demand
(million
gallons)
488
1721
3080
1805
29
153
409
342
229
Fraction of
Total
0.059
0.208
0.373
0.219
0.004
0.019
0.050
0.041
0.028
2000-2008
Multiplicative
Growth Factor
1.105
1.063
1.039
1.138
1.083
1.000
1.074
1.074
1.238
Consumption
Weighted
Multiplicative
Growth Factor
0.065
0.222
0.388
0.249
0.004
0.019
0.053
0.044
0.034
1.078
% Simple
Annual
Growth Rate
—
—
—
—
—
—
—
0.97
Source of Energy Consumption
Diesel demand from Table 7.1-8;, Growth factor from AEO2002,
Table 2, Commercial, Distillate Fuel
Diesel demand from Table 7. 1-8; Growth factor from AEO2002,
Table 2, Industrial, Distillate Fuel
Diesel demand from Table 7. 1-8; Growth factor from AEO2002,
Table 32, Agriculture, Distillate Fuel
Diesel demand from Table 7. 1-8;, Growth factor from AEO2002,
Table 32, Construction, Distillate Fuel
Diesel demand from Table 7.1-8;, Growth factor from AEO2002,
Table 7, Energy Use by Mode, Railroad
Diesel demand from Table 7.1-8, Assumed no growth due to
base closings and no information suggesting long term increases
in training or emergency operations
Diesel demand from Table 7.1-8, Growth factor from Table 7.1-
15
Assumed same as Other Non-Highway
Diesel demand from Table 7.1-8; Growth factor from AEO2002,
Table 7, Energy Use by Mode, Freight Trucks (over 10,000 Ibs.
GVWR.
7-153
-------
Draft Regulatory Impact Analysis
Chapter 7 References
1. Energy Information Administration. Fuel Oil and Kerosene Sales 2000. DOE/EIA-0535(00).
Office of Oil and Gas, U.S. Department of Energy. Washington, D. C. September, 2001.
2. Energy Information Administration. Petroleum Supply Annual 2000, Volume 1. DOE/EIA-
0340(00)71. Office of Oil and Gas, U.S. Department of Energy. Washington, D. C. June, 2001b.
3. Office of Highway Policy Information, Highway Statistics 2000. FHWA-PL-01-1011. Federal
Highway Administration, U.S. Department of Transportation. Washington, D.C. November,
2001.
4. Energy Information Administration, Annual Energy Outlook 2002. DOE/EIA-0383(2002).
U.S. Department of Energy. Washington, D.C. December, 2001.
5. U.S. Environmental Protection Agency, Final Regulatory Impact Analysis, Control of
Emission for Marine Diesel Engines. Office of Transportation and Air Quality, EPA 420-R-99-
026. November 1999.
6. National Institute for Petroleum and Energy Research, Diesel Fuel Oils, (years 1996-1998).
NIPER-202 PPS (96-98)75. BDM Petroleum Technologies. Bartlesville, OK. 1997-1999.
7. TRW Petroleum Technologies, Diesel Fuel Oils, (years 1999-2001). TRW-217 PPS (1999-
2001)75. Bartlesville, OK, 1999-2001.
8. U.S. Environmental Protection Agency, Regulatory Impact Analysis, Heavy-Duty Engine and
Vehicle Standards and Highway Diesel Fuel Sulfur Control Requirements. EPA420-R-00-026.
Office of Air and Radiation. Washington, D.C. December, 2000.
9.Confidential Information Submission from Diesel Desulfurization Vendor A, August 1999.
10.UOP Information Submission to the National Petroleum Council, August 1999.
1 l."The Lower it Goes, The Tougher it Gets," Bjorklund, Bradford L., UOP, Presentation at the
National Petroleum Council Annual Meeting, March 2000.
12.U.S. Petroleum Refining Draft Report, Appendix H, National Petroleum Council, March 30,
2000.
13. Chemical Engineering Plant Cost Index, Chemical Engineering, February 2003.
14.Conversation with Jim Kennedy, Manager Project Sales, Distillate and Resid Technologies,
UOP, November 2000.
7-154
-------
Estimated Costs of Low-Sulfur Fuels
15. Conversation with Tim Heckel, Manager of Distillate Technologies Sales, UOP, March 2000.
16.Conversation with Tom W. Tippett et al, Refining Technology Division, Haldor Topsoe,
March 2000.
17. Very-Low-Sulfur Diesel Distribution Cost, Engine Manufacturers Association, August 1999.
IS.Moncrief, T. I, 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.
19.Christie, David A., Advanced Controls Improve Reformulated Fuel Yield and Quality, Fuel
Reformulation, July/August 1993.
20.Personal conversation with Debbie Pack, ABB Process Analytics Inc., November 1998.
21. SZorb Diesel, Process Overview and Operations and Economics, www.fuelstechnology.com.
Phillips Petroleum Company, 2001.
22. Slater, Peter N., et al, Phillips SZorb Diesel Sulfur Removal Technology, Technical Paper
AM-02-46 Presented at the 2002 National Petrochemical and Refiners Association Annual
Meeting, March 2002.
23.Ackerson, Michael; Skeds, Jon, Presentation to the Clean Diesel Independent Review Panel,
Process Dynamics and Linde Process Plants, July 30, 2002.
24. Conversation with Jon Skeds of Linde Process Plants at the 2002 National Petrochemical and
Refiners Association Question and Answer Meeting, October, 2002.
25.American Automobile Manufacturers Association Diesel Fuel Survey, Summer 1997.
26.Conversation with Cal Hodge, A Second Opinion, February 2000.
27.Worldwide Refining/Worldwide Production, Oil and Gas Journal, vol. 97, No. 51, December
20, 1999.
28.Final Report: 1996 American Petroleum Institute/National Petroleum Refiners Association
Survey of Refining Operations and Product Quality, July 1997.
29.Final Report: 1996 American Petroleum Institute/National Petroleum Refiners Association
Survey of Refining Operations and Product Quality, July 1997.
30.Gary, James H., Handwerk, Glenn E., Petroleum Refining: Technology and Economics,
Marcel Dekker, New York (1994).
7-155
-------
Draft Regulatory Impact Analysis
31.Conversation with Lyondel-Citgo refinery staff, April 2000.
32.Gary, James H., Handwerk, Glenn E., Petroleum Refining: Technology and Economics,
Marcel Dekker, New York (1994).
33.Peters, Max S., Timmerhaus, Klaus D., Plant Design and Economics for Chemical Engineers,
Third Edition, McGraw Hill Book Company, 1980.
34.Waguespack, Kevin, Review of DOE/Ensys Reformulated Diesel Study-Draft for Discussion,
Price-Waterhouse Coopers for the American Automobile Manufacturers, October 5, 2000.
35.U.S. Petroleum Refining, Meeting Requirements for Cleaner Fuels and Refineries, Volume V
- Refining Capability Appendix, National Petroleum Council, 1993.
36.Waguespack, Kevin, Review of DOE/Ensys Reformulated Diesel Study-Draft for Discussion,
Price-Waterhouse Coopers for the American Automobile Manufacturers, October 5, 2000.
37.Hadder, Gerry and Tallet, Martin; Documentation for the Oak Ridge National Laboratory
Refinery Yield Refinery Model (ORNL-RYM), 2001.
38.Petroleum Marketing Annual, Energy Information Administration, 1999.
39.Perry, Robert H., Chilton, Cecil H., Chemical Engineer's Handbook, McGraw Hill 1973.
40. 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).
41.1999 Worldwide Refining Survey, Oil and Gas Journal, December 20, 1999.
42.Peters, Max S., Timmerhaus, Klaus D., Plant Design and Economics for Chemical Engineers,
Third Edition, McGraw Hill Book Company, 1980.
43.Jena, Rabi, Take the PC-Based Approach to Process Control, Fuel Reformulation,
November/December 1995.
44.Sutton, IS., Integrated Management Systems Improve Plant Reliability, Hydrocarbon
Processing, January 1995.
45.King, M. J., Evans, H. N., Assessing your Competitors' Application of CIM/CIP,
Hydrocarbon Processing, July 1993.
46.U. S. Petroleum Refining, Assuring the Adequacy and Affordability of Cleaner Fuels, A
Report by the National Petroleum Council, June 2000.
7-156
-------
Estimated Costs of Low-Sulfur Fuels
47. "Refining Economics of Diesel Fuel Sulfur Standards," study performed for the Engine
Manufacturers Association by MathPro, Inc. October 5, 1999.
48. "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.
49. Phone conversation in mid-2002 with Massey's Truck and Tank Repair, Pheonix Arizona.
50. Regulatory Impact Analysis (RIA) for the Highway Diesel Fuel Final Rule, EPA Air Docket
A-99-06.
51. Regulatory Impact Analysis (RIA) for the Highway Diesel Fuel Final Rule, EPA Air Docket
A-99-06.
52. Regulatory Impact Analysis (RIA) for the Highway Diesel Fuel Final Rule, section V.C.4,
EPA Air Docket A-99-06.
7-157
-------
CHAPTER 8: Estimated Aggregate Cost and Cost per Ton of Reduced Emissions
8.1 Projected Sales and Cost Allocations 8-1
8.2 Aggregate Engine Costs 8-4
8.2.1 Aggregate Engine Fixed Costs 8-4
8.2.2 Aggregate Engine Variable Costs 8-6
8.3 Aggregate Equipment Costs 8-9
8.3.1 Aggregate Equipment Fixed Costs 8-9
8.3.2 Aggregate Equipment Variable Costs 8-11
8.4 Aggregate Fuel Costs and Other Operating Costs 8-13
8.4.1 Aggregate Fuel Costs 8-14
8.4.2 Aggregate Oil Change Maintenance Savings 8-16
8.4.3 Aggregate CDPF & CCV Maintenance Costs and CDPF Regeneration Costs 8-19
8.4.4 Summary of Aggregate Operating Costs 8-21
8.5 Summary of Total Aggregate Costs of the Proposed Program 8-23
8.6 Emission Reductions 8-26
8.7 Cost per Ton 8-28
8.7.1 Cost per Ton for the 500 ppm Fuel Program 8-28
8.7.2 Cost per Ton for the Proposed Program 8-30
-------
Estimated Aggregate Cost and Cost per Ton of Reduced Emissions
CHAPTER 8: Estimated Aggregate Cost and Cost per Ton
of Reduced Emissions
Our estimate of fixed and variable costs for new nonroad engines and equipment compliant
with today's proposal are detailed in Chapter 6. Chapter 6 also contains a discussion of the
operating savings and costs expected to result from the new low sulfur diesel fuels available to
nonroad engines, locomotive engines, and marine engines. Our estimates of the costs to meet the
proposed nonroad diesel fuel requirements can be found in Chapter 7. This chapter contains
information on how the incremental costs for engines, the incremental operating costs, the
incremental equipment costs, and the incremental fuel costs are aggregated to estimate the cost of
the proposed program. The detailed information on the calculation for the cost per ton per
pollutant is also included in this Chapter.
We have calculated the cost per ton of our proposed program 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 of the costs and emissions reductions
from our proposed program including those costs incurred and emissions reductions generated by
the existing fleet. The baseline (i.e., 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 chosen is meant to capture both the early period of the program when very few
new engines that meet the proposed standards would be in the fleet, and the later period when
essentially all engines would meet the proposed standards. Note that all costs and emission
reductions presented here are 30 year numbers (net present values are 2007 through 2036,
expressed in 2004). Elsewhere in this draft RIA we use numbers through 2030 (i.e., 2007
through 2030, expressed in 2004). As a result, net present values presented here will differ from
net present values in other sections of this draft RIA.
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.1'2 The PSR database has been used as the basis for
our current fleet mix - i.e., what equipment types were sold in 2000 and with what horsepower
engines. The sales estimates and growth rates used throughout this analysis are shown in Table
8.1-1.3
8-1
-------
Draft Regulatory Impact Analysis
Table 8.1-1
Estimated 2000 Engine Sales and Future Sales Growth
Horsepower 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
3.7%
2.8%
2.3%
2.3%
2.2%
2.4%
1.3%
1.9%
1.9%
2.5%
Because our proposed program would result in reductions in a number of important
pollutants (i.e., NOx, PM, NMHC, and SOx), we have attempted to identify what costs are
associated with what pollutant reductions. This apportionment of costs by pollutant allows us to
calculate the average cost per ton of emission reduction that would be realized by this program.
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. Similary, we have apportioned 100 percent
of the 500 ppm fuel program to SOx control even though it would also provide a significant PM
reduction. In our proposed fuel program, the cost for reducing sulfur from uncontrolled levels to
15 ppm are apportioned to NOx+NMHC and PM because the 15 ppm sulfur level has been
selected based on the needs of the control technology (i.e., NOx adsorber and CDPF). We have
done this even though a significant SOx reduction would also be realized by the new 15 ppm
fuel. We have noted throughout our discussion to which pollutant we have attributed costs, and
Table 8.1-2 presents a summary of these allocations.
8-2
-------
Estimated Aggregate Cost and Cost per Ton of Reduced Emissions
Table 8.1-2.
Summary of How Cost are Allocated Among Pollutants
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 - Redesign
Equipment Fixed Costs - Operator Manuals
3000 to 500 ppm fuel
3 000 to 15 ppm fuel
500 ppm fuel
1 5 ppm fuel
1 5 ppm fuel
1 5 ppm fuel
2008+
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
>75hp
<75 hp 2008
50-75 hp 20 13
25-50 hp 20 13
>75hp
<75hp
2008 standards
2011,2012, 20 13 standards
2008 standards
2011,2012,2013 standards
NOx+
NMHC
50%
50%
50%
100%
50%
100%
50%
67%
50%
100%
50%
50%
50%
50%
50%
PM
50%
50%
100%
100%
50%
100%
100%
50%
100%
50%
33%
100%
100%
50%
100%
100%
50%
100%
100%
50%
50%
100%
100%
50%
100%
50%
SOx
100%
100%
8-3
-------
Draft Regulatory Impact Analysis
8.2 Aggregate Engine Costs
This section presents aggregate engine fixed costs (recovered costs) and variable costs.
These costs were discussed in detail in Section 6.2.
8.2.1 Aggregate Engine Fixed Costs
In Chapter 6, Tables 6.2-4, 6.2-6, and 6.2-8 presented the aggregate engine fixed costs,A
along with our best estimate of how those costs might be recovered (i.e., on what engines), for
engine R&D, tooling, and certification, respectively. Table 8.2-1 presents the combined total of
all engine fixed costs in the each of the indicated years and also shows to what pollutant these
costs are attributed. 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 (e.g., costs associated with the proposed 2008 standards and costs
associated with the proposed 2013 standards for 50 to 75 horsepower engines). Therefore, the
costs presented in Table 8.2-1 for PM would 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 6.2 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.
A
We have estimated a "recovered" cost for all engine and equipment fixed costs to present a per unit analysis of the
cost of the proposal. In general, in environmental economics, it would be more conventional to simply count the total
cost of the program (i.e., opportunity costs) in the year they occur. However, this approach would not directly estimate a
per unit cost since fixed costs occur prior to implementation of the standards and, therefore, there are not yet any units
certified as complying with the new standards to which the fixed costs can be attributed. 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 rate of return to the manufacturer.
8-4
-------
Estimated Aggregate Cost and Cost per Ton of Reduced Emissions
Table 8.2-1
Aggregate Engine Fixed Costs (millions)
Year
2004
2005
2006
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
Total
NPV 2004-2036
Recovery of PM Costs
$0.0
$0.0
$0.0
$0.0
$17.4
$17.4
$17.4
$31.9
$37.1
$36.9
$37.1
$37.1
$22.6
$17.4
$0.2
$0.0
$0.0
$272.3
$210.7
Recovery of NOx+NMHC
Costs
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$14.7
$19.9
$21.2
$35.2
$35.2
$20.5
$15.3
$14.0
$0.0
$0.0
$175.9
$129.5
Recovery of Fixed Costs
$0.0
$0.0
$0.0
$0.0
$17.4
$17.4
$17.4
$46.6
$57.0
$58.0
$72.3
$72.3
$43.1
$32.7
$14.2
$0.0
$0.0
$448.3
$340.2
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. Where a phase-in exists
(e.g., for NOx standards on >75 horsepower engines), expenditures are assumed to occur over the
five year span preceding the first year NOx adsorbers would be introduced, and then to continue
during the phase-in years; the expenditures would be incurred in a manner 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 assume a seven
percent rate of return for all R&D to reflect the time value of money.
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 following implementation of the new
standard; all tooling costs include a seven percent rate of return to reflect the time value of
money.
8-5
-------
Draft Regulatory Impact Analysis
We have calculated the net present value of the engine fixed costs over the 30 year period
following implementation of the program as $340 million. This value assumes a three percent
social discount rate and a 2004 date for promulgation of the final standards.
8.2.2 Aggregate Engine Variable Costs
Engine variable costs are discussed in detail in Section 6.2.2 of this Draft RIA. 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 variable costs during the years 2008 through 2036 as shown in Tables 8.2-2 and 8.2-3
(refer to Table 8.1-2 for how costs have been allocated to PM and NOx). Because of their nature,
variable costs vary with sales. We have calculated the net present value of the variable costs over
the 30 year period following implementation of the program (2007 through 2036) as $13.9
billion. This value assumes a three percent social discount rate.
8-6
-------
Table 8.2-2
Aggregate Engine Variable Costs by Horsepower Category (millions)
Year
0750hp
Total
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
$0.0
$0.0
$0.0
$0.0
$0.0
$20.5
$20.0
$20.5
$21.0
$21.5
$22.0
$22.5
$23.0
$23.5
$24.1
$24.6
$25.1
$25.6
$26.1
$26.6
$27.1
$27.6
$28.2
$28.7
$29.2
$29.7
$30.2
$30.7
$31.2
$31.7
$32.3
$32.8
$33.3
$0.0
$0.0
$0.0
$0.0
$24.0
$24.6
$23.7
$24.2
$24.7
$152.2
$155.2
$84.9
$122.4
$124.6
$126.9
$129.2
$131.4
$133.7
$136.0
$138.2
$140.5
$142.8
$145.0
$147.3
$149.6
$151.8
$154.1
$156.4
$158.7
$160.9
$163.2
$165.5
$167.7
$0.0
$0.0
$0.0
$0.0
$18.8
$19.2
$18.4
$18.7
$19.1
$101.8
$103.5
$80.0
$81.3
$82.6
$83.9
$85.2
$86.5
$87.9
$89.2
$90.5
$91.8
$93.1
$94.4
$95.7
$97.0
$98.4
$99.7
$101.0
$102.3
$103.6
$104.9
$106.2
$107.6
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$99.7
$101.4
$102.1
$103.9
$105.6
$107.3
$109.0
$110.7
$112.4
$114.1
$115.8
$117.5
$119.2
$120.9
$122.6
$124.4
$126.1
$127.8
$129.5
$131.2
$132.9
$134.6
$136.3
$138.0
$139.7
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$197.6
$200.9
$200.4
$203.6
$206.8
$210.0
$213.3
$216.5
$219.7
$222.9
$226.1
$229.3
$232.5
$235.7
$239.0
$242.2
$245.4
$248.6
$251.8
$255.0
$258.2
$261.4
$264.6
$267.9
$271.1
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$157.1
$160.0
$126.2
$161.9
$164.7
$167.5
$170.3
$173.2
$176.0
$178.8
$181.6
$184.4
$187.2
$190.0
$192.8
$195.6
$198.4
$201.2
$204.0
$206.8
$209.6
$212.4
$215.2
$218.0
$220.8
$223.6
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$106.5
$107.7
$84.3
$107.3
$108.4
$109.6
$110.8
$111.9
$113.1
$114.3
$115.4
$116.6
$117.8
$118.9
$120.1
$121.2
$122.4
$123.6
$124.7
$125.9
$127.1
$128.2
$129.4
$130.6
$131.7
$132.9
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$17.1
$17.4
$13.6
$17.4
$17.6
$17.9
$18.1
$18.4
$18.6
$18.9
$19.1
$19.4
$19.6
$19.9
$20.1
$20.4
$20.6
$20.9
$21.1
$21.4
$21.7
$21.9
$22.2
$22.4
$22.7
$22.9
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$18.8
$19.1
$15.3
$31.0
$31.4
$31.9
$32.3
$32.8
$33.2
$33.7
$34.1
$34.6
$35.0
$35.5
$35.9
$36.4
$36.8
$37.3
$37.7
$38.2
$38.6
$39.1
$39.5
$40.0
$40.4
$40.9
$0.0
$0.0
$0.0
$0.0
$62.8
$64.2
$62.1
$362.9
$666.2
$817.2
$900.8
$817.1
$866.0
$879.6
$893.3
$907.0
$920.7
$934.4
$948.1
$961.7
$975.4
$989.1
$1,002.8
$1,016.5
$1,030.2
$1,043.9
$1,057.5
$1,071.2
$1,084.9
$1,098.6
$1,112.3
$1,126.0
$1,139.6
|NPV 2004-2036 $424.2 $1,983.0 $1,327.5 $1,652.8 $3,228.6 $2,734.7
$1,730.6
$287.8
$487.5
$13,874.3
-------
Table 8.2-3
Aggregate Engine Variable Costs by Technology and by Pollutant (millions)
Year
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
Fuel System
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$55.0
$56.0
$43.0
$43.7
$44.5
$45.3
$46.0
$46.8
$47.6
$48.3
$49.1
$49.9
$50.6
$51.4
$52.2
$53.0
$53.7
$54.5
$55.3
$56.0
$56.8
$57.6
$58.3
$59.1
Cooled EGR
$0
$0
$0
$0
$0.0
$0.0
$0.0
$1.2
$1.2
$18.3
$19.6
$15.4
$15.7
$16.0
$16.2
$16.5
$16.8
$17.1
$17.4
$17.6
$17.9
$18.2
$18.5
$18.8
$19.1
$19.3
$19.6
$19.9
$20.2
$20.5
$20.8
$21.0
$21.3
CCV
$0
$0
$0
$0
$0.6
$0.6
$0.5
$7.2
$13.8
$12.1
$10.7
$10.9
$11.0
$11.2
$11.4
$11.5
$11.7
$11.9
$12.0
$12.2
$12.4
$12.5
$12.7
$12.9
$13.0
$13.2
$13.4
$13.5
$13.7
$13.9
$14.0
$14.2
$14.4
DOC
$0
$0
$0
$0
$62.2
$63.6
$61.7
$63.0
$64.3
$21.5
$22.0
$22.5
$23.0
$23.5
$24.1
$24.6
$25.1
$25.6
$26.1
$26.6
$27.1
$27.6
$28.2
$28.7
$29.2
$29.7
$30.2
$30.7
$31.2
$31.7
$32.3
$32.8
$33.3
CDPF System
$0
$0
$0
$0
$0.0
$0.0
$0.0
$186.0
$361.2
$435.4
$408.3
$387.0
$393.1
$399.2
$405.3
$411.4
$417.5
$423.6
$429.7
$435.8
$441.9
$448.0
$454.0
$460.1
$466.2
$472.3
$478.4
$484.5
$490.6
$496.7
$502.8
$508.9
$515.0
CDPF Regen
System
$0
$0
$0
$0
$0.0
$0.0
$0.0
$28.9
$72.6
$134.4
$126.7
$76.8
$113.9
$115.8
$117.7
$119.5
$121.4
$123.3
$125.1
$127.0
$128.9
$130.8
$132.6
$134.5
$136.4
$138.3
$140.1
$142.0
$143.9
$145.7
$147.6
$149.5
$151.4
NOx Adsorber
System
$0
$0
$0
$0
$0.0
$0.0
$0.0
$76.6
$153.0
$140.5
$257.5
$261.5
$265.5
$269.5
$273.4
$277.4
$281.4
$285.4
$289.4
$293.4
$297.4
$301.4
$305.3
$309.3
$313.3
$317.3
$321.3
$325.3
$329.3
$333.3
$337.2
$341.2
$345.2
Total PM Costs
$0
$0
$0
$0
$62.5
$63.9
$61.9
$282.1
$505.7
$634.0
$600.2
$521.0
$565.3
$574.4
$583.4
$592.5
$601.6
$610.7
$619.8
$628.9
$638.0
$647.1
$656.1
$665.2
$674.3
$683.4
$692.5
$701.6
$710.7
$719.8
$728.9
$737.9
$747.0
lotal
NOx+NMHC
Costs
$0
$0
$0
$0
$0.3
$0.3
$0.2
$80.8
$160.5
$183.2
$300.6
$296.1
$300.7
$305.3
$309.9
$314.5
$319.1
$323.7
$328.3
$332.9
$337.5
$342.1
$346.7
$351.3
$355.8
$360.4
$365.0
$369.6
$374.2
$378.8
$383.4
$388.0
$392.6
Total Costs
$0
$0
$0
$0
$62.8
$64.2
$62.1
$362.9
$666.2
$817.2
$900.8
$817.1
$866.0
$879.6
$893.3
$907.0
$920.7
$934.4
$948.1
$961.7
$975.4
$989.1
$1,002.8
$1,016.5
$1,030.2
$1,043.9
$1,057.5
$1,071.2
$1,084.9
$1,098.6
$1,112.3
$1,126.0
$1,139.6
| NPV 2004-2036 $678.2 $244.2
$182.8
$620.3
$6,336.7
$1,788.2 $4,023.8 $9,297.9 $4,576.4 $13,874.3 |
-------
Estimated Aggregate Cost and Cost per Ton of Reduced Emissions
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 also shows to what
pollutant these costs are attributed. Note that the cost allocations shown in Table 8.3-1 are not
generated assuming any simple split of costs between NOx and PM control. Some equipment
fixed costs are solely attributed to PM control (e.g., costs associated with the proposed 2008
standards and costs associated with the proposed 2013 standards for 50 to 75 horsepower
engines). Therefore, the costs presented in Table 8.3-1 for PM would 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.
8-9
-------
Draft Regulatory Impact Analysis
Table 8.3-1
Aggregate Equipment Fixed Costs (millions)
Year
2004
2005
2006
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
Total
NPV 2004-2036
Recovery of PM
Costs
$0.0
$0.0
$0.0
$0.0
$4.9
$4.9
$4.9
$26.1
$39.8
$49.1
$58.5
$58.5
$58.5
$58.5
$53.6
$53.6
$53.6
$32.4
$18.7
$9.4
$0.0
$584.9
$408.7
Recovery of
NOx+NMHC Costs
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$21.2
$34.9
$34.9
$44.3
$44.3
$44.3
$44.3
$44.3
$44.3
$44.3
$23.1
$9.4
$9.4
$0.0
$443.1
$308.0
Recovery of Fixed
Costs
$0.0
$0.0
$0.0
$0.0
$4.9
$4.9
$4.9
$47.3
$74.7
$84.1
$102.8
$102.8
$102.8
$102.8
$98.0
$98.0
$98.0
$55.5
$28.1
$18.7
$0.0
$1,028.0
$716.7
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 (e.g., for NOx standards on >75 horsepower engines),
expenditures are assumed to occur over the two year span preceding the first year NOx adsorbers
would be introduced, and then to continue during the phase-in years; the expenditures would be
incurred in a manner 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 have assumed a seven percent rate of return for all equipment fixed costs to
reflect the time value of money.
8-10
-------
Estimated Aggregate Cost and Cost per Ton of Reduced Emissions
We have calculated the net present value of the equipment fixed costs over the 30 year period
following implementation of the program as $720 million. This value assumes a three percent
social discount rate and a 2004 date for promulgation of the final standards.
8.3.2 Aggregate Equipment Variable Costs
The equipment variable costs including sheet metal costs, mounting hardware, labor, etc.
were estimated by horsepower category in Section 6.3. The variable costs were aggregated by
multiplying the cost per unit within each horsepower category by the projected sales of that
horsepower category. The aggregate equipment variable costs through 2036 are presented in
Table 8.3-2. Table 8.3-3 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).
8-11
-------
Table 8.3-2
Aggregate Equipment Variable Costs by Horsepower Category (millions)
Year
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
7.036
NPV 2004-2036
0750hp
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.4
$0.4
$0.4
$0.8
$0.8
$0.8
$0.8
$0.8
$0.8
$0.8
$0.9
$0.9
$0.9
$0.9
$0.9
$0.9
$0.9
$0.9
$0.9
$1.0
$1.0
$1.0
$1.0
$1.0
$1.0
SI 0
$12.1
Total
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$11.5
$24.2
$29.9
$30.8
$31.2
$31.7
$32.2
$32.7
$33.2
$33.7
$34.1
$34.6
$35.1
$35.6
$36.1
$36.6
$37.1
$37.5
$38.0
$38.5
$39.0
$39.5
$40.0
$40.5
$40.9
-------
Estimated Aggregate Cost and Cost per Ton of Reduced Emissions
Table j
Aggregate Equipment Variable
.3-3
Costs by Pollutant (millions)
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
NPV 2004-2036
PM Costs
$0.0
$0.0
$0.0
$5.7
$12.1
$17.6
$18.1
$18.4
$18.7
$19.0
$19.3
$19.5
$19.8
$20.1
$20.4
$20.7
$21.0
$21.3
$21.6
$21.9
$22.2
$22.5
$22.8
$23.1
$23.3
$23.6
$23.9
$24.2
$24.5
$291.4
NOx+NMHC Costs
$0.0
$0.0
$0.0
$5.7
$12.1
$12.3
$12.7
$12.9
$13.1
$13.2
$13.4
$13.6
$13.8
$14.0
$14.2
$14.4
$14.6
$14.8
$15.0
$15.2
$15.4
$15.6
$15.8
$16.0
$16.1
$16.3
$16.5
$16.7
$16.9
$206.9
Total
$0.0
$0.0
$0.0
$11.5
$24.2
$29.9
$30.8
$31.2
$31.7
$32.2
$32.7
$33.2
$33.7
$34.1
$34.6
$35.1
$35.6
$36.1
$36.6
$37.1
$37.5
$38.0
$38.5
$39.0
$39.5
$40.0
$40.5
$40.9
$41.4
$498.3
8.4 Aggregate Fuel Costs and Other Operating Costs
Aggregate costs presented here are used in the calculation of costs per ton of emissions that
would be reduced by the proposed standards for nonroad fuel and engines. We are proposing a
500 ppm sulfur cap on nonroad, locomotive, and marine fuels beginning in 2007. In Section
8.4.2 we summarize the costs for this program as if it remained in place for 30 years, even though
it would be supplanted by the second step of our fuel program in 2010.
8-13
-------
Draft Regulatory Impact Analysis
We are also proposing a second step in the fuel program that would cap nonroad fuel sulfur
levels at 15 ppm beginning in 2010. This fuel program enables the introduction of advanced
emission control technologies including CDPFs and NOx adsorbers. The combination of the
two-step fuel program and the new diesel engine standards represents the total Tier 4 program for
nonroad diesel engines and fuel proposed today. In Section 8.4.1 we present our estimate of the
annual and total costs for this complete program beginning in 2007 and continuing for 30 years.
8.4.1 Aggregate Fuel Costs
Fuel costs are developed on a cents per gallon basis. Chapter 7 contains a description of the
development of fuel costs for the proposed fuel program. Table 8.4-1 contains a summary of
cent/gallon fuel costs, estimated fuel volumes for nonroad, locomotive, and marine, and the
aggregate fuel costs through 2036 for the proposed two-step fuel program. Table 8.4-2 shows the
same information assuming the proposed 500 ppm fuel program remained in place indefinitely
and no new engine standards were implemented.
8-14
-------
Table 8.4-1
Aggregate Fuel Costs of the Proposed Two-Step Fuel Program (millions)
Year
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 2004-2036
Aggregate Fuel Costs of Proposed Two-Step Fuel Program
Affected Volume for 500 ppm (gallons)*
2007-2010
at $/gal cost of
$0.025
0
0
0
5,449
9,504
9,671
4,099
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
25,206
2010-2014
at $/gal cost of
$0.026
0
0
0
0
0
0
2,892
5,034
5,088
5,137
2,162
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
16,077
2014+
at $/gal cost of
$0.024
0
0
0
0
0
0
0
0
0
0
2,369
4,093
4,136
4,170
4,203
4,238
4,268
4,311
4,354
4,397
4,441
4,485
4,530
4,575
4,621
4,668
4,715
4,763
4,811
4,859
4,908
4,958
5,008
54,791
Affected Volume for 15 ppm (gallons)*
2010-2014
at $/gal cost of
$0.048
0
0
0
0
0
0
3,495
6,124
6,256
6,389
2,717
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
19,763
2014+
at $/gal cost of
$0.048
0
0
0
0
0
0
0
0
0
0
4,461
7,803
7,957
8,111
8,265
8,419
8,573
8,727
8,881
9,035
9,190
9,344
9,497
9,650
9,803
9,956
10,110
10,263
10,416
10,569
10,723
10,876
1 1 ,029
112,851
Aggregate Fuel Costs
500 ppm
$0
$0
$0
$136
$238
$242
$178
$131
$132
$134
$113
$98
$99
$100
$101
$102
$102
$103
$104
$106
$107
$108
$109
$110
$111
$112
$113
$114
$115
$117
$118
$119
$120
$2,363
15 ppm
$0
$0
$0
$0
$0
$0
$168
$294
$300
$307
$345
$375
$382
$389
$397
$404
$412
$419
$426
$434
$441
$448
$456
$463
$471
$478
$485
$493
$500
$507
$515
$522
$529
$6,366
Total
$0
$0
$0
$136
$238
$242
$345
$425
$433
$440
$458
$473
$481
$489
$498
$506
$514
$522
$531
$539
$548
$556
$565
$573
$581
$590
$598
$607
$615
$624
$632
$641
$650
$8,729
Note that "Affected Volumes" for 500 ppm and 15 ppm are taken from Table 7.1-35.
-------
Draft Regulatory Impact Analysis
Table 8.4-2
Aggregate Fuel Costs for the 500 ppm Fuel Program (millions)
Year
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 2004-2036
Aggregate Fuel Costs of 500 ppm Fuel Program
Affected Volume for 500 ppm (gallons)*
2007-2010
at $/gal cost of
$0.025
0
0
0
5,449
9,504
9,671
4,099
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
25,206
2010+
at $/gal cost of
$0.024
0
0
0
0
0
0
6,387
11,158
11,344
11,526
11,709
11,895
12,092
12,281
12,468
12,657
12,841
13,038
13,235
13,432
13,630
13,829
14,027
14,226
14,425
14,624
14,825
15,025
15,227
15,429
15,631
15,834
16,037
203,483
Aggregate Fuel Costs
Total
$0
$0
$0
$136
$238
$242
$256
$268
$272
$277
$281
$285
$290
$295
$299
$304
$308
$313
$318
$322
$327
$332
$337
$341
$346
$351
$356
$361
$365
$370
$375
$380
$385
$5,514
* Note that the 2010+ gallons shown here are the summation of the 2010-2014 and 2014+ columns
for both 500 ppm and 15 ppm fuel shown in Table 8.4-1 because there would be no introduction of 15
ppm fuel under the 500 ppm fuel program.
8.4.2 Aggregate Oil Change Maintenance Savings
Maintenance savings associated with extended oil change intervals are developed on a cents
per gallon basis. Section 6.2.3 contains a description of the development of maintenance savings
for the proposed program. Table 8.4-3 contains a summary of the maintenance savings and
8-16
-------
Estimated Aggregate Cost and Cost per Ton of Reduced Emissions
estimated fuel volumes for nonroad, locomotive, and marine through 2036 for the proposed two-
step fuel program. Also presented in Table 8.4-3 are the maintenance savings of the 500 ppm
fuel program assuming it remained in place indefinitely and no new engine standards were
implemented. Note that the nonroad volumes shown in Table 8.4-3 under the 500 ppm fuel
program are the summation of the two columns of nonroad volumes shown under the proposed
fuel program (500 ppm during 2007 through part of 2010 and 15 ppm for the remainder of 2010
and beyond). The cent per gallon savings for locomotive and marine shown in Table 8.4-3 are
taken from Table 6.2-26 (see also Table 6.4-3) while the cent per gallon savings shown for
nonroad are taken from the discussion in section 6.2.3.1 (these values can also be derived using
data shown in Table 6.2-26 by weighting the cent per gallon savings within each horsepower
range by the fuel use weighting for that horsepower range - the summation of these weighted
values for all horsepower ranges is 3.3 cents per gallon for 15 ppm fuel and 3.0 cents per gallon
for 500 ppm fuel).
8-17
-------
Table 8.4-3
Oil Change Maintenance Savings Associated with the Proposed Two-Step and the 500 ppm Fuel Programs (millions)
Year
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 2004-2036
Aggregate Maintenance Savings of Proposed Fuel Program
500 ppm Volumes
Nonroad Volume*
at $/gal savings of
$0.030
0
0
0
3,298
5,788
5,923
2,524
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
15,383
Loco&Marine Volume*
at $/gal savings of
$0.011
0
0
0
1,780
3,073
3,100
3,127
3,416
3,441
3,462
3,484
3,484
3,510
3,548
3,576
3,604
3,634
3,658
3,695
3,732
3,770
3,808
3,846
3,885
3,924
3,964
4,005
4,046
4,087
4,129
4,171
4,214
4,257
65,449
15 ppm Volume
Nonroad Volume*
at $/gal savings of
$0.033
0
0
0
0
0
0
3,533
6,025
6,156
6,288
6,419
7,545
7,700
7,853
8,006
8,159
8,312
8,465
8,619
8,772
8,925
9,078
9,231
9,383
9,535
9,687
9,839
9,992
10,144
10,296
10,448
10,600
10,752
128,702
Aggregate Maintenance Savings
500 ppm
$0
$0
$0
-$118
-$206
-$210
-$110
-$38
-$38
-$39
-$39
-$39
-$39
-$40
-$40
-$40
-$40
-$41
-$41
-$42
-$42
-$42
-$43
-$43
-$44
-$44
-$45
-$45
-$46
-$46
-$46
-$47
-$47
-$1,186
15 ppm
$0
$0
$0
$0
$0
$0
-$115
-$196
-$200
-$205
-$209
-$246
-$251
-$256
-$261
-$266
-$270
-$275
-$280
-$285
-$290
-$295
-$300
-$305
-$310
-$315
-$320
-$325
-$330
-$335
-$340
-$345
-$350
-$4,188
Total
$0
$0
$0
-$118
-$206
-$210
-$225
-$234
-$239
-$243
-$248
-$284
-$290
-$295
-$300
-$306
-$311
-$316
-$322
-$327
-$332
-$338
-$343
-$349
-$354
-$359
-$365
-$370
-$376
-$381
-$386
-$392
-$397
-$5,374
Aggregate Maintenance Savings of 500 ppm Fuel Program
500 ppm Volume
Nonroad Volume*
at $/gal savings of
$0.030
0
0
0
3,298
5,788
5,923
6,057
6,025
6,156
6,288
6,419
7,545
7,700
7,853
8,006
8,159
8,312
8,465
8,619
8,772
8,925
9,078
9,231
9,383
9,535
9,687
9,839
9,992
10,144
10,296
10,448
10,600
10,752
144,086
Loco&Marine Volume*
at $/gal savings of
$0.011
0
0
0
1,780
3,073
3,100
3,127
3,416
3,441
3,462
3,484
3,484
3,510
3,548
3,576
3,604
3,634
3,658
3,695
3,732
3,770
3,808
3,846
3,885
3,924
3,964
4,005
4,046
4,087
4,129
4,171
4,214
4,257
65,449
Aggregate Maintenance Savings
Total
$0
$0
$0
-$118
-$206
-$210
-$215
-$217
-$221
-$225
-$229
-$263
-$268
-$273
-$278
-$283
-$287
-$292
-$297
-$302
-$307
-$312
-$317
-$322
-$327
-$332
-$337
-$342
-$347
-$352
-$357
-$362
-$367
-$5,009
* Note that volumes are taken from Table 7.1-34 and are expressed in millions of gallons (volumes here do not include highway spillover volumes). Factors of 5/12 (Jan through May)
and 7/12 (Jun through Dec) have been used in this table during transition years.
-------
Estimated Aggregate Cost and Cost per Ton of Reduced Emissions
8.4.3 Aggregate CDPF & CCV Maintenance Costs and CDPF Regeneration Costs
Maintenance costs associated with CDPF maintenance and CCV maintenance are developed
on a cents per gallon basis. Section 6.2.3 contains a description of the development of
maintenance costs for the proposed program. Table 8.4-4 contains a summary of the
maintenance costs and estimated fuel volumes for CDPF and CCV equipped engines through
2036 for the proposed two-step fuel program. Note that there are no maintenance costs or CDPF
regeneration costs associated with the 500 ppm fuel program because that program has no new
engine standards and, therefore, neither CDPF nor CCV hardware would be added to new
engines. The fuel volumes shown in Table 8.4-4 differ from those shown in Tables 8.4-1
through 8.4-3 because the volumes of importance for maintenance costs are not volumes
consumed by nonroad engines but rather volumes consumed by CDPF and CCV equipped
engines (i.e., new engines that comply with the proposed standards). The CDPF volumes shown
in Table 8.4-4 contain some highway spillover volumes so they do not match the nonroad
volumes of Tables 8.4-1 through Table 8.4-3 even in the later years. The volumes shown for <75
horsepower engines are used only for CCV maintenance costs during the years from 2008
through 2012 after which time their volumes are captured by the CDPF volumes shown in the
Table.
The cent per gallon costs shown for CDPF maintenance are taken from Table 6.2-27. The
value can be derived using data shown in Table 6.2-26 by weighting the cent per gallon savings
within each horsepower range by the fuel use weighting for that horsepower range - the
summation of these weighted values for all horsepower ranges is 0.6 cents per gallon; note that
weighting these values to reflect the phase-in schedule is not necessary since the volume of fuel
already reflects this weighting. The cent per gallon costs shown for CCV maintenance are taken
from Table 6.2-28, again doing a weighting for each horsepower range using the fuel use
weightings shown in Table 6.2-28 (note that these fuel use weightings include the turbo-charged
fraction shown in Table 6.2-28 because only turbo-charged engines would be adding the CCV
system). Because the volumes shown for <75 horsepower engines includes volumes burned by
non-turbo charged engines, it is necessary to weight the cent per gallon value by the phase-in
schedule of the proposed standards; as a result, this value does not remain constant until 2013
and beyond. The cent per gallon costs shown for CDPF regeneration are taken from Table 6.4-3.
Because of the phase-in schedule of the standards and the different fuel economy impacts of
engines equipped with both a CDPF and a NOx adsorber (1 percent) versus a CDPF-only (2%),
this value varies during the phase-in years before leveling off at a fleetwide average of 0.77 cents
per gallon.
8-19
-------
Table 8.4-4
CDPF & CCV Maintenance Costs and CDPF Regeneration Costs Associated with the Proposed Two-Step Fuel Program (millions)
Year
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 2004-2036
Fuel Volume in
CDPF engines*
0
0
0
0
0
0
0
450
1,310
2,255
3,249
4,230
5,172
6,064
6,880
7,628
8,314
8,936
9,511
10,041
10,540
11,011
1 1 ,445
1 1 ,834
12,183
12,516
12,836
13,165
13,502
13,848
14,203
14,567
14,940
124,535
Fuel Volume in <75hp
engines during 2008-
2012*
0
0
0
0
163
316
477
644
818
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1,986
CDPF Maintenance
CDPF-fleet weighted
$/gal cost
$0.0060
$0.0060
$0.0060
$0.0060
$0.0060
$0.0060
$0.0060
$0.0060
$0.0060
$0.0060
$0.0060
$0.0060
$0.0060
$0.0060
$0.0060
$0.0060
$0.0060
$0.0060
$0.0060
$0.0060
$0.0060
$0.0060
$0.0060
$0.0060
$0.0060
$0.0060
CDPF Regeneration
CDPF-fleet weighted
$/gal cost**
$0.0047
$0.0073
$0.0099
$0.0077
$0.0077
$0.0077
$0.0077
$0.0077
$0.0077
$0.0077
$0.0077
$0.0077
$0.0077
$0.0077
$0.0077
$0.0077
$0.0077
$0.0077
$0.0077
$0.0077
$0.0077
$0.0077
$0.0077
$0.0077
$0.0077
$0.0077
CCV Maintenance
CCV-fleet weighted
$/gal cost
$0.0004
$0.0004
$0.0004
$0.0012
$0.0017
$0.0017
$0.0017
$0.0017
$0.0017
$0.0017
$0.0017
$0.0017
$0.0017
$0.0017
$0.0017
$0.0017
$0.0017
$0.0017
$0.0017
$0.0017
$0.0017
$0.0017
$0.0017
$0.0017
$0.0017
$0.0017
$0.0017
$0.0017
$0.0017
Aggregate Maintenance Costs of Proposed Fuel Program
CDPF
Maintenance
$0
$0
$0
$0
$0
$0
$0
$3
$8
$13
$19
$25
$31
$36
$41
$46
$50
$53
$57
$60
$63
$66
$68
$71
$73
$75
$77
$79
$81
$83
$85
$87
$89
$745
CDPF
Regeneration
$0
$0
$0
$0
$0
$0
$0
$2
$10
$22
$25
$33
$40
$47
$53
$59
$64
$69
$74
$78
$82
$85
$89
$92
$94
$97
$99
$102
$105
$107
$110
$113
$116
$966
CCV
Maintenance
$0
$0
$0
$0
$0
$0
$0
$1
$4
$4
$6
$7
$9
$11
$12
$13
$14
$16
$17
$17
$18
$19
$20
$21
$21
$22
$22
$23
$23
$24
$25
$25
$26
$218
Total
$0
$0
$0
$0
$0
$0
$0
$6
$21
$39
$50
$65
$80
$94
$106
$118
$129
$138
$147
$155
$163
$170
$177
$183
$188
$194
$198
$204
$209
$214
$220
$225
$231
$1,929
*Note that fuel used in CDPF engines includes some highway spillover fuel. Refer to Table 7.1-34fortotoal nonroad volumes including highway spillover volumes; CDPF volumes are less than that total volume
even in 2036 (14,940 vs. 15,626 million gallons) because some pre-control engines would still remain in the fleet. Note that fuel volumes in <75hp engines are used for CCV maintenance costs from 2008 to
2012, after which time their volumes are captured in the CDPF volumes.
hase-in schedules of the proposed standards.
-------
Estimated Aggregate Cost and Cost per Ton of Reduced Emissions
8.4.4 Summary of Aggregate Operating Costs
The net operating costs include the incremental costs for fuel (Table 8.4-1), costs for oil
change maintenance savings (Table 8.4-3 for the proposed fuel program), and costs for CDPF
maintenance, CCV maintenance, and CDPF regeneration (Table 8.4-4). The results of this
summation for the proposed two-step fuel program are shown in Table 8.4-5. The oil change
maintenance savings, CDPF and CCV maintenance costs, and CDPF regeneration costs are
added together in Table 8.4-5 and presented as "Net Maintenance Costs." The net maintenance
costs are presented as negative values, thus, they represent a net savings. The "Net Operating
Cost" is the sum of the incremental fuel costs (both the 500 ppm fuel for nonroad, locomotive,
and marine from 2007 through 2010, 500 ppm fuel for locomotive and marine in 2010 and
beyond, and 15 ppm fuel for nonroad in 2010 and beyond) and the net maintenance costs. Table
8.4-5 also presents the allocation of these costs to each 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. As shown in the table, the net
present value during the period 2004 through 2036 of the proposed two-step fuel program is $5.3
billion using a three percent social discount rate, consisting of $8.7 billion in incremental fuel
costs and $3.4 billion in net maintenance savings.
Table 8.4-6 presents the net operating costs associated with the 500 ppm fuel program. The
costs presented in Table 8.4-6 include the incremental costs for fuel (Table 8.4-2), and costs for
oil change maintenance savings (Table 8.4-3 for the 500 ppm fuel program). The oil change
maintenance savings are presented in the table as "Net Maintenance Costs," and, thus, represent a
net savings. The "Net Operating Cost" is the sum of the incremental fuel costs and the net
maintenance costs. Table 8.4-6 also presents the allocation of these costs to each pollutant (refer
to Table 8.1-2 for how these costs have been allocated). The costs shown in Table 8.4-6 assume
the 500 ppm fuel program remains in place indefinitely. As a result, no new NOx and/or PM
standards would be implemented. The 500 ppm fuel would result in large SOx reductions and,
by comparison, smaller but still important PM reductions. Since the largest reduction is in SOx
emissions, we have simply attributed all costs for the 500 ppm fuel program to SOx for our cost
per ton calculations. Note that our emissions inventory projections and our benefits analysis
include these PM reductions. We have also presented the SOx costs with maintenance savings
(i.e., the net total cost) and without maintenance savings (i.e., the incremental fuel cost) so that a
cost per ton can be calculated with and without the maintenance savings.
8-21
-------
Table 8.4-5
Aggregate Net Operating Costs Associated with the Proposed Two-Step Fuel Program (millions)
Year
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 2004-2036
Aggregate Net Operating Costs of the Proposed Two-Step Fuel Program
500 ppm Fuel Costs
$0
$0
$0
$136
$238
$242
$178
$131
$132
$134
$113
$98
$99
$100
$101
$102
$102
$103
$104
$106
$107
$108
$109
$110
$111
$112
$113
$114
$115
$117
$118
$119
$120
$2,363
15 ppm Fuel Costs
$0
$0
$0
$0
$0
$0
$168
$294
$300
$307
$345
$375
$382
$389
$397
$404
$412
$419
$426
$434
$441
$448
$456
$463
$471
$478
$485
$493
$500
$507
$515
$522
$529
$6,366
Total Fuel Costs
$0
$0
$0
$136
$238
$242
$345
$425
$433
$440
$458
$473
$481
$489
$498
$506
$514
$522
$531
$539
$548
$556
$565
$573
$581
$590
$598
$607
$615
$624
$632
$641
$650
$8,729
Net Maintenance Costs
$0
$0
$0
-$118
-$206
-$210
-$225
-$228
-$218
-$204
-$197
-$219
-$210
-$201
-$194
-$188
-$182
-$178
-$175
-$172
-$169
-$168
-$166
-$166
-$166
-$166
-$166
-$167
-$167
-$167
-$167
-$167
-$166
-$3,445
Net Operating Cost
$0
$0
$0
$18
$32
$31
$121
$197
$215
$236
$260
$254
$271
$288
$304
$318
$332
$344
$356
$367
$378
$389
$398
$407
$416
$424
$432
$440
$449
$457
$466
$474
$483
$5,284
SOx Cost
$0
$0
$0
$18
$31
$31
$68
$93
$94
$95
$74
$59
$60
$61
$61
$62
$62
$63
$63
$64
$65
$65
$66
$67
$67
$68
$69
$69
$70
$71
$71
$72
$73
$1,177
PM Cost
$0
$0
$0
$0
$0
$0
$26
$54
$69
$89
$115
$126
$141
$155
$168
$181
$192
$202
$212
$221
$229
$237
$245
$252
$258
$264
$270
$276
$282
$288
$295
$301
$308
$2,909
NOx+NMHC Cost
$0
$0
$0
$0
$0
$0
$26
$50
$52
$53
$71
$68
$70
$72
$74
$76
$78
$79
$81
$83
$85
$86
$88
$89
$91
$92
$94
$95
$97
$98
$100
$101
$103
$1,198
Note: for fuel costs see Table 8.4-1; for maintenance costs see Tables 8.4-3 and 8.4-4; for cost allocations by pollutant see Table 8.1-2.
-------
Estimated Aggregate Cost and Cost per Ton of Reduced Emissions
Table 8.4-6
Aggregate Net Operating Costs Associated with the 500 ppm Fuel Program (millions)
Year
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 2004-2036
Aggregate Net Operating Costs of 500 ppm Fuel Program
Total Fuel
Costs
$0
$0
$0
$136
$238
$242
$256
$268
$272
$277
$281
$285
$290
$295
$299
$304
$308
$313
$318
$322
$327
$332
$337
$341
$346
$351
$356
$361
$365
$370
$375
$380
$385
$5,514
Net Maintenance
Costs
$0
$0
$0
-$118
-$206
-$210
-$215
-$217
-$221
-$225
-$229
-$263
-$268
-$273
-$278
-$283
-$287
-$292
-$297
-$302
-$307
-$312
-$317
-$322
-$327
-$332
-$337
-$342
-$347
-$352
-$357
-$362
-$367
-$5,009
Net Operatina Cost
$0
$0
$0
$18
$31
$31
$41
$51
$51
$51
$52
$23
$22
$22
$22
$21
$21
$21
$20
$20
$20
$20
$20
$19
$19
$19
$19
$19
$19
$18
$18
$18
$18
$505
SOx Cost w/
Maintenance Savinas
$0
$0
$0
$18
$31
$31
$41
$51
$51
$51
$52
$23
$22
$22
$22
$21
$21
$21
$20
$20
$20
$20
$20
$19
$19
$19
$19
$19
$19
$18
$18
$18
$18
$505
SOx Cost w/o
Maintenance Savinas
$0
$0
$0
$136
$238
$242
$256
$268
$272
$277
$281
$285
$290
$295
$299
$304
$308
$313
$318
$322
$327
$332
$337
$341
$346
$351
$356
$361
$365
$370
$375
$380
$385
$5,514
Note: for fuel costs see Table 8.4-2; for cost allocations by pollutant see Table 8.1-2.
8.5 Summary of Total Aggregate Costs of the Proposed Program
Table 8.5-1 presents a summary of all the costs presented above for the proposed program.
Engine costs are the summation of costs presented in Tables 8.2-1 and 8.2-2, equipment costs are
the summation of costs presented in Tables 8.3-1 and 8.3-2, and fuel costs and net maintenance
costs are presented in Table 8.4-5. The "Total Program Costs" are the summation of engine
costs, equipment costs, and net fuel costs. Table 8.5-2 presents the summary of all the costs
8-23
-------
Draft Regulatory Impact Analysis
presented above for the proposed program by pollutant (refer to Table 8.1-2 for how we have
allocated costs among the various pollutants). We did the cost analysis using a 3% discount rate.
We will also be conducting a similar analysis using a 7% discount rate and including this
information in the docket.
Note that the total aggregate costs associated with the 500 ppm fuel program are presented in
full in Table 8.4-6 since there are no new engine or equipment costs associated with that
program.
Table 8.5-1
Summary of Aggregate Costs for the Proposed Two-Step Fuel and Engine Program by Segment
(millions)
Year
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 2004-2036
Engine Costs
$0
$0
$0
$0
$80
$82
$79
$410
$723
$875
$973
$889
$909
$912
$908
$907
$921
$934
$948
$962
$975
$989
$1,003
$1,016
$1,030
$1,044
$1,058
$1,071
$1,085
$1,099
$1,112
$1,126
$1,140
$14,215
Equipment
Costs
$0
$0
$0
$0
$5
$5
$5
$59
$99
$114
$134
$134
$135
$135
$131
$131
$132
$90
$63
$54
$36
$36
$37
$37
$38
$38
$39
$39
$39
$40
$40
$41
$41
$1,215
Fuel Costs
$0
$0
$0
$136
$238
$242
$345
$425
$433
$440
$458
$473
$481
$489
$498
$506
$514
$522
$531
$539
$548
$556
$565
$573
$581
$590
$598
$607
$615
$624
$632
$641
$650
$8,729
Net Maintenance
Costs
$0
$0
$0
-$118
-$206
-$210
-$225
-$228
-$218
-$204
-$197
-$219
-$210
-$201
-$194
-$188
-$182
-$178
-$175
-$172
-$169
-$168
-$166
-$166
-$166
-$166
-$166
-$167
-$167
-$167
-$167
-$167
-$166
($3,445)
Net Operating
Costs
$0
$0
$0
$18
$32
$31
$121
$197
$215
$236
$260
$254
$271
$288
$304
$318
$332
$344
$356
$367
$378
$389
$398
$407
$416
$424
$432
$440
$449
$457
$466
$474
$483
$5,284
Total Costs
$0
$0
$0
$18
$117
$118
$205
$665
$1,037
$1,226
$1,367
$1,277
$1,315
$1,335
$1,342
$1,356
$1,384
$1,368
$1,367
$1,383
$1,389
$1,414
$1,438
$1,461
$1,484
$1,506
$1,528
$1,550
$1,573
$1,596
$1,618
$1,641
$1,664
$20,713
8-24
-------
Estimated Aggregate Cost and Cost per Ton of Reduced Emissions
Table 8.5-2
Summary of Aggregate Costs for the Proposed Two-Step Fuel and Engine Program by Pollutant
(millions)
Year
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 2004-2036
PM Costs
$0
$0
$0
$0
$85
$86
$111
$400
$664
$826
$829
$761
$806
$825
$825
$846
$867
$865
$871
$880
$888
$906
$922
$939
$954
$970
$985
$1,000
$1,016
$1,032
$1,047
$1,063
$1,079
$13,117
NOx+NMHC Costs
$0
$0
$0
$0
$0
$0
$27
$172
$279
$304
$463
$457
$449
$450
$456
$448
$455
$440
$433
$440
$437
$443
$449
$456
$462
$468
$475
$481
$487
$493
$500
$506
$512
$6,419
SOx Costs
$0
$0
$0
$18
$31
$31
$68
$93
$94
$95
$74
$59
$60
$61
$61
$62
$62
$63
$63
$64
$65
$65
$66
$67
$67
$68
$69
$69
$70
$71
$71
$72
$73
$1,177
Total Costs
$0
$0
$0
$18
$117
$118
$205
$665
$1,037
$1,226
$1,367
$1,277
$1,315
$1,335
$1,342
$1,356
$1,384
$1,368
$1,367
$1,383
$1,389
$1,414
$1,438
$1,461
$1,484
$1,506
$1,528
$1,550
$1,573
$1,596
$1,618
$1,641
$1,664
$20,713
8-25
-------
Draft Regulatory Impact Analysis
8.6 Emission Reductions
Table 8.6-1 presents the emission reductions estimated to result from the proposed two-step
fuel program in conjunction with the proposed engine program. Also presented are reductions
associated with the 500 ppm fuel program. A complete discussion of these emission reductions
and how they were generated can be found in Chapter 3 of this draft RIA.
8-26
-------
Estimated Aggregate Cost and Cost per Ton of Reduced Emissions
Table 8.6-1
Emission Reductions Associated with the Proposed Two-Step Fuel and Engine Program (tons)"
Year
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
sTPV 2004-2036
Proposed Two-Step Fuel Program and Engine Program
NOx+NMHC
0
0
0
0
331
679
1,100
21,527
54,771
92,356
159,869
227,196
293,259
356,969
416,003
471,893
523,758
569,840
611,898
651,009
687,113
721,134
752,239
780,753
807,161
831,947
854,392
875,480
895,923
915,611
934,719
953,145
971,043
7.909.477
PM
0
0
0
11,636
20,911
21,936
23,976
27,947
34,221
41,592
49,424
57,447
65,327
72,981
80,207
86,972
93,290
99,189
104,709
109,897
114,844
119,594
124,061
128,162
131,868
135,400
138,813
141,991
145,065
148,083
151,025
153,851
156,591
1.501.011
SOx
0
0
0
144,298
252,100
256,935
273,470
287,583
292,817
297,975
303,138
308,386
313,862
319,130
324,374
329,641
334,799
340,233
345,674
351,122
356,578
362,041
367,483
372,933
378,391
383,859
389,337
394,825
400,323
405,832
411,352
416,883
422,425
5.977.653
500 ppm Fuel Program
SOxb
0
0
0
144,298
252,100
256,935
261,786
267,117
271,879
276,559
281,244
286,014
291,018
295,815
300,591
305,392
310,086
315,058
320,039
325,028
330,025
335,032
340,021
345,019
350,027
355,045
360,073
365,112
370,162
375,223
380,295
385,378
390,472
5.585.742
a Note that values shown here are emissions reductions. Chapter 3 presents emissions inventories. The values
here are the differences between the baseline inventory values and the proposal inventory values presented in
Chapters.
b Note that the SOx reductions for the two-step fuel program and the 500 ppm fuel program are identical during
the years 2007 through 2010 because only 500 ppm fuel is available during those years. The introduction of 15
ppm fuel in 2010 under the two-step fuel program results in slightly greater SOx reductions for that program
relative to the 500 ppm fuel program.
8-27
-------
Draft Regulatory Impact Analysis
8.7 Cost per Ton
We have calculated the cost per ton of our proposed program 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 of the costs and emissions reductions
from our proposed program including those costs incurred and emissions reductions generated by
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 chosen is meant to capture both the early period of the program when very
few new engines that meet the proposed standards would be in the fleet, and the later period
when essentially all engines would meet the proposed standards. The proposed program also
would require reductions in sulfur content of nonroad diesel fuel (and also locomotive and
marine diesel fuel). We are proposing a 500 ppm sulfur cap on nonroad, locomotive, and marine
fuels beginning in 2007. This fuel program, the first step in our two-step fuel program, provides
significant air quality benefits through reduced SOX emissions, see Tables 3.5-4 and 3.5-5, and
PM emissions, see Tables 3.5-1 and 3.5-2, from both new and existing nonroad, locomotive, and
marine engines. In Table 8.4-6 we summarized the cost for this program as if it remained in
place for 30 years, even though it would be supplanted by the second step of our fuel program in
2010. In Table 8.6-1, we presented the SOX emission reductions expected from this program.
Here we provide an analysis of the cost per ton for the SOX reductions that would be realized by
the 500 ppm fuel program for the same 30 year time window. In this way, the cost per ton of the
SOX reductions realized by the 500 ppm fuel program can be compared to other available means
to control SOX emissions. The PM reductions are not accounted for in the relative cost per ton
estimate, but are listed in Section 3.5.
We are also proposing a second step in the fuel program that would cap nonroad fuel sulfur
levels at 15 ppm beginning in 2010. This fuel program enables the introduction of advanced
emission control technologies including CDPFs and NOx adsorbers. The combination of the
two-step fuel program and the new diesel engine standards represents the total Tier 4 program for
nonroad diesel engines and fuel proposed today. In Table 8.5-2, we presented our estimate of the
annual costs for this complete program by pollutant beginning in 2007 and continuing for 30
years. In Table 8.6-1, we presented the estimated emission reductions for this proposed program.
Here we include an estimate of the cost per ton of emissions reductions realized by this program
forNOx+NMHC, PM, and SOX.
8.7.1 Cost per Ton for the 500 ppm Fuel Program
Table 8.4-6 contains the aggregate fuel costs and net maintenance costs for the 500 ppm fuel
program from 2007-20036. Cost estimates in Table 8.4-6 differ from those in Table 8.5-1 (which
presents total aggregate costs of the two-step fuel program plus the engine program) because the
costs of the second fuel step, related engine standards, equipment modifications and associated
maintenance costs are not part of the 500 ppm fuel program. Figure 8.7-1 presents the total
8-28
-------
Estimated Aggregate Cost and Cost per Ton of Reduced Emissions
annual costs associated with the 500 ppm fuel program through 2036; this is a graphical
representation of the data presented in Table 8.4-6
As can be seen in Figure 8.7-1, the costs for refining and distributing the 500 ppm fuel range
from $250 million in 2008 to nearly $400 million in 2036. These control costs are largely offset
by the maintenance savings that range from $200 million in 2008 to $380 million in 2036. As a
result, the net cost of the program in each year is essentially zero, ranging from $50 million in the
early years to only $18 million in 2036. The net present value of the net costs and savings
associated with the proposed 500 ppm fuel program during the years 2007 to 2036 is estimated at
$510 million.
Figure 8.7-1
Annual Costs of the 500 ppm Fuel Program
$500
2036
($400)
($500)
Year
•Fuel Costs •
•Net Maintenance Costs •
•Net Costs
The 500 ppm fuel program would result in significant reductions in SOX and PM emissions.
For the existing fleet, approximately 98 percent of fuel sulfur is converted to SOX in the engine
with the remaining two percent being emitted in the exhaust as sulfate PM. Because the majority
of the emissions reductions associated with this program would be SOX, we have attributed all the
8-29
-------
Draft Regulatory Impact Analysis
control costs to SOX in calculating the cost per ton for the 500 ppm fuel program. Table 8.6-1
presents the SOX reductions for the 500 ppm fuel program from 2007-2036 which are shown in
the table to have a net present value of 5.6 million tons. The PM reductions for the 500 ppm fuel
program are listed in Section 3.5.
Table 8.7-1 shows the cost per ton of emissions reduced as a result of the proposed 500 ppm
fuel program. The cost per ton numbers include costs and emission reductions that would occur
from both the new and the existing fleet (i.e., those pieces of nonroad equipment that were sold
into the market prior to the proposed emission standards) of nonroad, locomotive, and marine
engines. The long term cost per ton is actually negative. This occurs because nonroad engines
would experience a net savings due to the 500 ppm fuel program (2.4 cents per gallon for the fuel
and 3 cents per gallon for maintenance savings) while locomotive and marine engines would
experience a net cost (2.4 cents per gallon for the fuel and 1.1 cents per gallon for maintenance
savings). Higher growth in nonroad fuel consumption relative to locomotive and marine fuel
consumption eventually results in net negative costs of the 500 ppm fuel program (see Table 8.4-
3 or Figure 8.7-1).
Table 8.7-1
Aggregate Cost per Ton for the 500 ppm Fuel Program
2004-2036 Net Present Values at 3% Discount Rate ($2001)
Item
500 ppm gallons at $0.025/gal (2007-2010)
500 ppm gallons at $0.024/gal (2010+)
Fuel Cost
Net Maintenance Cost
SOx Reduction
Cost per Ton
(with Maintenance Savings)
Cost per Ton
(without Maintenance Savings)
Millions (except $/ton values)
25,206
203,483
$5,514
-$5,009
5,586
$90
$990
Source
Table 8.4-2
Table 8.4-2
Table 8.4-6
Table 8.4-6
Table 8.6-1
Calculated
Calculated
8.7.2 Cost per Ton for the Proposed Program
The proposed program contains a two-step fuel program which is a reduction in sulfur levels
for nonroad diesel fuel from current uncontrolled levels ultimately to 15 ppm, though we are
proposing an interim cap of 500 ppm. Beginning June 1, 2007, refiners would therefore be
required to produce nonroad, locomotive, and marine diesel fuel that meets a maximum sulfur
level of 500 ppm. Then, beginning in June 1, 2010, fuel used for nonroad diesel applications
(excluding locomotive and marine engines) is proposed to meet a maximum sulfur level of 15
8-30
-------
Estimated Aggregate Cost and Cost per Ton of Reduced Emissions
ppm, since all 2011 and later model year nonroad diesel-fueled engines with aftertreatment must
be refueled with this new low sulfur diesel fuel.
The costs of the proposal include costs associated with both steps in the fuel program and
costs for the engine standards including equipment modifications. Maintenance costs and
savings realized by both the existing fleet (nonroad, locomotive, and marine) and the new fleet of
engines complying with the proposed standards are included. Figure 8.7-2 presents in graphic
form the cost of the proposed program. These costs are also summarized in Table 8.5-1. The
cost streams include the amortized capital (fixed) costs and variable costs.
Figure 8.7-2
Costs of the Proposed Fuel and Engine Program
$2,000
2036
-$500
Year
•Engine Costs "^Equipment Costs —^Fuel Costs
Maintenance Costs ^^ Total Program Costs
Figure 8.7-2 shows that total annual costs are estimated to be $120 million in the first year
the new engine standards apply, increasing to $1.7 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 net present value of a 30 year window from 2007 to 2036 is $20.7
billion.
8-31
-------
Draft Regulatory Impact Analysis
The calculations of cost per ton of each emission reduced for the total 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 $6.4 billion for NOx+NMHC, $13.1 billion for
PM and $1.2 billion for SOX. The 30 year net present value, with a three percent discount rate, of
emission reductions are 7.9 million tons for NOx+NMHC, 1.5 million tons for PM and 6.0
million tons for SOX. Our air quality analysis and benefits analysis are found in Chapter 3 and
Chapter 9, respectively.
The cost per ton of emissions reduced associated with the proposed engine and fuel program
are 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 8.7-2.
8-32
-------
Estimated Aggregate Cost and Cost per Ton of Reduced Emissions
Table 8.7-2
Aggregate Cost per Ton for the Proposed Two-Step Fuel Program and Engine Program
2004-2036 Net Present Values at 3% Discount Rate ($2001)
Item
500 ppm gallons at $0.025/gal (2007-2010)
500 ppm gallons at $0.026/gal (2010-2014)
500 ppm gallons at $0.024/gal (2014+)
15 ppm gallons at $0.048/gal (2010-2014)
15 ppm gallons at $0.048/gal (2014+)
500 ppm Fuel Cost
15 ppm Fuel Cost
Net Maintenance Cost
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 SOx
Millions (except $/ton values)
25,206
16,077
54,791
19,763
112,851
$2,363
$6,366
-$3,445
$14,215
$1,215
$20,713
$6,419
$13,117
$1,177
7,909
1,501
5,978
$810
$8,700
$200*
Source
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 8.4-5
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
* This result does not match that in Table 8.4-2 because the nonroad portion of the fuel is reduced to 15 ppm and does not
stay at 500 (locomotive and marine portions are kept at SOOppm). The costs to reduce fuel sulfur from uncontrolled
levels to 15ppm were assigned 50/50 to NOx+NMHC and PM because the reduction to 15 ppm is to enable
aftertreatment technology.
We have also calculated the cost per ton of emissions in the year 2036 using the annual
costs and emission reductions in that year alone. This number, shown in Table 8.7-3, 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
-------
Draft Regulatory Impact Analysis
Table 8.7-3
Long Term Cost per Ton of the Proposed Two-Step Fuel Program and Engine Program
Annual Values without Discounting ($2001)
Pollutant
NOx+NMHC
PM
SOX
Long-Term Cost per Ton
in 2036
$530
$6,900
$170
8-34
-------
Estimated Aggregate Cost and Cost per Ton of Reduced Emissions
Chapter 8 References
1. Power Systems Research, OELink Sales Version, 2002.
2. Nonroad Engine Growth Estimate, Report No. NR-008b, Docket Item II-A-32.
3. "Engine Sales Used in Proposed Nonroad Tier 4 Cost Analysis," memorandum from Todd
Sherwood to Public Docket No. A-2001-28.
8-35
-------
CHAPTER 9: Cost-Benefit Analysis
9.1 Time Path of Emission Changes for the Proposed Standards 9-7
9.2 Development of Benefits Scaling Factors Based on Differences in Emission Impacts
Between Proposed and Modeled Preliminary Control Options 9-9
9.3 Summary of Modeled Benefits and Apportionment Method 9-10
9.3.1 Overview of Analytical Approach 9-11
9.3.2 Air Quality Modeling 9-12
9.3.2.1 PM Air Quality Modeling with REMSAD 9-12
9.3.2.2 Ozone Air Quality Modeling with CAMx 9-13
9.3.3 Health Effect Concentration-Response Functions 9-14
9.3.4 Economic Values for Health Outcomes 9-17
9.3.5 Welfare Effects 9-18
9.3.5.1 Visibility Benefits 9-18
9.3.5.2 Agricultural Benefits 9-19
9.3.5.3 Other Welfare Benefits 9-20
9.3.6 Treatment of Uncertainty 9-22
9.3.7 Model Results 9-23
9.3.8 Apportionment of Benefits to NOx, SO2, and PM Emissions Reductions 9-35
9.4 Estimated Benefits of Proposed Nonroad Diesel Engine Standards in 2020 and 2030
9-37
9.5 Development of Intertemporal Scaling Factors and Calculation of Benefits Over Time
9-40
9.6 Comparison of Costs and Benefits 9-45
APPENDIX 9A: Benefits Analysis of Modeled Preliminary Control Option 9-63
APPENDIX 9B: Sensitivity Analyses of Key Parameters in the Benefits Analysis .... 9-182
APPENDIX 9C: Visibility Benefits Estimates for Individual Class I Areas 9-201
-------
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 proposed 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 proposed 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 proposed regulation. In Chapter 12, we provide an analysis of the benefits of several
alternatives to the proposed standards to examine their relative benefits and costs.
Due to the time requirements for running the sophisticated emissions and air quality
models needed to obtain estimates of the changes in air quality expected to result from
implementation of emission controls, it is often necessary to select a set of preliminary control
options for the purposes of emissions and air quality modeling. The standards we are proposing
in this rulemaking are slightly different in the amount of emission reductions expected to be
achieved in 2020 and 2030 relative to the preliminary control options that we modeled. EPA has
used the best available information and tools of analysis to quantify the expected changes in
public health, environmental and economic benefits of the preliminary control options. We
summarize the results of that analysis in section 9.3, and present details in Appendix 9A, directly
following this chapter. However, we determined that additional analysis was necessary to reflect
the differences in emission reductions between the modeled and proposed standards. The results
of that additional analysis are the focus of this chapter.
In order to characterize the benefits attributable to the proposed 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 proposed standards is similar to that used to estimate
benefits in the recent analysis of the Large Si/Recreational Vehicles standards (see RIA, Docket
A-2000-01). A similar method has also been used in recent benefits analyses for the proposed
Industrial Boilers and Process Heaters MACT standards and the Reciprocating Internal
Combustion Engines MACT standards. One significant limitation to this method is the inability
AEmissions from nonroad diesel engines include directly emitted particles as well as gaseous pollutants that react
in the atmosphere to form fine particles. This proposed 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.
9-1
-------
Draft Regulatory Impact Analysis
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 proposed
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. Note that this table categorizes ozone-related benefits as unquantified effects. We have
quantified ozone-related benefits in our analysis of a set of preliminary benefits, summarized in
Section 9.3 and detailed in Appendix 9A. However, as noted above, we are unable to quantify
ozone-related benefits for the rule we are proposing. 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 proposed rule at this time.
The benefit analysis that we performed for our proposed rule can be thought of as having
seven parts, each of which will be discussed separately in the Sections that follow. These seven
steps are:
1. Identification of proposed standards and calculation of the impact that the proposed
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 proposed
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).
4. Application of scaling factors to apportioned modeled benefits associated with NOx,
SO2, and PM in 2020 and 2030.
5. Development of intertemporal scaling factors based on 2020 and 2030 modeled air
quality and benefits results.
9-2
-------
Cost-Benefit Analysis
6. Application of intertemporal scaling factors to the yearly emission changes expected
to result from the proposed standards from 2010 through 2030 to obtain yearly
monetized benefits.
7. Calculation of present value of stream of benefits.
This primary analysis presents estimates of the potential benefits from the proposed
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 addition, we have also evaluated an alternative, more conservative estimate that can
provide useful insight into the potential impacts of the key elements underlying estimates of the
benefits of reducing NOx and PM emissions from this rule through calculated alternative benefits
for mortality and chronic bronchitis. The alternative approach uses different data on valuation
and makes adjustments relating to the health status and potential longevity of the populations
most likely affected by PM. We are continuing to examine the merits of applying this alternative
approach to the calculation of benefits. Some of the issues that warrant further investigation are
described later in this chapter.
In general, the chapter is organized around the 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 ratios of
proposed to preliminary emissions 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 proposed 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 proposed standards, based
on scaling of the benefits in 2020 and 2030. Finally, in 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 proposed standards.
9-3
-------
Table 9-1
Health and Welfare Effects of Pollutants Affected by the Proposed Nonroad Diesel Engine Rule
Pollutant/Effect
Quantified and Monetized in Base
and Alternative Estimates
Quantified and/or Monetized Effects in
Sensitivity Analyses
Unquantified Effects
PM/Health
Premature mortality - long term
exposures
Premature mortality - short term
exposures
Bronchitis - chronic and acute
Hospital admissions - respiratory
and cardiovascular
Emergency room visits for asthma
Non-fatal heart attacks (myocardial
infarction)
Lower and upper respiratory illness
Minor restricted activity days
Work loss days
Asthma attacks (asthmatic population)
Respiratory symptoms (asthmatic
population)
Infant mortality
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
9-4
-------
Pollutant/Effect
Quantified and Monetized in Base
and Alternative Estimates
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
SO7/Health
Hospital admissions for respiratory and cardiac diseases
Respiratory symptoms in asthmatics
9-5
-------
Pollutant/Effect
Quantified and Monetized in Base
and Alternative Estimates
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 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, et
al, 1999).
b Premature mortality associated with ozone is not separately included in the calculation of total monetized benefits.
0 All non-methane hydrocarbons (NMHCs) listed in the table are also hazardous air pollutants listed in Section 112(b) of the Clean Air Act.
9-6
-------
9.1 Time Path of Emission Changes for the Proposed Standards
The proposed 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 proposed standards, implementation will occur in two stages: reduction in sulfur
content of nonroad diesel fuel and 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 proposed 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 2010 and extends 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 proposed 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 proposed regulations 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 and summarized
in Table 9-3 for the species that form the inputs to the benefits modeling. 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.
9-7
-------
Draft Regulatory Impact Analysis
Table 9-2
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 (TPM)
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.
Summary of 48-State Baseline Emissions for
Nonroad Diesel Engines for Key Emission Species"
2000
2005
2010
2015
2020
2025
2030
Annual Tons
NOx
1,591,801
1,509,081
1,319,917
1,199,235
1,175,544
1,211,002
1,273,245
SO2
243,333
273,331
288,617
315,367
341,941
369,475
397,109
VOC
191,136
155,943
122,996
101,641
93,241
91,709
93,899
PM2.5
218,311
194,554
179,213
178,559
183,250
191,976
201,567
' Excludes Alaska and Hawaii.
9-8
-------
Cost-Benefit Analysis
Table 9-4 summarizes the expected changes in emissions of key species. SO2 emissions are
expected to be reduced by over 90 percent within the first two years of implementation.
Emissions of NOx, NMHC, and PM2 5 are expected to be reduced gradually over the period of
implementation from 2007 to 2030. Overall, NOx, SO2, NMHC, and PM25 emissions are
expected to decline by 65, 97, 30, and 63 percent, respectively, over the 2007 to 2030
implementation period.
Table 9-4
Summary of Reduction in 48-State Emissions
Attributable to Proposed Nonroad Diesel Engine Standards
2010
2015
2020
2025
2030
Tons Reduced
(% of baseline)
NOx
1,007
0.2%
217,575
18.1%
503,701
42.8%
693,857
57.2%
821,911
64.5%
SO2
270,977
93.9%
305,639
96.9%
331,840
97.0%
358,863
97.1%
385,932
97.2%
voc
90
0.4%
8,788
8.5%
18,033
18.6%
24,624
25.6%
29,487
29.9%
PM25
21,864
11.8%
52,476
28.7%
85,254
46.0%
109,325
56.6%
126,910
62.8%
9.2 Development of Benefits Scaling Factors Based on Differences in
Emission Impacts Between Proposed 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 models. The
emissions changes modeled for these two years are similar to those in the proposed standards,
9-9
-------
Draft Regulatory Impact Analysis
differing in the treatment of smaller engines and fuel requirements8. Table 9-5 summarizes the
reductions in emissions of NOx, SO2, and PM2 5 from baseline for the preliminary and proposed
standards, the difference between the two, and the ratio of emissions reductions from the
proposed 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.
Table 9-5
Comparison of 48-state Emission Reductions
in 2020 and 2030 Between Preliminary and Proposed Standards
Emissions Species
2020
NOx
S02
PM2.5
2030
NOx
S02
PM2.5
Reduction from Baseline
Preliminary
663,618
414,692
98,121
1,009,744
483,401
138,208
Proposed
503,701
331,840
85,254
821,911
385,932
126,910
Difference in
Reductions
(Proposed-
Preliminary)
-159,917
-82,852
-12,867
-187,833
-97,469
-11,298
Ratio of
Reductions
(Proposed/
Preliminary)
0.759
0.800
0.869
0.814
0.798
0.918
9.3 Summary of Modeled Benefits and Apportionment Method
Based on the emissions inventories developed 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
PM2.5. Based on the availability of air quality and exposure models, this summary focuses on
reporting the health and welfare benefits of reductions in ambient particulate matter (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).
BEmissions 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 has revised the control scenario. For the reasons discussed in the preamble, EPA has decided
not to propose standards based on aftertreatment for certain of the smallest engine sizes. 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 proposed baseline and control scenarios.
9-10
-------
Cost-Benefit Analysis
The reductions in emissions of NOx, SO2, and PM2 5 from nonroad engines in the United
States are expected to result in wide-spread overall reductions in ambient concentrations of
ozone and PM2 5C. 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.
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 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; 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
c 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.
9-11
-------
Draft Regulatory Impact Analysis
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 infant mortality, 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
Document (U.S. EPA, 1996). 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
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
concentrations0, 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:
DIn 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).
9-12
-------
Cost-Benefit Analysis
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.
3. 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.
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 proposed 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., OTAG, On-highway Tier-2, and HD Engine/Diesel Fuel).
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.
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
9-13
-------
Draft Regulatory Impact Analysis
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 Effect Concentration-Response Functions
Health benefits for this analysis are based on health effect incidence changes due to
predicted air quality changes in the years 2020 and 2030. Integral to the estimation of such
benefits is a reasonable estimate of future population projections. The underlying data used to
create county-level 2020 and 2030 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, which account for patterns of economic
growth and migration. Growth factors are calculated using the Woods and Poole data and are
applied to 2000 U.S. Census data.
Fundamental to the estimation of health benefits was our utilization of the PM and ozone
epidemiology literature. We rely upon C-R functions derived from published epidemiological
studies that relate health effects to ambient concentrations of PM and ozone. The specific studies
from which C-R functions are drawn are listed in Table 9-5. While a broad range of serious
health effects have been associated with exposure to elevated PM and ozone levels, we include
only a subset of health effects in this benefit analysis due to limitations in available C-R
functions and concerns about double-counting of overlapping effects (U.S. EPA, 1996). Since
the analysis of the Heavy Duty Engine rulemaking, we have added a number of new endpoints,
which are described in detail in Appendix 9B.
To generate health outcomes, projected changes in ambient PM and ozone
concentrations were put into the Criteria Air Pollutant Modeling System (CAPMS), a customized
GIS-based program. CAPMS aggregates population to air quality model grids and calculates
changes in air pollution metrics (e.g., daily averages) for input into C-R functions. CAPMS uses
grid cell level population data and changes in pollutant concentrations to estimate changes in
health outcomes for each grid cell. Details on the application of CAPMS for this analysis are
provided in a separate report (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 C-R
functions and the application of those functions in estimating impacts on health outcomes. In
9-14
-------
Cost-Benefit Analysis
general, we selected C-R functions that 1) most closely match the pollutants of interest, i.e.
PM2.5 and ozone, 2) 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 C-R functions 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.
While there is a consistent body of evidence supporting a relationship between a number
of adverse health effects and PM and ozone exposure, 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 C-R function to more completely capture
health benefits in the general population. Estimates for subpopulations are provided in Appendix
9A.
Based on a review of the recent literature on health effects of PM exposure (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 C-R functions are continuous and
differentiable down to natural background levels. In addition, we explore this important
assumption in a sensitivity analysis described in Appendix 9B.
Premature Mortality
As in the Kunzli et al. (2000) analysis, we focus on the prospective cohort long-term
exposure studies in deriving the C-R 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 a C-R function from the re-analysis
of the American Cancer Society (ACS) study conducted for the Health Effects Institute (Pope et
al., 1995; Krewski et al; 2000)E. This C-R function relates annual mean PM25 levels and all-
cause mortality in adults 30 and older. The selected C-R function relates premature mortality and
mean PM2 5 levels rather than median levels as used in the original ACS analysis. For policy
analysis purposes, functions based on the mean air quality levels may be preferable to functions
based on the median air quality levels because changes in the mean more accurately reflect the
changes in peak values targeted by many policies than do changes in the median.
EA recent analysis (Pope et al, 2002) reexamines the ACS cohort using a longer follow-up period. We have
examined how using alternative C-R functions derived from this new study impact our results in a sensitivity analysis
presented in Appendix 9B.
9-15
-------
Draft Regulatory Impact Analysis
To reflect concerns about the inherent limitations in the number of studies supporting a
causal association between long-term exposure and mortality, an Alternative benefit estimate for
premature mortality was derived from the large number of time-series studies that have
established a likely causal relationship between short-term measures of PM and daily mortality
statistics. The Alternative Estimate assumes that there is no mortality effect of chronic
exposures to fine particles. Instead, it assumes that the full impact of fine particles on premature
mortality can be captured using a concentration-response function relating daily mortality to
short-term fine particle levels. This will clearly provide a lower bound to the mortality impacts
of fine particle exposure, as it omits any additional mortality impacts from longer term
exposures. Specifically, concentration- response functions based on Schwartz et al. (1996) are
employed, with an adjustment to account for recent evidence that daily mortality is associated
with particle levels from a number of previous days (Schwartz, 2000). The size of the effect
estimates from these models suggests consistency between the findings of studies that examine
assocations premature mortality impacts of short-term and long-term exposures. Additional
research may be necessary to confirm this trend. Two C-R functions are used for the alternative
estimate, one relating short-term PM2 5 levels to daily COPD mortality for all ages, and one
relating short-term PM2 5 levels to non-COPD mortality for all ages.
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;
9-16
-------
Cost-Benefit Analysis
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 is 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
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.
It is currently unknown whether there is a delay between changes in chronic PM
exposures and changes in mortality rates. The existence 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 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 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 base estimate, 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. No lag adjustment is necessary for the alternative estimate, which focuses
on premature mortality occurring within a few days of the PM exposure.
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
9-17
-------
Draft Regulatory Impact Analysis
used to account for projected growth in real income from 1990 to 2030 are 1.09 for minor health
effects, 1.33 for severe and chronic health effects, 1.29 for premature mortality, and 1.79 for
recreational visibility. Adjustment factors for 2020 are 1.08 for minor health effects, 1.30 for
severe and chronic health effects, 1.26 for premature mortality, and 1.70 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.
For two endpoints, premature mortality and chronic bronchitis, we provide both a base
valuation estimate, reflecting the best available scientific literature and methods, and an
alternative estimate, reflecting different assumptions about the value of reducing risks of
premature death and chronic bronchitis. Following the advice of the Environmental Economics
Advisory Committee of the Science Advisory Board, the base estimate uses the VSL approach in
calculating the primary estimate of mortality benefits, because we believe this calculation to
provide the most reasonable single estimate of an individual's willingness to trade off money for
reductions in mortality risk (EPA-SAB-EEAC-00-013). The mean value of avoiding one
statistical death (the VSL) is estimated to be $6.3 million in constant 2000 dollars. This
represents an intermediate value from a variety of estimates that appear in the economics
literature, and it is a value EPA has frequently used in RIAs for other rules and in the Section 812
Reports to Congress. The Alternative Estimate reflects the impact of changes to key assumptions
associated with the valuation of mortality. These include: l)an alternaive interpretation of the
literature on monetary valuation of VSL, 2) the use of a value of a statistical life years rather than
a VSL approach, and 3) the degree of prematurity (number of statistical life years lost) for
mortalities from air pollution.
9.3.5 Welfare Effects
Our analysis examines 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.
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.
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,
9-18
-------
Cost-Benefit Analysis
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.F
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 (Abt Associates, 2003). 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.
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 AGSJJVI® agricultural benefits model (Taylor, et al., 1993). AGSEVI®
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
F For details of the visibility estimates discussed in this section, please refer to the benefits technical support
document for this RIA (Abt Associates 2003).
9-19
-------
Draft Regulatory Impact Analysis
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 AGSEVI® 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
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 FID 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 physical 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.
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 proposed 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.
9-20
-------
Cost-Benefit Analysis
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 9B.
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 proposed 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.
9-21
-------
Draft Regulatory Impact Analysis
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
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 base estimate of the total benefits, based on our interpretation of the best
available scientific literature and methods and supported by the SAB and the NAS (NRC, 2002).
In addition, we provide an alternative estimate based on several important alternative
assumptions about the estimation and valuation of reductions in premature mortality and chronic
bronchitis. We also 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
and have analyzed a full Alternative Estimate 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 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. We plan to better characterize some of this
uncertainty, especially regarding mortality-related benefits in the RIA to accompany the final
rule.
9-22
-------
Cost-Benefit Analysis
9.3.7 Model Results
Full implementation of the modeled preliminary control options is projected in 2020 to
reduce 48-state emissions of NOx by 663,600 tons (58 percent of landbased nonroad emissions),
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 options are 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 129,000 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.
Applying the C-R functions described in Table 9-5 to the estimated changes in PM2 5 and
ozone yields estimates of the number of avoided incidences for each health outcome. These
estimates are presented in Table 9-10 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 9-10, 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 are presented in Table 9-11.
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 in our base
estimate and over 60 percent of total monetized benefits in the alternative estimate. 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 in the
base estimate. 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 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). However, by 2030, there is a net decrease in ozone-related MRAD
consistent with widespread reductions in ozone concentrations from the increased NOx
9-23
-------
Draft Regulatory Impact Analysis
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. 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
$2 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.
9-24
-------
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
Base - Long-term
exposure
Alternative - Short-
term exposure
PM25
PM25
>29 years
all ages
Krewski, et al. (2000)
Schwartz et al. (1996) adjusted
using ratio of distributed lag to
single day coefficients from
Schwartz et al. (2000)
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
03
03
PM25
PM25
PM25
> 64 years
< 2 years
>64 years
20-64 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 (2000) - ICD 490-
496 (COPD)
Lippman et al. (2000) - ICD
490-496 (COPD)
Moolgavkar (2000) - ICD 490-
496 (COPD)
Lippman et al. (2000) - ICD
480-486 (pneumonia)
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
9-25
-------
Draft Regulatory Impact Analysis
Table 9-6
Endpoints and Studies Used to Calculate Total Monetized Health Benefits
Endpoint
Cardiovascular
Asthma-Related ER
Visits
Pollutant
PM25
PM25
PM25
03
PM25
Applied
Population
< 65 years
> 64 years
20-64 years
All ages
0-18 years
Source of Effect Estimate(s)
Sheppard, etal. (1999)-ICD
493 (asthma)
Pooled estimate:
Moolgavkar (2000) - ICD 390-
429 (all cardiovascular)
Lippman et al. (2000) - 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 etal. (1992),
Stieb etal. (1996)
Norrisetal. (1999)
Source of Baseline
Incidence
1999 NHDS public use
data files
1999 NHDS public use
data files
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
Upper Respiratory
Symptoms
Lower Respiratory
Symptoms
Work Loss Days
School Absence Days
Worker Productivity
Minor Restricted
Activity Days
PM25
PM10
PM25
PM25
03
03
PM2 5, O3
8-12 years
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)
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)
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)
9-26
-------
Table 9-7
Unit Values Used for Economic Valuation of Health Endpoints (2000$)
Health
Endpoint
Premature Mortality
Base Estimate (VSL)
Alternative Estimate (VSLY)
3% discount rate
Under 65
65 and older
7% discount rate
Under 65
65 and older
Chronic Bronchitis (CB)
Base Estimate
Alternative Estimate
3% discount rate
Age 27-44
Age 45-64
Age 65 and older
7% discount rate
Age 27-44
Age 45-64
Age 65 and older
Central Estimate of Value Per Statistical Incidence
1990 Income
Level
$6,300,000
$172,000
$434,000
$286,000
$527,000
$340,000
$150,542
$97,610
$11,088
$86,026
$72,261
$9,030
2020 Income
Level
$8,000,000
$217,000
$547,000
$360,000
$664,000
$430,000
$150,542
$97,610
$11,088
$86,026
$72,261
$9,030
2030 Income
Level
$8,100,000
$221,000
$559,000
$368,000
$678,000
$440,000
$150,542
$97,610
$11,088
$86,026
$72,261
$9,030
Derivation of Estimates
Base value is the mean of VSL estimates from 26 studies (5
contingent valuation and 21 labor market studies) reviewed for the
Section 812 Costs and Benefits of the Clean Air Act, 1990-2010 (US
EPA, 1999).
Alternative VSLY estimates are derived from a VSL based on the
mean of VSL estimates from the 5 contingent valuation studies
referenced above. VSLY for populations under 65 are based on 35
years of assumed average remaining life expectancy. VSLY for
populations 65 and older are based on 10 years of assumed average
remaining life expectancy.
Base value is the mean of a generated distribution of WTP to avoid a
case of pollution-related CB. WTP to avoid a case of pollution-
related CB is derived by adjusting WTP (as described in Viscusi et
al., 1991) to avoid a severe case of CB for the difference in severity
and taking into account the elasticity of WTP with respect to severity
ofCB.
Alternative value is a cost of illness (COI) estimate based on
Cropper and Krupnick (1990). Includes both medical costs and
opportunity cost from age of onset to expected age of death (assumes
that chronic bronchitis does not change life expectancy).
-------
Health
Endpoint
Non-fatal 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
Central Estimate of Value Per Statistical Incidence
1990 Income
Level
$66,902
$74,676
$78,834
$140,649
$66,902
$65,293
$73,149
$76,871
$132,214
$65,293
2020 Income
Level
$66,902
$74,676
$78,834
$140,649
$66,902
$65,293
$73,149
$76,871
$132,214
$65,293
2030 Income
Level
$66,902
$74,676
$78,834
$140,649
$66,902
$65,293
$73,149
$76,871
$132,214
$65,293
Derivation of Estimates
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:
aae 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)
Pneumonia
(ICD codes 480-487)
Asthma admissions
$12,378
$14,693
$6,634
$12,378
$14,693
$6,634
$12,378
$14,693
$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 COPD
category illnesses) reported in Agency for Healthcare Research and
Quality, 2000 (www.ahrq.gov).
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
pneumonia category illnesses) reported in Agency for Healthcare
Research and Quality, 2000 (www.ahrq.gov).
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).
-------
Health
Endpoint
All Cardiovascular
(ICD codes 390-429)
Emergency room visits for asthma
Central Estimate of Value Per Statistical Incidence
1990 Income
Level
$18,387
$286
2020 Income
Level
$18,387
$286
2030 Income
Level
$18,387
$286
Derivation of Estimates
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
cardiovascular category illnesses) reported in Agency for Healthcare
Research and Quality, 2000 (www.ahrq.gov).
Simple average of two unit COI values:
(1) $311.55, from Smith et al., 1997, and
(2) $260.67, from Stanford et al., 1999.
Respiratory Ailments Not Requiring Hospitalization
Upper Respiratory Symptoms (URS)
Lower Respiratory Symptoms (LRS)
Acute Bronchitis
$25
$16
$360
$27
$17
$390
$27
$17
$390
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.
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.
Assumes a 6 day episode, with daily value equal to the average of
low and high values for related respiratory symptoms recommended
in Neumann, et al. 1994.
Restricted Activity and Work/School Loss Days
Work Loss Days (WLDs)
Variable
(national
median =
$115)
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.
-------
Health
Endpoint
School Absence Days
Worker Productivity
Minor Restricted Activity Days
(MRADs)
Central Estimate of Value Per Statistical Incidence
1990 Income
Level
$75
$0.95 per
worker per 10%
change in ozone
per day
$51
2020 Income
Level
$75
$0.95 per
worker per 10%
change in ozone
per day
$55
2030 Income
Level
$75
$0.95 per
worker per 10%
change in ozone
per day
$56
Derivation of Estimates
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). This median 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.
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).
Median WTP estimate to avoid one MRAD from Tolley, et al.
(1986) .
-------
Table 9-8
Primary Sources of Uncertainty in the Benefit Analysis
1. Uncertainties Associated With Concentration-Response Functions
The value of the ozone- or PM-coefficient in each C-R function.
Application of a single C-R function to pollutant changes and populations in all locations.
Similarity of future year C-R relationships to current C-R relationships.
Correct functional form of each C-R relationship.
Extrapolation of C-R relationships beyond the range of ozone or PM concentrations observed in the study.
Application of C-R relationships 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 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 C-R functions. Thus, unquantified or
unmonetized benefits are not included.
-------
Draft Regulatory Impact Analysis
Table 9-9
Distribution of PM2.5 Air Quality Improvements Over Population
Due to Nonroad Engine/Diesel Fuel Standards: 2020 and 2030
Change in Annual Mean PM2 s
Concentrations (ng/m3)
0 > A PM2 5 Cone < 0.25
0.25>APM25Conc < 0.5
0.5 > A PM2 5 Cone < 0.75
0.75>APM25Conc < 1.0
1.0>APM25Conc < 1.25
1.25>APM25Conc < 1.5
1.5> APM25Conc < 1.75
APM25Conc>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%
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%
The change is defined as the control case value minus the base case value.
9-32
-------
Cost-Benefit Analysis
Table 9-10
Reductions in Incidence of Adverse Health Effects Associated with Reductions in Particulate
Matter and Ozone Due to the Modeled Preliminary Nonroad Engine Standards
Endpoint
PM-related Endpoints
Premature mortality5 -
Base estimate: Long-term exposure (adults, 30 and over)
Alternative estimate: Short-term exposure (all ages)
Chronic bronchitis (adults, 26 and over)
Non-fatal myocardial infarctions (adults, 1 8 and older)
Hospital admissions - Respiratory (all ages)c
Hospital admissions - Cardiovascular (adults, 20 and older)D
Emergency Room Visits for Asthma (18 and younger)
Acute bronchitis (children, 8-12)
Lower respiratory symptoms (children, 7-14)
Upper respiratory symptoms (asthmatic children, 9-1 1)
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 6-1 1)
Avoided IncidenceA
(cases/year)
2020
6,200
3,700
4,300
11,000
3,100
3,300
4,300
10,000
110,000
92,000
780,000
4,600,000
370
150
93
(2,400)
65,000
2030
11,000
6,600
6,500
18,000
5,500
5,700
6,500
16,000
170,000
120,000
1,100,000
6,500,000
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.
9-33
-------
Draft Regulatory Impact Analysis
Table 9-11
Results of Human Health and Welfare Benefits
Valuation for the Modeled Preliminary Nonroad Diesel Engine Standards
Endpoint
Premature mortality0
Base estimate: Long-term exposure, (adults, 30 and over)
3% discount rate
7% discount rate
Alternative estimate: Short-term exposure, (all ages)
3% discount rate
7% discount rate
Chronic bronchitis (adults, 26 and over)
Base estimate: Willingness-to-pay
Alternative estimate: Cost-of-illness
3% discount rate
7% discount rate
Non-fatal myocardial infarctions
3% discount rate
7% discount rate
Hospital Admissions from Respiratory Causes
Hospital Admissions from Cardiovascular Causes
Emergency Room Visits for Asthma
Acute bronchitis (children, 8-12)
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 Total11
Base estimate
3% discount rate
7% discount rate
Alternative estimate
3% discount rate
7% discount rate
Pollutant
PM
PM
PM
O3 and PM
PM
O3 and PM
PM
PM
PM
PM
O3 and PM
03
03
PM
03
O3 and PM
Monetary Benefits^13
(millions 2000$, Adjusted for Income
Growth)
2020
$47,000
$44,000
$7,200
$8,200
$1,900
$420
$270
$900
$870
$55
$72
$1
$4
$2
$2
$110
$250
$5
$4
$1,400
$89
$52,000+B
$49,000+B
$11,000+B
$11,000+B
2030
$85,000
$80,000
$13,000
$15,000
$3,000
$600
$390
$1,400
$1,400
$110
$120
$2
$6
$3
$3
$150
$370
$10
$7
$2,200
$140
$92,000+B
$87,000+B
$19,000+B
$20,000+B
A 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
9-34
-------
Cost-Benefit Analysis
assumes the 5 year distributed lag structure described earlier. Results reflect the use of two different discount rates; a 3% rate which is
recommended by EPA's Guidelines for Preparing Economic Analyses (US EPA, 2000c), and 7% which is recommended by OMB Circular A-94
(OMB, 1992).
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.
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.
9.3.8 Apportionment of Benefits to NOx, SO2, and PM Emissions Reductions
As noted in the introduction to this chapter, the proposed standards 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 proposed 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.
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, SO2,
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 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
particle 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
SO2 alone is reduced, ammonia is freed to react with any NOx that has not been used in forming
9-35
-------
Draft Regulatory Impact Analysis
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 particle mass that is most related to health benefits is
the population-weighted change in PM2 5 ng/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 9-12. The sulfate percentage of total change is used to represent the SO2 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 Table
9-10 and 9-11 and combined with the emission scaling factors developed in section 9.2 to
estimate benefits for the proposed standards.
Table 9-12. Apportionment of Population Weighted Change in Ambient PM2.5 to Nitrate,
Sulfate, and Primary Particles
2020 2030
Population- Percent of Total Population-
weighted Change Change weighted Change
(|j.g/m3) (M.g/m3)
Total PM2.5 0.316 0.438
Sulfate 0.071 22.5% 0.090
Nitrate 0.041 13.1% 0.073
Primary PM 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
9-36
-------
Cost-Benefit Analysis
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):
&PEXT = [3F(r/0* 1.375 * ATSO4]+ [3F(r/j)* 1.29 * APNO3]+ 10 * APEC + 4 * ATOA + APMFINE + 0.6 * APMCOARSE
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 * ATS04]/A/3EXT . The proportion of the total change in light extinction
associated with changes in nitrate particles is \?F(rh) * 1.29 * APNOl]/A(3EXT Finally, the
proportion of the total change in light extinction associated with the change in directly emitted
particles is [10 * APEC + 4 * ATOA + APMFINE + 0.6 * APMCOARSE]/A/3EXT .
We calculate these proportions for each park to apportion park specific benefits between SO2,
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 9C.
9.4 Estimated Benefits of Proposed Nonroad Diesel Engine Standards in
2020 and 2030
To estimate the benefits of the NOx, SO2, and direct PM emission reductions from the
proposed 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 9-10 and 9-11. 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 proposed 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 *
where Ri is the emissions ratio for emission species i from Table 9-4, and Ai is the health
benefits apportionment factor for emission species i, from Table 9-12. 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:
9-37
-------
Draft Regulatory Impact Analysis
Scaled Premature Mortality Incidence = 6,200 * (0.759*0.129 + 0.800*0.224 + 0.869*0.647) =
5,200
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 proposed 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. 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.
Table 9-13.
Reductions in Incidence of PM-related Adverse
Health Effects Associated with the Proposed Nonroad Diesel Engine Standards
Endpoint
Premature mortality8 -
Base estimate: Long-term exposure (adults, 30 and over)
Alternative estimate: Short-term exposure (all ages)
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)
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
5,200
3,100
3,600
9,200
2,400
1,900
3,600
8,400
92,000
77,000
650,000
3,900,000
2030
9,600
5,800
5,700
16,000
4,500
3,800
5,700
14,000
150,000
110,000
960,000
5,700,000
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.
9-38
-------
Table 9-14. Results of Human Health and Welfare Benefits
Valuation for the Proposed Nonroad Diesel Engine Standards
Endpoint
Premature mortality0
Base estimate: Long-term exposure, (adults, 30 and over)
3% discount rate (over 5 year cessation lag)
7% discount rate (over 5 year cessation lag)
Alternative estimate: Short-term exposure, (all ages)
3% discount rate
7% discount rate
Chronic bronchitis (adults, 26 and over)D
Base estimate: Willingness-to-pay
Alternative estimate: Cost-of-illness
3% discount rate (over lifetime with disease)
7% discount rate (over lifetime with disease)
Non-fatal myocardial infarctionsE
3% discount rate (over 5 year follow up)
7% discount rate (over 5 year follow up)
Hospital Admissions from Respiratory CausesF
Hospital Admissions from Cardiovascular Causes0
Emergency Room Visits for Asthma
Acute bronchitis (children, 8-12)
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 TotalH
Base estimate
3% discount rate
7% discount rate
Alternative estimate
3% discount rate
7% discount rate
Monetary BenefitsA B
(millions 2000$, Adjusted for Income
Growth)
2020
$39,000
$37,000
$6,100
$6,800
$1,600
$350
$220
$750
$730
$38
$40
$1
$3
$2
$2
$90
$210
$1,200
$43,000+B
$41,000+B
$8,700+B
$9,300+B
2030
$74,000
$70,000
$12,000
$13,000
$2,600
$530
$340
$1,300
$1,200
$74
$80
$2
$5
$3
$3
$130
$320
$1,900
$81,000+B
$76,000+B
$16,000+B
$17,000+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 5 year distributed lag structure described earlier. Valuation of alternative estimate
assumes value of a statistical life year derived from amortization of $3.7 million value of statistical life over age group-specific remaining life
expectancy. Results reflect the use of two different discount rates; a 3% rate which is recommended by EPA's Guidelines for Preparing
Economic Analyses (US EPA, 2000c), and 7% which is recommended by OMB Circular A-94 (OMB, 1992).
D Alternative estimate assumes costs of illness and lost earnings in later life years are discounted using either 3 or 7 percent.
E Estimates assume costs of illness and lost earnings in later life years are discounted using either 3 or 7 percent
F Respiratory hospital admissions for PM includes admissions for COPD, pneumonia, and asthma.
0 Cardiovascular hospital admissions for PM includes total cardiovascular and subcategories for ischemic heart disease, dysrhythmias, and heart
failure.
H 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.
-------
Draft Regulatory Impact 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 proposed 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 PM2 5, 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 SO2 emission reductions may change in a non-linear fashion. It is not
clear what overall biases these nonlinearities will introduce into the scaling exercise.
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.
9-40
-------
Cost-Benefit Analysis
Table 9-15.
Projected Marginal Contribution of Reductions
in Emission Species to Reductions in Ambient PM2.5
Change in PM2.5 species (population weighted ng/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 target year of 2010, there are 1,007 tons of NOx reductions, 270,977
tons of SO2 reductions, and 21,864 tons of PM reductions. These are associated with a populated
9-41
-------
Draft Regulatory Impact Analysis
weighted change in total PM25 of 0.089, calculated from Table 9-15. The ratio of this change to
the change in the 2030 base year is 0.202. The age group apportionment factors (based on the
Base estimate using a 3% discount rate for 2030) are 0.02% for 0 to 18, 19.4% for 19 to 64, and
80.6% for 65 and older. The age group population growth ratios for 2010 relative to 2030 are
0.88 for 0 to 18, 0.96 for 19 to 64, and 0.55 for 65 and older. The income growth adjustment
ratios for 2015 are 0.85 for mortality endpoints and 0.84 for morbidity endpoints. Mortality
accounts for 93 percent of total health benefits and morbidity accounts for 7 percent of health
benefits. Combining these elements with the total Base estimate of PM health benefits in 2030
of $89.8 billion , total PM health benefits in 2010 for the proposed standards are calculated as:
Total PM health benefits (2010) =
$89.8 billion * 0.203*(0.0002*0.876+0.194*0.961+0.806*0.552)*(0.93*.855+0.07*.838) = $9.8
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
PM25, 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
SO2 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.
Table 9-16 provides undiscounted estimates of the time stream of benefits for the
proposed standards for the Base and Alternative estimates using 3 and 7 percent concurrent
9-42
-------
discount rates0. Figure 9-1 shows the undiscounted time stream for the Base estimate 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 proposed standards for the base estimate is approximately $550 billion for the time period
2007 to 2030, using either a 3 percent concurrent discount rate or $520 billion using a 7 percent
concurrent discount rate. For the alternative estimate, the present value using a 3 percent
intertemporal discount rate is approximately $90 billion using either a 3 or 7 percent concurrent
discount rate. Annualized benefits using a 3 percent intertemporal discount rate for the base
estimate are approximately $30 billion using either a 3 or 7 percent concurrent discount rate.
Annualized benefits using a 3 percent intertemporal discount rate for the alternative estimate are
approximately $5 billion using either a 3 or 7 percent concurrent discount rate.
Using a 7 percent intertemporal discount rate, the present value in 2004 of the benefits of
the proposed standards for the base estimate is approximately $290 billion for the time period
2007 to 2030, using a 3 percent concurrent discount rate or $270 billion using a 7 percent
concurrent discount rate. For the alternative estimate, the present value using a 7 percent
intertemporal discount rate is approximately $45 billion using a 3 percent concurrent discount
rate or $48 billion using a 7 percent concurrent discount rate.
GWe refer to discounting that occurs during the calculation of benefits for individual years as concurrent
discounting. This is distinct from discounting that occurs over the time stream of benefits, which is referred to as
intertemporal discounting.
-------
Table 9-16. Time Stream of Benefits for Proposed Nonroad Diesel Engine Standards
A,B
Year
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
Base Estimate
(Million 2000$)
3% Concurrent
Discount Rate
$4,700
$8,600
$9,100
$10,000
$12,500
$14,800
$17,600
$20,800
$24,300
$27,900
$31,700
$35,500
$39,300
$43,200
$47,100
$51,000
$55,000
$58,900
$62,700
$66,400
$69,900
$73,500
$77,100
$80,600
7% Concurrent
Discount Rate
$4,400
$7,900
$8,400
$9,300
$11,500
$13,600
$16,200
$19,200
$22,400
$25,800
$29,200
$32,700
$36,300
$39,900
$43,500
$47,100
$50,700
$54,400
$57,800
$61,200
$64,600
$67,900
$71,200
$74,500
Alternative Estimate
(Million 2000$)
3% Concurrent
Discount Rate
$950
$1,800
$1,800
$2,000
$2,600
$3,100
$3,600
$4,300
$5,000
$5,700
$6,500
$7,200
$8,000
$8,800
$9,600
$10,000
$11,000
$12,000
$13,000
$13,000
$14,000
$15,000
$15,000
$16,000
7% Concurrent
Discount Rate
$1,000
$1,900
$2,000
$2,200
$2,700
$3,300
$3,900
$4,600
$5,300
$6,100
$6,900
$7,800
$8,600
$9,400
$10,000
$11,000
$12,000
$13,000
$14,000
$14,000
$15,000
$16,000
$17,000
$17,000
Present Value in 2004
3% Intertemporal
Discount Rate
7% Intertemporal
Discount Rate
$550,000
$290,000
$510,000
$270,000
$110,000
$58,000
$120,000
$63,000
A All dollar estimates rounded to two significant digits.
B Results reflect the use of two different discount rates; a 3% rate which is recommended by EPA's Guidelines for Preparing Economic Analyses
(US EPA, 2000c), and 7% which is recommended by OMB Circular A-94 (OMB, 1992).
-------
Cost-Benefit Analysis
Figure 9-1.
Base Estimate of the Stream of Annual Benefits for the Proposed Nonroad Diesel Engine
Standards: 2007 to 2030
$100,000
2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022 2023 2024 2025 2026 2027 2028 2029 2030
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 $1.5 billion (2000$). Thus, the net
benefit (social benefits minus social costs) of the program at full implementation is
approximately $79 + 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 proposed 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 proposed rule. Figure 9-2
displays the stream of benefits, costs, and net benefits of the Nonroad Diesel Engine and Fuel
Standards from 2007 to 2030. In addition, Table 9-18 presents the present value of the stream of
benefits, costs, and net benefits associated with the rule for this 23 year period (using a three
percent discount rate). The total present value of the stream of monetized net benefits (benefits
minus costs) is $540 billion.
9-45
-------
Draft Regulatory Impact Analysis
Table 9-17.
Summary of Monetized Benefits, Costs, and Net Benefits of the
Proposed Nonroad Diesel Engine and Fuel StandardsA
Social Costsc
Social Benefits13*:
CO, VOC, Air Toxic-related benefits
Ozone-related benefits
PM-related Welfare benefits
PM-related Health benefits
Net Benefits (Benefits-Costs)13 E
Base Estimate15
2020
(Billions of
2000 dollars)
$1.4
Not
monetized
Not
monetized
$1.2
$42 + B
$42 + B
2030
(Billions of
2000 dollars)
$1.5
Not
monetized
Not
monetized
$1.9
$79 + B
$79 + 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 SO2 reductions.
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 benefits are presented using two different discount rates. Results calculated using 3 percent discount
rate are recommended by EPA's Guidelines for Preparing Economic Analyses (U.S. EPA, 2000c). Results
calculated using 7 percent discount rate are recommended by OMB Circular A-94 (OMB, 1992).
9-46
-------
Cost-Benefit Analysis
Figure 9-2.
Stream of Benefits, Costs, and Net Benefits of the
Proposed Nonroad Diesel Engine and Fuel Standards
$100,000
$80,000
$60,000
$40,000
$20,000
Net Present Value = $540 billion
t> OO O\ O i—i
o o o o o
(N (N (N (N (N
(N (N (N (N
OO O\ O i—i
S(N
O
(N (N (N (N
(N m
S(N
o
(N (N
S(N (N (N
O O O
(N (N (N (N
Table 9-18.
Present Value in 2004 of the Stream of
Benefits, Costs, and Net Benefits for the
Proposed Nonroad Diesel Engine and Fuel Standards
(Billions of 2000$)a
Social Costs
Social Benefits
Net Benefits
Base
$17
$550
$530
' Rounded to two significant digits
Two key inputs to our benefit-cost analysis are the social costs and emission
reductions associated with the proposed program. Each of these elements also has associated
uncertainty which contributes to the overall uncertainty in our analysis of benefit-cost.
9-47
-------
Draft Regulatory Impact Analysis
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.
Some uncertainty is also introduced when translating engineering cost into 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 projected to be very small, with a
maximum impact of less than one percent.
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 new standards went into effect.11 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
HFor this proposal, 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. 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 information that could be used in trying to estimate cost uncertainty for nonroad diesel
engine technologies.
9-48
-------
Cost-Benefit Analysis
changes observed in the market place. Table 9-19 presents a summary of results for the fuel and
vehicle rules reviewed in the paper.
Table 9-19.
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
l.lc/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 9-19 would lead us to believe that cost uncertainty is largely a risk of
overestimation by EPA. However, given the uncertainty in constructing the comparison in
Anderson and Sherwood plus the increasing sophistication of our cost analyses as time goes on,
we believe that a more conservative approach is appropriate. 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 -20. This uncertainty has virtually no impact on our estimates of the
net benefits of the proposed rule, given the large magnitude by which benefits exceed costs.
Table 9-20.
Estimated Uncertainty for Social Cost of Proposal
Year
2010
2020
2030
Social Cost Estimate
$0.26 billion
$1.4 billion
$1.5 billion
Uncertainty Range (-20 to +20 percent)
$0.21 -$0.31 billion
$1.1 -$1.6 billion
$1.2 -$1.8 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
9-49
-------
Draft Regulatory Impact Analysis
aspects of emissions estimations. Uncertainty exists in the estimates of emissions from the
nonroad sources affected by this proposal, 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 proposal 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.
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
modeling of nonroad emissions, please refer to the discussions in Chapter 3 of this RIA.
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. In an attempt to estimate the effect of a reasonable
amount of uncertainty, we have performed an analysis of the effect of a plus or minus five
percent change in the amount of emission reduction produced by our proposal. Table 9-21
presents the results of this analysis for 2030 (where the largest effect would be seen).
9-50
-------
Cost-Benefit Analysis
Table 9-21.
Estimated Effect of Emissions Uncertainty on 2030 Benefits Estimates
Case Examined
-5%-+5%forNOx
-5% to + 5% for SO2
-5% to +5% for PM
-5% to +5% for all emissions
Range
$80
$80
$78
$76
of 2030 Benefit
-$81 billion
-$81 billion
- $83 billion
- $85 billion
The effect of this analysis shows the final benefit value changing a maximum of the full
five percent sensitivity to a value of less than one percent, depending on which pollutant or
pollutants were affected. In the real world, each of these three pollutants would not necessarily
have the same uncertainty or see errors in the same direction at the same time.
9-51
-------
Draft 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. Proposed Nonroad 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, 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.
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.
9-52
-------
Cost-Benefit Analysis
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.
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
9-53
-------
Draft Regulatory Impact Analysis
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. MedCare. 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 I.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 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.
9-54
-------
Cost-Benefit Analysis
Evans, William N., and W. Kip Viscusi. 1993. Income Effects and the Value of Health. Journal
of Human Resources 28(3):497-518.
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. HI. 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 Dlinios 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(ll):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
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 (ffic). 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.
9-55
-------
Draft 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.
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 Particulate 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 J Epidemiol 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.
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., 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, 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.
9-56
-------
Cost-Benefit Analysis
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, Christian! DC. 2001. Association of
heart rate variability with occupational and environmental exposure to paniculate 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, 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, 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.
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).
9-57
-------
Draft Regulatory Impact Analysis
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.
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 Med 54(8): 53 5-40.
Pope, C. A. 2000. Invited Commentary: Particulate Matter-Mortality Expsoure-Response
Relations and Thresholds. American Journal of Epidemiology, 152: 407-412.
Pope, C.A., HI, 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., HI, 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., HI, 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.
9-58
-------
Cost-Benefit Analysis
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, B.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,
Mortality and Air Pollution in the United States. Research Report No. 94, Part II. Health
Effects Institute, Cambridge MA, June 2000.
Schwartz, J., 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 Manag Assoc. 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.
9-59
-------
Draft Regulatory Impact Analysis
Cooperative Institute for Research in the Atmosphere, Colorado State University; Fort
Collins, CO July.
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 JRespir Crit Care Med. 156(3
Pt 1): 787-93.
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 H. 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 Paniculate Matter: Assessment of Scientific and Technical Information.
9-60
-------
Cost-Benefit 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.
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
9-61
-------
Draft Regulatory Impact Analysis
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., 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.
Weisel, C.P., R.P. Cody and PJ. 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.
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.
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.
9-62
-------
APPENDIX 9A: Benefits Analysis of Modeled Preliminary Control Option
9A.1 Summary of Emissions Inventories and Modeled Changes in Emissions from Nonroad
Engines 9-74
9A.2 Air Quality Impacts 9-77
9A.2.1 PM Air Quality Estimates 9-78
9A.2.1.1 Modeling Domain 9-79
9A.2.1.2 Simulation Periods 9-80
9A.2.1.3 Model Inputs 9-83
9A.2.1.4 Converting REMSAD Outputs to Benefits Inputs 9-83
9A.2.1.5 PM Air Quality Results 9-84
9A.2.2 Ozone Air Quality Estimates 9-87
9A.2.2.1 Modeling Domain 9-88
9A.2.2.2 Simulation Periods 9-90
9A.2.2.3 Converting CAMx Outputs to Full-Season Profiles for Benefits Analysis
9-90
9A.2.2.4 Ozone Air Quality Results 9-91
9A.2.3 Visibility Degradation Estimates 9-95
9A.2.3.1 Residential Visibility Improvements 9-96
9A.2.3.2 Recreational Visibility Improvements 9-97
9A.3 Benefit Analysis- Data and Methods 9-99
9A.3.1 Valuation Concepts 9-100
9A.3.2 Growth in WTP Reflecting National Income Growth Over Time 9-102
9A.3.3 Methods for Describing Uncertainty 9-105
9A.3.4 Demographic Projections 9-109
9A.3.5 Health Benefits Assessment Methods 9-111
9A.3.5.1 Selecting Concentration-Response Functions 9-111
9A.3.5.2 Uncertainties Associated with Concentration-Response Functions . . . 9-127
9A.3.5.3 Baseline Health Effect Incidence Rates 9-130
9A.3.5.4 Accounting for Potential Health Effect Thresholds 9-136
9A.3.5.5 Selecting Unit Values for Monetizing Health Endpoints 9-138
9A.3.5.6 Unquantified Health Effects 9-155
9A.3.6 Human Welfare Impact Assessment 9-157
9A.3.6.1 Visibility Benefits 9-157
9A.3.6.2 Agricultural, Forestry and other Vegetation Related Benefits 9-160
9A.3.6.3 Benefits from Reductions in Materials Damage 9-162
9A.3.6.4 Benefits from Reduced Ecosystem Damage 9-163
9A.4 Benefits Analysis—Results 9-163
9-63
-------
Draft Regulatory Impact Analysis
This appendix details the models and methods used to generate the benefits estimates
from which the benefits of the proposed 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
Clear Skies Act (U.S. EPA, 2002). 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.
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 EPA's approach for estimating the health benefits of
regulations designed to reduce concentrations of airborne particulate matter (PM).
In its report, the NAS said that EPA has generally used a reasonable framework for
analyzing the health benefits of PM-control measures. It recommended, however, that the
9-64
-------
Cost-Benefit Analysis
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 five 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 proposed 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;
- 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.
Although the NAS made a number of recommendations for improvement in EPA's
approach, it found that the studies selected by EPA for use in its benefits analysis were generally
reasonable choices. In particular, the NAS agreed with 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 should be evaluated by the Agency.
EPA has addressed many of the NAS comments in our analysis of the proposed rule. We
provide benefits estimates for each year over the rule implementation period for a wide range of
regulatory alternatives, in addition to our proposed emission control program. We use the
estimated time path of benefits and costs to calculate the net present value of benefits of the rule.
In the RIA, we provide baseline statistics for air emissions, air quality, population, and health
outcomes. We have examined how our benefits estimates might be impacted by expanding the
age ranges to which epidemiological studies are applied, and we have added several new health
endpoints, including non-fatal heart attacks, which are supported by both U.S. studies and studies
conducted in Europe. We have also improved the documentation of our methods and provided
additional details about model assumptions.
Several of the NAS recommendations addressed the issue of uncertainty and how the
Agency can better analyze and communicate the uncertainties associated with its benefits
assessments. In particular, the Committee expressed concern about the Agency's reliance on a
single value from its analysis and suggested that EPA develop a probabilistic approach for
analyzing the health benefits of proposed regulatory actions. The Agency agrees with this
suggestion and is working to develop such an approach for use in future rulemakings. EPA plans
to hold a meeting of its Science Advisory Board (SAB) in early Summer 2003 to review its plans
for addressing uncertainty in its analyses. Our likely approach will incorporate short-term
elements intended to provide interim methods in time for the final Nonroad rule to address
uncertainty in important analytical parameters such as the concentration-response relationship for
PM-related premature mortality. Our approach will also include longer-term elements intended
9-65
-------
Draft Regulatory Impact Analysis
to provide scientifically sound, peer-reviewed characterizations of the uncertainty surrounding a
broader set of analytical parameters and assumptions, including but not limited to emissions and
air quality modeling, demographic projections, population health status, concentration-response
functions, and valuation estimates.
Our primary approach, generating our Base Estimate is a peer-reviewed method
developed for previous risk and benefit-cost assessments carried out by the Environmental
Protection Agency. It is the method used in the regulatory assessments of the Heavy Duty
Diesel and Tier n (light duty engine) Rules and the Section 812 Report to Congress. Following
the approach of these earlier assessments, along with the results of the Base Estimate, we present
various sensitivity analyses on the Base Estimate that alter select subsets of variables, such as the
concentration-response function for premature mortality.
Many of the techniques applied in analyzing the benefits of the proposed rule have also
been reviewed by EPA's independent Science Advisory Board (SAB). We have relied heavily
on the advice of the SAB in determining the health and welfare effects considered in the benefits
analysis and in establishing the most scientifically valid measurement and valuation techniques.
Since the publication of the final HD Engine/Diesel Fuel RIA, we have updated some of the
assumptions and methods used in our analysis to reflect SAB and NAS recommendations, as
well as advances in data and methods in air quality modeling, epidemiology, and economics.
Changes to the methodology are described fully in the following sections and in the benefits
technical support document (Abt Associates, 2003) and include the following:
Demographic/population data:
We have updated our base population data from 1990 to Census 2000 block level data
- We have developed future year population projections based on Woods and Poole
Economics, Inc. 2001 Regional Projections of county population.
Health effects incidence/prevalence data:
We have updated county-level mortality rates (all-cause, non-accidental,
cardiopulmonary, lung cancer, COPD) from 1994-1996 to 1996-1998 using the CDC
Wonder database.
- We have updated hospitalization rates from 1994 to 1999 and switched from national
rates to regional rates using 1999 National Hospital Discharge Survey results.
We have developed regional emergency room visit rates using results of the 2000
National Hospital Ambulatory Medical Care Survey.
- We have updated prevalence of asthma and chronic bronchitis to 1999 using results of
the National Health Interview Survey (HIS), as reported by the American Lung
Association (ALA), 2002
We have developed non-fatal heart attack incidence rates based on National Hospital
Discharge Survey results.
We have updated the national acute bronchitis incidence rate using HIS data as
reported in ALA, 2002, Table 11.
We have updated the work loss days rate using the 1996 HIS data, as reported in
Adams, et al. 1999, Table 41
9-66
-------
Cost-Benefit Analysis
We have developed school absence rates using data from the National Center for
Education Statistics and the 1996 HIS, as reported in Adams, et al., 1999, Table 46.
We have developed baseline incidence rates for respiratory symptoms in asthmatics,
based on epidemiological studies (Ostro et al. 2001; Vedal et al. 1998; Yu et al; 2000;
McConnell et al., 1999; Pope et al., 1991).
Concentration-Response Functions
- We have added several new endpoints to the analysis, including
> hospital admissions for all cardiovascular causes in adults 20-64, PM
(Moolgavkar et al., 2000)
> ER visits for asthma in children 0-18, PM (Norris et al., 1999)
> non-fatal heart attacks, adults over 30, PM (Peters et al, 2001)
> school loss days, Ozone (Gilliland et al, 2001; Chen et al, 2000)
> hospital admissions for all respiratory causes in children under 2, Ozone (Burnett
etal., 2001)
- We have changed the sources for concentration-response functions for hospital
admission for pneumonia, COPD, and total cardiovascular from Samet et al, 2000 (a
PM10 study), to Lippmann et al, 2000 and Moolgavkar, 2000 (PM2 5 studies)
We have added a separate table with incidence estimates for the asthmatic
subpopulation, based on studies by Ostro et al, 2001; Yu et al, 2000; Vedal et al,
1998; Pope et al., 1991; Ostro et al., 1991; and McConnell et al., 1999.
- We have added a separate table showing age specific impacts, as well as the impact of
extending the population covered by a C-R function to additional ages, i.e. extending
lower respiratory symptoms to all children, rather than only children aged 7-14.
Valuation of Changes in Health Outcomes:
We have developed a value for school absence days by determining the proportion of
families with two working families, multiplying that proportion by the number of
school loss days, and multiplying the resulting number of school loss days resulting in
a parent staying home (or requiring purchase of a caregivers time) by the average
daily wage.
- We have developed age-specific values for non-fatal heart attacks using cost-of-
illness methods, based on direct cost estimates reported in Wittels et al (1990) and
Russell et al (1998) and lost earnings estimates reported in Cropper and Krupnick
(1990). These estimates include expected medical costs in the 5 years following a
myocardial infarction, as well as the lost earnings over that period.
We have corrected a previous error in the valuation of acute bronchitis episodes.
Previously, episodes were valued as if they lasted only a single day. We have
corrected this value to account for multiday duration of episodes.
Air Quality:
- PM air quality modeling results are used to develop adjustment factors which will be
applied to ambient monitoring data to estimate future base and control ambient PM
9-67
-------
Draft Regulatory Impact Analysis
levels (consistent with past practice for ozone modeling). This change is due to the
recent availability of sufficient ambient PM2 5 monitoring data.
We have changed the ozone air quality model from the Urban Airshed Model to
CAM-X, modeled using 30 episode days in 1995 for the Eastern U.S. and 19 episode
days in 1996 for the Western U.S. (note that in the HD Engine/Diesel Fuel analysis,
we did not use ozone modeling results for the Western U.S.). For both Eastern and
Western domains, a nested grid structure was used, with a 36 km outer resolution, and
a 12 km inner resolution over urban areas.
We have updated the PM air quality model, REMSAD, to version 7.3, run at 36 km
grid resolution.
In addition to the above changes, for the proposed rule, the Agency has used an interim
approach that shows the impact of several important alternative assumptions about the estimation
and valuation of reductions in premature mortality and chronic bronchitis. This approach, which
was developed in the context of the Agency's Clear Skies analysis, provides an alternative
estimate of health benefits using the time series studies in place of cohort studies, as well as
alternative valuation methods for mortality and chronic bronchitis risk reductions.
All such benefit estimates are subject to a number of assumptions and uncertainties,
which are discussed throughout the appendix. For example key assumptions underlying the Base
and Alternative Estimates for the 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. While 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 fine
particles directly emitted from diesel engines are chemically different from fine particles
resulting from both utility sources and industrial facilities, but no clear scientific grounds exist
for supporting differential effects estimates by particle type. (3) The concentration-response
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 paniculate matter, 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 proposal.
In addition to the quantified and monetized benefits summarized above, there are 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 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
9-68
-------
Cost-Benefit Analysis
PM and ozone for which appropriate concentration-response functions are not available or which
do not provide easily interpretable outcomes (i.e. changes in forced expiratory volume (FEV1)).
As a result, both the Base and Alternative monetized benefits may underestimate the total
benefits attributable to the preliminary control options.
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.
Table 9A.1. Summary of Results: Estimated Benefits
of the Modeled Preliminary Control Option
Estimation Method
Total BenefitsA B
(Billions 2000$)
2020
2030
Base Estimate0:
Using a 3% discount rate
Using a 7% discount rate
$52+B
$49+B
$92+B
$87+B
Alternative Estimate0:
Using a 3% discount rate
Using a 7% discount rate
$11+B
$11+B
$19+B
$20+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, unquantified
benefits are indicated with a "B" to represent the sum of additional monetary benefits and disbenefits. A detailed listing of
unquantified health and welfare effects is provided in Table 9A-2.
B Results reflect the use of two different discount rates; a 3% rate which is recommended by EPA's Guidelines for Preparing
Economic Analyses (US EPA, 2000c), and 7% which is recommended by OMB Circular A-94 (OMB, 1992). Results are rounded to
two significant digits.
c 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.
D Alternative Estimate reflects premature mortality based on application of concentration-response function derived from short-term
exposures to PM25, valuation using the value of statistical life-years saved apporach, assumption of 0.5 life years saved for each
COPD related premature mortality avoided and 5 years for all other causes of death, and a cost-of-illness approach for valuing
chronic bronchitis incidence.
9-69
-------
Draft Regulatory Impact 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)
i 1
Concentration-
Response Functions
Incidence and
Prevalence Rates
Population and
Demographic Data
i 1
Valuation Functions
i i
GDP Projections !
Income Elasticities !
Model baseline and post-
control ambient air quality
(REMSAD, CAM-X)
Model Population Exposure
to Changes in Ambient
Concentrations
(CAPMS)
Estimate Expected
Changes in Human
Health Outcomes
(CAPMS)
Estimate Expected
Changes inWelfare
Estimate Monetary
Value of Changes in
Human Health
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
(AGSIM)
Valuation Functions
9-70
-------
Table 9A.2.
Human Health and Welfare Effects of Pollutants Affected by the Proposed Nonroad Diesel Engine Rule
Pollutant/Effect
Quantified and Monetized in Base
and Alternative EstimatesA
Quantified and/or Monetized Effects
in Sensitivity Analyses B
Unquantified Effects
Ozone/Health
Hospital admissions - respiratory
Emergency room visits for asthma
Minor restricted activity days
School loss days
Chronic Asthma0
Asthma attacks
Cardiovascular emergency room visits
Premature mortality - acute
exposures0
Acute respiratory symptoms
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
Ozone/Welfare
Decreased outdoor worker
productivity
Decreased yields for commercial
crops (selected species)
Decreased Eastern commercial forest
productivity (selected
species)
Decreased Western commercial forest productivity
Decreased Eastern commercial forest productivity
(other species)
Decreased yields for fruits and vegetables
Decreased yields for other commercial and
non-commercial crops
Damage to urban ornamental plants
Impacts on recreational demand from damaged
forest aesthetics
Damage to ecosystem functions
PM/Health
Premature mortality - long term
exposures
Bronchitis - chronic and acute
Hospital admissions - respiratory and
cardiovascular
Emergency room visits for asthma
Non-fatal heart attacks (myocardial
infarction)
Lower and upper respiratory illness
Minor restricted activity days
Work loss days
Premature mortality - short term
exposures
Asthma attacks (asthmatic population)
Respiratory symptoms (asthmatic
population)
Infant mortality
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
-------
Pollutant/Effect
Quantified and Monetized in Base
and Alternative EstimatesA
Quantified and/or Monetized Effects
in Sensitivity Analyses B
Unquantified Effects
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
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
SO,/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
Non-asthma respiratory ER visits
-------
Pollutant/Effect
NMHCs E
Health
NMHCs E
Welfare
Quantified and Monetized in Base
and Alternative EstimatesA
Quantified and/or Monetized Effects
in Sensitivity Analyses B
Unquantified Effects
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)
Direct toxic effects to animals
Bioaccumlation in the food chain
Reduction in odors
A Primary quantified and monetized effects are those included when determining the primary estimate of total monetized benefits of the Noroad Diesel Engine rule. See Section
C-2 for a more complete discussion of presentation of benefits estimates.
B Alternative quantified and/or monetized effects are those presented as alternatives to the primary estimates or in addition to the primary estimates, but not included in the
primary estimate of total monetized benefits.
c 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, et al, 1999).
D Premature mortality associated with ozone is not separately included in the primary analysis. It is assumed that the American Cancer Society (ACS)/ Krewski, et al., 2000 C-R
function we use for premature mortality captures both PM mortality benefits and any mortality benefits associated with other air pollutants (ACS/ Krewski, et al., 2000).
E All non-methane hydrocarbons (NMHCs) listed in the table are also hazardous air pollutants listed in the Clean Air Act.
-------
Draft Regulatory Impact 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 stages:
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 proposed 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 PM25 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
regarding the technical feasibility of the standards, and has revised the control scenario. 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 proposed
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 proposed rule. The
emission sources and the basis for current and future-year inventories are listed in Table 9A-5.
9-74
-------
Cost-Benefit Analysis
Table 9A-3
Summary of Baseline Emissions for Preliminary Nonroad Engine Control Options
Source
Pollutant Emissions (tons)
NOX
S02
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
9-75
-------
Draft Regulatory Impact 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 SOX and PM2 5 reductions from recreational marine diesel engines, commercial marine diesel
engines, and locomotives due to control of diesel fuel sulfur levels.
9-76
-------
Cost-Benefit 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 particulate 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.
Although we expect reductions in airborne sulfur and nitrogen deposition, these air quality
impacts have not been quantified for this proposed rule nor have the associated benefits been
estimated.
9-77
-------
Draft Regulatory Impact Analysis
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 particulate 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 proposed
pollution control measures that affect ozone, PM and deposition of pollutants to the surface.a
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 proposed 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 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.
A 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.
9-78
-------
Cost-Benefit Analysis
• 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.
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.
9-79
-------
Draft Regulatory Impact Analysis
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.
9-80
-------
Cost-Benefit Analysis
Figure 9A-2
REMSAD Modeling Domain for Continental United States
Note: Gray markings define individual grid-cells in the REMSAD model.
9-81
-------
Draft Regulatory Impact Analysis
49
42
104
Figure 9A-3. Example of REMSAD 36 x 36km Grid-cells for Maryland Area
9-82
-------
Cost-Benefit 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 C-R functions of the benefits analysis.13 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.
BBased on AIRS, there were 1,071 FRM PM monitors with valid data as defined as more than 11 observations
per season.
9-83
-------
Draft Regulatory Impact Analysis
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.c 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 PM25 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 (PM2 5) with little or no reduction in
coarse particles (PM10less PM25). Therefore, the observed changes in PM10are composed
primarily of changes in PM2 5. 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
PM25 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 proposed rule generates greater absolute air quality improvements in more
populated, urban areas.
GThis 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).
9-84
-------
Cost-Benefit Analysis
Table 9A-6.
Summary of Base Case PM Air Quality
and Changes Due to Nonroad Engine/Diesel Fuel Standards: 2020 and 2030
Statistic
2020
Base Case
Change"
Percent
Change
PM2.5 (ug/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%
2030
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 reductions in annual mean PM2 5 concentrations resulting from the proposed 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.
9-85
-------
Draft Regulatory Impact Analysis
Table 9A-7
Distribution of PM2.5 Air Quality Improvements Over Population Due to Nonroad Engine/Diesel
Fuel Standards: 2020 and 2030
Change in Annual Mean PM2 5
Concentrations (jug/m3)
0 > A PM2 5 Cone < 0.25
0.25 > A PM2.5 Cone < 0.5
0.5 > A PM2 5 Cone < 0.75
0.75 > A PM2.5 Cone < 1.0
1.0>APM25Conc < 1.25
1.25>APM25Conc < 1.5
1.5> APM25Conc < 1.75
APM25Conc> 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%
2030 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%
The change is defined as the control case value minus the base case value.
Table 9 A-7 provides additional insights on the changes in PM air quality resulting from
the proposed 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 PM25 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
proposed 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.
9-86
-------
Cost-Benefit Analysis
Table 9A-8.
Summary of Absolute and Relative Changes in PM Air Quality Due to Nonroad
Engine/Diesel Fuel Standards: 2020 and 2030
Statistic
2020
PM2 5 Annual Mean
2030
PM2 5 Annual Mean
Absolute Change from Base Case (fig/m3)"
Minimum
Maximum
Average
Median
Population- Weighted Average c
-0.02
-1.36
-0.20
-0.19
-0.42
-0.02
-2.03
-0.28
-0.26
-0.59
Relative Change from Base Case (%)b
Minimum
Maximum
Average
Median
Population- Weighted Average c
-0.33%
-12.24%
-2.44%
-2.33%
-3.28%
-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 proposed 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., OTAG, On-highway Tier-2, and HD Engine/Diesel Fuel).
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
9-87
-------
Draft Regulatory Impact Analysis
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,
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.
-------
Cost-Benefit Analysis
Figure 9A-4. CAMx Eastern 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.
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.
9-89
-------
Draft 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.d'e 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., the Criteria Air Pollutant Modeling System (CAPMS).
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
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
D 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.
EBased 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.
9-90
-------
Cost-Benefit Analysis
daily average. f'g 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 proposed 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 proposed rule results in average
reductions of roughly 1.5 percent, or between 0.57 to 0.52 ppb, in the daily average ozone
concentration 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.
FThe 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.
GThis 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).
9-91
-------
Draft Regulatory Impact Analysis
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.
9-92
-------
Table 9A-9.
Summary of CAMx Derived Ozone Air Quality Metrics Due to Nonroad Engine/Diesel Fuel Standards
for Health Benefits EndPoints: Eastern U.S.
Statistic a
2020
Base Case
Change
Percent Change
2030
Base Case
Change b
Percent Change
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 '
Maximum c
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 c
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 c
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%
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.
-------
Table 9A-10.
Summary of CAMx Derived Ozone Air Quality Metrics Due to Nonroad Engine/Diesel Fuel Standards
for Health Benefits EndPoints: Western U.S.
Statistic "
2020
Base Case
Change *
Percent Change
2030
Base Case
Change *
Percent Change
Daily 1 -Hour Maximum Concentration (ppb)
Minimum c
Maximum c
Average
Median
Population-Weighted Average d
Daily 5-Hour Average Concentration (ppb)
Minimum c
Maximum c
Average
Median
Population-Weighted Average d
27.48
201.28
47.02
46.10
63.80
24.20
163.41
41.11
40.48
53.56
-0.01
4.87
-0.62
-0.56
0.34
-0.01
2.55
-0.52
-0.40
0.45
-0.03%
2.42%
-1.31%
-1.19%
0.54%
-0.04%
1.56%
-1.26%
-1.04%
0.84%
27.48
208.02
47.04
46.06
64.23
24.21
168.89
41.13
40.46
53.89
-0.01
6.26
-0.97
-0.66
0.38
-0.01
6.04
-0.82
-0.69
0.55
-0.05%
3.01%
-2.07%
-1.43%
0.58%
-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
Daily 12-Hour Average Concentration (ppb)
Minimum c
Maximum c
Average
Median
Population-Weighted Average d
23.77
157.49
40.68
40.11
51.96
22.13
140.48
39.30
38.85
47.68
-0.01
1.33
-0.51
-0.36
0.46
0.31
1.65
-0.48
-0.38
0.49
-0.04%
0.84%
-1.25%
-1.03%
0.88%
1.39%
1.18%
-1.23%
-0.97%
1.02%
23.77
161.92
40.69
40.09
52.29
22.09
143.59
39.31
38.82
47.99
-0.01
5.94
-0.81
-0.72
0.57
0.44
1.78
-0.77
-0.58
0.63
-0.05%
3.67%
-1.99%
-1.79%
1.10%
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%
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 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 Nonroad Engine/Diesel Fuel Standards
for Welfare Benefits Endpoints: 2020 and 2030
Statistic "
2020
Base Case
Change b
Percent
Change b
2030
Base Case
Change b
Percent
Change b
Eastern U.S.
Sum06 (ppm)
Minimum c
Maximum c
Average
Median
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%
-
* 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
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.
9-95
-------
Draft Regulatory Impact Analysis
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 proposed 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 Nonroad Diesel
Engine Standards: 2020 and 2030
Improvements in Visibility a
(annual average deciviews)
0 > A Deciview < 0.2
0.2 > A Deciview < 0.4
0.4 > A Deciview < 0.6
0.6 > A Deciviews 0.8
0.8 > A Deciviews 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%
a The change is defined as the control case deciview level minus the base case deciview level.
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),
9-96
-------
Cost-Benefit Analysis
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 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.
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.
9-97
-------
Draft Regulatory Impact 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.
9-98
-------
Cost-Benefit Analysis
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), 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).
9-99
-------
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 paniculate 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. 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, 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
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
-------
Cost-Benefit Analysis
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
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.11 Stated preference valuation studies form the basis for valuing a number of health
HConcerns 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
9-101
-------
Draft Regulatory Impact Analysis
and welfare endpoints, including the value of 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 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
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 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
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.
'Income elasticity is a common economic measure equal to the percentage change in WTP for a one percent
change in income.
9-102
-------
Cost-Benefit Analysis
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
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.
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 Growth"*
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
9-103
-------
Draft Regulatory Impact Analysis
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 Neumann (1999)
for the years 1990 to 2010k. We use projections of real GDP (in chained 1996 dollars) provided
by Standard and Poor's1 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 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.
JU.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)
KU.S. Bureau of Economic Analysis, Table 2A (1992$). (Available on the internet at
http://www.bea.doc. gov/bea/dn/0897nip2/tab2a.htm) 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.
LStandard and Poor's. 2000. "The U.S. Economy: The 25 Year Focus." Winter.
9-104
-------
Cost-Benefit Analysis
Table 9A-16. Adjustment Factors Used to Account for Projected Real Income GrowthA
Benefit Category
Minor Health Effect
Severe and Chronic Health Effects
Premature Mortality
Visibility
2020
1.084
1.299
1.262
1.704
2030B
1.092
1.329
1.287
1.787
A Based on elasticity values reported in Table 9A-11, US Census population projections, and projections of real gross domestic
product per capita.
B 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."1 This analysis is no exception. As outlined
both in this and preceding chapters, there are many inputs used to derive the final estimate of
benefits, including emission inventories, air quality models (with their associated parameters and
inputs), epidemiological estimates of concentration-response (C-R) functions, 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;
- Variability in estimated relationships, such as C-R functions, introduced through
differences in study design and statistical modeling;
- Errors in measurement and projection for variables such as population growth rates;
M 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.
9-105
-------
Draft Regulatory Impact Analysis
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 the 26 studies
that underlie the Base Estimate represent a quantifiable uncertainty.
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, 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. For example, the C-R function used to estimate
avoided premature mortality has an associated standard error which represents the sampling error
around the pollution coefficient in the estimated C-R function. It is possible to report a
confidence interval around the estimated incidences of avoided premature mortality based on this
standard error. However, this would omit the contribution of air quality changes, baseline
population incidences, projected populations exposed, and transferability of the C-R function to
diverse locations to uncertainty about premature mortality. Thus, a confidence interval based on
the standard error gives a misleading picture about the overall uncertainty in the estimates.
Information on the uncertainty surrounding particular C-R and valuation functions is provided in
the benefits TSD for this RIA (Abt Associates, 2003). But, this information should be
interpreted within the context of the larger uncertainty surrounding the entire analysis.
Our approach to characterizing model uncertainty is to present a primary estimate of the
benefits, based on the best available scientific literature and methods, and to then provide
sensitivity analyses to illustrate the effects of uncertainty about key analytical assumptions. Our
9-106
-------
Cost-Benefit Analysis
analysis of the preliminary control options has not included formal integrated uncertainty
analyses, although we have conducted several sensitivity tests and have analyzed a full
Alternative Estimate 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 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. At this time, we simply demonstrate the sensitivity of our benefits
results to key parameters which may be uncertain. Sensitivity estimates are presented 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
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.
9-107
-------
Draft Regulatory Impact Analysis
Table 9A-17. Primary Sources of Uncertainty in the Benefit Analysis
1. Uncertainties Associated With Concentration-Response Functions
The value of the ozone- or PM-coefficient in each C-R function.
Application of a single C-R function to pollutant changes and populations in all locations.
Similarity of future year C-R relationships to current C-R relationships.
Correct functional form of each C-R relationship.
Extrapolation of C-R relationships beyond the range of ozone or PM concentrations observed in the study.
Application of C-R relationships 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 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 and forestry products are uncertain.
7. Uncertainties Associated With Aggregation of Monetized Benefits
- Health and welfare benefits estimates are limited to the available C-R functions. Thus, unquantified or
unmonetized benefits are not included.
9-108
-------
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.
Figure 9A-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.
-------
400.0
350.0
300.0
o 250.0
c
•-S 200.0
(o
Q.
O
Q.
1 150.0
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% •§
0)
-- 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
-------
Cost-Benefit Analysis
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. 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." 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. EPA's Criteria
Documents for ozone and PM list numerous health effects known to be linked to ambient
concentrations of these pollutants (US EPA, 1996a and 1996b). As illustrated in Figure 9A. 1,
quantification of health impacts requires several inputs, including concentration-response
functions, 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 C-R functions 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.
9A.3.5.1 Selecting Concentration-Response Functions
Quantifiable health benefits of the modeled preliminary control options 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, non-fatal heart attacks, asthma emergency room
visits, chronic bronchitis, acute bronchitis, upper and lower respiratory symptoms, and work loss
N US Bureau of Economic Analysis, Table 2A (1992$). (Available on the internet at
http://www.bea.doc.go^ea/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.
0 Standard and Poor's. 2000. "The U.S. Economy: The 25 Year Focus." Winter 2000.
9-111
-------
Draft Regulatory Impact Analysis
days.p Health effects related to both PM and ozone include hospital admissions and minor
restricted activity days.
We relied on the available published scientific literature to ascertain the relationship
between particulate matter 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.
Recently, 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 Chapter 9, 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 EPA and to CASAC (Greenbaum, 2002), researchers working on the
National Morbidity, Mortality, and Air Pollution Study (NMMAPS) found problems in the
default "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. These and other scientists have begun to reanalyze the results of
several important time series studies with alternative approaches that address these issues and
have found a downward revision of some results. For example, the mortality risk estimates for
short-term exposure to PM10 from NMMAPS were overestimated (this study was not used in this
benefits analysis of fine particle effects). However, both the relative magnitude and the
direction of bias introduced by the convergence issue is case-specific. In most cases, the
concentration-response relationship may be overestimated; in other cases, it may be
underestimated. The preliminary reanalyses of the mortality and morbidity components of
NMMAPS suggest that analyses reporting the lowest relative risks appear to be affected more
greatly by this error than studies reporting higher relative risks (Domenici et al, 2002).
p Some evidence has been found linking both PM and ozone exposures with premature mortality. The SAB has
raised concerns that mortality-related benefits of air pollution reductions may be overstated if separate pollutant-
specific estimates, some of which may have been obtained from models excluding the other pollutants, are
aggregated. In addition, there may be important interactions between pollutants and their effect on mortality (EPA-
SAB-Council-ADV-99-012, 1999).
Because of concern about overstating of benefits and because the evidence associating mortality with exposure
to PM is currently stronger than for ozone, only the benefits related to the long-term exposure study (ACS/Krewkski,
et al, 2000) of mortality are included in the total primary benefits estimate. The benefits associated with ozone
reductions are presented as a sensitivity analysis in Appendix 9-B but are not included in the estimate of total
benefits.
9-112
-------
Cost-Benefit Analysis
During the compilation of the draft Air Quality Criteria Document, 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 in both the Base and
Alternative Estimates, reduced lower respiratory symptoms in both the Base and Alternative
Estimates, and reduced premature mortality due to short-term PM exposures in the Alternative
Estimate. While resolution of these issues is likely to take some time, the preliminary results
from ongoing reanalyses of some of the studies used in our analyses (Dominici et al, 2002;
Schwartz and Zanobetti, 2002; Schwartz, personal communication 2002) suggest a more modest
effect of the S-plus error than reported for the NMMAPS PM10 mortality study. In December
2002, a number of researchers submitted reanalysis reports, and the HEI is currently coordinating
review of these reports by a peer review panel. The final report on these reanalyses is expected
by the end of April 2003, and the results will be incorporated in the fourth external review draft
of the Criteria Document that will be released in summer 2003. While we wait for further
clarification from the scientific community, we have chosen not to remove these results from the
Nonroad Diesel benefits estimates, nor have we elected to apply any interim adjustment factor
based on the preliminary reanalyses EPA will continue to monitor the progress of this concern,
and make appropriate adjustments as further information is made available.
While a broad range of serious health effects have 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 (US 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: (i) the possibility of double counting (such as hospital admissions for specific
respiratory diseases); (ii) uncertainties in applying effect relationships based on clinical studies to
the affected population; or (iii) a lack of an established C-R relationship.
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 which make 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 due to 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 C-R relationship.
When several estimated C-R relationships between 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 TSD provides details of the procedures used to
combine multiple C-R 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 studies 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
9-113
-------
Draft Regulatory Impact Analysis
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.q Pooled C-R functions are used to estimate hospital admissions related
to PM and asthma-related emergency room visits related to ozone. For more details on methods
used to pool incidence estimates, see the benefits TSD (Abt Associates, 2003).
Concentration-response relationships between a pollutant and a given health endpoint are
applied consistently across all locations nationwide. This applies to both C-R relationships
defined by a single C-R function and those defined by a pooling of multiple C-R functions.
Although the C-R relationship may, in fact, vary from one location to another (for example, due
to differences in population susceptibilities or differences in the composition of PM),
location-specific C-R functions are generally not available.
The specific studies from which C-R functions for calculating the Base and Alternative
estimates are drawn are included in Table 9A-14. A complete discussion of the C-R functions
used for this analysis and information about each endpoint are contained in the benefits TSD for
this RIA (Abt Associates, 2003). Basic information on each endpoint is presented below.
QThe 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, and therefore there may be a
number of different underlying pollutant coefficients.
9-114
-------
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 PM
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 tecnically appropriate measures of benefits would be based on C-R functions that cover the
entire sensitive population, but allow for heterogeneity across age or other relevant demographic
factors. In the absence of C-R functions specific to age, sex, preexisting condition status, or other
relevant factors, it may be appropriate to select C-R functions that cover the broadest popuation, 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 C-R 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, C-R functions based on PM2 5 are preferred to PM10 because reductions in emissions
from diesel engines are expected to reduce fine particles and not have much impact on coarse particles.
Where PM2 5 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.
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. Including emergency room visits in a benefits analysis that
already considers hospital admissions, for example, will result in double counting of some benefits if
the category "hospital admissions" includes emergency room visits.
9-115
-------
Draft Regulatory Impact Analysis
Table 9A-19.
Endpoints and Studies Used to Calculate Total Monetized Health Benefits
1 ml point
Premature Mortality
Base - Long-term exposure
Alternative - Short-term exposure8
Chronic Illness
Chronic Bronchitis
Non-fatal Heart Attacks
Hospital Admissions
Respiratory
Cardiovascular
Asthma-Related ER Visits
Other Health Endpoints
Acute Bronchitis
Upper Respiratory Symptoms
Lower Respiratory Symptoms
Work Loss Days
School Absence Days
Worker Productivity
Minor Restricted Activity Days
Pollutant
PM25
PM25
PM25
PM25
Ozone
Ozone
PM25
PM25
PM25
PM25
PM25
PM25
Ozone
PM25
PM25
PM10
PM25
PM25
Ozone
Ozone
PM2 5, Ozone
Study
Krewski, et al. (2000)A
Schwartz et al. (1996) adjusted using ratio of distributed lag to
single day coefficients from Schwartz et al. (2000)
Abbey, et al. (1995)
Peters etal. (2001)
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 (2000) - ICD 490-496 (COPD)
Lippman et al. (2000) - ICD 490-496 (COPD)
Moolgavkar (2000) - ICD 490-496 (COPD)
Lippman et al. (2000) - ICD 480-486 (pneumonia)
Sheppard, et al. (1999) - ICD 493 (asthma)
Pooled estimate:
Moolgavkar (2000) - ICD 390-429 (all cardiovascular)
Lippman et al. (2000) - 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), Stieb
etal. (1996)
Norris et al. (1999)
Dockeryetal. (1996)
Pope etal. (1991)
Pooled estimate: Schwartz et al. (1994); 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)
Study Population
>29 years
all ages
> 26 years
Adults
> 64 years
< 2 years
> 64 years
20-64 years
> 64 years
< 65 years
> 64 years
20-64 years
All ages
0-1 8 years
8- 12 years
Asthmatics, 9-11
years
7- 14 years
18-65 years
9-10 years
6-11 years
Outdoor workers, 18-
65
18-65 years
1 Estimate derived from Table 31, PM2.5(DC), All Causes Model (Relative Risk =1.12 for a 24.5 |o.g/m3 increase in mean PM25).
9-116
-------
Cost-Benefit Analysis
9-117
-------
Draft Regulatory Impact Analysis
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 mortality risk estimates
from these epidemiological studies, the serious nature of the effect itself, and the high monetary
value ascribed to prolonging life make mortality risk reduction the most important health
endpoint quantified in this analysis. Because of the importance of this endpoint and the
considerable uncertainty among economists and policymakers as to the appropriate way to value
reductions in mortality risks, both a base and an alternative estimate are provided. As in the
Kunzli et al. (2000) analysis, we focus on the prospective cohort long-term exposure studies in
deriving the C-R function for our base estimate of premature mortality.
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). The alternative estimate is based on time-
series studies demonstrating the effect of short-term exposures. This section discusses some of
the issues surrounding the estimation of premature mortality.
Base Estimate
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
9-118
-------
Cost-Benefit Analysis
"American Cancer Society or ACS study" (Pope et al., 1995); these studies have found
consistent relationships between fine particle indicators and premature mortality across multiple
locations in the U.S. 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
mortality in men. Results from this cohort, however, have been inconsistent and the air quality
results are not geographically representative of most of the US. More recently, a cohort of adult
male veterans diagnosed with hypertension has been examined (Lipfert et al., 2000). The
characteristics of this group differ from the cohorts in the ACS, Six-Cities, and 7th Day Adventist
studies with respect to income, race, and smoking status. Unlike previous long-term analyses,
this study found some associations between mortality and ozone but found inconsistent results
for PM indicators.
Given their consistent results and broad geographic coverage, the Six-City and ACS data
have been of particular importance 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 commisioned by the Health Effects Institute (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 as well as highlighting
sensitivities concerning (a) the relative impact of various pollutants, (b) the potential role of
education in mediating the association between pollution and mortality, and (c) the influence of
spatial 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 mortality risk over the years, EPA has consulted with a panel of the Science
Advisory Board. That panel recommended use of long-term prospective cohort studies in
estimating 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)." More specifically, 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 (U.S. 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 particular specification yielded a relative risk based on
changes in mean levels of PM2 5, as opposed to the specification in the original study, which
9-119
-------
Draft Regulatory Impact Analysis
reported a relative risk based on median levels1. The Krewski et al analysis also includes a
broader geographic scope than the original study (63 cities versus 50). Specifically, the relative
risk from which the Base estimate is derived is 1.12 per 24.5 |j.g/m3 for all-cause mortality
(Krewski, et al. 2000, Part II, page 173, Table 31). The SAB has recently agreed with EPA's
selection of this specification for use in analyzing mortality benefits of PM reductions
(EPA-SAB-COUNCIL-ADV-01-004, 2001).
Alternative Estimate
To reflect concerns about the more limited number of cohort studies that examine the
association between long-term exposure and mortality and the inherent limitations for drawing
conclusions regarding causality from these studies, especially the ecological measure of exposure
used, a plausible alternative to the base benefit estimate is provided. This estimate was derived
from the larger number of time-series studies, the body of which have established a likely causal
relationship between short-term measures of PM and daily mortality statistics. A particular
strength of the design of these studies for drawing conclusions about causality is the fact that
potential confounding variables such as socio-economic status, occupation, and smoking do not
vary on a day-to-day basis in an individual area. A number of multi-city and other types of
studies strongly suggest that these short term PM exposure-premature mortality relationships
cannot be explained by weather, statistical approaches, or other pollutants.
The fact that the PM-mortality coefficients from the cohort studies are far larger than the
coefficients derived from the daily time-series studies provides some evidence for an independent
chronic effect of PM pollution on health. Indeed, the Base Estimate presumes that the larger
coefficients represent a more complete accounting of mortality effects, including both the
cumulative total of short-term mortality as well as an additional chronic effect. This is, however,
not the only possible interpretation of the disparity. Various reviewers have argued that 1) the
long-term estimates may be biased high and/or 2) the short-term estimates may be biased low.
In this view, the two study types could be measuring the same underlying relationship.
With respect to possible sources of upward bias in the long-term studies, HEI reviewers
have noted that the less robust estimates based on the Six-Cities Study are significantly higher
than those based on the more broadly distributed ACS data sets. Some reviewers have also noted
that the observed mortality associations from the 1980's and 90's may reflect higher pollution
exposures from the 1950's to 1960's. Such an argument is consistent with the dramatic decrease
in PM levels over the last 50 years, as long as the relative differences in PM among the cities did
not change. Indeed, Pope et al (2002) demonstrated that the relative differences in pollution
levels among the cities was similar between the years 1979-1980 and the years 1999-1980. If the
lower PM exposures today pose disproportionately less risk than the exposures of the 1950's-
RFor policy analysis purposes, functions based on the mean air quality levels may be preferable to functions
based on the median air quality levels because changes in the mean more accurately reflect changes in peak values
than do changes in the median. Policies which affect peak PM days more than average PM days will result in a
larger change in the mean than in the median. In these cases, all else being equal, C-R functions based on median
PM2 5 will lead to lower estimates of avoided incidences of premature mortality than C-R functions based on mean
PM25.
9-120
-------
Cost-Benefit Analysis
1960's, then the base mortality estimate may be biased upwards. While this would bias estimates
based on more recent pollution levels upwards, it also would imply a truly long-term chronic
effect of pollution, at least for the higher exposure levels.
With regard to possible sources of downward bias, it is of note that the recent studies
suggest that the single day time series studies may understate the short-term effect on the order of
a factor of two (Zanobetti et al, 2002). Previous daily mortality studies (Schwartz et al., 1996)
examined the impact of PM25 on mortality on a single day or over the average of two or more
days. Although the risk estimates from the vast majority of the short-term studies include the
effects of only one or two-day exposure to air pollution, more recently, several studies have
found that the practice of examining the effects on a single day basis may significantly understate
the risk of short-term exposures (Schwartz, 2000; Zanobetti et al, 2002). These studies suggest
that the short-term risk can double when the single-day effects are combined with the cumulative
impact of exposures over multiple days to weeks prior to a mortality event. Multi-day models are
often referred to as "distributed lag" models because they assume that mortality following a PM
event will be distributed over a number of days following or "lagging" the PM event. The size of
the effect estimates from these models suggests consistency between the findings of studies that
examine premature mortality impacts of short-term and long-term exposures. Additional
research may be necessary to confirm this trend.
The United Kingdom's Committee on the Medical Effects of Air Pollution's evaluated
the various models proposed by Krewski et al., 2000 (COMEAP 2001 Annual Report Annex C).
In the judgment of the COMEAP, as published in it's "Statement on the Long-term Effects of
Particles on Mortality," it is more appropriate to develop a "range of estimates along with
comments on their confidence in them" than selecting a single estimate of possible effects. The
inclusion of an Alternative Estimate, as well as the sensitivity analyses presented in Appendix 9B
is an appropriate response to this suggestion.
These considerations provide a basis for considering an Alternative Estimate using the
most recent estimates from the wealth of time-series studies, in addition to the Base Estimate
based on the long-term cohort studies. In essence, the Alternative Estimate offers an approach to
characterizing some of the uncertainties in the relationship between premature mortality and
exposures to ambient levels of fine particles by assuming that there is no mortality effect of
chronic exposures to fine particles. Instead, it assumes that the full impact of fine particles on
premature mortality is captured using a concentration-response function relating daily mortality
to short-term fine particle levels. This will clearly provide a lower bound to the mortality impacts
of fine particle exposure, as it omits any additional mortality impacts from longer term
exposures.
There are no PM25 daily mortality studies which report numeric estimates of relative risks
from distributed lag models; only PM10 studies are available. Daily mortality C-R functions for
PM10 are consistently lower in magnitude than PM2 5-mortality C-R functions, because fine
particles are believed to be more closely associated with mortality than the coarse fraction of PM.
Given that the emissions reductions under the Nonroad Diesel Engine program result primarily in
9-121
-------
Draft Regulatory Impact Analysis
reduced ambient concentrations of PM25, use of a PM10 based C-R function results in a
significant downward bias in the estimated reductions in mortality.
The Alternative Estimate is based on a concentration- response function derived from
Schwartz et al. (1996), with an adjustment to account for recent evidence that daily mortality is
associated with particle levels from a number of previous days (Schwartz, 2000). Specifically, to
account for the full potential multi-day mortality impact of acute PM2 5 events, we use the
distributed lag model for PM10 reported in Schwartz (2000) to develop an adjustment factor
which we then apply to the PM2 5 based C-R function reported in Schwartz et al. (1996).
If most of the increase in mortality is expected to be associated with the fine fraction of
PM10, then it is reasonable to assume that the same proportional increase in risk would be
observed if a distributed lag model were applied to the PM2 5 data. The distributed lag
adjustment factor is constructed as the ratio of the estimated coefficient from the unconstrained
distributed lag model to the estimated coefficient from the single-lag model reported in Schwartz
(2000). The unconstrained distributed lag model coefficient estimate is 0.0012818 and the
single-lag model coefficient estimate is 0.0006479. The ratio of these estimates is 1.9784. This
adjustment factor is then multiplied by the estimated coefficients from the Schwartz et al. (1996)
study. There are two relevant coefficients from the Schwartz et al. (1996) study, one
corresponding to all-cause mortality, and one corresponding to chronic obstructive pulmonary
disease (COPD) mortality (separation by cause is necessary to implement the life years lost
approach detailed below). The adjusted estimates for these two C-R functions are:
All cause mortality = 0.001489 * 1.9784 = 0.002946
COPD mortality = 0.003246 * 1.9784 = 0.006422
Note that these estimates, while approximating the full impact of daily pollution levels on
daily death counts, do not capture any impacts of long-term exposure to air pollution. As
discussed earlier, EPA's Science Advisory Board, while acknowledging the uncertainties in
estimation of a PM-mortality relationship, has repeatedly recommended the use of a study that
does reflect the impacts of long-term exposure. This recommendation has been confirmed by the
recent NRC report on estimating health benefits of air pollution regulations. The omission of
long-term impacts accounts for approximately a 40 percent reduction in the estimate of avoided
premature mortality in the Alternative Estimate relative to the Base Estimate. For comparison,
an estimate calculated using the lower confidence interval of the Base estimate C-R function
coefficient would fall between these two estimates (i.e., the lower confidence interval on the RR
of 1.12 used in the Base Estimate is 1.06, translating to a coefficient estimate of 0.002).s In
sThe 40% smaller estimate is also consistent with a judgment offered by COMEAP based on its review of the
use of U.S. cohort studies in a European context. The alternative estimate of premature mortality is similar in
magnitude to their judgement of a "likely" estimate, based on a sensitivity analysis included in the Krewski et al
(2000) reanalysis of the ACS cohort data. That sensitivity analysis was based on a much smaller set of cities and
included SO2 and other variables as controls. In addition to PM2.5, this specification found a significant impact of
SO2 on mortality. SO2 and PM2.5 levels are at least somewhat correlated, so it is not clear in a multi-pollutant
9-122
-------
Cost-Benefit Analysis
summary, the alternative estimate has a technical foundation in both a plausible interpretation of
the cohort studies and the time-series studies that incorporate the longer lag periods .
Chronic bronchitis
Chronic bronchitis is characterized by mucus in the lungs and a persistent wet cough for
at least three months a year for several years in a row. Chronic bronchitis affects an estimated
five percent of the U.S. population (American Lung Association, 1999). There are a limited
number of studies that 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 U.S. Because the nonroad
standards 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 PM2 5 and new incidences
of chronic bronchitis.
Non-fatal myocardial infarctions (heart attacks)
Non-fatal heart attacks have been linked with short term exposures to PM2.5 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 PM2.5 and
non-fatal 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 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 is consistent with an increased risk of heart
attacks.
Hospital and emergency room admissions
specification how much of the SO2 effect is capturing some of the PM2.5 signal and vice versa. An appropriate
comparison would be between the COMEAP estimate and the total mortality impact of the nonroad rule predicted
using both the PM2.5 and SO2 changes. Comparing PM2.5 related mortality generated from a single pollutant
mortality function with a multi-pollutant specification ignores the implied benefits of SO2 reductions under the rule.
Thus it is likely that the COMEAP estimate would understate the total mortality impacts likely to be associated with
the proposed nonroad rule.
9-123
-------
Draft Regulatory Impact Analysis
Due to 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 will use pooled C-R
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 U.S. are asthma-related
visits.
To estimate avoided incidences of cardiovascular hospital admissions associated with
PM2.5, we use studies by Moolgavkar (2000) and Lippmann et al (2000). 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 estimated a C-R function for populations over 65,
allowing us to pool the C-R functions for this age group. Only Moolgavkar (2000) estimated a
separate C-R function for populations 20 to 64. 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. In order to avoid double counting benefits from reductions in MI when applying the
C-R function for cardiovascular hospital admissions, we first adjusted the baseline cardiovascular
hospital admissions to remove admissions for MI.
To estimate total avoided incidences of respiratory hospital admissions, we use C-R
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
hospital admissions. We use only those focusing on PM2.5. Both Moolgavkar (2000) and
Lippmann et al (2000) estimated C-R functions for COPD in populations over 65, allowing us to
pool the C-R functions for this group. Only Moolgavkar (2000) estimated a separate C-R
function for populations 20 to 64. 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 Lippmann et al (2000) estimated pneumonia, and only for the population 65 and older. In
addition, Sheppard, et al. (1999) estimated a C-R function 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.
9-124
-------
Cost-Benefit Analysis
To estimate the effects of PM air pollution reductions on asthma-related ER visits, we use
the C-R function based on 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 PM2.5. We selected the Norris et al. (1999) C-R
function 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 PM2.5 on asthma ER visits in populations under 65, although there may still be
significant impacts in the adult population under 65. Because we are estimating ER visits as
well as hospital admissions for asthma, we must avoid counting twice the ER visits for asthma
that are subsequently admitted to the hospital. To avoid double-counting, the baseline incidence
rate for ER visits is adjusted by subtracting the percentage of patients that are admitted into the
hospital.
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 C-R functions for 2 different cities
relating ozone and hospital admissions for all respiratory causes (defined as ICD codes 460-519).
These C-R functions 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 C-R functions for Minneapolis are pooled together first, and
the resulting C-R function is then pooled with the C-R 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. In order 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 a C-R function for
respiratory hospital admissions in children under two.
Minor Illnesses, Restricted Activity Days, and School/Work Loss Days
As indicated in Table 9A-1, in addition to mortality, chronic illness, and hospital
admissions, there are a number of acute health effects not requiring hospitalization that are
associated with exposure to ambient levels of ozone and PM. The sources for the C-R functions
used to quantify these effects are described below.
Around four percent of U.S. children between ages five and seventeen 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 encyclopedia1, with the exception of
T See http://www.nlm.nih.gov/medlineplus/ency/article/000124.htm. accessed January 2002
9-125
-------
Draft Regulatory Impact Analysis
cough, most acute bronchitis symptoms abate within 7 to 10 days. Incidence of episodes of acute
bronchitis in children between the ages of five and seventeen are estimated using a C-R function
developed from Dockery, et al. (1996).
Incidences of lower respiratory symptoms (i.e., wheezing, deep cough) in children aged
seven to fourteen are estimated using a C-R function developed from Schwartz, et al. (1994).
Because asthmatics have greater sensitivity to stimuli (including air pollution), children
with asthma can be more susceptible to a variety of upper respiratory symptoms (i.e., runny or
stuffy nose; wet cough; and burning, aching, or red eyes). Research on the effects of air pollution
on upper respiratory symptoms have thus focused on effects in asthmatics. Incidences of upper
respiratory symptoms in asthmatic children aged nine to eleven are estimated using a C-R
function 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 a C-R function 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.
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 a C-R function
derived from Ostro and Rothschild (1989).
The Agency is currently evaluating how air pollution related symptoms in the asthmatic
population should be incorporated into the overall benefits analysis. Clearly, 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 is included in the calculation of total benefits. However,
other endpoints, such as lower respiratory symptoms, are estimated for the total population of
children. Given the increased susceptibility of the asthmatic population, it is of interest to
9-126
-------
Cost-Benefit Analysis
understand better the specific impacts on asthmatics. We are providing a separate set of
estimated health impacts for asthmatic populations, listed it Table 9A-20, with the caveat that
these are not additive, nor can they be easily combined with other endpoints to derive total
benefits. They are provided only to highlight the potential impacts on a susceptible population.
Table 9A.20.
Studies Examining Health Impacts in the Asthmatic Population
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
PM2.5
PM2.5
PM2.5
PM10,
PM,.o
PM10
Ostroetal. (2001)
Ostroetal. (2001)
Ostroetal. (2001)
Yuetal.(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
PM2.5
PM2.5
PM2.5
PM2.5,
ozone
Popeetal. (1991)
Ostroetal. (1991)
McConnelletal. (1999)
McConnelletal. (1999)
Whittemore and Korn
(1980)
Asthmatics 9-11
Asthmatics, all ages
Asthmatics, 9-15*
Asthmatics, 9-15*
Asthmatics, all ages
9A.3.5.2 Uncertainties Associated with Concentration-Response 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 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
9-127
-------
Draft Regulatory Impact Analysis
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.
Application of C-R Relationship Nationwide
Whether this analysis estimated the C-R relationship between a pollutant and a given
health endpoint using a single function from a single study or using multiple C-R functions from
several studies, each C-R relationship was applied uniformly throughout the U.S. to generate
health benefit estimates. However, to the extent that pollutant/health effect relationships are
region-specific, applying a location-specific C-R function at all locations in the U.S. 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 C-R function to
the entire U.S. 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".
Extrapolation of C-R Relationship Across Populations
Epidemiological studies often focus on specific age ranges, either due to data availability
limitations (for example, most hospital admission data comes from Medicare records, which are
"Although we are not able to use region-specific C-R functions, 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.
9-128
-------
Cost-Benefit Analysis
limited to populations 65 and older), or to simplify data collection (for example, 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 C-R functions should be applied only to those
population with ages that strictly match the populations in the underlying epidemiological
studies. 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. We provide a set of expanded incidence estimates to show the effect of this
assumption.
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 mortality rates. However, there is much about this
relationship that is still uncertain. These uncertainties include:
Causality. A substantial number of published epidemiological studies recognize a
correlation between elevated PM concentrations and increased mortality rates;
however these epidemiological studies, by design, can not definitively prove
causation. For the analysis of the Nonroad Diesel Engine 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 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 mortality effects
independent of PM, but we do not view the evidence as conclusive at this time.
To the extent that the C-R functions we use to evaluate the preliminary control
options in fact capture 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 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 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 mortality may be mis-estimated.
9-129
-------
Draft Regulatory Impact Analysis
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 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 studies suggest 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 five-year lag structure.
This approach assumes that 25 percent of PM-related premature deaths occur in
each of the first two years after the exposure and the rest occur in equal parts
(approximately 17%) in each of the ensuing three 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 |j.g/m3 decrease in daily PM25 levels might
decrease hospital admissions by three 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
9-130
-------
Cost-Benefit Analysis
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, i.e. asthma. For both baseline incidence and prevalence data, we use age-
specific rates where available. Concentration-response 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, due to 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 is
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
working to develop methods to use these data to project future incidence rates. However, for our
primary benefits analysis of the proposed nonroad rule, we will continue to use current incidence
rates. We will examine the impact of using projected mortality rates and asthma prevalence in
sensitivity analyses.
Table 9A-2 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 three 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 25-34 year
old death count and population by one-half and then generated a population-weighted mortality
rate using data for the older age groups.
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-21 lists the baseline incidence rates and their sources for asthma symptom
endpoints. 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
9-131
-------
Draft Regulatory Impact Analysis
change in prevalence rates in future years. As noted above, we are investigating methods for
projecting asthma prevalence rates in future years.
9-132
-------
Cost-Benefit Analysis
Table 9A-21.
Baseline Incidence Rates and Population Prevalence Rates for Use in C-R Functions, General
Population
Endpoint
Mortality
Hospital izations
Asthma ER
visits
Chronic
Bronchitis
Nonfatal MI
(heart attacks)
Acute
Bronchitis
Lower
Respiratory
Symptoms
Upper
Respiratory
Symptoms
Work Loss Days
Minor
Restricted
Activity Days
School Loss
Days5
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
Annual bronchitis incidence
rate, children
Daily lower respiratory
symptom incidence among
children4
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 person5
Northeast
Midwest
South
Southwest
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
0.043
0.0012
0.3419
0.00540
0.00678
0.00492
0.02137
0.055
0.0136
0.0146
0.0142
0.0206
Source1
CDC Wonder (1996- 1998)
1999 NHDS public use data files2
2000 NHAMCS public use data
files3; 1999 NHDS public use data
files2
1999 HIS (American Lung
Association, 2002b, Table 4)
Abbey et al. (1993, Table 3)
1999 NHDS public use data files2;
adjusted by 0.93 for prob. of
surviving after 28 days (Rosamond
etal., 1999)
American Lung Association
(2002a, Table 11)
Schwartz (1994, Table 2)
Pope etal. (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
9-133
-------
Draft Regulatory Impact Analysis
Endpoint
Parameter
Daily respiratory illness-
related school absence rate per
person
Northeast
Midwest
South
West
Rates
Value
0.0073
0.0092
0.0061
0.0124
Source1
1996 HIS (Adams et al., 1999,
Table 47); estimate of 180 school
days per year
1. 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.
2. See ftp://ftp.cdc.gov/pub/Health Statistics/NCHS/Datasets/NHDS/
3. See ftp://ftp.cdc.gov/pub/Health Statistics/NCHS/Datasets/NHAMCS/
4. Lower Respiratory Symptoms are defined as >2 of the following: cough, chest pain, phlegm, wheeze
5. 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.
9-134
-------
Cost-Benefit Analysis
Table 9A-22.
Baseline Incidence Rates and Population Prevalence Rates of Asthma Symptoms for use in C-R
Functions, Asthmatic Population.
Endpoint
Asthma
Exacerbation, wheeze
Asthma
Exacerbation, cough
Asthma
Exacerbation,
dyspnea
Asthma
Exacerbation, one or
more
Asthma Attacks
Acute/Chronic
Bronchitis
Chronic Phlegm
Upper Respiratory
Symptoms
Parameter
Daily wheeze incidence among asthmatic children
(African-American)
Daily wheeze prevalence among asthmatic
children (African-American)
Daily wheeze prevalence among asthmatic
children
Daily cough incidence among asthmatic children
(African-American)
Daily cough prevalence among asthmatic children
(African-American)
Daily cough prevalence among asthmatic children
Daily dyspnea incidence among asthmatic children
(African-American)
Daily dyspnea prevalence among asthmatic
children (African-American)
Daily dyspnea prevalence among asthmatic
children
Daily prevalence among asthmatic children of at
least one of the following symptoms: wheeze,
cough, chest tightness, shortness of breath.
Daily incidence of asthma attacks
Annual bronchitis incidence rate among
asthmatic children
Annual phlegm incidence rate among
asthmatic children
Daily upper respiratory symptom incidence
among asthmatic children*
Rates
Value
0.076
0.173
0.038
0.067
0.145
0.086
0.037
0.074
0.045
0.60
0.055
0.326
0.257
0.3419
Source1
Ostroetal. (2001,p. 202)
Ostroetal. (2001,p. 202)
Vedal etal. (1998, Table 1)
Ostroetal. (2001, p. 202)
Ostroetal. (2001, p. 202)
Vedal etal. (1998, Table 1)
Ostroetal. (2001, p. 202)
Ostroetal. (2001, p. 202)
Vedal etal. (1998, Table 1)
Yu et al. (2000, Table 2)
HIS 1999
McConnell et al.(1999, Table 2)
McConnell et al.(1999, Table 2)
Pope etal. (1991, Table 2)
1. 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.
* Upper Respiratory Symptoms are defined as > 1 of the following: runny or stuffy nose; wet cough; burning, aching, or red eyes.
9-135
-------
Draft Regulatory Impact Analysis
Table 9A-24.
Asthma Prevalence Rates Used to Estimate Asthmatic Populations in C-R 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 files1
American Lung Association (2002c, Table 9)- based
on 1999 HIS
American Lung Association (2002c, Table 9)- based
on 1999 HIS
1. See ftp://ftp.cdc.gov/pub/Health_Statistics/NCHS/Datasets/HIS/2000/
9A.3.5.4 Accounting for Potential Health Effect Thresholds
When conducting clinical (chamber) and epidemiological studies, C-R 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 non-zero 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 Science Advisory Board 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 Clean Air Act, has advised EPA
that there is currently no scientific basis for selecting a threshold of 15 |j.g/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) which finds
in most cases no evidence of a non-linear concentration-response relationship and certainly does
not find a distinct threshold for health effects. The most recent draft of the EPA Air Quality
Criteria for Particulate Matter (U.S. EPA, 2002) reports only one study, analyzing data from
9-136
-------
Cost-Benefit Analysis
Phoenix, AZ, that reported even limited evidence suggestive of a possible threshold for PM2.5
(Smith et al., 2000).
Recent cohort analyses by the Health Effects Institute (Krewski et al., 2000) and Pope et
al. (2002) provide additional evidence of a quasi-linear concentration-response relationship
between long-term exposures to PM2 5 and 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 (ag/ms. 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
concentration-response relationships relating PM2.5 and mortality "were not significantly
different from linear associations."
Daniels et al. (2000) examined the presence of threshold in PM10 concentration-response
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 PMio on the types of mortality of main interest. Schwartz and Zanobetti
(2000) investigated the presence of threshold by simulation and actual data analysis of 10 U.S.
cities. In the analysis of real data from 10 cities, the combined concentration-response curve did
not show evidence of a threshold in the PMio-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 |J.g/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 PM2 5 and PMi0-2.5 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 |j.g/m3. However, the concentration-response curve for PM2.5 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
9-137
-------
Draft Regulatory Impact Analysis
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 of PM.
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
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 costs of
illness (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-15 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, 2020 and 2030. 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. Details of the derivation
of values for hospital admissions and other endpoints can be found in the benefits TSD for this
RIA (Abt Associates, 2003). A discussion of the valuation methods for premature mortality and
chronic bronchitis is provided here due to the relative importance of these effects. Discussions of
the methods used to value non-fatal myocardial infarctions (heart attacks) and school absence
days are provided because these endpoints have not been included in previous analyses 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-15.
9-138
-------
Table 9A-25. Unit Values Used for Economic Valuation of Health Endpoints (2000$)
Health
Endpoint
Premature Mortality
Base Estimate (VSL)
Alternative Estimate (VSLY)
3% discount rate
Under 65
65 and older
7% discount rate
Under 65
65 and older
Chronic Bronchitis (CB)
Base Estimate
Alternative Estimate
3% discount rate
Age 27-44
Age 45-64
Age 65+
7% discount rate
Age 27-44
Age 45-64
Age 65+
Central Estimate of Value Per Statistical
Incidence
1990 Income
Level
$6,300,000
$172,000
$434,000
$286,000
$527,000
$340,000
$150,542
$97,610
$11,088
$86,026
$72,261
$9,030
2020 Income
Level
$8,000,000
$217,000
$547,000
$360,000
$664,000
$430,000
$150,542
$97,610
$11,088
$86,026
$72,261
$9,030
2030 Income
Level
$8,100,000
$221,000
$559,000
$368,000
$678,000
$440,000
$150,542
$97,610
$11,088
$86,026
$72,261
$9,030
Derivation of Estimates
Base value is the mean of VSL estimates from 26 studies (5
contingent valuation and 21 labor market studies) reviewed for the
Section 812 Costs and Benefits of the Clean Air Act, 1990-2010 (US
EPA, 1999).
Alternative VSLY estimates are derived from a VSL based on the
mean of VSL estimates from the 5 contingent valuation studies
referenced above. VSLY for populations under 65 are based on 35
years of assumed average remaining life expectancy. VSLY for
populations 65 and older are based on 10 years of assumed average
remaining life expectancy.
Base value is the mean of a generated distribution of WTP to avoid a
case of pollution-related CB. WTP to avoid a case of pollution-
related CB is derived by adjusting WTP (as described in Viscusi et
al, 1991) to avoid a severe case of CB for the difference in severity
and taking into account the elasticity of WTP with respect to severity
ofCB.
Alternative value is a cost of illness (COI) estimate based on
Cropper and Krupnick (1990). Includes both medical costs and
opportunity cost from age of onset to expected age of death (assumes
that chronic bronchitis does not change life expectancy).
-------
Table 9A-25. Unit Values Used for Economic Valuation of Health Endpoints (2000$)
Health
Endpoint
Non-fatal 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
Central Estimate of Value Per Statistical
Incidence
1990 Income
Level
$66,902
$74,676
$78,834
$140,649
$66,902
$65,293
$73,149
$76,871
$132,214
$65,293
2020 Income
Level
$66,902
$74,676
$78,834
$140,649
$66,902
$65,293
$73,149
$76,871
$132,214
$65,293
2030 Income
Level
$66,902
$74,676
$78,834
$140,649
$66,902
$65,293
$73,149
$76,871
$132,214
$65,293
Derivation of Estimates
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 (1 990). 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 etal., 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)
Pneumonia
(ICD codes 480-487)
Asthma admissions
$12,378
$14,693
$6,634
$12,378
$14,693
$6,634
$12,378
$14,693
$6,634
The COI estimates (lost earnings plus direct medical costs) are based
on ICD-9 code level information (e.g., average hospital care costs,
average length of hospital stay, and weighted share of total COPD
category illnesses) reported in Agency for Healthcare Research and
Quality, 2000 (www.ahrq.gov).
The COI estimates (lost earnings plus direct medical costs) are based on
ICD-9 code level information (e.g., average hospital care costs, average
length of hospital stay, and weighted share of total pneumonia category
illnesses) reported in Agency for Healthcare Research and Quality, 2000
(www.ahrq.gov).
The COI estimates (lost earnings plus direct medical costs) are based
on ICD-9 code level information (e.g., average hospital care costs,
average length of hospital stay, and weighted share of total asthma
category illnesses) reported in Agency for Healthcare Research and
Quality, 2000 (www.ahrq.gov).
-------
Table 9A-25. Unit Values Used for Economic Valuation of Health Endpoints (2000$)
Health
Endpoint
All Cardiovascular
(ICD codes 390-429)
Emergency room visits for asthma
Central Estimate of Value Per Statistical
Incidence
1990 Income
Level
$18,387
$286
2020 Income
Level
$18,387
$286
2030 Income
Level
$18,387
$286
Derivation of Estimates
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
cardiovascular category illnesses) reported in Agency for Healthcare
Research and Quality, 2000 (www.ahrq.gov).
Simple average of two unit COI values:
(1) $31 1.55, from Smith et al., 1997, and
(2) $260.67, from Stanford et al., 1999.
-------
Table 9A-25. Unit Values Used for Economic Valuation of Health Endpoints (2000$)
Health
Endpoint
Central Estimate of Value Per Statistical
Incidence
1990 Income
Level
2020 Income
Level
2030 Income
Level
Derivation of Estimates
Respiratory Ailments Not Requiring Hospitalization
Upper Respiratory Symptoms (URS)
Lower Respiratory Symptoms (LRS)
Acute Bronchitis
$25
$16
$360
$27
$17
$390
$27
$17
$390
Combinations of the 3 symptoms for which WTP estimates are
available that closely match those listed by Pope, et al. result in 7
different "symptom clusters," each describing a "type" of URS. A
dollar value was derived for each type of URS, using mid-range
estimates of WTP (lEc, 1 994) to avoid each symptom in the cluster
and assuming additivity of WTPs. The dollar value for URS is the
average of the dollar values for the 7 different types of URS.
Combinations of the 4 symptoms for which WTP estimates are
available that closely match those listed by Schwartz, et al. result in
1 1 different "symptom clusters," each describing a "type" of LRS. A
dollar value was derived for each type of LRS, using mid-range
estimates of WTP (lEc, 1 994) to avoid each symptom in the cluster
and assuming additivity of WTPs. The dollar value for LRS is the
average of the dollar values for the 1 1 different types of LRS.
Assumes a 6 day episode, with daily value equal to the average of
low and high values for related respiratory symptoms recommended
in Neumann, et al. 1994.
Restricted Activity and Work/School Loss Days
Work Loss Days (WLDs)
Variable
(national
median = )
County-specific median annual wages divided by 50 (assuming 2 weeks of
vacation) and then by 5 - to get median daily wage. U.S. Year 2000
Census, compiled by Geolytics, Inc.
-------
Table 9A-25. Unit Values Used for Economic Valuation of Health Endpoints (2000$)
Health
Endpoint
School Absence Days
Worker Productivity
Minor Restricted Activity Days
(MRADs)
Central Estimate of Value Per Statistical
Incidence
1990 Income
Level
$75
$0.95 per
worker per 10%
change in
ozone per day
$51
2020 Income
Level
$75
$0.95 per
worker per 10%
change in
ozone per day
$55
2030 Income
Level
$75
$0.95 per
worker per 10%
change in
ozone per day
$56
Derivation of Estimates
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).
This median 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.
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: 1 985 to 2000")
(Source of data in table: U.S. Bureau of Labor Statistics, Bulletin 2307
and Employment and Earnings, monthly).
Median WTP estimate to avoid one MRAD from Tolley, et al.
(1986) .
-------
Draft Regulatory Impact Analysis
9A. 3.5.5.1 Valuing Reductions in Premature Mortality Risk
Base Estimate
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 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 estimated to
be $6 million in 1999 dollars. This represents an intermediate value from a variety of estimates
that appear in the economics literature, and it is a value EPA has frequently used in RIAs for
other rules and in the Section 812 Reports to Congress.
This estimate is the mean of a distribution fitted to the estimates from 26 value-of-life
studies identified in the Section 812 reports as "applicable to policy analysis." The approach and
set of selected studies mirrors that of Viscusi (1992) (with the addition of two studies), and uses
the same criteria as Viscusi in his review of value-of-life studies. The $6.3 million estimate is
consistent with Viscusi's conclusion (updated to 2000$) that "most of the reasonable estimates of
the value of life are clustered in the $3.8 to $8.9 million range." Five of the 26 studies are
contingent valuation (CV) studies, which directly solicit WTP information from subjects; the rest
are wage-risk studies, which base WTP estimates on estimates of the additional compensation
demanded in the labor market for riskier jobs. 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
five years following exposure. To take this into account in the valuation of reductions in
premature mortality, we apply an annual three percent discount rate to the value of premature
mortality occurring in future years.v
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 economic and
public policy analysis community. Regardless of the theoretical economic considerations, 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.
v The choice of a discount rate, and its associated conceptual basis, is a topic of ongoing discussion within the
federal government. 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 inEPA's Guidelines for Preparing Economic Analyses (in
press).
9-144
-------
Cost-Benefit Analysis
Following the advice of the EEAC of the SAB, EPA currently uses the VSL approach in
calculating the primary estimate of mortality benefits, because we believe this calculation to
provide the most reasonable single estimate of an individual's willingness to trade off money for
reductions in mortality risk (EPA-SAB-EEAC-00-013). While there are several differences
between the labor market studies EPA uses to derive a VSL estimate and the particulate matter
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-17 lists some of these
differences and the expected effect on the VSL estimate for air pollution-related mortality. For
example, adjusting for age differences may imply the need to adjust the $6.3 million VSL
downward, but the involuntary nature of air pollution-related risks and the lower level of risk-
aversion of the manual laborers in the labor market studies may imply the need for upward
adjustments. In the absence of a comprehensive and balanced set of adjustment factors, EPA
believes it is reasonable to continue to use the $6.3 million value while acknowledging the
significant limitations and uncertainties in the available literature.
Some economists emphasize that the value of a statistical life is not a single number
relevant for all situations. Indeed, the VSL estimate of $6.3 million (2000 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 populations used in the labor market studies, as is the case here,
some economists prefer to adjust the VSL estimate to reflect those differences. Some of the
alternative approaches that have been proposed for valuing reductions in mortality risk are
discussed in Figure 9A-6.
There is general agreement that the value to an individual of a reduction in mortality risk
can vary based on several factors, including the age of the individual, the type of risk, the level of
control the individual has over the risk, the individual's attitudes towards risk, and the health
status of the individual. While the empirical basis for adjusting the $6.3 million VSL for many
of these factors does not yet exist, a thorough discussion of these factors is contained in the
benefits TSD for this RIA (Abt Associates, 2003). EPA recognizes the need for investigation by
the scientific community to develop additional empirical support for adjustments to VSL for the
factors mentioned above.
The SAB-EEAC advised in their recent report 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, U.S. EPA, 2000b). In
developing our primary estimate of the benefits of premature mortality reductions, we have
discounted over the lag period between exposure and premature mortality. However, in
accordance with the SAB advice, we use the VSL in our primary estimate and present sensitivity
estimates reflecting age-specific VSL.
9-145
-------
Draft Regulatory Impact Analysis
Table 9A-26. 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
Alternative Estimate
The Alternative Estimate reflects the impact of changes to key assumptions associated with
the valuation of mortality. These include: 1) the impact of using wage-risk and contingent
valuation-based value of statistical life estimates in valuing risk reductions from air pollution as
opposed to contingent valuation-based estimates alone, 2) the use of a value of statistical life
years approach as opposed to a VSL approach, and 3) the degree of prematurity (number of life
years lost) for mortalities from air pollution.
The Alternative Estimate addresses the first issue by using an estimate of the value of
statistical life that is based only on the set of five contingent valuation studies included in the
larger set of 26 studies recommended by Viscusi (1992) as applicable to policy analysis. The
mean of the five contingent valuation based VSL estimates is $3.7 million (2000$), which is
approximately 60 percent of the mean value of the full set of 26 studies. Note that because these
are deaths associated with short-term exposures to PM2 5, it is assumed that there is no lag
between reduced exposure and reduced risk of mortality. Sensitivity analyses exploring the
implications for the alternative estimate of using different starting VSL estimates are presented in
Appendix 9B.
While the base estimate is based on a VSL approach, the alternative estimate is based on
the number of years of life saved and economic value of saving a statistical life year (VSLY).
The VSLY approach has been developed in the peer-reviewed economics literature (e.g., Viscusi
and Moore, 1988) and has been applied for many years by the U.S. Food and Drug
Administration (US FDA 1995, 1996, 1997, 1998, 1999, 2000, 2001). Some recent analyses,
however, have raised concerns about the use of this method to value reductions in premature
mortality in an environmental context (Science Advisory Board, 1999; Krupnick et al., 2002).
The VSLY approach applied in this RIA recognizes that each year late in the life span may have
a higher monetary value than the average life year saved in the middle of the life span. The non-
constant VSLY, rising later in the lifespan, is qualitatively compatible with theoretical economic
models of an individual's demand for lifesaving as a function of age (Shepard and Zeckhauser,
9-146
-------
Cost-Benefit Analysis
1984). The conceptonal rationale for a premium on VSLY among the elderly is that they have
saved through their working lifetimes and accumulated assets that can be devoted to health
protection, and have rising baseline risks, which increase the marginal value of risk reductions.
(Pratt and Zeckhauser 1996).
Under the alternative approach, the value of a life year for younger individuals is calculated
as if they had an average life expectancy. However, instead of attempting to estimate the
remaining life expectancy for different age groups, we have assumed that everyone who dies
from exposure to air pollution loses five years of life. Because we assume that younger
individuals do not have the accumulated assets or do not adjust the value of life years to reflect
reductions in life expectancy, this approach implies that the total value of a five-year loss in life
years is greater for the elderly than for younger individuals. An additional limitation of this
approach is the discontinuity at age 65. A more complex approach would produce a continuous
VSLY curve; however, the empirical data required to specify these models are not available.
There is no latency period assumed in the alternative analysis since the premature deaths
are assumed to occur primarily among persons with chronic disease who experience short-term
elevations in daily air pollution levels. Even the latency periods associated with the distributed
lag models are too short to be of significance in the valuation process.
In order to implement the non-constant VSLY approach, we begin by using a VSL of of
$3.7 million based on five contingent valuation studies which were also considered as part of the
base estimate. This smaller VSL is also consistent with an alternative interpreation of the wage-
risk literature (Mrozek and Taylor 2002). For persons under age 65, the $3.7 million VSL is
assumed to reflect an average loss of 35 years. The VSLY associated with $3.7 million VSL is
$172,000, annualized using a 3 percent discount rate, or $286,000, annualized using a 7 percent
discount rate. Note that the larger discount rate increases the VSLY because at a higher discount
rate, a larger stream of VSLY is required to yield a VSL of $3.7 million. For those over age 65,
the VSLY is derived from a $3.7 million VSL and an assumed 10-year life expectancy. This
gives a VSLY of $434,000 at a 3 % discount rate of a $527,000 at a 7% discount rate.
The alternative estimate also assumes that deaths from chronic obstructive pulmonary
disease (COPD) are advanced by 6 months, and deaths from all other causes are advanced by 5
years. As a first approximation, these reductions in life years lost are applied regardless of the
age at death. Actuarial evidence suggests that individuals with serious preexisting cardiovascular
conditions have an average remaining life expectancy of around 5 years. While many deaths
from daily exposure to PM may occur in individuals with cardiovascular disease, studies have
shown relationships between all cause mortality and PM, and between PM and mortality from
pneumonia (Schwartz, 2000). In addition, recent studies have shown a relationship between PM
and non-fatal heart attacks, which suggests that some of the deaths due to PM may be due to fatal
heart attacks (Peters et al., 2001). And, a recent meta-analysis has shown little effect of age on
the relative risk from PM exposure (Stieb et al., 2002). The alternative estimate suggests that the
number of deaths in non-elderly populations that 21 percent of non-COPD premature deaths
avoided are in populations under 65 (with the possibility for greater loss of life years in this age
group). Thus, while the assumption of 5 years of life lost may be appropriate for a subset of total
9-147
-------
Draft Regulatory Impact Analysis
avoided premature mortalities, it may overestimate or underestimate the degree of life shortening
attributable to PM for the remaining deaths. Sensitivity analyses of the alternative estimate using
different assumptions about the degree of life shortening are presented in Appendix 9B.
Monetized estimates of the benefits of a reduction in premature mortality are calculated as
follows. First, the expected reduction in premature mortality by age category (over 65 and under
65) is multiplied by the assumed gain in additional discounted life years for the two disease
categories (COPD and non-COPD) to obtain an estimate of the additional discounted life-years
associated with the reduction in PM exposures. No adjustment is made for the quality of life-
years saved. The monetized benefit estimate for a reduction in premature mortality, then, is the
product of the additional discounted life-years times the calculated VSLY.
For the alternative analysis, a sensitivity analysis (Table 9B-2) was performed to reflect
plausible changes in the numeric values of uncertain inputs to the mortality evaluation. The key
uncertain inputs are the average number of life years lost per premature death and the starting
VSL assumption. Results are presented for discount rates of 3 %.
Uncertainties Specific to Premature Mortality Valuation
The economic benefits associated with premature mortality are the largest category of
monetized benefits of the Nonroad Diesel Engine rule. In addition, in prior analyses EPA has
identified valuation of mortality benefits as the largest contributor to the range of uncertainty in
monetized benefits (see U.S. 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 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 mortality, or reflect differences in the nature of
the risk presented by air pollution relative to the risks studied in the relevant economic 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. At risk individuals include those who have
suffered strokes or are suffering from cardiovascular disease and angina (Rowlatt, et al. 1998).
An ideal benefits estimate of mortality risk reduction would reflect these human characteristics,
in addition to an individual's willingness to pay (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
9-148
-------
Cost-Benefit Analysis
individuals because survival curves are dependent 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
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 value of statistical life approach in the
Base Estimate, supplemented by valuation based on the life-year method in the Alternative
Estimate.
Other uncertainties specific to premature mortality valuation include the following:
Across-study Variation: The analytical procedure used in the main analysis to estimate the
monetary benefits of avoided premature mortality assumes that the appropriate economic
value for each incidence is a value from the currently accepted range of the value of a
statistical life. This estimate is based on 26 studies of the value of mortal risks. There is
considerable uncertainty as to whether the 26 studies on the value of a statistical life
provide adequate estimates of the value of a statistical life saved by air pollution reduction.
Although there is considerable variation in the analytical designs and data used in the 26
underlying studies, 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
willingness to pay (WTP) to reduce the risk. The appropriateness of a distribution of WTP
estimates from the 26 studies for valuing the 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 (1) the extent to which the
risks being valued are similar, and (2) 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 value of a statistical life
from the 26 studies to the Nonroad Diesel Engine 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 ten times the risk reduction valued
in the study). Under the assumption of linearity, the estimate of the value of a statistical
life 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).
9-149
-------
Draft Regulatory Impact Analysis
Voluntariness of risks evaluated. Although there may be several ways in which job-related
mortality risks differ 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. There is some evidence 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.
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.
9^4.3.5.5.2 Valuing Reductions in the Risk of Chronic Bronchitis
Base Estimate
The best available estimate of WTP to avoid a case of chronic bronchitis (CB) comes from
Viscusi, et al. (1991). The Viscusi, et al. study, however, describes a severe case of CB to the
survey respondents. We therefore employ an estimate of WTP to avoid a pollution-related case of
CB, 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 CB 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 this RIA (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 CB in this analysis. The distribution incorporates
uncertainty from three sources: (1) the WTP to avoid a case of severe CB, as described by
Viscusi, et al.; (2) the severity level of an average pollution-related case of CB (relative to that of
the case described by Viscusi, et al.); and (3) the elasticity of WTP with respect to severity of the
9-150
-------
Cost-Benefit Analysis
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 CB by
statistical uncertainty analysis techniques. The expected 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 CB.
Alternative Estimate
For the Alternative Estimate, a cost-of illness value is used in place of willingness-to-pay to
reflect uncertainty about the value of reductions in incidences of chronic bronchitis. In the Base
Estimate, the willingness-to-pay estimate was derived from two contingent valuation studies
(Viscusi et al., 1991; Krupnick and Cropper, 1992). These studies were experimental studies
intended to examine new methodologies for eliciting values for morbidity endpoints. Although
these studies were not specifically designed for policy analysis, the SAB (EPA-SAB-COUNCIL-
ADV-00-002, 1999) has indicated that the severity-adjusted values from this study provide
reasonable estimates of the WTP for avoidance of chronic bronchitis. As with other contingent
valuation studies, the reliability of the WTP estimates depends on the methods used to obtain the
WTP values. In order to investigate the impact of using the CV based WTP estimates, the
Alternative Estimate relies on estimates of lost earnings and medical costs. Using age-specific
annual lost earnings and medical costs estimated by Cropper and Krupnick (1990) and a three
percent discount rate, we estimated a lifetime present discounted value (in 2000$) due to chronic
bronchitis of $150,542 for someone between the ages of 27 and 44; $97,610 for someone
between the ages of 45 and 64; and $11,088 for someone over 65. The corresponding age-
specific estimates of lifetime present discounted value (in 2000$) using a seven percent discount
rate are $86,026, $72,261, and assuming $9,030, respectively. These estimates assumed that 1)
lost earnings continue only until age 65, 2) medical expenditures are incurred until death, and 3)
life expectancy is unchanged by chronic bronchitis.
9A. 3.5.5.3 Valuing Reductions in Non-FatalMyocardial Infarctions (Heart Attacks)
The Agency has not previously estimated the impact of its programs on reductions in the
expected number of non-fatal heart attacks, although it has examined the impact of reductions in
other related cardiovascular endpoints™. We were not able to identify a suitable WTP value for
reductions in the risk of non-fatal heart attacks. Instead, we propose a cost-of-illness unit value
with two components: the direct medical costs and the opportunity cost (lost earnings) associated
with the illness event. Because the costs associated with an MI extend beyond the initial event
itself, we consider costs incurred over several years. Using age-specific annual lost earnings
estimated by Cropper and Krupnick (1990), and a three percent discount rate, we estimated a
present discounted value in lost earnings (in 2000$) over 5 years due to an MI 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 seven percent discount rate are $7,855, $11,578, and
$66,920, respectively. Cropper and Krupnick (1990) do not provide lost earnings estimates for
9-151
-------
Draft Regulatory Impact Analysis
populations under 25 or over 65. As such we do not include lost earnings in the cost estimates
for these age groups.
We have found three possible sources in the literature of estimates of the direct medical costs of
MI:
Wittels et al. (1990) estimated expected total medical costs of MI 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 MI
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 (LOS) for MI may have decreased from 1983 to
the present. The average LOS for ICD code 410 (MI) 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 non-fatal MI cases), whose LOS was therefore substantially
shorter than it would be if they hadn't died.
Eisenstein et al. (2001) estimated 10-year costs of $44,663, in 1997$, or $49,651 in
2000$ for MI 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 MI 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 $29,568 for a ten-year
period.
In summary, the three different studies provided significantly different values:
9-152
-------
Cost-Benefit Analysis
Table 9A-27.
Alternative Direct Medical Cost of Illness Estimates for Nonfatal Heart Attacks
Study
Wittels et al., 1990
Russell et al., 1998
Eisenstein et al., 2001
Russell et al., 1998
Direct Medical Costs (2000$)
$109,474*
$22,331**
$49,651**
$27,242**
Over an x-year period, for x
5
5
10
10
*Wittels et al. did not appear to discount costs incurred in future years.
**Using a 3 percent discount rate.
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 will 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 will use a simple average of the two 5-year estimates, or
$65,902, and add it to the 5-year opportunity cost estimate. The resulting estimates are given in
the table below.
Table 9A-28.
Estimated Costs Over a 5-Year Period (in 2000$) of a Non-Fatal Myocardial Infarction
Age Group
0-24
25-44
45-54
55-65
>65
Opportunity Cost
$0
$8,774*
$12,253*
$70,619*
$0
Medical Cost**
$65,902
$65,902
$65,902
$65,902
$65,902
Total Cost
$65,902
$74,676
$78,834
$140,649
$65,902
*From Cropper and Krupnick, 1990, using a 3% discount rate.
**An average of the 5-year costs estimated by Wittels et al., 1990, and Russell et al., 1998.
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
9-153
-------
Draft Regulatory Impact Analysis
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: (1) 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 (2)
value the 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 two significant problems
with this method, however. First, it omits WTP to avoid the symptoms/illness which 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). 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, conservative 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 the table below.
9-154
-------
Cost-Benefit Analysis
Table 9A-29.
Women with Children: Number and Percent
in the Labor Force, 2000, and Weighted Average Participation Rate*
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%
*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.
**Widowed, divorced, or separated.
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.
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.
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 with infant mortality (Woodruff, et al., 1997) and cancer
(US EPA, 1996b), 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), 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
9-155
-------
Draft Regulatory Impact Analysis
to be PM or ozone-induced, C-R data 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.
Another category of potential effects that may change in response to ozone reduction
strategies results from the shielding provided by ozone against the harmful effects of ultraviolet
radiation (UV-B) derived from the sun. The great majority of this shielding results from
naturally occurring ozone in the stratosphere, but the 10 percent of total "column"ozone present
in the troposphere also contributes (NAS, 1991). A variable portion of this tropospheric fraction
of UV-B shielding is derived from ground level or "smog" ozone related to anthropogenic air
pollution. Therefore, strategies that reduce ground level ozone could, in some small measure,
increase exposure to UV-B from the sun.
While it is possible to provide quantitative estimates of benefits associated with globally
based strategies to restore the far larger and more spatially uniform stratospheric ozone layer, the
changes in UV-B exposures associated with ground level ozone reduction strategies are much
more complicated and uncertain. Smog ozone strategies, such as mobile source controls, are
focused on decreasing peak ground level ozone concentrations, and it is reasonable to conclude
that they produce a far more complex and heterogeneous spatial and temporal pattern of ozone
concentration and UV-B exposure changes than do stratospheric ozone protection programs. In
addition, the changes in long-term total column ozone concentrations are far smaller from
ground-level programs. To properly estimate the change in exposure and impacts, it would be
necessary to match the spatial and temporal distribution of the changes in ground-level ozone to
the spatial and temporal distribution of exposure to ground level ozone and sunlight. More
importantly, it is long-term exposure to UV-B that is associated with effects. Intermittent, short-
term, and relatively small changes in ground-level ozone and UV-B are not likely to measurably
change long-term risks of these adverse effects.
For all of these reasons, we were unable to provide reliable estimates of the changes in
UV-B shielding associated with ground-level ozone changes. This inability lends an upward bias
to the net monetized benefits presented in this analysis. It is likely that the adverse health effects
associated with increases in UV-B exposure from decreased tropospheric ozone would, however,
be relatively very small from a public health perspective because 1) the expected long-term
ozone change resulting from this rule is small relative to total anthropogenic tropospheric ozone,
which in turn is small in comparison to total column natural stratospheric and tropospheric
ozone; 2) air quality management strategies are focused on decreasing peak ozone concentrations
and thus may change exposures over limited areas for limited times; 3) people often receive peak
exposures to UV-B in coastal areas where sea or lake breezes reduce ground level pollution
concentrations regardless of strategy; and 4) ozone concentration changes are greatest in urban
areas and areas immediately downwind of urban areas. In these areas, people are more likely to
spend most of their time indoors or in the shade of buildings, trees or vehicles.
9-156
-------
Cost-Benefit 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 paniculate 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
models were used to predict the change in visibility, measured in deciviews, of the areas affected
by the preliminary control options."
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
x 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 in visibility 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 in visibility which when considered together amount to perceptible
changes in visibility.
9-157
-------
Draft Regulatory Impact Analysis
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/ 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.2
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 basis for monetary
estimates of the benefits of visibility changes in recreational areas.aa 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-
COUNCIL-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
Y The Clean Air Act designates 156 national parks and wilderness areas as Class I areas for visibility protection.
z For details of the visibility estimates discussed in this chapter, please refer to the benefits technical support
document for this RIA (Abt Associates 2003).
^ An SAB advisory letter indicates thaf'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).
9-158
-------
Cost-Benefit Analysis
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
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.
9-159
-------
Draft Regulatory Impact Analysis
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.
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 AGSEVI®
agricultural benefits model (Taylor, et al., 1993). AGSEVI® 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.bb
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.cc Dollar
BBAGSIM° 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.
cc 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.
9-160
-------
Cost-Benefit Analysis
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.dd 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.
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 FID 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.
DD The total value for these crops in 1998 was $47 billion.
9-161
-------
Draft Regulatory Impact Analysis
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
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
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 9B.
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
9-162
-------
Cost-Benefit Analysis
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.
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.
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.
9-163
-------
Draft Regulatory Impact Analysis
Both the Base and Alternative estimates are dominated by benefits of mortality risk
reductions. The Base estimate projects that the modeled preliminary control options will result
in 6,200 avoided premature deaths in 2020 and 11,000 avoided premature deaths in 2030. The
Alternative estimate projects that reductions in short-term PM2 5 exposures alone will result in
3,700 avoided premature deaths in 2020 and 6,600 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. The omission of possible long-term impacts of particulate matter on mortality
accounts for an approximately 40 percent reduction in the estimate of avoided premature
mortality in the Alternative Estimate relative to the Base Estimate.
Our Base estimate of total monetized benefits in 2030 for the modeled nonroad preliminary
control options is $92 billion using a 3 percent discount rate and $87 billion using a 7 percent
discount rate. In 2020, the base monetized benefits are estimated at $52 billion using a 3 percent
discount rate and $47 billion using a 7 percent discount rate. Health benefits account for 94
percent of total benefits. The monetized benefit associated with reductions in the risk of
premature mortality, which accounts for $85 billion in 2030 and $47 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.
The alternative estimate of total monetized benefits in 2030 for the modeled preliminary
control option is $19 billion using a 3 percent discount rate and $20 billion using a 7 percent
discount rate. In 2020, the alternative monetized benefits are estimated at $11 billion using a 3
percent discount rate and $11 billion using a 7 percent discount rate. Health benefits account for
around 80 percent of the total alternative benefits estimates. The 40 percent reduction in
mortality under the Alternative Estimate and the difference in valuation of premature mortality
and chronic bronchitis explain the difference in benefits between these two approaches.
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 of 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
9-164
-------
Cost-Benefit Analysis
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.
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 mortality5 -
Base estimate: Long-term exposure (adults, 30 and over)
Alternative estimate: Short-term exposure (all ages)
Chronic bronchitis (adults, 26 and over)
Non-fatal myocardial infarctions (adults, 1 8 and older)
Hospital admissions - Respiratory (all ages)c
Hospital admissions - Cardiovascular (adults, 20 and older)D
Emergency Room Visits for Asthma (18 and younger)
Acute bronchitis (children, 8-12)
Lower respiratory symptoms (children, 7-14)
Upper respiratory symptoms (asthmatic children, 9-1 1)
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 6-1 1)
Avoided IncidenceA
(cases/year)
2020
6,200
3,700
4,300
11,000
3,100
3,300
4,300
10,000
110,000
92,000
780,000
4,600,000
370
150
93
(2,400)
65,000
2030
11,000
6,600
6,500
18,000
5,500
5,700
6,500
16,000
170,000
120,000
1,100,000
6,500,000
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.
9-165
-------
Draft Regulatory Impact Analysis
Table 9A.31
Results of Human Health and Welfare Benefits Valuation for the Modeled Preliminary
Nonroad Diesel Engine Standards
Endpoint
Premature mortality0
Base estimate: Long-term exposure, (adults, 30 and over)
3% discount rate
7% discount rate
Alternative estimate: Short-term exposure, (all ages)
3% discount rate
7% discount rate
Chronic bronchitis (adults, 26 and over)
Base estimate: Willingness-to-pay
Alternative estimate: Cost-of-illness
3% discount rate
7% discount rate
Non-fatal myocardial infarctions
3% discount rate
7% discount rate
Hospital Admissions from Respiratory Causes
Hospital Admissions from Cardiovascular Causes
Emergency Room Visits for Asthma
Acute bronchitis (children, 8-12)
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
Base estimate
3% discount rate
7% discount rate
Alternative estimate
3% discount rate
7% discount rate
Pollutant
PM
PM
PM
O3 and PM
PM
O3 and PM
PM
PM
PM
PM
O3 and PM
03
03
PM
03
O3 and PM
Monetary BenefitsA'B
(millions 2000$, Adjusted for
Income Growth)
2020
$47,000
$44,000
$7,200
$8,200
$1,900
$420
$270
$900
$870
$55
$72
$1
$4
$2
$2
$110
$250
$5
$4
$1,400
$89
$52,000+B
$49,000+B
$11,000+B
$11,000+B
2030
$85,000
$80,000
$13,000
$15,000
$3,000
$600
$390
$1,400
$1,400
$110
$120
$2
$6
$3
$3
$150
$370
$10
$7
$2,200
$140
$92,000+B
$87,000+B
$19,000+B
$20,000+B
Monetary benefits are rounded to two significant digits.
! Monetary benefits are adjusted to account for growth in real GDP per capita between 1990 and the analysis year (2020 or 2030).
9-166
-------
Cost-Benefit Analysis
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 of two different discount rates; a 3% rate which is
recommended by EPA's Guidelines for Preparing Economic Analyses (US EPA, 2000c), and 7% which is recommended by OMB Circular A-94
(OMB, 1992).
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.
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 XI-B. 1.
To gain further understanding into the public health impact of the modeled change in air
quality associated with the preliminary control options, we examined the incidence of health
effects occurring in three age groups: children (0-17), adults (18-64), and elderly adults (65 and
older). Certain endpoints occur only in a subset of age groups, so not all endpoints are reported
for all age groups. Two sets of age group estimates were calculated. The first is based on the
specific age ranges examined in the epidemiological studies, for example, the Dockery et al
(1996) acute bronchitis study focused on a sample population aged 8 to 12. These are the
estimates that were used in deriving total incidences as reported in Table 9A.9. In many cases
however, the study populations were defined as a matter of convenience or due to data
availability, rather than due to any biological factor that would restrict the effect to the specific
age group. In order to gain a more complete understanding of the potential magnitude of the
health impact in the entire population, we calculate a separate estimate including the health
impact on all population within an age group. The two sets of age specific incidence estimates
are provided in Table 9A-32. Note that for premature mortality, we chose not to extend the
estimates based on long-term exposure to children, even though there is some evidence that PM
exposure has mortality impacts in this age group (see Woodruff et al., 1997). The short-term
exposure studies used in the alternative estimate include all ages, and thus provide an estimate of
mortality benefits occuring in children.
We also estimated respiratory symptoms and attacks occurring the asthmatic population,
based on the studies defined in Table 9A-22. As with the age group specific estimates, we
provide two sets of calculations, one based on applying the C-R function only to the specific
population subgroup included in a study's sample population, and another based on applying the
C-R function to all populations within a broader population. The two sets of asthma symptom
incidences are provided in Table 9A-33. As noted earlier in this appendix, the asthma symptom
estimates provided in Table 9A-33 are not additive to the total benefits presented in Table 9A-31.
They are provided to show the specific impacts on an especially susceptible subpopulations.
Also note that the estimates are not additive even within the table. We have grouped the
estimates based on the type of symptoms measured, but there is the potential for considerable
overlap. However, these estimates provide an illustration of the consistency of the effects across
studies and populations of asthmatics.
9-167
-------
Draft Regulatory Impact Analysis
Table 9A-32.
Reductions in Incidence of Health Endpoints by Age GroupA
Endpoint/Age Group
Children, 0-17
Premature mortality8 -
Alternative estimate: Short-term exposure
Hospital Admissions - Respiratory Causes0
Emergency Room Visits for Asthma
Acute bronchitis
Lower respiratory symptoms
Upper respiratory symptoms in asthmatic children
School absence days (children, age 6-11)
Adults, 18-64
Premature mortality8 -
Base estimate: Long-term exposure
Alternative estimate: Short-term exposure
Chronic bronchitis
Non-fatal myocardial infarctions
Hospital admissions - Cardiovascular1'
Hospital admissions - Respirator}^
Work loss days
Minor restricted activity days
Adults, 65 and older
Premature mortality8 -
Base estimate: Long-term exposure
Alternative estimate: Short-term exposure
Chronic Bronchitis
Non-fatal Myocardial Infarctions
Hospital Admissions - Cardiovascular Causes
Hospital Admissions - Respiratory Causes
Pollutants
PM
O3 and PM
O3 and PM
PM
PM
PM
03
PM
PM
PM
PM
PM
PM
PM
O3 and PM
PM
PM
PM
PM
PM
O3 and PM
Avoided Incidence - Study
Population Only (cases/year)
2020
20
240
4,300
10,000
110,000
92,000
1,400
770
7,600
3,900
1,100
490
780,000
4,600,000
4,900
2,900
1,000
6,600
2,300
2,700
2030
30
570
6,500
16,000
170,000
120,000
1,800
1,000
11,000
5,300
1,450
660
1,100,000
6,600,000
9,100
5,500
1,900
12,000
4,300
5,700
Avoided Incidence - Total Age
Group Population
(cases/year)
2020
20
240
4,300
31,000
220,000
430,000
1,500
770
8,300
3,900
1,100
490
780,000
4,600,000
4,900
2,900
1,000
6,600
2,300
2,700
2030
30
570
6,500
47,000
330,000
660,000
1,900
1,000
12,000
5,300
1,450
660
1,100,000
6,600,000
9,100
5,500
1,900
12,000
4,300
5,700
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 children include ICD codes 493, 464.4, 466, and 480-486).
D Cardiovascular hospital admissions for adults includes total cardiovascular and subcategories for ischemic heart
disease, dysrhythmias, and heart failure.
E Respiratory hospital admissions for adults include admissions for all respiratory causes and subcategories for
COPD and pneumonia, and asthma.
9-168
-------
Cost-Benefit Analysis
Table 9A-33.
Reductions in Incidence of Respiratory Symptoms in the Asthmatic Population
Endpoint
(Study population)
Study
Pollutant
Avoided Incidence - Study
Population Only (cases/year)
2020
2030
Avoided Incidence - Total Age
Group Population
(cases/year)
2020
2030
Asthma Attack Indicators*
Shortness of Breath
(African American
asthmatics, 8-13)
Cough
(African American
asthmatics, 8-13)
Wheeze
(African American
asthmatics, 8-13)
Asthma Exacerbation -
one or more symptoms
(Asthmatics, 5-13)
Cough
(Asthmatics, 6-13)
Ostroetal. (2001)
Ostroetal. (2001)
Ostroetal. (2001)
Yu et al. (2000)
Vedaletal. (1998)
PM
PM
PM
PM
PM
10,000
21,000
16,000
400,000
180,000
15,000
31,000
24,000
530,000
240,000
30,000
63,000
49,000
630,000
320,000
45,000
94,000
74,000
950,000
490,000
Other symptoms/illness endpoints
Upper Respiratory
Symptoms
(Asthmatics 9-11)
Moderate or Worse
Asthma
(Asthmatics, all ages)
Acute Bronchitis
(Asthmatics, 9-15)
Chronic Phlegm
(Asthmatics, 9-15)
Asthma Attacks
(Asthmatics, all ages)
Popeetal. (1991)
Ostro et al.
(1991)
McConnell et al.
(1999)
McConnell et al.
(1999)
Whittemore and
Korn (1980)
PM
PM
PM
PM
PM
92,000
86,000
3,000
7,500
130,000
120,000
121,000
4,700
12,000
160,000
430,000
86,000
7,000
18,000
130,000
660,000
121,000
11,000
27,000
160,000
A Note that these are not necessarily independent symptoms. Combinations of these symptoms may occur in the
same individuals, so that the sum of the avoided incidences is not necessarily equal to the sum of the affected
populations. Also, some studies cover the same or similar endpoints in overlapping populations. For example, the
Vedal et al (1998) and Ostro et al (2000) studies both examine cough. The Ostro et al (2000) estimate examines a
more restricted population than Veal et al (1998), so estimates should be combined with caution.
9-169
-------
Draft 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 11,000 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 9B 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.
9-170
-------
Cost-Benefit Analysis
Appendix 9A 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. Proposed Nonroad 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.
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.
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.
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.
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 Journal53: 977-
984.
9-171
-------
Draft Regulatory Impact Analysis
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.
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.
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 J Epidemiol 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.
9-172
-------
Cost-Benefit Analysis
Dominici, F.; McDermott, A.; Daniels, M.; et al. (2002) Report to the Health Effects Institute:
reanalyses of the NMMAPS Database. [Accessed at
www.biostat.jhsph.edu/~fominic/HEI/nmmaps.html.]
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, PJ. 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, KJ. 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. MedCare. 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 I.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 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.
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. IJI. 1993. The Measurement of Environmental and Re source Values: Theory and
Methods. Resources for the Future, Washington, D.C.
9-173
-------
Draft Regulatory Impact Analysis
Garcia, P., Dixon, B. and Mjelde, J. (1986): Measuring the benefits of environmental change
using a duality approach: The case of Ozone and Dlinios cash grain farms. Journal of
Environmental Economics and Management.
Gilliland, F.D., K. Berhane, E.B. Rappaport, D.C. Thomas, E. Avol, WJ. 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(ll):1267-73
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,
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.
Hammitt, JK (2002). Understanding Differences in Estimates of the Value of Mortality Risk.
Journal of Policy Analysis and Management. Vol 21. No2: 271-273.
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
Paper No. 38, Population Projections Branch, Population Division, U.S. Census Bureau,
Department of Commerce. January.
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 (ffic). 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.
9-174
-------
Cost-Benefit Analysis
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
O
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, 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.
Krupnick A (2002). The Value of Reducing Risk of Death: A Policy Perspective. Journal of
Policy Analysis and Management. Vol 2: 275-282.
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 J Epidemiol 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.
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., 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, 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
9-175
-------
Draft Regulatory Impact Analysis
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, Christian! 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, 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, 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.
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
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.
9-176
-------
Cost-Benefit Analysis
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.
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 Med 54(8): 53 5-40.
Pope, C. A. 2000. Invited Commentary: Particulate Matter-Mortality Expsoure-Response
Relations and Thresholds. American Journal of Epidemiology, 152: 407-412.
Pope, C.A., HI, 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., HI, 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., HI, 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, JW and Zeckhauser, RJ (1996). Willingness to Pay and the Distribution of Risk and
Wealth. Journal of Political Economy, Vol. 104: 747-763.
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, J.E. Kelsall, J. Xu and L.S. Kalkstein. 1997. Air Pollution, Weather,
and Mortality in Philadelphia 1973-1988. Health Effects Institute. Cambridge, MA.
March.
9-177
-------
Draft Regulatory Impact Analysis
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.
Schwartz, J., 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 Manag Assoc. 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.
Sheppard, DC and Zeckhauser, RJ. (1984). Survival Versus Consumption. Management
Science. Vol.30. No. 4, April 1984.
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.
Smith, V. K., G. Van Houtven, and S.K. Pattanayak. 2002. Benefit Transfer via Preference
Calibration. Land Economics. 78: 132-152.
9-178
-------
Cost-Benefit Analysis
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, JA.
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 H. 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 P articulate Matter: Assessment of Scientific and Technical Information.
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.
9-179
-------
Draft Regulatory Impact Analysis
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
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.
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., 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.
Viscusi, WK and More MJ (1989). Rates of Time Preference and Valuations of the Duration of
Life. Journal of Public Economics. 38:297-317.
Viscusi WK and Aldy, JE (2003, forthcoming). The Value of A Statistical Life: A Critical
Review of Market Estimates Throughout the World. Journal of Risk and Uncertainty.
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.
9-180
-------
Cost-Benefit Analysis
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.
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.
9-181
-------
APPENDIX 9B: Sensitivity Analyses of Key Parameters in the Benefits
Analysis
9B.1 Premature Mortality—Long term exposure 9-183
9B.1.1 Alternative C-RFunctions 9-184
9B.1.2 Alternative Lag Structures 9-184
9B.1.3 AgeandVSL 9-185
9B.1.4 Thresholds 9-186
9B.2 Premature Mortality—Short term exposure 9-189
9B.3 Other Health Endpoint Sensitivity Analyses 9-190
9B.3.1 Overlapping Endpoints 9-190
9B.3.2 Alternative and Supplementary Estimates 9-192
9B.4 Income Elasticity of Willingness to Pay 9-196
9-182
-------
Cost-Benefit Analysis
The Base Estimate 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 reasonable alternative assumptions exist
for some inputs to the analysis, such as the mortality C-R functions. In Chapter 9 and Appendix
9A, we provide an Alternative estimate to show the impact of combining several alternative
assumptions about the estimation and valuation of mortality impacts, as well as the valuation of
chronic bronchitis. This appendix provides additional senstivity analyses, relative to both the
Base and Alternative estimates.
We supplement our Base Estimate of benefits with a series of sensitivity calculations that
make use of other sources of concentration-response and valuation data for key benefits
categories. These sensitivity calculations are conducted relative to the Base Estimate and not for
the Alternative Estimate. The sensitivity estimates can be used to answer questions like "What
would total benefits be if we were to value avoided incidences of premature mortality using the
age-dependent VSL rather than the age-independent VSL approach?" These estimates examine
sensitivity to both valuation issues (e.g. the correct value for a statistical life saved) 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. Additional sensitivity estimates are provided in the benefits
TSD (Abt Associates, 2003).
We supplement the Alternative Estimate of benefits with a set of senstivity analyes that
explore the impacts of changing two elements: the starting point value of a stastistical life used to
derive the value of a statistical life year, and the assumed number of life years gained for
premature mortalities avoided from reductions in short-term exposures to PM2 5.
9B.1 Premature Mortality—Long term exposure
Given current evidence regarding their value, reductions in the risk of premature
mortality is the most important PM-related health outcome in terms of contribution to dollar
benefits. There are four 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,
the relationship between age and VSL, and effect thresholds. Results of this set of sensitivity
analyses are presented in Table 9B.1.
9-183
-------
Draft Regulatory Impact Analysis
9B.1.1 Alternative C-R Functions
Although we used the Krewski, et al. (2000) mean-based ("PM2.5(DC), All Causes")
model exclusively to derive our Base Estimate of avoided premature mortality, this analysis also
examined the sensitivity of the benefit results to the selection of alternative C-R functions for
premature mortality. We used two sources of alternative C-R functions for this sensitivity
analysis: (1) an extended analysis of the American Cancer Society data, reported in Table 2 of
Pope et al. (2002); and (2) the Krewski et al. "Harvard Six Cities" estimate. The Pope et al
(2002) analysis provides estimates of the relative risk for all-cause, cardiopulmonary, and lung
cancer mortality, using a longer followup period relative to the original data examined in
Krewski et al (2000). 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). However, the Krewski-Harvard Six Cities study
had a more limited geographic scope (and a smaller study population) than the Krewski-ACS
study. 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 Krewski-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). The HEI commentary notes that
"the inherent limitations of using only six cities, understood by the original investigators, should
be taken into account when interpreting the results of the Six Cities Study." We emphasize, that
based on our understanding of the relative merits of the two datasets, the Krewski, et al. (2000)
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 primary
estimate of this important health effect. In addition to these alternative C-R functions, a broader
set of alternative mortality C-R functions is examined in the benefits TSD (Abt Associates,
2003).
9B.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 a 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 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
9-184
-------
Cost-Benefit Analysis
year lag is not supported by any scientific literature on PM-related mortality (NRC 2002). Rather
it is intended to be a best guess at the appropriate distribution of avoided incidences of PM-
related mortality.
Although the 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, respectively66. 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. 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 Krewski et al ACS C-R
function) are presented in Table 9B.2. These estimates are based on the value of statistical lives
saved approach, i.e. $6 million per incidence, and are presented for both a 3 and 7 percent
discount rate over the lag period. Even with an extreme lag assumption of 15 years, benefits are
reduced by less than half relative to the no lag and primary (5-year distributed lag) benefit
estimates.
9B.1.3 Age and VSL
The relationship between age and willingness to pay for mortality risk reductions has
been the subject of much research over the past several years. Recent research in the U.S. has not
found a significant reduction in WTP for risk reductions in older populations (Smith et al. 2002;
Alberini et al., 2002; Schultze, 2002). Studies outside of the U.S. have found a signficant
reduction in WTP for older individuals, ranging from 10 percent (Jones-Lee, 1993) to around 35
percent (Alberini et al. 2002) for a 70 year old, relative to a 40 year old. Around 80 percent of
the deaths projected to be avoided from reduced exposure to PM in 2020 and 2030 are in
populations over 65. As such, the assumption that populations of all ages have the same VSL
can have a significant impact on the total benefits. For this sensitivity analysis, the method we
use to account for age differences is to adjust the base $6.1 million VSL based on ratios of VSL's
for specific ages to the VSL for a 40 year old individual. There are several potential sources for
these ratios.
EEAlthough 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.
9-185
-------
Draft Regulatory Impact Analysis
Two Jones-Lee studies to provide evidence of strong and weak age effects on WTP for
mortality risk reductions. The ratios based on Jones-Lee (1989), as summarized in U.S. EPA
(2000), suggest a steep inverted U shape between age and VSL, with the VSL for a 70 year old at
63 percent of that for a 40 year old, and the VSL for an 85 year old at 7 percent of that for 40 year
old. The ratios based on Jones-Lee (1993) and summarized in U.S. EPA (2000), suggest a much
flatter inverted U shape, with the VSL for a 70 year old at 92 percent of that for a 40 year old,
and the VSL for an 85 year old at 82 percent of that for a 40 year old. Recent analyses conducted
in Canada and the U.S. (Alberini et al, 2002; Krupnick et al, 2002) found mixed results. The
Canadian analysis found around a 35 percent reduction in VSL for respondents over age 70, but
the U.S. analysis found no significant differences in VSL across ages. The wide range of age-
adjustment ratios, especially at older ages demonstrates the difficulty in making these kinds of
adjustments. We select the recent Krupnick et al results for Canada as the basis for calculating
age-specific VSL, because it uses state of the art stated preference methods and reflects more
current preferences. Krupnick (2002) may understate the effect of age because they only control
for income and do not control for wealth. While there is no empirical evidence to support or
reject hypotheses regarding wealth and observed WTP, WTP for additional life years by the
elderly may in part reflect their wealth position vis a vis middle age respondents.
We note that our Base estimate is the most consistent with current evidence on U.S.
preferences for risk reduction in older populations. To calculate benefits using the age-adjusted
VSL, we first calculate the number of avoided premature mortalities in each age category, and
then apply the age adjusted VSL to the appropriate incidences in each age category.
9B.1.4 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 relationship. 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. EPA does not necessarily
endorse any particular threshold and, as discussed in section 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 PM25 standard of 15 |J.g/m3, our
estimate of the total number of avoided PM-related premature mortalities in 2030 would be
reduced by approximately 70 percent, from approximately 11,000 annually to approximately
3,200 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)"
9-186
-------
Cost-Benefit Analysis
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.
As noted in earlier sections, EPA is developing a formal uncertainty analysis processs which we
intend to at least partially implement for the analysis of the final rule.
9-187
-------
Draft Regulatory Impact Analysis
Table 9B-1.
Sensitivity of Estimates to Alternative Assumptions Regarding Quantification of Mortality
Benefits
Description of Sensitivity Analysis
Avoided IncidencesA
2020
2030
Value (million 2000$)B
2020
2030
Alternative Concentration-Response Functions for PM-related Premature Mortality
Pope/ACS Study (2002)
All Cause
Lung Cancer
Cardiopulmonary
Krewski/Harvard Six-city Study
5,400
740
4,000
18,000
9,500
1,300
7,200
32,000
$41,000
$5,600
$30,000
$140,000
$74,000
$9,900
$55,000
$240,000
Alternative Lag Structures for PM-related Premature Mortality (3% discount rate)
None Incidences all occur in the first year
8-year Incidences all occur in the 8th year
15 -year Incidences all occur in the 15th year
6,200
6,200
6,200
11,000
11,000
11,000
$49,000
$40,000
$33,000
$89,000
$72,000
$59,000
Alternative Mortality Risk Valuation Based on Age Specific VSL
VSL applied to statistical deaths avoided in
populations 70 and over equal to 65% of VSL for
avoided deaths in populations under 70
6,200
11,000
$36,000
$63,000
Alternative Thresholds
No Threshold (base estimate)
5
10
15
20
25
6,200
6,200
5,000
1,300
500
150
11,000
11,000
9,400
3,200
1,000
430
$47,000
$47,000
$38,000
$10,000
$3,800
$1,100
$85,000
$85,000
$72,000
$25,000
$8,000
$3,300
1 Incidences rounded to two significant digits.
! Dollar values rounded to two significant digits.
The results of these sensitivity analysis demonstrate that choice of C-R function can have
a large impact on benefits, potentially doubling the effect estimate if the C-R function 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. If no lag is assumed, benefits are increased by around five percent. The threshold
9-188
-------
Cost-Benefit Analysis
analysis indicates that approximately 80 percent of the premature mortality related benefits are
due to changes in PM25 concentrations occurring above 10 ng/m3, and around 20 percent are due
to changes above 15 ng/m3, the current PM25 standard.
9B.2 Premature Mortality—Short term exposure
The Alternative estimate is based on several key parameters, including the starting point
value of a statistical life used to calculate the value of a statistical life year and the number of life
years gained for each premature death from air pollution avoided. This set of senstivity analyses
examines how changes to each of these assumptions will impact the Alternative Estimate. Two
alternative values are examined for each parameter. For the starting VSL, values of $1 million
and $10 million are used. For the number of life years gained, values of 1 year and 14 years are
used. Results are presented in Table 9B-2. We performed the analysis below using a 3%
discount rate. We will also be conducting a similar analysis using a 7% discount rate and
including this information in the public docket.
Table 9B-2.
Impacts of VSL and Life Years Gained Assumptions on Alternative Benefits Estimates
Alternative Calculation
1
2
3
4
$1 million VSL
$10 million VSL
1 life year gained
14 life years
gained
Description of Estimate
Derivation of VSLY based on starting
VSL of $1 million
Derivation of VSLY based on starting
VSL of $10 million
Assumes each premature mortality avoided
due to reductions in short-term exposures
to PM2 5 results in 1 life year gained.
Assumes each premature mortality avoided
due to reductions in short-term exposures
to PM2 5 results in 14 life years gained.
Impact on Alternative Benefit
Estimate (3% discount rate)
(Billion 2000$)
2020
-$5.3 (-48%)
+$12 (+112%)
-$5.5 (-50%)
+$13 (+116%)
2030
-$9.8 (-52%)
+$23 (+121%)
-$10 (-54%)
+$24 (+126%)
9-189
-------
Draft Regulatory Impact Analysis
9B.3 Other Health Endpoint Sensitivity Analyses
9B.3.1 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 concentration-response (C-R) function that linked a larger set of effects to a change in
pollution, rather than using C-R functions for individual effects. For example, for premature
mortality, we selected a C-R 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 C-R 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 analyses are
presented in Table 9B-3.
The benefit estimates presented in the Alternative estimate in Tables 9A-30 and 9A-31 of
Appendix 9A do not capture any additional short-term mortality impacts related to changes in
exposure to ambient ozone. 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. The authors 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. Weather is a potentially important
confounder of the ozone/mortality effect, and Thurston and Ito found that earlier studies using
less sophisticated approaches to controlling for weather consistently under-predicted the
ozone/mortality effect. They 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 Ito and Thurston only controlled for PM10 or
9-190
-------
Cost-Benefit Analysis
broader measures of particles and did not directly control for PM2 5. As such, there may still be
potential for confounding of PM2 5 and ozone mortality effects, as ozone and PM25 are highly
correlated during summer months in some areas. 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). Thus, recent evidence suggests that by not
including an estimate of reductions in short-term mortality due to changes in ambient ozone, both
the Base and Alternative Estimates 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 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 9B-2.
Sensitivity Estimates for Potentially Overlapping EndpointsA
Description of Sensitivity Analysis
Avoided Incidences
2020
2030
Monetized Value
(Million 2000$)
2020
2030
Mortality from Short-term Ozone Exposure8
Ito and Thurston (1996)
Kinney etal. (1995)
Moolgavkar et al. (1995)
Samet etal. (1997)
Pooled estimate (random effects weights)
440
0
77
120
94
1,000
0
240
360
280
$3,500
$0
$620
$960
$750
$8,100
$0
$1,900
$2,900
$2,300
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
$38
$340
$71
$490
A All estimates rounded to two significant digits.
B Mortality valued using Base estimate of $6.3 million per premature statistical death, adjusted for income growth.
9-191
-------
Draft Regulatory Impact Analysis
9B.3.2 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 9B.3, we show the impact of alternative assumptions about other
parameters, including infant mortality associated with exposure to PM, treatment of reversals in
chronic bronchitis as lowest severity cases, effects of ozone on new incidences of chronic
asthma, alternative C-R function for chronic bronchitis, alternative C-R 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.
9-192
-------
Cost-Benefit Analysis
Table 9B-3.
Additional Parameter Sensitivity Analyses
Alternative Calculation
1
2
3
4
5
6
7
Infant Mortality
Chronic Asthmaa
Reversals 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
Household soiling
damage
Description of Estimate
Avoided incidences of mortality in infants
are estimated using the Woodruff et al
(1997) C-R function. The number of
avoided incidences of infant mortality is
35 in 2020 and 52 in 2030
Avoided incidences of chronic asthma are
estimated using the McDonnell, et al.
(1999) C-R function relating annual
average ozone levels to new incidences of
asthma in adult males over the age of 27.
The number of avoided incidences of
chronic asthma is 1,200 in 2020 and 2,400
in 2030
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 et al. (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 et al. (1990).
Value of decreases in expenditures on
cleaning are estimated using values
derived from Manuel, et al. (1983).
Impact on Base Benefit Estimate
(3% discount rate)
(million 2000$)
2020
+$270 (+0.5%)
+$36 (+0.1%)
+$730 (+1.4%)
+$640 (+1.2%)
+$700 (+1.3%)
+$530 (+1.0%)
+$170 (+0.3%)
2030
+$400 (+0.4%)
+$74 (+0.1%)
+$1,100 (+1.2%)
+$970 (+1.1%)
+$1,100 (+1.1%)
+$830 (+0.9%)
+$260 (+0.3%)
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, et al., 1999).
9-193
-------
Draft Regulatory Impact Analysis
The estimated effect of PM exposure on premature mortality in post neo-natal infants
(row 1 of Table 9B.3) is based on a single U.S. study (Woodruff et al.,1997) which, on SAB
advice, was deemed too uncertain to include in the primary analysis. Adding this endpoint to the
primary benefits estimate would result in an increase in total benefits. The infant mortality
estimate indicates that exclusion of this endpoint does not have a large relative impact, either in
terms of incidences (35 in 2020 and 52 in 2030) or monetary value (approximately $270 million
in 2020 and $400 million in 2030).
The alternative calculation for the development of chronic asthma (row 2 of Table 9B.3)
is estimated using a recent study by McDonnell, et al. (1999) which found a statistical association
between ozone and the development of asthma in adult white, non-Hispanic males. Other studies
have not identified an association between air quality and the onset of asthma. The McDonnell, et
al. prospective cohort study found a statistically significant effect for adult males, but none for
adult females. EPA believes it to be appropriate to apply the C-R function to all adult males over
age 27 because no evidence exists to suggest that non-white adult males have a lower
responsiveness to air-pollution. For other health effects such as shortness of breath, where the
study population was limited to a specific group potentially more sensitive to air pollution than
the general population (Ostro et al., 1991), EPA has applied the C-R function only to the limited
population.
Some commentors have raised questions about the statistical validity of the associations
found in this study and the appropriateness of transferring the estimated C-R function from the
study populations (white, non-Hispanic males) to other male populations (i.e. African-American
males). Some of these concerns include the following: 1) no significant association was
observed for female study participants also exposed to ozone; 2) the estimated C-R function is
based on a cross-sectional comparison of ozone levels, rather than incorporating information on
ozone levels over time; 3) information on the accuracy of self-reported incidence of chronic
asthma was collected but not used in estimating the C-R function; 4) the study may not be
representative of the general population because it included only those individuals living 10 years
or longer within 5 miles of their residence at the time of the study; and 5) the study had a
significant number of study participants drop out, either through death, loss of contact, or failure
to provide complete or consistent information. EPA believes that while these issues may result
in increased uncertainty about this effect, none can be identified with a specific directional bias in
the estimates. In addition, the SAB reviewed the study and deemed it appropriate for
quantification of changes in ozone concentrations in benefits analyses (EPA-SAB-COUNCIL-
ADV-00-001, 1999). EPA recognizes the need for further investigation by the scientific
community to confirm the statistical association identified in the McDonnell, et al. study.
Following SAB advice (EPA-SAB-COUNCIL-ADV-00-001, 1999) and consistent with
the Section 812 Prospective Report, we quantify this endpoint for the RIA. However, it should
be noted that it is not clear that the intermittent, short-term, and relatively small changes in
annual average ozone concentrations resulting from this rule alone are likely to measurably
change long-term risks of asthma.
9-194
-------
Cost-Benefit Analysis
Similar to the valuation of chronic bronchitis, WTP to avoid chronic asthma is presented
as the net present value of what would potentially be a stream of costs and lower well-being
incurred over a lifetime. Estimates of WTP to avoid asthma are provided in two studies, one by
Blumenschein and Johannesson (1998) and one by O'Conor and Blomquist (1997). Both studies
use the contingent valuation method to solicit annual WTP for an asthma cure (or almost
complete cure) from individuals who have been diagnosed as asthmatics. The central estimate of
lifetime WTP to avoid a case of chronic asthma among adult males, approximately $25,000, is
the average of the present discounted value from the two studies. Details of the derivation of this
central estimate from the two studies is provided in the benefits TSD for this RIA (Abt
Associates, 2003).
Another important issue related to chronic conditions is the possible reversal in chronic
bronchitis incidences (row 3 of Table 9B.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 4 of Table 9B.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 5 and 6 of Table 9B.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 7 of Table 9B.3) is based on the
Manuel, et al. study of consumer expenditures on cleaning and household maintenance. This
9-195
-------
Draft Regulatory Impact Analysis
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.
9B.4 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 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 9B.4 lists the ranges
elasticity values used to calculate the income adjustement factors, while Table 9B.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 9B.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.
Table 9B-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.
9-196
-------
Cost-Benefit Analysis
Table 9B-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
1.023
1.156
1.047
-
2030
1.025
1.170
1.052
-
Upper Sensitivity Bound
2020
1.190
1.420
1.814
-
2030
1.208
1.464
1.914
-
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.
Table 9B-6.
Sensitivity Analysis of Alternative Income ElasticitiesA
Benefit Category
Minor Health Effect
Severe and Chronic Health Effects
(base estimate)
Premature Mortality (base estimate)
Visibility and Other Welfare EffectsA
Total Benefits
Lower Sensitivity Bound
2020
$1,400
$1,700
$38,000
$1,500
$43,000
2030
$2,200
$2,600
$67,000
$2,400
$75,000
Upper Sensitivity Bound
2020
$1,400
$2,100
$67,000
$1,500
$72,000
2030
$2,200
$3,300
$130,000
$2,400
$130,000
1 All estimates rounded to two significant digits.
! No range was applied for visibility because no ranges were available in the current published literature.
9-197
-------
Draft Regulatory Impact Analysis
Appendix 9B References
Abt Associates, Inc. 2003. Proposed Nonroad 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, andN.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 El gar 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 I.July.
EPA-SAB-COUNCIL-ADV-01-004. 2001. Review of the Draft Analytical Plan for EPAs
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.
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.
9-198
-------
Cost-Benefit Analysis
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, June
O
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., HI, 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: 1132-1141.
Samet, J.M., S.L. Zeger, J.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.
9-199
-------
Draft Regulatory Impact Analysis
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.
9-200
-------
APPENDIX 9C: Visibility Benefits Estimates for Individual Class I Areas
Table 9C-1
Apportionment Factors for 2020 Park Specific Visibility Benefits
PARK
shenandoah
\naconda-Pintlar W
Boundary Waters Canoe A
Breton W
sle Royale
arbidge W
Medicine Lake W
led Rock Lakes W
loosevelt Campobello IP
selway-Bitterroot W
>eney W
Nolf Island W
\gua Tibia W
Black Canyon of the Gun
Caribou W
^hiricahua
^ucamonga W
)ome Land W
7lat Tops W
jrand Canyon
-loover W
bhn Muir W
Caiser W
M Garita W
vlazatzal W
an Gabriel W
san Gorgino W
san Jacinto W
COUNTY
Lawrence Co
Cochise Co
Gila Co
Gila Co
Coconino Co
Apache Co
Apache Co
Graham Co
PimaCo
Maricopa Co
Coconino Co
Yavapai Co
Tuolumne Co
San Bernardino Co
Calaveras Co
Trinity Co
Fresno Co
Mono Co
Inyo Co
MarinCo
Los Angeles Co
Monterey Co
San Benito Co
Riverside Co
Siskiyou Co
San Bernardino Co
Del Norte Co
Shasta Co
Fresno Co
Lassen Co
Riverside Co
San Diego Co
Shasta Co
El Dorado Co
Mariposa Co
Fresno Co
Percent of 2020 Visibility Benefit Due to Changes in:
CTATF
si Air, SQ^ NQx direct PM
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
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
0.234
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
0.338
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
9-201
-------
Draft Regulatory Impact Analysis
PARK
san Rafael W
sequoia-Kings
sycamore Canyon W
/entana W
folla-Bolly-Middle-Eel
fosemite
Carlsbad Caverns
jilaW
oyce Kilmer-Slickrock
Calmiopsis W
^inville Gorge W
^ostwood W
3ecos W
'residential Range-Dry
Salt Creek W
shining Rock W
Vheeler Peak W
Wichita Mountains W
7itzpatrick W
jlacier Peak W
^/lount Adams W
Dolly Sods W
vforth Absaroka W
Olympic
_,ye Brook W
Bridger W
joat Rocks W
)tter Creek W
'asayten W
Bandelier
Bosque del Apache W
Brigantine W
Crater Lake
^/lount Hood W
^/lount Washington W
san Pedro Parks W
swanguarter W
Theodore Roosevelt
Maroon Bells-Snowmass W
^/lount Rainier
sforth Cascades
COUNTY
Tuolumne Co
Tulare Co
Siskiyou Co
Santa Barbara Co
Tulare Co
Modoc Co
San Juan Co
Garfield Co
Routt Co
Larimer Co
Pitkin Co
Alamosa Co
Gunnison Co
Montezuma Co
Montrose Co
Summit Co
Mineral Co
Larimer Co
Monroe Co
Wakulla Co
Citrus Co
Charlton Co
Mclntosh Co
Edmonson Co
Stone Co
Hyde Co
Haywood Co
Avery Co
Graham Co
Sandoval Co
Rio Arriba Co
Grant Co
Chaves Co
Mora Co
Eddy Co
Socorro Co
Taos Co
Lincoln Co
Elko Co
Polk Co
Blount Co
Percent of 2020 Visibility Benefit Due to Changes in:
CTATF
SO2 NOx direct PM
CA
CA
CA
CA
CA
CA
CO
CO
CO
CO
CO
CO
CO
CO
CO
CO
CO
CO
FL
FL
FL
GA
GA
KY
MS
NC
NC
NC
NC
NM
NM
NM
NM
NM
NM
NM
NM
NM
NV
TN
TN
0.090
0.052
0.469
0.111
0.052
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
0.580
0.478
0.220
0.156
0.478
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
0.330
0.470
0.311
0.733
0.470
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
9-202
-------
Cost-Benefit Analysis
PARK
Bob Marshall W
.rates of the Mountain W
jlacier
>t. Marks W
/oyageurs
Teton W
Yellowstone
jrand Teton NP
Vashakie W
COUNTY
San Juan Co
Grand Co
San Juan Co
Washington Co
Garfield Co
Botetourt Co
Madison Co
Grant Co
Tucker Co
Percent of 2020 Visibility Benefit Due to Changes in:
CTATF
SO2 NOx direct PM
UT
UT
UT
UT
UT
VA
VA
WV
WV
0.373
0.354
0.373
0.219
0.295
0.485
0.385
0.533
0.568
0.048
0.038
0.048
0.096
0.052
0.151
0.316
0.190
0.118
0.579
0.608
0.579
0.685
0.652
0.364
0.300
0.278
0.314
Table 9C-2. Apportionment Factors for 2030 Park Specific Visibility Benefits
PARK
shenandoah
\naconda-Pintlar W
Boundary Waters Canoe A
Breton W
sle Royale
arbidge W
Medicine Lake W
led Rock Lakes W
loosevelt Campobello IP
>elway-Bitterroot W
>eney W
Nolf Island W
\gua Tibia W
Black Canyon of the Gun
Caribou W
^hiricahua
^ucamonga W
)ome Land W
7lat Tops W
jrand Canyon
loover W
ohn Muir W
Caiser W
.a Garita W
^azatzal W
COUNTY
Lawrence Co
Cochise Co
Gila Co
Gila Co
Coconino Co
Apache Co
Apache Co
Graham Co
PimaCo
Maricopa Co
Coconino Co
Yavapai Co
Tuolumne Co
San Bernardino Co
Calaveras Co
Trinity Co
Fresno Co
Mono Co
Inyo Co
MarinCo
Los Angeles Co
Monterey Co
San Benito Co
Riverside Co
Siskiyou Co
Percent of 2030 Visibility Benefit Due to Changes in:
CTATF
SO2 NOx direct PM
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
0.376
0.313
0.277
0.293
0.342
0.429
0.429
0.188
0.207
0.342
0.057
0.293
0.055
0.226
0.065
0.129
0.039
0.046
0.070
0.070
0.049
0.033
0.049
0.049
0.116
0.297
0.075
0.048
0.089
0.107
0.069
0.069
0.173
0.072
0.107
0.019
0.089
0.571
0.407
0.191
0.111
0.520
0.493
0.616
0.616
0.109
0.376
0.109
0.109
0.518
0.327
0.612
0.675
0.619
0.551
0.503
0.503
0.639
0.721
0.551
0.924
0.619
0.375
0.368
0.745
0.759
0.441
0.461
0.314
0.314
0.842
0.591
0.842
0.842
0.366
9-203
-------
Draft Regulatory Impact Analysis
PARK
^lesa Verde
'etrified Forest
3ine Mountain W
•'innacles
3oint Reyes
lawah W
locky Mountain
saguaro
san Gabriel W
san Gorgino W
san Jacinto W
san Rafael W
sequoia-Kings
sycamore Canyon W
/entana W
folla-Bolly-Middle-Eel
fosemite
Carlsbad Caverns
jilaW
byce Kilmer-Slickrock
Calmiopsis W
^inville Gorge W
^ostwood W
3ecos W
'residential Range-Dry
salt Creek W
shining Rock W
Wheeler Peak W
iVichita Mountains W
7itzpatrick W
jlacier Peak W
vlount Adams W
Dolly Sods W
sTorth Absaroka W
Olympic
_,ye Brook W
Sridger W
joat Rocks W
)tter Creek W
3asayten W
Sandelier
COUNTY
San Bernardino Co
Del Norte Co
Shasta Co
Fresno Co
Lassen Co
Riverside Co
San Diego Co
Shasta Co
El Dorado Co
Mariposa Co
Fresno Co
Tuolumne Co
Tulare Co
Siskiyou Co
Santa Barbara Co
Tulare Co
Modoc Co
San Juan Co
Garfield Co
Routt Co
Larimer Co
Pitkin Co
Alamosa Co
Gunnison Co
Montezuma Co
Montrose Co
Summit Co
Mineral Co
Larimer Co
Monroe Co
Wakulla Co
Citrus Co
Charlton Co
Mclntosh Co
Edmonson Co
Stone Co
Hyde Co
Haywood Co
Avery Co
Graham Co
Sandoval Co
Percent of 2030 Visibility Benefit Due to Changes in:
CTATF
SO2 NOx direct PM
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CO
CO
CO
CO
CO
CO
CO
CO
CO
CO
CO
CO
FL
FL
FL
GA
GA
KY
MS
NC
NC
NC
NC
NM
0.411
0.411
0.158
0.043
0.047
0.053
0.468
0.090
0.065
0.033
0.099
0.046
0.225
0.116
0.059
0.321
0.073
0.312
0.464
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
0.270
0.270
0.344
0.535
0.663
0.588
0.133
0.175
0.191
0.376
0.179
0.493
0.452
0.518
0.593
0.292
0.400
0.203
0.087
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
0.320
0.320
0.498
0.422
0.289
0.360
0.399
0.735
0.745
0.591
0.722
0.461
0.323
0.366
0.348
0.386
0.527
0.485
0.449
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
9-204
-------
Cost-Benefit Analysis
PARK
Bosque del Apache W
Brigantine W
Crater Lake
^lount Hood W
^lount Washington W
>an Pedro Parks W
swanguarter W
Theodore Roosevelt
Maroon Bells-Snowmass W
^/lount Rainier
sforth Cascades
Bob Marshall W
.rates of the Mountain W
jlacier
It. Marks W
/oyageurs
Teton W
Yellowstone
jrand Teton NP
Vashakie W
COUNTY
Rio Arriba Co
Grant Co
Chaves Co
Mora Co
Eddy Co
Socorro Co
Taos Co
Lincoln Co
Elko Co
Polk Co
Blount Co
San Juan Co
Grand Co
San Juan Co
Washington Co
Garfield Co
Botetourt Co
Madison Co
Grant Co
Tucker Co
Percent of 2030 Visibility Benefit Due to Changes in:
CTATF
SO2 NOx direct PM
NM
NM
NM
NM
NM
NM
NM
NM
NV
TN
TN
UT
UT
UT
UT
UT
VA
VA
WV
WV
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
0.037
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
0.589
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
9-205
-------
CHAPTER 10: Economic Impact Analysis
10.1 Overview of Results 10-1
10.1.1 What is an Economic Impact Analysis? 10-1
10.1.2 What is EPA's Economic Analysis Approach for this Proposal? 10-1
10.1.3 What are the key features of the NDEIM? 10-4
10.1.4 Summary of Economic Analysis 10-8
10.1.4.1 What are the Expected Market Impacts of this Proposal? 10-9
10.1.4.2 What are the Expected Social Costs of this Proposal? 10-13
10.2 Economic Methodology 10-19
10.2.1 Behavioral Economic Models 10-20
10.2.2 Conceptual Economic Approach 10-20
10.2.2.1 Types of Models: Partial vs. General Equilibrium Modeling Approaches 10-20
10.2.2.2 Market Equilibrium in a Single Commodity Market 10-22
10.2.2.3 Incorporating Multimarket Interactions 10-23
10.2.3 Key Modeling Elements 10-28
10.2.3.1 Perfect vs. Imperfect Competition 10-28
10.2.3.2 Short- vs. Long-Run Models 10-29
10.2.3.3 Variable vs. Fixed Regulatory Costs 10-33
10.2.3.4 Estimation of Social Costs 10-36
10.3 Economic Impact Modeling 10-39
10.3.1 Operational Economic Model 10-39
10.3.2 Baseline Economic Data 10-40
10.3.2.1 Baseline Population 10-40
10.3.2.2 Baseline Prices 10-43
10.3.3 Market Linkages 10-44
10.3.3.1 Engine Markets 10-45
10.3.3.2 Equipment Markets 10-46
10.3.3.3 Application Markets 10-47
10.3.3.4 Diesel Fuel Markets 10-48
10.3.3.5 Calibrating the Spillover Baseline (Impacts Relative to Highway Rule) 10-50
10.3.4 Compliance Costs 10-50
10.3.4.1 Engine and Equipment Compliance Costs 10-51
10.3.4.2 Nonroad Diesel Fuel Compliance Costs 10-52
10.3.4.3 Changes in Operating Costs 10-57
10.3.4.4 Fuel Marker Costs 10-58
10.3.5 Supply and Demand Elasticity Estimates 10-59
10.3.6 Model Solution Algorithm 10-62
APPENDIX 10A: Impacts on the Engine Market and Engine Manufacturers 10-66
APPENDIX 10B: Impacts on Equipment Market and Equipment Manufacturers 10-75
APPENDIX IOC: Impacts on Application Market Producers and Consumers 10-84
APPENDIX 10D: Impacts on the Nonroad Fuel Market 10-88
APPENDIX 10E: Time Series of Social Cost 10-93
APPENDIX 10F: Model Equations 10-96
APPENDIX 10G: Elasticity Parameters for Economic Impact Modeling 10-101
APPENDIX 10H: Derivation of Supply Elasticity 10-117
APPENDIX 101: Sensitivity Analysis 10-118
-------
Economic Impact Analysis
CHAPTER 10: Economic Impact Analysis
An Economic Impact Analysis (EIA) was prepared to estimate the economic impacts of this
proposal on producers and consumers of nonroad engines and equipment and related industries.
The Nonroad Diesel Economic Impact Model (NDEIM), developed for this analysis, was used 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 (RTI, 2003).
10.1 Overview of Results
This section provides a summary of the EPA economic analysis approach and presents an
overview of its results. As described below, the overall economic impact of the proposed
emission control program on society should be minimal. According to this analysis, the average
price of goods and services produced using equipment and fuel affected by the proposal is
expected to increase by about 0.02 percent. A more detailed description of this analysis is
presented in the following sections of Chapter 10 and the corresponding appendices.
10.1.1 What is an Economic Impact Analysis?
Regulatory agencies conduct economic impact analyses of potential regulatory actions to
inform decision makers about the effects of a proposed regulation on society's current and future
well-being. In addition to informing decision makers within the Agency, economic impact
analyses are conducted to meet the statutory and administrative requirements imposed by
Congress and the Executive office. The Clean Air Act requires an economic impact analysis
under section 317, while Executive Order 12866-Regulatory Planning and Review requires
Executive Branch agencies to perform benefit-costs analysis of all rules it deems to be
"significant" (typically over $100 million annual social costs) and submit these analysis to the
Office of Management and Budget (OMB) for review. This economic impact analysis estimates
the potential market impacts of the proposed rule's compliance costs and provides the associated
social costs and their distribution across stakeholders for comparison with social benefits (as
presented in Chapter 9).
10.1.2 What is EPA's Economic Analysis Approach for this Proposal?
The underlying objective of an EIA is to evaluate the effect of a proposed regulation on the
welfare of affected stakeholders and society in general. Using information on the expected
compliance costs of the proposed program as presented in Chapters 6 and 7, this EIA explores
how the companies that produce nonroad diesel engines, equipment, or fuel may change their
production behavior in response to the costs of complying with the standards. It also explores
how the consumers who use the affected products may change their purchasing decisions. For
10-1
-------
Draft Regulatory Impact Analysis
example, the construction industry may reduce purchases if the prices of nonroad diesel
equipment increase, thereby reducing the volume of equipment sold (or market demand) for such
equipment. Alternatively, the construction industry may pass along these additional costs to the
consumers of their final goods and services by increasing prices, which would mitigate the
potential impacts on the purchases of nonroad diesel equipment.
The Nonroad Diesel Economic Impact Model (NDEEVI) developed for this analysis evaluates
how producers and consumers are expected to respond to the regulatory costs associated with the
proposed emission control program. The conceptual approach is to link significantly affected
markets to mimic how compliance costs will potentially ripple through the economy. The
NDEEVI employs a multi-market partial equilibrium framework to track changes in price and
quantity for over 50 integrated product markets. Figure 10.1-1 illustrates the industry segments
included in the model and the flow of compliance costs through the economic system.
As shown in Figure 10.1-1, the compliance costs will be directly borne by engine
manufacturers, equipment manufacturers, and petroleum refineries. Depending on market
characteristics, some or all of these compliance costs will be passed on through the supply chain
in the form of higher prices extending to producers and consumers in the application markets
(i.e., construction, agriculture, and manufacturing). In this way the proposed rule indirectly
affects producers and consumers in all of the related markets included in Figure 10.1-1. For
example, the proposed rule will increase the cost of producing nonroad diesel engines. Engine
manufacturers will attempt to pass these increased costs on to equipment manufacturers in the
form of higher diesel engine prices. Similarly, equipment manufacturers will attempt to pass
their direct compliance costs and the increased cost of engines to application manufacturers
through higher diesel equipment prices. Petroleum refiners will also attempt to pass their direct
compliance costs on to application manufacturers through higher prices for diesel fuel. Finally,
application manufacturers will look to pass on the increased equipment and diesel fuel costs to
consumers of final application products and services. The NDEEVI explicitly models these
linkages and estimates the behavioral responses that lead to new equilibrium prices and output
for all related markets and the resulting distribution of social costs across affected stakeholders.
10-2
-------
Economic Impact Analysis
Figure 10.1-1
Market Linkages Included in Economic Model
Application Consumers
c
Agriculture
Application Markets
• Construction • Manufacturing
Application Suppliers
Diesel Equipment Markets (by hp size)
• Agriculture • Refrigeration • Pumps and
Construction
• Generator Sets
• Lawn and Garden
Compressors
• Industrial
Diesel Equipment
Manufacturers
Diesel Engine Markets
<25hp • 101-175 hp
• 26-50 hp .176-600 hp
51-75hp »>601 hp
•76-100 hp
Diesel Engine
Manufacturers
/Diesel Fuel
/ (by PADD Region) \
I • 500 ppm sulfur content I
V • 15 ppm sulfur content /
10-2
-------
Draft Regulatory Impact Analysis
10.1.3 What are the key features 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
(RTI, 2003). The model methodology is firmly rooted in applied microeconomic theory and was
developed following the OAQPS Economic Analysis Resource Document (EPA, 1999). Table
10.1-1 provides a summary of the markets included in the NDEIM, including their baseline
characterization and behavioral response parameters (i.e., supply and demand elasticities). These
market characteristics are described in more detail in Section 10.3. Based on the specified
market linkages, shown in Figure 10.1-1, 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 NDEIM uses a multi-market partial equilibrium approach to track changes in price and
quantity for 60 integrated product markets, 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)
• 42 diesel equipment markets (7 horsepower categories within 7 application categories:
construction, agricultural, 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 (construction, agriculture, and manufacturing)
• 8 nonroad diesel fuel markets (2 sulfur content levels of 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.
The economic impacts of the proposed rule are largely determined by behavioral response
parameters within the model (i.e., the supply and demand elasticities). For most markets, as
summarized in Table 10.1-1, the supply and demand elasticities were either obtained from the
professional literature or econometrically estimated. Details on sources and estimation method
are provided in Section 10.3 and Appendix 10G. Demand responses in the equipment, engine,
and diesel fuel markets are derived internally as a function of changes in output levels in the
applications markets (i.e., derived demand specification). Therefore, parameter values are not
required for demand elasticities in these markets.
The actual economic impacts of the proposed rule will be determined by the ways in which
producers and consumers of the engines, equipment, and fuels affected by the proposal change
their behavior in response to the costs incurred in complying with the standards. In the NDEIM,
these behaviors are modeled by the demand and supply elasticities. Table 10.1-1 summarizes the
sources of the demand and supply elasticities used in the model; more information can be found
in section 10.3-5, below. As noted in Table 10.1-1, the supply elasticities for the engine and
10-4
-------
Economic Impact Analysis
equipment markets and the demand elasticities for the application markets were estimated using
econometric methods. The procedures and results are reported in Appendix 10.1 of this draft
RIA. Literature-based estimates were used for the supply elasticities in the application and fuel
markets. There are two ways to handle the demand elasticities for the engine, equipment, and
fuel markets. The approach in NDEIM internally derives these elasticities based on the specified
market linkages, i.e., the demand for engines, equipment, and fuel are modeled as directly related
to the supply and demand of goods and services supplied by the final application markets. In
other words, the supply of those goods and services determines the demand for equipment and
fuel, and the supply of equipment determines the demand for engines. An alternative approach
could be used in which the demand elasticities for the equipment, engine, and fuel markets are
estimated outside the model.
The estimated supply and demand elasticities used in this analysis for the application markets
and the supply elasticity for the diesel fuel market are inelastic or unit elastic. This means that
the quantities of goods and services demanded/supplied are expected to be fairly insensitive to
price changes (inelastic) or that the quantity demanded/supplied is expected to vary directly with
changes in prices. In other words, price changes are not expected to have a large impact on the
level of consumption in these application markets. For the agricultural application market, the
inelastic supply and demand elasticities reflects 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 flexibility to adjust production levels. The estimated supply elasticity
for the diesel fuel market is also inelastic, because most refineries operate near capacity and are
therefore less responsive to fluctuations in market prices. The supply elasticities used in this
analysis for the engine and equipment markets, on the other hand, are fairly elastic. This means
that quantities supplied in these markets are expected to be very responsive to price changes, that
manufacturers are more likely (better able) to change production levels in response to price
changes. The demand elasticities for the diesel engine and equipment markets and for the diesel
fuel market are not explicitly specified because these demand levels are derived as part of the
modeled outcomes for the application markets. 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.
10-5
-------
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
proposed standard*
Perfectly
competitive
Power Systems
Research (PSR)
database for 2000
as modified by
EPAf
EPA's nonroad
model
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
population
Based on engine
growth
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
Published
econometric
estimate
(inelastic)
Econometric
estimate
(inelastic)
No direct
compliance costs
but higher prices
for diesel
equipment and
fuel cause shift in
supply function
Horsepower categories are 0-25, 26-50, 51-75, 76-100, 101-175, 176-600, and 601 and greater; the EIA
includes more horsepower categories than the standards, allowing more efficient use of the engine compliance
cost estimates developed for this proposal.
Engine categories are agricultural (SIC 3523), construction (SIC 3531), pumps and compressors (SIC 3561 and
3563), generator and welder sets (SIC 3548), refrigeration and air conditioning (SIC 3585), general industrial
(SIC 3537), and lawn and garden (SIC 3524).
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). It should be noted that PADD 5 includes Alaska
and Hawaii.
Application market categories are construction, agriculture, and manufacturing.
See Section 8.1 in Chapter 8 of this draft RIA for an explanation of how the engines were allocated to the seven
categories.
-------
Economic Impact Analysis
Because the elasticity estimates are a key input to the model, a sensitivity analysis for supply
and demand elasticity parameters used in the model was also 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
because equipment and diesel fuel costs are a relatively small share of total production costs in
the construction, agriculture, and manufacturing industries. As a result compliance costs are
expected to have little influence on production in these application markets, and the derived
demand for equipment, engines and fuel are minimally affected.
With regard to the compliance costs that are used to shock the model, the NDEEVI uses the
expected increase in variable costs associated with the proposed engine emission standards and
the sum of variable and fixed costs associated with the fuel standards. Fixed costs associated
with the engine emission standards are not included in the market analysis reported in Table
10.1-2. This is 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, 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. Therefore, fixed costs
are not likely to affect the prices of engines or equipment. This assumption is described in
greater detail below in Section 10.2.3.3. 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 10.1-3 as additional costs on producers.
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
their behavior based on the R&D required for compliance with the standards. A sensitivity
analysis is included in Chapter 10 of the draft RIA for this proposal that reflects this approach.
In addition to the variable and fixed costs described above, there are three additional costs
components that are included in the total social cost estimates of the proposed regulation but that
are not explicitly included in the NDEEVI. These are operating savings (costs), fuel marker costs,
and spillover from 15 ppm fuel to higher sulfur fuel.
Operating savings (costs) 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 engines, and marine and locomotive engines, that are already
in use as well as new nonroad engines that will comply with the proposed standards. These
savings (costs) also include any extra operating costs associated with the new PM emission
control technology which may accrue to new engines that use this new technology. These
savings (costs) are not included directly in the model because some of the savings accrue to
10-7
-------
Draft Regulatory Impact Analysis
existing engines and because these savings (costs) are not expected to affect consumer decisions
with respect to new engines. Instead, they are added into the estimated welfare impacts as
additional costs to the application markets, since it is the users of these engines that will see these
savings (costs). Nevertheless, a sensitivity analysis was also performed in which these savings
(costs) are included as inputs to the NDEEVI, where they are modeled as benefits accruing to the
application producers. The results of this analysis are presented in Appendix 10.1.
Fuel marker costs refers to costs associated with marking high sulfur diesel fuel in the
locomotive, marine, and heating oil markets between 2007 and 2014. Marker costs are not
included in the market analysis because locomotive, marine, and heating oil markets are not
explicitly modeled in the NDEEVI. Similar to the operating savings (costs), marker costs are
added into the estimated welfare impacts separately.
The costs of fuel that spills over from the 15 ppm market to higher grade sulfur fuel are also
not included in the NDEEVI but, instead, are added into the estimated welfare impacts separately.
As described in Chapter 7, refiners are expected to produce more 15 ppm fuel than is required for
the nonroad diesel fuel market. This excess 15 ppm fuel will be sold into markets that allow fuel
with a higher sulfur level (e.g., locomotive, marine diesel, or home heating fuel). Because this
spillover fuel will meet the 15 ppm limit, it is necessary to count the costs of sulfur reduction
processes against those fuels.
Consistent with the engine and equipment cost discussion in Chapter 6, this EIA does not
include any cost savings associated with the proposed equipment transition flexibility program or
the proposed nonroad engine ABT program. As a result, the results of this EIA can be viewed as
somewhat conservative.
10.1.4 Summary of Economic Analysis
The economic analysis consists of two parts: a market analysis and welfare analysis. The
market analysis looks at expected changes in prices and quantities for directly and indirectly
affected market commodities as shown in Figure 10.1-1. The welfare analysis looks at economic
impacts in terms of annual and present value changes in social costs. For this proposed rule, the
social costs are computed as the sum of market surplus offset by operating cost savings. Market
surplus is equal to the aggregate change in consumer and producer surplus based on the estimated
market impacts associated with the proposed rule. Operating cost savings are associated with the
decreased sulfur content of diesel fuel. These include maintenance savings (cost reductions) and
changes in fuel efficiency. Increased maintenance costs may also be incurred for some
technologies. Operating costs are not included in the market analysis but are instead listed as a
separate category in the social cost results tables.
As noted in Chapter 6, engine and equipment costs vary over time because fixed costs are
recovered over five to ten year periods while variable costs, despite learning effects that serve to
reduce costs on a per unit basis, continue to increase in total at a rate consistent with new sales.
Similarly, engine operating costs also vary over time because oil change maintenance savings,
10-8
-------
Economic Impact Analysis
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. Fuel related compliance costs (costs for refining and distributing the proposed
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. The
total fuel costs do 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. The
distribution costs decrease in 2014 as it would no longer be necessary to use a marker.
Economic impact results for 2013, 2020, and 2030 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 8-7-2, aggregate program
costs peak in 2014; increases in costs after that year are due increases in the population of
engines over time. The other years, 2020 and 2030, correspond to years analyzed in our benefits
analysis. Detailed results for all years are included in Appendix 10E for this chapter.
10.1.4.1 What are the Expected Market Impacts of this Proposal?
The market impacts of this rule suggest that the overall economic impact of the proposed
emission control program on society is expected be small, on average. According to this
analysis, the average price of goods and services produced using equipment and fuel affected by
the proposal is expected to increase by about 0.02 percent. The estimated price increases and
quantity reductions for engines and equipment vary depending on compliance costs. In general,
price increases would be expected to be higher (lower) as a result of a high (low) relative level of
compliance costs to market price. The change in price would also be expected to be highest
when compliance costs are highest.
This analysis indicates that most of the direct compliance costs for engine, equipment, and
fuel producers will be passed through to the application markets in the form of higher prices to
the consumers of final construction, agricultural, and manufactured goods and services. This is
expected to occur because the demand for nonroad diesel equipment (and hence the derived
demand for diesel engines and fuel) is estimated to be relatively price inelastic. The demand for
nonroad diesel equipment is inelastic because of the following:
1) Nonroad diesel equipment and fuel expenditures are a relatively small share of total
production costs for the products and services that use this equipment and fuel as inputs.
2) There are limited substitutes for nonroad diesel equipment and fuel.
The suppliers to the application markets are thus not expected to respond very much to increases
in the price of nonroad diesel equipment and fuel because these factors represent a small share of
total production costs. Furthermore, to the extent these increased costs might be significant
enough to cause a response, there are few substitutes available to these suppliers. Therefore, the
NDEEVI predicts a small decrease in demand for diesel equipment and fuel. This would allow
10-9
-------
Draft Regulatory Impact Analysis
engine, equipment and fuel producers to pass through compliance costs in the form of higher
prices.
The estimated market impacts for 2013, 2020, and 2030 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 different horsepower size categories. The 23 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, equipment impacts presented in Table 10.1-2 are
weighted averages of 42 equipment-application markets, such as small (< 25hp) agricultural
equipment and large (>600hp) industrial equipment. It should be noted that price increases and
quantity decreases for specific types of engines, equipment, application sectors, or diesel fuel
markets are likely to be different. But the data in this table provide a broad overview of the
expected market impacts that is useful when considering the impacts of the proposal on the
economy as a whole. Individual market-level impacts are presented in Appendix 10A through
Appendix 10D.
Engine Market Results: Most of the variable costs associated with the proposed rule are
passed along in the form of higher prices. The average price increase in 2013 for engines is
estimated to be about 23 percent. This percentage is expected to decrease to about 19.5 percent
for 2020 and later. 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 year of greatest
compliance costs overall, the largest expected percent price increase is for engines between 25
and 50 hp: 34 percent or $852; the average price for an engine in this category is about $2,500.
However, this price increase is expected to drop to 26 percent, or about $647, for 2016 and later.
The smallest expected percent 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 5.6 percent in 2014 and beyond. The expected price increase for these
engines is about $4,211 in 2013, increasing to about $6,950 in 2014 and later, for engines that
cost on average about $125,000.
The market impact model predicts that even with these increase in engine prices, total
demand is not expected to change very much. The expected average change in quantity is only
about 69 engines per year in 2013, 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. The average decrease in the quantity of all engines produced as a result of the regulation
is estimated to be about 0.013 percent. This decrease ranges from 0.010 percent for engines less
than 25 hp to 0.016 percent for engines 175 to 600 hp.
10-10
-------
Economic Impact Analysis
Equipment Market Results: Estimated price changes for the equipment markets reflect both
the direct costs of the proposed standards on equipment production and the indirect cost through
increased engine prices. In 2013, the average price increase for nonroad diesel equipment is
estimated to be about 5.2 percent. This percentage is expected to decrease to about 4.5 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 1.4 percent, or $1,835 for
equipment with an average cost of $130,000. This compares with an estimated engine price
increase of about $1,754 for engines of that size. The largest expected price increase in 2013 for
equipment is $4,335, or 4.9 percent, for pumps and compressors over 600 hp. This compares
with an estimated engine price increase of about $4,211 for engines of that size. The smallest
expected price increase in 2013 for equipment is $125, or 3.6 percent, for construction equipment
less than 25 hp. This compares with an estimated engine price increase of about $124 for
engines of that size. The price changes for the equipment are less than that for engines because
the engine is only one input in the production of equipment.
The output reduction for nonroad diesel equipment is estimated to be very small and to
average about 0.014 percent for all years. This decrease ranges from 0.005 percent for general
manufacturing equipment to 0.019 percent for construction equipment. The largest expected
decrease in quantity in 2013 is 13 units of construction equipment per year for construction
equipment between 100 and 175 hp, out of about 62,800 units. The smallest expected decrease
in quantity in 2013 is less than one unit per year in all hp categories of pumps and compressors.
10-11
-------
Draft Regulatory Impact Analysis
Table 10.1-2
Summary of Market Impacts ($2001)
Market
Engineering Cost
Per Unit
Change in Price
Absolute Percent
($million)
Change in Quantity
Absolute Percent
2013
Engines
Equipment
Application Markets'1
No. 2 Distillate Nonroad
$1,087
$1,021
$0.039
$840 22.9
$1,017 5.2
0.02
$0.038 4.1
-69a -0.013
-118 -0.014
-0.010
-1.38° -0.013
2020
Engines
Equipment
Application Markets1"
No. 2 Distillate Nonroad
$1,028
$1,018
$0.039
$779 19.5
$1,013 4.4
0.02
$0.039 4.1
-79a -0.013
-135 -0.014
-0.010
-1.58° -0.014
2030
Engines
Equipment
Application Markets*
No. 2 Distillate Nonroad
$1,027
$1,004
$0.039
$768 19.4
$999 4.5
0.02
$0.039 4.1
-92a -0.013
-156 -0.014
-0.010
-1.84° -0.014
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.
0 Units are in million of gallons.
Application Market Results'. The estimated price increase associated with the proposed
standards in all three of the application markets is very small and averages about 0.02 percent for
all years. In other words, on average, the prices of goods and services produced using the
engines, equipment, and fuel affected by this proposal are expected to increase only negligibly.
This is because in all of the application markets the compliance costs passed on through price
increases represent a very small share of total production costs. For example, the construction
industry realizes an increase in production costs of approximately $468 million in 2013 because
10-12
-------
Economic Impact Analysis
of the price increases for diesel equipment and fuel. However, this represents only 0.03 percent
of the $1,392 billion value of shipments in the construction industry in 2001. The estimated
average commodity price increase in 2013 ranges from 0.06 percent in the agricultural
application market to about 0.01 percent in the manufacturing application market. The
percentage change in output is also estimated to be very small and averages about 0.01 percent.
This reduction ranges from less than a 0.01 percent decrease in manufacturing to about a 0.02
percent decrease in construction. 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 4 percent for all years. For 15 ppm fuel, the estimated price increase for 2013 ranges from
3.2 percent in the East Coast region (PADD 1&3) to 9.3 percent in the mountain region (PADD
4). The average national output decrease for all fuel is estimated to be about 0.01 percent for all
years, and is relatively constant across all four regional fuel markets.
10.1.4.2 What are the Expected Social Costs of this Proposal?
Social costs include the changes in market surplus estimated by the NDEEVI and changes in
operating costs and marker costs associated with the regulation. Table 10.1-3 shows the time
series of engineering compliance costs and social cost estimates for 2007 through 2030. As
shown, these estimates are of similar magnitude for each year of the analysis. However, the
distribution of costs across the affected stakeholders is very different. This is highlighted by the
comparison of Figure 10.1-3a and Figure 10.1-3b, which show the way in which the estimated
engineering compliance costs and the estimated social costs are distributed across stakeholders,
for 2013. Figure 10. l-3a shows that the direct compliance costs are borne relatively evenly
across engine, equipment, and fuel producers, with each bearing about one-third of the costs. In
contrast, as shown in Figure 10.1-3b, most of the social costs are borne by producers and
consumers in the application markets (about 89 percent when the operating savings (costs) are
not considered) due to the increased prices for diesel engines, equipment, and fuel. Engine
producers are able to pass on 94 percent their compliance costs through higher prices. The
remaining 6 percent are primarily fixed R&D costs that are internalized by engine manufacturers
and not passed into the market. Equipment manufacturers retain a slightly higher share of
compliance costs because they have greater fixed costs. Diesel fuel refiners pass over 98 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.
10-13
-------
Draft Regulatory Impact Analysis
Table 10.1-3
National Engineering Compliance Costs and
Social Costs Estimates for the Proposed Rule (2004 - 2030)
($2001; 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
NPV at 3%
NPV at 7%
$0.00
$0.00
$0.00
$39.61
$130.41
$132.25
$262.02
$641.12
$1,010.37
$1,202.52
$1,329.14
$1,260.74
$1,298.40
$1,318.75
$1,325.02
$1,339.30
$1,366.79
$1,351.08
$1,349.58
$1,365.53
$1,371.60
$1,395.98
$1,419.79
$1,442.91
$1,465.41
$1,487.68
$1,509.77
$16,524.29
$9,894.02
Total Social Costs
$0.00
$0.00
$0.00
$39.61
$130.40
$132.25
$262.01
$641.07
$1,010.27
$1,202.40
$1,329.01
$1,260.62
$1,298.27
$1,318.62
$1,324.89
$1,339.16
$1,366.66
$1,350.94
$1,349.44
$1,365.38
$1,371.45
$1,395.83
$1,419.64
$1,442.76
$1,465.26
$1,487.53
$1,509.61
$16,522.66
$9,893.06
10-14
-------
Economic Impact Analysis
Figure 10.1-2 shows the time series of total social costs from 2007 through 2030. Social
costs increase rapidly between 2007 and 2013 as engine, equipment and fuel costs are phased
into the regulation. Estimated net annual social costs (including operating savings (cost) and
marker costs) in 2013 are about $1,202 million. After 2013, 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 2014 to 2030. The estimated net present value of social costs over the time period 2004
through 2030 based on a social discount rate of 3 percent is reported in Table 10.1-3 and is about
$16.5 billion. The present value over this same period based on a social discount rate of 7
percent is about $9.9 billion.
Figure 10.1-2
Total Social Costs (2004-2030)
$1,600.00
$1,400.00
$1,200.00
« $1,000.00
o
$800.00
$600.00
$CO.OO
$200.00
$0.00
2007
2027
10-15
-------
Draft Regulatory Impact Analysis
Figure 10.1-3
Comparison of the Distribution of Engineering Compliance
and Social Cost Estimates by Industry Segment (2013)
$1,385.9 million
Application Producers and Consumers
0.0%
Fuel Refiners
28.6%
Engine Producers
35.2%
Equipment Producers
36.2%
a) Engineering Cost Distribution3
$1,385.8 million
Application Producers
and Consumers
88.9%
Engine Producers
2.2%
Equipment Producers
8.4%
Fuel Refiners
0.6%
b) Social Cost Distribution3
a Costs do not include operating cost savings, nonroad spillover, or marker costs, which represent negative
183.4 million in costs (i.e., benefits).
10-16
-------
Economic Impact Analysis
Estimated social costs are disaggregated by market in Table 10.1-4, for 2013, 2020, and
2030. A more detailed time series from 2007 to 2030 provided is in Appendix 10.E. The data in
Table 10.1-4 shows that in 2013, social costs are expected to be about $1,202.4 million ($2001).
About 82 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 are expected to be passed on in
the form of higher prices. When these estimated impacts are broken down, about 58 percent of
the social costs are expected to be borne by consumers in the application markets and about 42
percent are expected to be borne by producers in the application markets. Equipment
manufacturers are expected to bear about 10 percent of the total social costs. Engine
manufacturers and diesel fuel refineries are expected to bear 2.5 percent and 0.5 percent,
respectively. The remaining 5.0 percent is accounted for by fuel marker costs and the additional
costs of 15 ppm fuel being sold in to markets such as marine diesel, locomotive, and home
heating fuel that do not require it.
In 2030, the total social costs are projected to be about $1,509.6 million ($2001). The
increase is due to the projected annual growth in the engine and equipment populations. As in
earlier years, producers and consumers in the application markets are expected to bear the large
majority of the costs, approximately 94 percent. This is consistent with economic theory, which
states that, in the long run, all costs are passed on to the consumers of goods and services.
10-17
-------
Table 10.1-4
Summary of Social Costs Estimates Associated with Primary Program: 2013, 2020, and 2030 ($million)a'b
Maximum Cost
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
PADD II
PADD IV
PADDV
Nonroad Spillover
Marker Costs
Total
Market
Surplus
($106)
$30.2
$116.1
$53.0
$39.9
$23.2
$1,231.8
$515.7
$716.1
$468.3
$348.7
$414.8
$7.8
$3.6
$2.9
$0.8
$0.5
$1,385.8
Operating
Savings
($106)
($241.9)
($77.9)
($44.7)
($119.3)
$51.2
$7.3
($183.4)
Year (20 13)
Total Percent
$30.2 2.5%
$116.1 9.7%
$53.0
$39.9
$23.2
$989.8 82.3%
$390.4
$304.0
$295.5
$7.8 0.6%
$3.6
$2.9
$0.8
$0.5
$4.3
0.6%
$1,202.4 100%
Market
Surplus
($106)
$0.1
$102.6
$48.2
$33.2
$21.2
$1,386.5
$583.4
$803.1
$550.4
$339.2
$436.8
$9.0
$4.1
$3.3
$0.9
$0.6
$1,498.2
Year
Operating
Savings
($106)
($190.1)
($61.2)
($35.2)
($93.8)
$58.6
-
($131.5)
2020
Total Percent
$0.1 0.01%
$102.6 7.5%
$48.2
$33.2
$21.2
$1,196.3 87.5%
$489.3
$364.0
$343.0
$9.0 0.7%
$4.1
$3.3
$0.9
$0.6
$4.3
0.0%
$1,366.7 100%
Market
Surplus
($106)
$0.1
$5.3
$3.8
$1.3
$0.2
$1,598.9
$672.9
$926.0
$635.7
$416.5
$501.8
$10.5
$4.8
$3.9
$1.0
$0.8
$1,614.9
Final Year
Operating
Savings
($106)
($174.5)
($56.1)
($32.3)
($86.0)
$69.2
-
($105.3)
(2030)
Total
$0.1
$5.3
$3.8
$1.3
$0.2
$1,424.5
$579.5
$429.2
$415.7
$10.5
$4.8
$3.9
$1.0
$0.8
$1,509.6
Percent
0.0%
0.3%
94.4%
0.7%
4.6%
0.0%
100%
a Figures are in 2001 dollars.
b Operating savings are shown as negative costs.
10-18
-------
Economic Impact Analysis
Table 10.1-5
Summary of Social Costs Estimates Associated with Primary Program:
NPV, 3%, 2004-2030 ($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
Nonroad Spillover
Marker Costs
Total
Market Surplus
($106)
$190.0
$927.4
$433.6
$306.7
$187.1
$17,744.2
$7,450.7
$10,293.5
$6,923.5
$5,050.4
$5,770.3
$113.9
$52.3
$41.9
$11.5
$8.1
$18,975.5
Fuel
Maintenance
($106) Total
$190.0
$927.4
$433.6
$306.7
$187.1
($3,402.4) $14,341.8
($1,094.9) $5,828.6
($629.3) $4,421.1
($1,678.1) $4,092.2
$113.9
$52.3
$41.9
$11.5
$8.1
$886.5
$63.0
($2,452.8) $16,522.7
Percent
1.1%
5.6%
86.8%
0.7%
5.4%
0.4%
100%
a Figures are in 2001 dollars.
b Operating savings are shown as negative costs.
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
10-19
-------
Draft Regulatory Impact Analysis
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.
10.2.1 Behavioral Economic Models
Models incorporating different levels of economic decision making can generally be
categorized as w/Y/z-behavior responses or without-behavior 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 proposed regulation. An overview of the
basic economic theory used to develop the model to estimate the potential effect of the proposed
program 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
proposed regulation 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
10-20
-------
Economic Impact Analysis
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
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 NDEEVI 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).
10-21
-------
Draft Regulatory Impact Analysis
Figure 10.2-1
Market Equilibrium without and with Regulation
= P
DM
Q
Domestic Supply
Foreign Supply
a) Baseline Equilibrium
Market
S'.
P'
P
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 10.2-1 (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
import supply, by the estimated compliance costs. As a result of the upward shift in the supply
10-22
-------
Economic Impact Analysis
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 the proposed rule, 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 proposed 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.A
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
AThis one-to-one relationship holds for engines sold on the market and for engines consumed
internally by integrated engine/equipment manufacturers.
10-23
-------
Draft 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.
10-24
-------
Economic Impact Analysis
Figure 10.2-2
Derived Demand for Construction Equipment
Unit Price of
Construction
AQ,
Construction
Output
Price Equipment
AP
t
AQr
Derived
Demand
Equipment
Output
APrice
Equipment
Upward Shift
Construction
Supply Curve
AQ,
AQr
10-25
-------
Draft 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 (c) 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).
10-26
-------
Economic Impact Analysis
Figure 10.2-3
Derived Demand for Engines
SV
Unit Price of
Construction
AQ,
Q - Construction
Price
Equipment
APr
Derived
Demand
AQr
Q - Equipment
Price
Engines
tAPeng
eng
Derived
Demand
AQeng
AQE=AQeng
Q - Engines
10-27
-------
Draft 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 would be increased by inducing the monopolist to increase production.
Social cost estimates associated with a proposed regulation are larger with monopolistic
market structures because the regulation exacerbates an already 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. For this reason, 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.
10-28
-------
Economic Impact Analysis
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).
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 would be
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,
10-29
-------
Draft Regulatory Impact Analysis
indicating that the marginal and average costs of production are constant with respect to output.8
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.
BThe constancy of marginal costs reflects an underlying assumption of constant returns to scale
of production, which may or may not apply in all cases.
10-30
-------
Economic Impact Analysis
Figure 10.2-4
Full-Cost Absorption of Regulatory Costs
Price
Output
Figure 10.2-5
Full-Cost Pass-Through of Regulatory Costs
Sr With Regulation
Price / 1
Increase I
p
po
\N
> Unit Cost Increase
a
\S • Wthoi
D
Output
10-31
-------
Draft Regulatory Impact Analysis
Figure 10.2-6
Partial Cost Pass-Through of Regulatory Costs
$ g
r With Regulation
Unit Cost Increase
SQ: 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
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.
10-32
-------
Economic Impact Analysis
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.0 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
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.
cAs 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."
10-33
-------
Draft Regulatory Impact Analysis
Figure 10.2-7
Modeling Fixed Costs
$/q
MC'
prrf
pm J
,m
(a) Upward-sloping supply function
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. This is
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,
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 would 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
10-34
-------
Economic Impact Analysis
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 fixed costs associated with the proposed
regulation 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 which are "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
their behavior based on the R&D required for compliance with the standards. A sensitivity
analysis is included in Chapter 10 of the draft RIA for this proposal that reflects this approach.
Fixed costs on the refiner side are treated differently in the NDEEVI. 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.
DThe fixed R&D costs capture the lost opportunity of forgone investments to the firm.
10-35
-------
Draft Regulatory Impact Analysis
The first, shift by average total cost (variable + fixed), is the primary scenario and is included in
the NDEEVI. The other two are investigated using sensitivity analyses. These supply shifts are
discussed further in sensitivity analysis presented in AppendixlOI.
10.2.3.4 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).
10-36
-------
Economic Impact Analysis
Figure 10.2-8
Market Surplus Changes with Regulation: Consumer and Producer Surplus
$/Q
Q2 Q1
(a) Change in Consumer Surplus with
Regulation
Q/t
$/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
10-37
-------
Draft Regulatory Impact Analysis
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 Q,. 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.
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(c) shows
the net (negative) change in economic welfare associated with the regulation as area D.E
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.
EHowever, 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.
10-38
-------
Economic Impact Analysis
In addition, two additional compliance cost components are included in the total social cost
estimates but not integrated in to the market analysis:
• Operating Costs: 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 the proposed standards. These savings (costs) also include
any extra operating costs associated with the new PM emission control technology which
may accrue to new engines that use this new technology.
• Marker costs: Costs associated with marking high sulfur diesel fuel in the locomotive,
marine, and heating oil markets between 2007 and 2014.
Operating costs are not included directly in the model because some of the savings accrue to
existing engines and because these savings (costs) are not expected to affect consumer decisions
with respect to new engines. Instead, they are added into the estimated welfare impacts as
additional costs to the application markets, since it is the users of these engines that will see these
savings (costs). Marker costs are not include in the market analysis because locomotive, marine,
and heating oil markets are not explicitly modeled in the NDEEVI. Similar to the operating
savings (costs), marker costs are added into the estimated welfare impacts separately.
Nevertheless, a sensitivity analysis was also performed in which these savings (costs) are
included as inputs to the NDEEVI, where they are modeled as benefits accruing to the application
producers. The results of this analysis are presented in Appendix 10.1.
10.3 Economic Impact Modeling
The impact of a regulatory action can be measured by the change in social costs that it
generates. Producers will experience economic impacts due to changes in production costs
(direct regulatory costs and indirect input price changes) and changes in the market price they
receive for their products. Consumers will experience economic impacts due to the adjustments
in market prices and their consumption levels.
The previous section described the economic theory that underpins this EIA. This section
focuses on the markets and linkages included in the NDEEVI. This is followed by a description of
the supply and demand elasticities used in the model and an overview of the baseline population
data used in the analysis. Finally, the steps used to operationalize the computer model are
presented.
10.3.1 Operational Economic Model
The Nonroad Diesel Economic Impact Model simulates the economic impacts using a
computer model comprising a series of spreadsheet modules that:
10-39
-------
Draft Regulatory Impact Analysis
• 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 determine an estimated new, with-regulation equilibrium price
and quantity for all markets; and
• estimate the change in producer and consumer surplus in all markets included in the
model.
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
below, that adjusts price upward (downward) by a given percentage in response to excess
demand (excess supply).
The remainder of this section describes elements of the NDEEVI including baseline
characteristics, compliance cost inputs, model elasticity parameters, and the model solution
algorithm.
10.3.2 Baseline Economic Data
This section describes the data needed to run the model. The major components are the
baseline data needed to establish the without-regulation equilibrium and the engineering
compliance costs that are used to "shock" the model to estimate the with-regulation equilibrium.
10.3.2.1 Baseline Population
The PSR sales data were the primary source for the population for diesel engines used in
domestically consumed nonroad diesel equipment (See Chapter 1). Sales data is used as a proxy
for production data in the NDEEVI because detailed production data by horsepower and
equipment application are not available. In addition, modeling inventory decisions of engine and
equipment manufacturers is beyond the scope of the NDEEVI. EPA adjusted the Power Systems
Research (PSR) population to reflect the population units affected by the regulation.11 Table 10.3-
1 lists sales data for affected diesel nonroad equipment consumed domestically in 2000 by engine
horsepower and equipment type. The population distribution by size and application is the same
FSee Section 8.1 in Chapter 8 of this draft RIA for an explanation of how the engines were
allocated to the seven categories.
10-40
-------
Economic Impact Analysis
for engines and equipment because of the one-to-one relationship between engines and
equipment.
Baseline nonroad diesel fuel consumption is provided in Table 10.3-2. Fuel consumption is
broken out by region (PADD) and application market (construction, agriculture, and
manufacturing). Nonroad diesel fuel consumption is further disaggregated into spillover and
nonspillover (referred to hereafter as simply nonroad). As described below, spillover fuel is
highway grade diesel fuel consumed by nonroad equipment. Spillover fuel is affected by the
diesel highway rule and is not affected by this regulation. The economic impact associated with
lowering the sulfur content of spillover fuel consumed by nonroad diesel equipment is calibrated
into the baseline prior to estimating the economic impacts of the nonroad regulation.0
GSpillover and nonspillover fuels consumed by nonroad diesel equipment are modeled as two
commodities and markets. Thus, in calibrating the baseline, the increased costs associated
with the highway rule are used to shock the supply curve for spillover diesel fuel. This results
in an increased cost of production in the application markets leading to a slight decrease in
application market output. This in turn ripples through the supply chain leading to a very
small adjustment (decrease) in the baseline equipment and engine output. The impact of the
nonroad rule is then estimated relative to this adjusted baseline.
10-41
-------
Table 10.3-1
Engine/Equipment Sales in 2000
Engine Market
0 600 hp
Grand Total
Construction
17,043
30,233
30,919
30,146
49,503
42,126
4,945
204,915
Agricultural
Equipment
13,195
38,303
19,156
11,788
35,226
41,678
—
159,347
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
1
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
5,538
628,543
10-42
-------
Economic Impact Analysis
Table 10.3-2
Nonroad Diesel Equipment, Locomotive and Marine Fuel Consumption in 2001
3 ADD I&III
3ADDII
3ADDIV
3ADDV
Total
Nonroad
Spillover
Nonroad
Spillover
Nonroad
Spillover
Nonroad
Spillover
Nonroad
Spillover
Construction
(million gallons)
1,700
359
622
222
124
142
268
59
2,714
782
Agriculture
(million gallons)
449
95
992
355
92
105
59
13
1,592
568
Manufacturing
(million gallons)
2,778
1,180
1,338
928
164
400
373
151
4,653
2,659
Total
(million gallons)
4,927
1,634
2,952
1,505
380
647
700
223
8,959
4,008
10.3.2.2 Baseline Prices
Prototypical engine and equipment prices were collected for engines by hp size and for diesel
equipment by application and horsepower size. Average prices were developed by the Agency
based on a review of publicly available market transactions and information listed in the PSR
database. Table 10.3-3 provides the prices for the seven engine categories, and Table 10.3-4
provides prices for the 42 diesel equipment categories used in the model.
Table 10.3-3
Baseline Engine Prices
Power Range
0 600 hp
Estimated Price
$1,500
$2,500
$3,000
$4,000
$5,500
$20,000
$125,000
10-43
-------
Draft Regulatory Impact Analysis
Table 10.3-4
Baseline Prices of Nonroad Diesel Equipment
Application
Construction Equip
Agricultural Equip
3umps & Compressors
GenSets & Welders
Refrigeration & A/C
General Industrial
!.awn & Garden
<25hp
$3,500
$3,000
$1,500
$3,500
$1,500
$3,500
$3,000
26-50 hp
$13,500
$6,000
$3,000
$6,000
$3,000
$13,500
$6,000
51-75hp
$25,000
$23,500
$11,000
$25,000
$11,000
$25,000
$23,500
76-100 hp
$50,000
$47,000
$21,500
$50,000
N/A
$50,000
$47,000
101-175 hp
$100,000
$70,000
$32,000
$75,000
N/A
$100,000
$70,000
176-600 hp
$575,000
$130,000
$60,000
$140,000
N/A
$575,000
$130,000
>600 hp
$700,000
N/A
$88,000
N/A
N/A
$700,000
N/A
10.3.3 Market Linkages
Figure 10.3-1 illustrates the sectoral linkages and the market interactions between producers
and consumers that are explicitly accounted for in the NDEEVI. This section provides a brief
discussion of each of these related markets and important linkages. A detailed description of the
market model equations (supply and demand functions, equilibrium conditions) is provided in
Appendix 10F.
One of the key features of the NDEEVI is that a subset of related markets is modeled together,
with sector linkages; hence, selected interaction effects, are explicitly specified and accounted for
in the model. A brief discussion of the markets and important linkages are highlighted in this
section. Detailed specifications of the market model equations (supply and demand functions,
equilibrium conditions) are provided in Appendix 10F.
10-44
-------
Economic Impact Analysis
Figure 10.3-1
Multimarket Linkages in the Engine/Equipment/Fuel Supply Chain
Domestic Application Consumers
Application
Imports
c
H Applicatio
"
1 Application IV
>
^
Equipmer
• Agriculture • Refrige
• Construction • Genera
• Lawn a
n Markets 1 » APP|ication
| " Exports
anufacturer- 1 W Nonroad Diesel
| M Fue| Markets
A
ration • Pumps and \
tor Sets Compressors 1 Refmene
nd narden * Industrial /
s
t
k >
Nonintegrated Equipment
Manufacturers
>
k
r^x
Diesel Engine Market i
^ • Size j
t
k
Merchant Engine
Manufacturers
/
\
k
Integrated
Equipment
Manufacturers
Captive
Engine
Manufacturers
10.3.3.1 Engine Markets
The engine markets are the markets associated with the production and consumption of
engines. 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 proposal):
• less than 25 hp,
• 26 to 50 hp,
• 51 to 75 hp,
• 76tolOOhp,
• 101 to 175 hp,
10-45
-------
Draft Regulatory Impact Analysis
• 176 to 600 hp, and
• greater than 601 hp.
An important feature of the engine and equipment markets is that many equipment
manufacturers also produce engines. These equipment manufacturers are referred to as
integrated manufacturers, and their facilities produce engines to consume internally (in the
nonroad equipment they produce) and to supply to the engine markets (to other equipment
manufacturers). An important modeling distinction is that all compliance costs for internally
consumed engines are absorbed into the equipment costs of integrated suppliers. In contrast,
nonintegrated equipment suppliers pay some portion of the engine compliance costs that is
determined by the incremental market price for engines. As long as engine demand is not
perfectly inelastic, the increased market price for engines will reflect only a partial pass through
of engine compliance costs. For the purposes of this analysis, engines sold on the market are
referred to as "merchant" engines, and engines consumed internally are referred to as "captive"
engines.
Because the impact of the regulation is not directly proportional to engine price, the relative
supply shift in each of the engine size markets varies. For example, the ratio of control costs to
market price for small engines (less than 25 hp) is approximately 12 percent, and the ratio of
control costs to market price for large engines (greater than 600 hp) is approximately 8 percent.
These different ratios lead to different relative shifts in the supply curves and larger percentage
changes in market price and quantity in the small engine markets. The impacts on the engine
market and engine manufacturers can be found in Appendix 10A.
10.3.3.2 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:
• construction,
• agricultural,
• pumps and compressors,
• generators and welder sets,
• refrigeration and air conditioning,
• general industrial, and
• lawn and garden.
These categories were identified by reviewing the "application" field in the PSR database.
Approximately 60 different equipment "applications" are listed in the database. These were
aggregated into these seven equipment categories to obtain a manageable number of individual
markets to be included in the NDEEVI.H For each of these equipment types, up to seven
HSee Section 8.1 in Chapter 8 of this draft RIA for an explanation of how the engines were
allocated to the seven categories.
10-46
-------
Economic Impact Analysis
horsepower size category markets are included in the model, for a total of 42 individual
equipment markets.1
Equipment manufacturers consume engines in their production processes and then supply
diesel equipment to the application markets. The demand for engines is determined by the
production levels in the application markets. Equipment is assumed to be a fixed factor of
production in the application markets. Thus, for example, 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 on 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. The impacts on the equipment market and manufacturers
can be found in Appendix 10B.
10.3.3.3 Application Markets
The application markets consist of the producers and consumers of products and services that
employ the diesel engines, equipment, and fuel affected by this proposal. Therefore, these
economic entities are indirectly affected by the proposal, through potential changes in equipment
and fuel prices. For the purpose of this analysis, application markets are grouped into three
categories:
• construction
• agricultural, and
• manufacturing.
These three application markets were selected because they encompass the majority of the
final products and services that incorporate diesel engines in their production process. In
addition, these three application markets represent a manageable number of markets to be
included in the NDEEVI and have well-established census data. The impacts on the equipment
market and manufacturers can be found in Appendix IOC.
The seven equipment categories are mapped into the three application markets as described in
Table 10.3-5.
'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.
10-47
-------
Draft Regulatory Impact Analysis
Table 10.3-5
Mapping from Equipment Markets to Application Markets
Application Market
Construction
Agricultural
Manufacturing
Equipment Market
Construction equipment
Pumps and compressors
Gen sets and welding equipment
Agricultural equipment
Refrigeration
Lawn and garden
General industrial
For example, mining equipment is included in the general industrial equipment categories.
This is linked to the manufacturing applications market/
10.3.3.4 Diesel Fuel Markets
The analysis estimates the economic impact of increasing the cost of production for
nonroad diesel, locomotive, and marine fuels. Nonroad diesel fuel cost increases are linked to
application markets (users of diesel engines and equipment) to estimate how the compliance
costs on refineries are linked to the application markets. For example, although locomotive and
commercial marine engines and equipment are not directly affected by the proposed rule, the
users of this equipment in the application markets are affected by the higher diesel fuel costs, and
these impacts are included in the model.
As shown in Figure 10.2-8, equipment users are the suppliers in the application markets and
are also the demanders of nonroad diesel fuel. Thus, the fuel markets are linked with the engine
and equipment markets through the application markets using the derived-demand framework
described above.
One can think of these relationships as the conceptual equivalent of the derived-demand
relationship between equipment and engines. For example, the demand for No. 2 distillate will
be specified as a function of the production and consumption decisions made in the construction,
agricultural, and manufacturer application markets. In this way 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
JA full mapping from PSR applications to the NDEEVI equipment categories and then to the
NDEEVI application markets can be found in a Memorandum from M. Gallaher, RTI, to Todd
Sherwood, Clarifications on Several Modeling Issues (March 24, 2003).
10-48
-------
Economic Impact Analysis
decrease in production in the application markets and hence a decrease in the demand for inputs
(fuel and equipment).
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, and
• 15 ppm nonroad diesel fuel.
The four regional nonroad diesel fuel markets are
• PADD 1 and 3,
• PADD 2,
• PADD 4, and
• PADD 5 (includes Alaska and Hawaii)
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. 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. The impacts on the nonroad fuel market can be found
in Appendix 10D.
As discussed in Section 10.2, all the engine and equipment markets are modeled as
competitive: it is assumed that no individual firm can affect the market price. In this case the
average compliance cost is used to shift the market supply curve. In this scenario, the fuel
markets are also modeled as competitive, and each regional supply curve is shifted by the average
total (variable + fixed) regional cost of the regulation. This fuel market scenario (referred to as
average total cost) is also 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. 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
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.
Locomotive and Marine Diesel. Locomotive and marine fuels are modeled as being
consumed by the manufacturers. Thus, these fuels are included in the total volume of diesel fuel
consumed by the manufacturing application market and their per unit (gallon) costs are included
in the refinery supply function shifts. Inclusion of locomotive and marine diesel fuel in the
market analysis has two main impacts:
10-49
-------
Draft Regulatory Impact Analysis
• It affects the magnitude of the supply shift because their per unit costs are slightly
different from the nonroad diesel per-unit fuel costs
• It increases the quantity of affected diesel fuel purchased by the manufacturing market
and thus increases the total compliance costs passed into the manufacturing sector from
refiners. This leads to a greater shift of the supply curve in the manufacturing application
market and thus a larger decrease in quantity in the manufacturing market that ripples
back through the fuel, equipment, and engine markets.
10.3.3.5 Calibrating the Spillover Baseline (Impacts Relative to Highway Rule)
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 the model, the increased cost of "spillover" fuel consumed by nonroad equipment is built
into the baseline. In effect, current market projections are "shocked" by the highway rule and a
new set of baseline prices and quantities is estimated for all linked markets. This then becomes
the new baseline 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.
10.3.4 Compliance Costs
Social costs capture the full range of economic impacts associated with the proposed
regulation. For this economic analysis, the sources of compliance costs are grouped in to the
following categories:
• Fixed and variable costs for diesel engines
• Fixed and variable costs for diesel equipment
• Fixed and variable costs for nonroad diesel fuel
• Changes in operating costs of diesel equipment
• Marker costs for locomotive and marine diesel fuel and heating oil.
10-50
-------
Economic Impact Analysis
All of the above compliance impact are included in the social cost estimates. The majority
are included in the market analysis using the NDEEVI. However, as discussed above, not all of
the compliance costs are incorporated in to the market analysis. Table 10.3-6 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 10.3-6
How Compliance Costs are Accounted for in the Economic Analysis
Compliance Costs used to
Shock the Market Model
Compliance Costs added after
Market Analysis
Variable costs for diesel engines
Variable costs for diesel equipment
Fixed and variable costs for nonroad
diesel fuel
Fixed costs for diesel engines
Fixed costs for diesel equipment
Changes in operating costs of diesel equipment
Marker costs for locomotive and marine diesel fuel
and heating oil
The compliance costs described in Chapters 6 and 7 were used to determine the regulation's
impacts on each industry sector. The compliance cost per unit varied over time and by industry
sector (engine, equipment, or fuel producer). All costs are presented in 2001 dollars and most are
broken out by variable and fixed costs.
10.3.4.1 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 per unit compliance costs are weighted average costs within the
appropriate horsepower range (refer to Chapter 6 for how we have estimated engine and
equipment costs; refer to Chapter 8 for aggregate costs and projected sales; per unit costs within
each horsepower range are the engine and equipment aggregate costs for that horsepower range
divided by the projected sales for that horsepower range). As shown in Table 10.3-7, 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. In 2013, there is a projected fourfold cost increase for engines in
the range of 25 hp to less than 75 hp, which then decreases over time. Because these engines
represent over 35 percent of the overall engine population, this cost increase contributes to the
year 2013 having the largest average cost per unit impact.
For nonroad equipment, the majority of the projected compliance cost increases are due to the
need to redesign the equipment. 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.
10-51
-------
Draft Regulatory Impact Analysis
The equipment sector compliance costs are broken out by horsepower category and impact
year in Table 10.3-8. 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. Table 10.3-8 shows a significant compliance cost
increase for equipment in the range of 25 hp to less than 75 hp in the year 2013.
10.3.4.2 Nonroad Diesel Fuel Compliance Costs
In the fuel market, the desulfurization (compliance) costs per gallon of diesel fuel differ
according to PADD and according to impact year as shown in Table 10.3-9a,b,c. Sulfur fuel
requirements are phased in a two-step process. From 2007 to 2010, both the nonroad sector and
the locomotive and marine sectors are required to meet the sulfur standard of SOOppm. The costs
for this combined SOOppm market are shown in Table 10.3-9a. Variable and fixed costs per
gallon are presented for the average cost refiner and the maximum cost refiner in each PADD.
Beginning in 2010, the costs diverge between these two groups. 2010 is the target year set
for nonroad diesel fuel to meet a 15 ppm capacity sulfur standard, while the sulfur standard for
marine and locomotive diesel fuel will remain at 500 ppm. Therefore, nonroad diesel fuel is
estimated to experience a higher increase in cost than locomotive and marine diesel fuel, after
2010, as shown in Tables 10.3-9b and 10.3-9c, respectively.
10-52
-------
Table 10.3-7
Compliance Costs per Engine"
HP Cateeorv
0600hp
Cost Tvoes
Variable
Fixed
Total
Variable
Fixed
Total
Variable
Fixed
Total
Variable
Fixed
Total
Variable
Fixed
Total
Variable
Fixed
Total
Variable
Fixed
Total
2008
$131
$30
$161
$149
$47
$196
$171
$48
$218
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
2009
$131
$29
$161
$149
$46
$195
$171
$47
$217
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
2010
$124
$29
$153
$141
$45
$186
$161
$46
$207
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
2011
$124
$28
$152
$141
$44
$185
$161
$45
$206
$0
$0
$0
$0
$0
$0
$2,266
$199
$2,466
$5,402
$904
$6,306
2012
$124
$27
$152
$141
$43
$184
$161
$44
$205
$1,150
$53
$1,2046
$1,410
$52
$1,461
$1,265
$188
$2,453
$5,402
$825
$6,228
2013
$124
$0
$124
$852
$61
$913
$845
$62
$9073
$1,150
$52
$1,203
$1,410
$51
$1,461
$1,755
$185
$1,939
$4,216
$813
$5,030
2014
$124
$0
$124
$852
$60
$912
$845
$61
$906
$1,139
$70
$1,209
$1,384
$68
$1,452
$2,209
$240
$2,450
$6,952
$1,222
$8,175
2015
$124
$0
$124
$457
$59
$516
$642
$60
$702
$1,139
$69
$1,208
$1,384
$67
$1,450
$2,209
$236
$2,445
$6,952
$1,205
$8,157
2016
$124
$0
$124
$647
$58
$705
$642
$59
$701
$1,139
$55
$1,194
$1,384
$53
$1,436
$2,208
$66
$2,2742
$6,953
$479
$7,432
2017
$124
$0
$124
$647
$57
$704
$642
$58
$700
$1,139
$18
$1,157
$1,384
$17
$1,401
$2,207
$56
$2,262
$6,953
$403
$7,356
2018
$124
$0
$124
$647
$0
$647
$642
$0
$642
$1,139
$17
$1,157
$1,384
$17
$1,400
$2,206
$55
$2,261
$6,953
$398
$7,351
(continued)
10-53
-------
Table 10.3-7 (continued)
Compliance Costs per Engine"
HP Cateeorv
0600hp
Cost Tvrjes
Variable
Fixed
Total
Variable
Fixed
Total
Variable
Fixed
Total
Variable
Fixed
Total
Variable
Fixed
Total
Variable
Fixed
Total
Variable
Fixed
Total
2019
$124
$0
$124
$647
$0
$647
$642
$0
$642
$1,139
$0
$1,139
$1,384
$0
$1,384
$2,205
$0
$2,205
$6,953
$0
$6.953
2020
$124
$0
$124
$647
$0
$647
$642
$0
$642
$1,139
$0
$1,139
$1,384
$0
$1,384
$2,204
$0
$2,2042
$6,953
$0
$6.953
2021
$124
$0
$124
$647
$0
$647
$642
$0
$642
$1,139
$0
$1,139
$1,384
$0
$1,384
$2,203
$0
$2,203
$6,953
$0
$6.953
2022
$124
$0
$124
$647
$0
$647
$642
$0
$642
$1,139
$0
$1,139
$1,384
$0
$1,384
$2,202
$0
$2,202
$6,953
$0
$6.953
2023
$124
$0
$124
$647
$0
$647
$642
$0
$642
$1,139
$0
$1,139
$1,384
$0
$1,384
$2,202
$0
$2,202
$6,953
$0
$6.953
2024
$124
$0
$124
$647
$0
$647
$642
$0
$642
$1,139
$0
$1,139
$1,384
$0
$1,384
$2,201
$0
$2,201
$6,953
$0
$6.953
2025
$124
$0
$124
$647
$0
$647
$642
$0
$642
$1,139
$0
$1,139
$1,384
$0
$1,384
$2,200
$0
$2,200
$6,953
$0
$6.953
2026
$124
$0
$124
$647
$0
$647
$642
$0
$642
$1,139
$0
$1,139
$1,384
$0
$1,384
$2,200
$0
$2,200
$6,953
$0
$6.953
2027
$124
$0
$124
$647
$0
$647
$642
$0
$642
$1,139
$0
$1,139
$1,384
$0
$1,384
$2,199
$0
$2,199
$6,953
$0
$6.953
2028
$124
$0
$124
$647
$0
$647
$642
$0
$642
$1,139
$0
$1,139
$1,384
$0
$1,384
$2,198
$0
$2,198
$6,953
$0
$6.953
2029
$124
$0
$124
$647
$0
$647
$642
$0
$642
$1,139
$0
$1,139
$1,384
$0
$1,384
$2,198
$0
$2,198
$6,953
$0
$6.953
2030
$124
$0
$124
$647
$0
$647
$642
$0
$642
$1,139
$0
$1,139
$1,384
$0
$1,384
$2,197
$0
$2,197
$6,954
$0
$6.954
'2001 dollars
10-54
-------
Table 10.3-8
Costs per Piece of Equipment
HP Cateeorv
0600hp
Cost Tvoes
Variable
Fixed
Total
Variable
Fixed
Total
Variable
Fixed
Total
Variable
Fixed
Total
Variable
Fixed
Total
Variable
Fixed
Total
Variable
Fixed
Total
2008
$0
$10
$10
$0
$12
$12
$0
$12
$12
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
2009
$0
$10
$10
$0
$12
$12
$0
$12
$12
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
2010
$0
$10
$10
$0
$12
$12
$0
$12
$12
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
2011
$0
$9
$9
$0
$11
$11
$0
$12
$12
$0
$0
$0
$0
$0
$0
$92
$322
$414
$125
$743
$868
2012
$0
$9
$9
$0
$11
$11
$0
$12
$12
$55
$90
$145
$55
$140
$196
$91
$317
$409
$125
$732
$857
2013
$0
$9
$9
$18
$41
$58
$18
$45
$62
$55
$89
$143
$55
$138
$193
$91
$312
$404
$125
$721
$846
2014
$0
$9
$9
$18
$40
$58
$18
$44
$62
$55
$109
$164
$55
$170
$225
$91
$384
$476
$181
$1,071
$1.252
2015
$0
$9
$9
$18
$39
$57
$18
$43
$61
$55
$107
$162
$55
$167
$223
$91
$379
$470
$181
$1,056
$1.237
2016
$0
$8
$8
$18
$38
$56
$18
$43
$60
$55
$105
$160
$55
$164
$220
$91
$373
$464
$181
$1,041
$1.222
2017
$0
$8
$8
$18
$38
$56
$18
$42
$60
$55
$104
$159
$55
$162
$217
$91
$368
$459
$181
$1,026
$1.207
2018
$0
$0
$0
$18
$27
$45
$18
$31
$48
$55
$102
$157
$55
$159
$215
$91
$362
$453
$181
$1,012
$1.193
(continued)
10-55
-------
Table 10.3-8 (continued)
Costs per Piece of Equipment
HP Cateeorv
0600hp
Cost Tvoes
Variable
Fixed
Total
Variable
Fixed
Total
Variable
Fixed
Total
Variable
Fixed
Total
Variable
Fixed
Total
Variable
Fixed
Total
Variable
Fixed
Total
2019
$0
$0
$0
$18
$27
$44
$18
$30
$48
$55
$101
$155
$55
$157
$213
$91
$357
$448
$181
$998
$1.180
2020
$0
$0
$0
$18
$26
$44
$18
$30
$48
$55
$99
$154
$55
$155
$210
$91
$352
$443
$181
$985
$1.166
2021
$0
$0
$0
$18
$26
$44
$18
$29
$47
$55
$98
$152
$55
$153
$208
$91
$69
$160
$181
$327
$508
2022
$0
$0
$0
$18
$25
$43
$18
$29
$47
$55
$19
$74
$55
$30
$86
$91
$69
$159
$181
$323
$504
2023
$0
$0
$0
$18
$0
$18
$18
$0
$18
$55
$19
$74
$55
$30
$85
$90
$68
$158
$181
$319
$500
2024
$0
$0
$0
$18
$0
$18
$18
$0
$18
$55
$0
$55
$55
$0
$55
$90
$0
$90
$181
$0
$181
2025
$0
$0
$0
$18
$0
$18
$18
$0
$18
$55
$0
$55
$55
$0
$55
$90
$0
$90
$181
$0
$181
2026
$0
$0
$0
$18
$0
$18
$18
$0
$18
$55
$0
$55
$55
$0
$55
$90
$0
$90
$181
$0
$181
2027
$0
$0
$0
$18
$0
$18
$18
$0
$18
$55
$0
$55
$55
$0
$55
$90
$0
$90
$181
$0
$181
2028
$0
$0
$0
$18
$0
$18
$18
$0
$18
$55
$0
$55
$55
$0
$55
$90
$0
$90
$181
$0
$181
2029
$0
$0
$0
$18
$0
$18
$18
$0
$18
$55
$0
$55
$55
$0
$55
$90
$0
$90
$181
$0
$181
2030
$0
$0
$0
$18
$0
$18
$18
$0
$18
$55
$0
$55
$55
$0
$55
$90
$0
$90
$181
$0
$181
10-56
-------
Economic Impact Analysis
Table 10.3-9a
Desulfurization Costs for Nonroad, Locomotive, and Marine Diesel Fuel by PADD Prior 2010
PADD I and III
PADD II
PADD IV
PADD V
Average Cost
Variable Costs
($/gallon) Fixed
0.0089
0.0143
0.0144
0.0089
Costs ($/gallon)
0.0063
0.0158
0.0268
0.0165
Maximum Cost
Variable Costs
($/gallon)
0.0129
0.0228
0.0174
0.0097
Fixed Costs ($/gallon
0.0207
0.0254
0.0403
0.0296
Table 10.3-9b
Desulfurization Costs for Nonroad Diesel Fuel by PADD Starting in 2010
PADD I and III
PADD II
PADD IV
PADD V
Average Cost
Variable Costs
($/gallon) Fixed
0.0184
0.0247
0.0280
0.0194
Costs ($/gallon)
0.0117
0.0364
0.0611
0.0391
Maximum Cost
Variable Costs
($/gallon)
0.0251
0.0285
0.0301
0.0191
Fixed Costs ($/gallon
0.0287
0.0459
0.0624
0.0649
Table 10.3-9c
Desulfurization Costs for Marine and Locomotive Diesel Fuel by PADD Starting in 2010
PADD I and III
PADD II
PADD IV
PADDV
Average Cost
Variable Costs
($/gallon) Fixed
0.0088
0.0169
0.0111
0.0080
Costs ($/gallon)
0.0071
0.0188
0.0225
0.0137
Maximum Cost
Variable Costs
($/gallon)
0.0089
0.0228
0.0114
0.0064
Fixed Costs ($/gallon
0.0247
0.0254
0.0254
0.0154
10.3.4.3 Changes in Operating Costs
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. Equipment operating saving are
generated as a result of the decreased sulfur content of diesel fuel. These savings will accrue to
all equipment users that use 500 ppm or 15 ppm sulfur fuel, regardless of whether the equipment
10-57
-------
Draft Regulatory Impact Analysis
has a compliant engine or not. In addition, there may be some operating costs associated with the
new PM emission reduction technology. These costs will accrue to engines that use these new
technologies. Both of these impacts are discussed in more detail in Chapter 4 and 5. These costs
are not included in the market analysis and are instead listed as a separate category in the social
cost results tables. In Appendix 101, a sensitivity analysis is presented where operating cost
savings are introduced into the market analysis as a downward shift in the application supply
functions.
The net impact is projected to be operating savings of between 1 to 17 cents per gallon
consumed by nonroad diesel equipment. Operating savings vary depending on the horsepower
size of the equipment (smaller engines have greater savings) and whether the equipment has
emission controls (existing noncontrolled fleet will have greater savings). Table 10.3-10 lists the
new operating savings by horsepower category and by existing versus new (emission controls)
fleet. Average cost savings per gallon for nonroad applications will vary by year as the existing
fleet of diesel equipment is replaced over time. EPA estimates that approximately 90 percent of
the existing fleet will be replaced by 2030.
Table 10.3-10
Net Change in Operating Cost"
Engine Size/Type
0600
Locomotive
Marine
Net Operating Cost
Per Gallon — Existing Fleet
-$0.160
-$0.076
-$0.066
-$0.030
-$0.017
-$0.011
-$0.011
-$0.011
Net Operating Cost
Per Gallon — New Fleet
-$0.175
-$0.041
-$0.036
-$0.014
-$0.010
-$0.006
N/A
N/A
^Changes in operating costs are shown as negative values to indicate savings (benefits).
10.3.4.4 Fuel Marker Costs
Fuel marker costs will be needed to identify high-sulfur diesel fuel in the locomotive, marine,
and heating oil markets as the proposed regulation is phased in between 2007 and 2014. These
are also added as a separate category in the social cost result tables. Marker costs are estimated
to be 0.2 cents per gallon. The affected fuel volume is presented in Table 10.3-11.
10-58
-------
Economic Impact Analysis
Table 10.3-11
Fuel Volume Affected by Marker Costs of 0.2 Cents per Gallon
Year
2007
2008
2009
2010
2011
2012
2013
2014
Locomotive and Marine (MMgals/yr)
2082
3621
3647
3670
1539
Heating Oil (MMgals/yr)
4371
7563
7633
3210
10.3.5 Supply and Demand Elasticity Estimates
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 10.3-12 and 10.3-13 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.K The general methodologies for estimating the supply and
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
KA 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
accepted economic theory. For this reason, literature estimates were used for the supply
elasticities in the construction and manufacturing application markets.
10-59
-------
Draft Regulatory Impact Analysis
changes. In the long run, supply and demand are expected to be more elastic since more
substitutes may become available.
Table 10.3-12
Summary of Market Demand Elasticities Used in the NDEIM
Market
Estimate
Source
Method
Input Data Summary
Applications
Construction
Agriculture
-0.96 EPA econometric
estimate
-0.20 EPA econometric
estimate
Simultaneous equation
(log-log) approach
Productivity shift
approach (Morgenstern,
Pizer, and Shih, 2002)
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)
Manufacturing
-0.58 EPA econometric
estimate
Simultaneous equation
(log-log) approach.
Annual time series from
1958- 1995 developed by
Jorgenson et al. (Jorgenson,
1990; Jorgenson, Gollop, and
Fraumeni, 1987)
Equipment
Construction
Agriculture
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
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
the production of products and services in the
application markets
into
10-60
-------
Economic Impact Analysis
Table 10.3-13
Summary of Market Supply Elasticities Used in the NDEIM
Markets
Applications
Construction
Agriculture
Manufacturing
Estimate Source
1.0 Literature-based
estimate
0.32 Literature-based
estimate
1.0 Literature-based
estimate
Method
Based on Topel and Rosen,
(1988).a
Production-weighted average
of individual crop estimates
ranging from 0.27 to 0.55.
(Lin etal., 2000)
Literature estimates are not
available so assumed same
value as for Construction
market
Input Data Summary
Census data, 1963 -
1983
Agricultural Census
data 1991 - 1995
Not applicable
Equipment
Construction
Agriculture
Pumps/
compressors
Generators/
Welder Sets
Refrigeration
Industrial
Lawn and
Garden
Engines
Diesel fuel
3.31
2.14
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 (2002).b
Census data 1958-
1996; SIC 3531
Census data 1958-
1996; SIC 3523
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 Group
(EIG); 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.
10-61
-------
Draft Regulatory Impact Analysis
10.3.6 Model Solution Algorithm
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.
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:
( V
Pi+i=Pi' — (10.1)
UJ
4. Recalculate market supply with new prices, accounting for fuel-switching choices
associated with new energy 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.
Section 10.1 presents a summary of the results of this modeling. More detailed information
is presented in the appendices to this chapter.
10-62
-------
Economic Impact Analysis
References for Chapter 10
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. n-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.
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.
M. Gallaher, 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.
10-63
-------
Draft Regulatory Impact Analysis
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/107319.pdf. A copy of this document is also
available in Docket A-2001-28, Document No. II-A-21.
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
Jack E. 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.
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/ulsd_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.
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. U-A-22.
10-64
-------
Economic Impact Analysis
Pizer, Bill. Communications between Mike Gallaher and Bill Pizer on November 5, 2002.
Docket A-2001-28, Document No. U-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.
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.
RTI. 2003. Economic Impact Analysis for Nonroad Diesel Tier 4 Rule. 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.
Topel, Robert and Sherwin Rosen. 1988. "Housing Investment in the United States." Journal of
Political Economy 96(4):718-40.
U.S. Census Bureau, 2002. "Annual Value of Construction Put in Place," C30 Table 101. As
accessed on November 12, 2002. . Docket
A-2001-28, Document No. U-B-17.
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.
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. U-A-01.
10-65
-------
Draft Regulatory Impact Analysis
APPENDIX 10A: Impacts on the Engine Market and Engine
Manufacturers
This appendix provides the time series of impacts from 2007 through 2030 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 proposal):
• less than 25 hp,
• 26 to 50 hp,
• 51 to 75 hp,
• 76tolOOhp,
• 101 to 175 hp,
• 176 to 600 hp, and
• 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 I 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 ($), and
• 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 $2001, 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 depreciated. The price increase after that time is driven by the per-engine variable
costs and remains relatively constant over time. We did the cost analysis using a 3% discount
10-66
-------
Economic Impact Analysis
rate. We will also be conducting a similar analysis using a 7% discount rate and including this
information in the docket.
For all the engine size categories, the majority of the cost of the regulation is passed along
through increased engine prices. Price increases range from $125 (8.3% increase) for small
(<25hp) engines to $6,950 (5.6% increase) for large (>600hp) engines. Even though 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.
10-67
-------
Draft Regulatory 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
NPVb
Engineering
Cost/Unit
—
$161.48
$160.69
$153.07
$152.35
$151.67
$124.47
$124.47
$124.47
$124.47
$124.47
$124.47
$124.47
$124.47
$124.47
$124.47
$124.47
$124.47
$124.47
$124.47
$124.47
$124.47
$124.47
$124.47
Absolute
Change in
Price
-$0.01
$131.33
$131.33
$124.45
$124.45
$124.44
$124.43
$124.43
$124.43
$124.43
$124.43
$124.43
$124.43
$124.43
$124.43
$124.43
$124.43
$124.43
$124.43
$124.43
$124.43
$124.43
$124.43
$124.43
Change in
Price (%)
0.00%
8.76%
8.76%
8.30%
8.30%
8.30%
8.30%
8.30%
8.30%
8.30%
8.30%
8.30%
8.30%
8.30%
8.30%
8.30%
8.30%
8.30%
8.30%
8.30%
8.30%
8.30%
8.30%
8.30%
Change in
Quantity
(%)
-0.002%
-0.003%
-0.003%
-0.004%
-0.006%
-0.009%
-0.010%
-0.010%
-0.010%
-0.010%
-0.010%
-0.010%
-0.010%
-0.010%
-0.010%
-0.010%
-0.010%
-0.010%
-0.010%
-0.010%
-0.010%
-0.010%
-0.010%
-0.010%
Total Change in Producer
Engineering Surplus for Engine
Costs (103) Manufacturers (103)
$-
$19,922.9
$20,361.4
$19,906.8
$20,322.4
$20,738.0
$17,434.5
$17,850.1
$18,265.7
$18,681.3
$19,096.9
$19,512.5
$19,928.1
$20,343.7
$20,759.3
$21,174.9
$21,590.5
$22,006.1
$22,421.7
$22,837.3
$23,252.8
$23,668.4
$24,084.0
$24,499.6
$308,900.8
-$1.0
-$3,720.4
-$3,720.4
-$3,721.1
-$3,722.4
-$3,723.8
-$5.5
-$5.9
-$5.6
-$5.9
-$6.0
-$6.1
-$6.3
-$6.4
-$6.5
-$6.6
-$6.8
-$6.9
-$7.0
-$7.1
-$7.3
-$7.4
-$7.5
-$7.6
-$15,668.6
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.
10-68
-------
Economic Impact Analysis
Table 10A-2. Impacts on the Engine Market and Engine Manufacturers
(Average Price per Engine = $2,500)a
: 26-50hp
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
NPVb
Engineering
Cost/Unit
—
$196.04
$195.05
$186.18
$185.27
$184.39
$913.48
$912.31
$516.20
$704.91
$703.86
$647.19
$647.19
$647.19
$647.19
$647.19
$647.19
$647.19
$647.19
$647.19
$647.19
$647.19
$647.19
$647.19
Absolute
Change in
Price
-$0.02
$149.22
$149.22
$141.30
$141.28
$141.26
$852.28
$852.27
$457.31
$647.10
$647.10
$647.10
$647.10
$647.10
$647.10
$647.10
$647.10
$647.10
$647.10
$647.10
$647.10
$647.10
$647.10
$647.10
Change in
Price (%)
0.00%
5.97%
5.97%
5.65%
5.65%
5.65%
34.09%
34.09%
18.29%
25.88%
25.88%
25.88%
25.88%
25.88%
25.88%
25.88%
25.88%
25.88%
25.88%
25.88%
25.88%
25.88%
25.88%
25.88%
Change in
Quantity
(%)
-0.002%
-0.003%
-0.003%
-0.005%
-0.008%
-0.012%
-0.013%
-0.014%
-0.013%
-0.013%
-0.013%
-0.013%
-0.013%
-0.013%
-0.013%
-0.013%
-0.013%
-0.013%
-0.013%
-0.013%
-0.013%
-0.013%
-0.013%
-0.013%
Total Change in Producer
Engineering Surplus for Engine
Costs (103) Manufacturers (103)
$-
$26,120.5
$26,553.3
$25,887.0
$26,296.9
$26,706.8
$134,957.6
$137,429.7
$79,257.9
$110,276.3
$112,153.3
$105,000.1
$106,877.1
$108,754.1
$110,631.1
$112,508.1
$114,385.1
$116,262.1
$118,139.1
$120,016.1
$121,893.1
$123,770.1
$125,647.0
$127,524.0
$1,363,271.2
-$2.0
-$6,238.9
-$6,238.9
-$6,240.5
-$6,244.0
-$6,247.3
-$9,042.7
-$9,043.7
-$9,043.0
-$9,043.6
-$9,043.8
-$13.8
-$14.1
-$14.3
-$14.5
-$14.8
-$15.0
-$15.3
-$15.5
-$15.7
-$16.0
-$16.2
-$16.5
-$16.7
-$58,965.6
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.
10-69
-------
Draft Regulatory 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
NPVb
Engineering
Cost/Unit
—
$218.10
$217.24
$206.87
$206.07
$205.29
$907.16
$906.11
$702.27
$701.29
$700.35
$641.85
$641.85
$641.85
$641.85
$641.85
$641.85
$641.85
$641.85
$641.85
$641.84
$641.84
$641.84
$641.84
Absolute
Change in
Price
-$0.02
$170.53
$170.53
$160.99
$160.96
$160.93
$844.59
$844.58
$641.75
$641.75
$641.75
$641.75
$641.75
$641.75
$641.74
$641.74
$641.74
$641.74
$641.74
$641.74
$641.74
$641.74
$641.74
$641.73
Change in
Price (%)
0.00%
5.68%
5.68%
5.37%
5.37%
5.36%
28.15%
28.15%
21.39%
21.39%
21.39%
21.39%
21.39%
21.39%
21.39%
21.39%
21.39%
21.39%
21.39%
21.39%
21.39%
21.39%
21.39%
21.39%
Change in
Quantity
(%)
-0.002%
-0.003%
-0.003%
-0.005%
-0.009%
-0.012%
-0.013%
-0.014%
-0.013%
-0.014%
-0.014%
-0.014%
-0.014%
-0.014%
-0.014%
-0.014%
-0.014%
-0.014%
-0.014%
-0.014%
-0.014%
-0.014%
-0.014%
-0.014%
Total Change in Producer
Engineering Surplus for Engine
Costs (103) Manufacturers (103)
—
$18,456.1
$18,723.8
$18,155.5
$18,408.3
$18,661.1
$83,885.0
$85,210.9
$67,144.2
$68,151.8
$69,159.3
$64,390.8
$65,398.3
$66,405.8
$67,413.4
$68,420.9
$69,428.4
$70,435.9
$71,443.5
$72,451.0
$73,458.5
$74,466.0
$75,473.6
$76,481.1
$855,626.8
-$1.5
-$4,025.4
-$4,025.4
-$4,026.7
-$4,029.4
-$4,032.0
-$5,785.8
-$5,786.6
-$5,786.0
-$5,786.4
-$5,786.6
-$10.8
-$10.9
-$11.1
-$11.2
-$11.4
-$11.6
-$11.7
-$11.9
-$12.1
-$12.2
-$12.4
-$12.5
-$12.7
-$37,885.0
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.
10-70
-------
Economic 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
NPVb
Engineering
Cost/Unit
—
—
—
—
—
$1,203.59
$1,202.69
$1,209.39
$1,208.24
$1,194.15
$1,157.02
$1,156.74
$1,139.37
$1,139.37
$1,139.37
$1,139.37
$1,139.38
$1,139.38
$1,139.38
$1,139.38
$1,139.38
$1,139.39
$1,139.39
$1,139.39
Absolute
Change in
Price
-$0.02
-$0.03
-$0.03
-$0.05
-$0.09
$1,150.18
$1,150.17
$1,139.20
$1,139.22
$1,139.22
$1,139.22
$1,139.22
$1,139.22
$1,139.23
$1,139.23
$1,139.23
$1,139.23
$1,139.23
$1,139.24
$1,139.24
$1,139.24
$1,139.24
$1,139.24
$1,139.25
Change in
Price (%)
0.00%
0.00%
0.00%
0.00%
0.00%
28.75%
28.75%
28.48%
28.48%
28.48%
28.48%
28.48%
28.48%
28.48%
28.48%
28.48%
28.48%
28.48%
28.48%
28.48%
28.48%
28.48%
28.48%
28.48%
Change in
Quantity
(%)
-0.002%
-0.003%
-0.003%
-0.005%
-0.009%
-0.012%
-0.013%
-0.014%
-0.013%
-0.014%
-0.014%
-0.014%
-0.014%
-0.014%
-0.014%
-0.014%
-0.014%
-0.014%
-0.014%
-0.014%
-0.014%
-0.014%
-0.014%
-0.014%
Total Change in Producer
Engineering Surplus for Engine
Costs (103) Manufacturers (103)
—
—
—
—
—
$68,915.0
$70,054.5
$71,643.2
$72,771.9
$73,105.8
$71,978.6
$73,107.3
$73,137.5
$74,266.2
$75,394.9
$76,523.6
$77,652.3
$78,781.0
$79,909.7
$81,038.4
$82,167.0
$83,295.7
$84,424.4
$85,553.1
$882,138.1
-$1.3
-$1.7
-$1.7
-$2.8
-$5.1
-$3,058.0
-$3,058.9
-$4,158.0
-$4,157.6
-$3,363.2
-$1,107.4
-$1,107.6
-$9.3
-$9.4
-$9.5
-$9.7
-$9.8
-$9.9
-$10.1
-$10.2
-$10.4
-$10.5
-$10.6
-$10.8
-$14,777.3
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.
10-71
-------
Draft Regulatory Impact Analysis
Table 10A-5. Impacts on the Engine Market and Engine Manufacturers:
(Average Price per Engine = $15,500)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
NPVb
Engineering
Cost/Unit
—
—
—
—
—
$1,461.38
$1,460.54
$1,451.52
$1,450.45
$1,436.25
$1,400.64
$1,400.38
$1,383.71
$1,383.71
$1,383.71
$1,383.71
$1,383.71
$1,383.71
$1,383.71
$1,383.71
$1,383.71
$1,383.71
$1,383.71
$1,383.71
Absolute
Change in
Price
-$0.04
-$0.04
-$0.05
-$0.07
-$0.14
$1,409.45
$1,409.44
$1,383.48
$1,383.50
$1,383.49
$1,383.49
$1,383.49
$1,383.49
$1,383.49
$1,383.49
$1,383.49
$1,383.49
$1,383.49
$1,383.49
$1,383.49
$1,383.49
$1,383.49
$1,383.49
$1,383.49
Change in
Price (%)
0.00%
0.00%
0.00%
0.00%
0.00%
25.63%
25.63%
25.15%
25.15%
25.15%
25.15%
25.15%
25.15%
25.15%
25.15%
25.15%
25.15%
25.15%
25.15%
25.15%
25.15%
25.15%
25.15%
25.15%
Change in
Quantity
(%)
-0.002%
-0.003%
-0.003%
-0.005%
-0.010%
-0.014%
-0.015%
-0.016%
-0.015%
-0.015%
-0.015%
-0.015%
-0.015%
-0.015%
-0.015%
-0.015%
-0.015%
-0.015%
-0.015%
-0.015%
-0.015%
-0.015%
-0.015%
-0.015%
Total Change in Producer
Engineering Surplus for Engine
Costs (103) Manufacturers (103)
—
—
—
—
—
$91,426.0
$92,886.1
$93,816.2
$95,249.4
$95,804.6
$94,879.5
$96,312.7
$96,599.4
$98,032.6
$99,465.8
$100,899.1
$102,332.3
$103,765.5
$105,198.8
$106,632.0
$108,065.2
$109,498.4
$110,931.7
$112,364.9
$1,161,715.6
-$2.0
-$2.6
-$2.7
-$4.5
-$8.6
-$3,248.7
-$3,250.0
-$4,397.7
-$4,397.0
-$3,519.5
-$1,161.4
-$1,161.6
-$15.3
-$15.5
-$15.7
-$15.9
-$16.2
-$16.4
-$16.6
-$16.8
-$17.0
-$17.3
-$17.5
-$17.7
-$15,656.9
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.
10-72
-------
Economic 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
NPVb
Engineering
Cost/Unit
—
—
—
—
$2,465.89
$2,452.90
$1,939.41
$2,449.55
$2,445.00
$2,273.55
$2,262.31
$2,260.63
$2,204.93
$2,204.09
$2,203.28
$2,202.49
$2,201.72
$2,200.97
$2,200.24
$2,199.53
$2,198.84
$2,198.16
$2,197.50
$2,196.86
Absolute
Change in
Price
$0.13
$0.17
$0.17
$0.28
$2,265.86
$2,264.55
$1,753.89
$2,208.60
$2,207.69
$2,206.73
$2,205.82
$2,204.94
$2,204.08
$2,203.25
$2,202.43
$2,201.64
$2,200.87
$2,200.13
$2,199.40
$2,198.69
$2,197.99
$2,197.32
$2,196.66
$2,196.02
Change in
Price (%)
0.00%
0.00%
0.00%
0.00%
11.33%
11.32%
8.77%
11.04%
11.04%
11.03%
11.03%
11.02%
11.02%
11.02%
11.01%
11.01%
11.00%
11.00%
11.00%
10.99%
10.99%
10.99%
10.98%
10.98%
Change in
Quantity
(%)
-0.003%
-0.003%
-0.003%
-0.005%
-0.010%
-0.015%
-0.016%
-0.017%
-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)
—
—
—
—
$99,063.4
$100,109.1
$80,391.8
$103,103.5
$104,474.2
$98,601.2
$99,559.5
$100,930.2
$99,852.3
$101,223.0
$102,593.7
$103,964.4
$105,335.1
$106,705.8
$108,076.5
$109,447.2
$110,817.9
$112,188.6
$113,559.4
$114,930.1
$1,280,605.9
-$5.0
-$6.4
-$6.6
-$11.2
-$8,035.8
-$7,687.3
-$7,690.4
-$10,142.1
-$10,140.4
-$2,897.9
-$2,486.1
-$2,486.6
-$38.4
-$38.9
-$39.5
-$40.0
-$40.5
-$41.0
-$41.6
-$42.1
-$42.6
-$43.2
-$43.7
-$44.2
-$39,033.9
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.
10-73
-------
Draft Regulatory Impact Analysis
Table 10A-7. Impacts on the Engine Market and Engine Manufacturers: >601hp
(Average Price per Engine = $125,000)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
NPVb
Engineering
Cost/Unit
—
—
—
—
$6,305.92
$6,227.66
$5,029.76
$8,174.82
$8,157.40
$7,432.01
$7,356.03
$7,350.56
$6,952.78
$6,952.86
$6,952.93
$6,953.00
$6,953.07
$6,953.13
$6,953.20
$6,953.26
$6,953.33
$6,953.39
$6,953.45
$6,953.50
Absolute
Change in
Price
-$0.80
-$1.00
-$1.00
-$1.65
$5,399.16
$5,397.83
$4,211.73
$6,947.29
$6,947.67
$6,947.66
$6,947.74
$6,947.82
$6,947.90
$6,947.98
$6,948.05
$6,948.13
$6,948.20
$6,948.27
$6,948.33
$6,948.40
$6,948.46
$6,948.53
$6,948.59
$6,948.65
Change in
Price (%)
0.00%
0.00%
0.00%
0.00%
4.32%
4.32%
3.37%
5.56%
5.56%
5.56%
5.56%
5.56%
5.56%
5.56%
5.56%
5.56%
5.56%
5.56%
5.56%
5.56%
5.56%
5.56%
5.56%
5.56%
Change in
Quantity
(%)
-0.002%
-0.003%
-0.003%
-0.005%
-0.009%
-0.013%
-0.014%
-0.016%
-0.015%
-0.015%
-0.015%
-0.015%
-0.015%
-0.015%
-0.015%
-0.015%
-0.015%
-0.015%
-0.015%
-0.015%
-0.015%
-0.015%
-0.015%
-0.015%
Total Change in Producer
Engineering Surplus for Engine
Costs (103) Manufacturers (103)
—
—
—
—
$10,293.4
$10,320.1
$8,459.7
$13,952.2
$14,124.8
$13,053.0
$13,102.0
$13,274.6
$12,728.6
$12,901.2
$13,073.7
$13,246.3
$13,418.9
$13,591.4
$13,764.0
$13,936.5
$14,109.1
$14,281.7
$14,454.2
$14,626.8
$160,049.3
-$1.2
-$1.6
-$1.6
-$2.6
-$1,480.2
-$1,375.2
-$1,375.9
-$2,095.1
-$2,094.7
-$850.7
-$727.2
-$727.3
-$8.9
-$9.0
-$9.2
-$9.3
-$9.4
-$9.5
-$9.6
-$9.7
-$9.9
-$10.0
-$10.1
-$10.2
-$8,057.4
a Figures are in 2001 dollars.
b Net present values are calculated using a social discount rate of 3 percent over the 2004 to 2030 time period.
10-74
-------
Economic Impact Analysis
APPENDIX 10B: Impacts on Equipment Market and Equipment
Manufacturers
This appendix provides the time series of impacts from 2007 through 2030 for the equipment
markets. The equipment markets are the markets associated with the production and
consumption of equipment that use nonroad diesel engines. Seven equipment types were
modeled:
• construction,
• agricultural,
• pumps and compressors,
• generators and welder sets,
• refrigeration and air conditioning,
• general industrial, and
• lawn and garden.
Forty-two equipment markets were modeled, representing 7 horsepower categories within 7
application categories. There are 7 horsepower/application categories that did not have sales in
2000 and are not included in the model, so the total number of diesel equipment markets is 42
rather than 49.
Tables 10B-1 through 10B-7 provide the time series of impacts for the seven equipment
markets included in the analysis. Each table includes the following:
• average equipment price,
• average engineering costs (variable and fixed) per piece of equipment,
- Note that in the engineering cost analysis, fixed costs for equipment manufacturers are
recovered in the first ten years (see Chapter 6)
• absolute change in the equipment market price ($),
- Note that the estimated absolute change in market price is based on variable costs only;
see Appendix I for a sensitivity analysis including fixed costs as well
• relative change in the equipment market price (%),
• relative change in the equipment market quantity (%),
• total engineering (regulatory) costs associated with each equipment market ($), and
• change in producer surplus for all equipment 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
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.
10-75
-------
Draft Regulatory Impact Analysis
All prices and costs are presented in $2001, and real equipment prices are assumed to be
constant. The engineering cost per piece of equipment peak around 2013 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. Price increases range from an average increase of
1.84 percent in the general industrial equipment market to 9.37 percent in the refrigeration and
air-conditioning market. 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.
10-76
-------
Economic Impact Analysis
Table 1 OB- 1.
Impacts on Agricultural Equipment Market and Manufacturers
(Average Price per Equipment = $55,396)a
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
NPVb
Engineering
Cost/Unit
-$0.01
$80.60
$80.52
$76.73
$796.90
$1,242.25
$1,386.50
$1,533.64
$1,405.61
$1,409.97
$1,396.93
$1,380.70
$1,362.36
$1,358.70
$1,286.32
$1,252.67
$1,241.18
$1,215.51
$1,214.27
$1,213.06
$1,211.88
$1,210.74
$1,209.63
$1,208.55
Absolute
Change in
Price
-$0.71
$67.81
$67.96
$63.67
$658.43
$1,060.42
$1,193.96
$1,300.28
$1,176.51
$1,222.04
$1,220.45
$1,218.91
$1,217.42
$1,215.97
$1,214.58
$1,213.22
$1,211.91
$1,210.63
$1,209.39
$1,208.19
$1,207.02
$1,205.88
$1,204.78
$1,203.71
Change in
Price (%)
0.00%
1.09%
1.09%
1.04%
1.49%
2.13%
5.41%
5.49%
3.76%
4.55%
4.55%
4.56%
4.56%
4.57%
4.57%
4.58%
4.58%
4.58%
4.59%
4.59%
4.59%
4.60%
4.60%
4.60%
3.76%
Change in
Quantity
(%)
-0.003%
-0.003%
-0.003%
-0.006%
-0.013%
-0.017%
-0.019%
-0.020%
-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%
-0.019%
£
Total
Engineering
Costs (103) IV
$-
$6,411.0
$6,499.3
$6,353.8
$132,042.2
$196,410.2
$199,850.0
$231,071.2
$224,459.1
$222,759.7
$223,659.0
$223,742.8
$223,224.0
$225,988.3
$212,477.8
$207,735.7
$208,148.7
$204,967.8
$207,732.1
$210,496.4
$213,260.7
$216,025.0
$218,789.4
$221,553.7
$2,632,706.9
"hange in Producer
Surplus for
Equipment
lanufacturers (103)
-$129.1
-$2,412.3
-$2,415.8
-$2,560.5
-$27,655.9
-$36,985.4
-$39,869.9
-$49,179.7
-$49,121.5
-$40,985.7
-$39,136.9
-$36,472.6
-$33,205.5
-$33,221.7
-$16,963.0
-$9,472.8
-$7,137.7
-$1,208.5
-$1,224.7
-$1,240.8
-$1,256.9
-$1,273.1
-$1,289.2
-$1,305.3
-$306,693.1
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.
10-77
-------
Draft Regulatory Impact Analysis
Table 10.B-2.
Impacts on Construction Equipment Market and Manufacturers
(Average Price per Equipment = $166,086)a
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
NPVb
Engineering
Cost/Unit
-$0.02
$66.22
$66.23
$62.97
$785.33
$1,359.19
$1,452.08
$1,647.81
$1,551.03
$1,537.78
$1,523.40
$1,512.99
$1,491.27
$1,487.37
$1,414.60
$1,372.58
$1,362.50
$1,331.08
$1,329.77
$1,328.51
$1,327.27
$1,326.07
$1,324.91
$1,323.77
Absolute
Change in
Price
-$1.57
$58.01
$58.14
$53.39
$648.33
$1,166.79
$1,252.19
$1,396.67
$1,304.51
$1,331.79
$1,330.14
$1,328.54
$1,326.99
$1,325.50
$1,324.04
$1,322.63
$1,321.27
$1,319.94
$1,318.65
$1,317.40
$1,316.18
$1,315.00
$1,313.85
$1,312.73
Change in
Price (%)
0.00%
0.61%
0.61%
0.58%
0.68%
1.38%
2.60%
2.62%
2.05%
2.27%
2.27%
2.28%
2.28%
2.28%
2.28%
2.29%
2.29%
2.29%
2.29%
2.29%
2.29%
2.30%
2.30%
2.30%
1.89%
Change in
Quantity
-0.003%
-0.004%
-0.004%
-0.007%
-0.014%
-0.020%
-0.021%
-0.023%
-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%
1
Total
Engineering
Costs (103) I
—
$2,983.7
$3,011.4
$2,954.9
$139,136.4
$230,575.9
$221,136.2
$266,895.3
$267,893.0
$260,863.2
$261,235.8
$262,618.9
$260,590.1
$263,663.6
$245,876.8
$236,856.8
$237,399.6
$231,100.5
$234,174.0
$237,247.5
$240,321.0
$243,394.5
$246,468.0
$249,541.6
$3,006,380.6
"hange in Producer
Surplus for
Equipment
Manufacturers (103)
-$366.2
-$1,985.4
-$1,994.8
-$2,409.1
-$35,069.1
-$50,141.3
-$53,022.7
-$67,782.3
-$67,679.2
-$57,507.7
-$54,852.6
-$53,208.1
-$48,151.8
-$48,197.6
-$27,383.1
-$15,335.5
-$12,850.7
-$3,524.0
-$3,569.9
-$3,615.7
-$3,661.6
-$3,707.4
-$3,753.2
-$3,799.0
-$433,600.5
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.
10-78
-------
Economic Impact Analysis
Table
10B-3. Impacts on Pumps and Compressor Equipment Market and Manufacturers
(Average Price per Equipment = $13,198)a
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
NPVb
Engineering
Cost/Unit
-$0.03
$107.28
$107.24
$101.82
$296.63
$637.09
$900.62
$940.17
$813.80
$851.85
$849.14
$840.21
$837.80
$835.47
$813.71
$790.85
$778.71
$767.62
$766.57
$765.55
$764.57
$763.61
$762.68
$761.78
Absolute
Change in
Price
-$0.12
$99.50
$99.63
$94.34
$265.84
$581.71
$833.50
$862.60
$737.70
$777.12
$775.75
$774.42
$773.14
$771.91
$770.71
$769.56
$768.44
$767.35
$766.30
$765.29
$764.30
$763.35
$762.42
$761.52
Change in
Price (%)
0.00%
3.69%
3.70%
3.51%
3.81%
5.18%
11.57%
11.61%
8.43%
9.80%
9.80%
9.81%
9.81%
9.82%
9.82%
9.82%
9.83%
9.83%
9.83%
9.84%
9.84%
9.84%
9.85%
9.85%
8.29%
Change in
Quantity
(%)
-0.002%
-0.002%
-0.002%
-0.003%
-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%
£
Total
Engineering
Costs (103) IV
—
$183.9
$183.9
$183.9
$949.8
$1,968.1
$2,506.4
$2,836.5
$2,848.8
$2,861.0
$2,873.2
$2,701.6
$2,713.8
$2,726.0
$2,140.5
$1,498.3
$1,172.3
$867.7
$880.0
$892.2
$904.4
$916.6
$928.9
$941.1
$24,133.3
"hange in Producer
Surplus for
Equipment
lanufacturers (103)
-$2.1
-$186.7
-$186.7
-$187.3
-$785.9
-$1,442.0
-$1,782.1
-$2,099.1
-$2,098.3
-$2,098.8
-$2,098.9
-$1,915.1
-$1,915.2
-$1,915.4
-$1,317.7
-$663.3
-$325.3
-$8.6
-$8.7
-$8.8
-$8.9
-$9.1
-$9.2
-$9.3
-$14,701.0
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.
10-79
-------
Draft Regulatory Impact Analysis
Table 10.B-4. Impacts on Generator Sets and Welding Equipment Market
(Average Price per Equipment = $14,483)a
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
NPVb
Engineering
Cost/Unit
-$0.01
$143.56
$143.18
$136.20
$175.95
$330.33
$607.35
$614.82
$482.77
$530.87
$527.13
$510.92
$508.22
$506.76
$501.54
$490.14
$479.37
$475.15
$474.37
$473.62
$472.89
$472.18
$471.50
$470.83
Absolute
Change in
Price
-$0.11
$124.60
$124.66
$118.07
$152.11
$292.14
$559.14
$562.28
$431.32
$482.22
$481.18
$480.17
$479.21
$478.28
$477.38
$476.51
$475.68
$474.87
$474.09
$473.34
$472.61
$471.90
$471.22
$470.55
Change in
Price (%)
0.00%
2.61%
2.62%
2.48%
2.51%
2.74%
6.37%
6.37%
4.48%
5.34%
5.34%
5.34%
5.34%
5.34%
5.34%
5.33%
5.33%
5.33%
5.33%
5.33%
5.33%
5.33%
5.33%
5.33%
4.59%
Welders
Change in
Quantity
(%)
-0.002%
-0.002%
-0.002%
-0.003%
-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%
Total
Engineering
Costs (103)
—
$7,583.3
$7,697.0
$7,536.0
$11,229.4
$25,160.7
$40,598.4
$42,375.4
$37,298.3
$39,279.4
$39,499.6
$37,501.4
$37,816.3
$38,339.1
$38,153.4
$36,762.9
$35,418.0
$35,285.3
$35,808.1
$36,330.8
$36,853.6
$37,376.3
$37,899.1
$38,421.8
$461,276.2
and Manufacturers
. Change in Producer
Surplus for
Equipment
Manufacturers (103)
-$13.5
-$2,696.7
-$2,696.9
-$2,700.7
-$3,630.6
-$5,944.6
-$7,666.3
-$8,531.3
-$8,526.4
-$8,225.8
-$7,924.0
-$5,403.8
-$5,196.6
-$5,197.4
-$4,489.7
-$2,577.2
-$710.3
-$55.7
-$56.4
-$57.2
-$58.0
-$58.8
-$59.5
-$60.3
-$59,177.6
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.
10-80
-------
Economic Impact Analysis
Table 10B-5. Impacts on Refrigeration and Air-Conditioning Equipment Market and
Manufacturers (Average Price per Equipment = $6,3 14)a
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
NPVb
Engineering
Cost/Unit
-$0.02
$166.67
$166.37
$157.66
$157.36
$157.08
$712.79
$711.32
$504.88
$556.92
$555.81
$545.06
$544.20
$543.37
$542.56
$541.78
$521.01
$520.60
$520.21
$519.82
$519.45
$519.09
$518.74
$518.40
Absolute
Change in
Price
-$0.06
$154.86
$154.80
$146.32
$146.24
$146.14
$678.12
$677.31
$471.52
$524.16
$523.64
$523.14
$522.66
$522.19
$521.74
$521.31
$520.88
$520.48
$520.08
$519.70
$519.32
$518.96
$518.61
$518.27
Conditioning
Change in
Change in Quantity
Price (%) (%)
0.00%
4.28%
4.29%
4.07%
4.08%
4.09%
13.88%
13.88%
9.34%
11.11%
11.12%
11.12%
11.12%
11.12%
11.13%
11.13%
11.13%
11.13%
11.14%
11.14%
11.14%
11.14%
11.14%
11.15%
9.37%
-0.002%
-0.002%
-0.002%
-0.003%
-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%
Total
Engineering
Costs (103)
—
$528.8
$528.8
$528.8
$528.8
$528.8
$2,382.7
$2,394.6
$2,406.5
$2,418.4
$2,430.4
$1,913.4
$1,925.4
$1,937.3
$1,949.2
$1,961.1
$782.0
$794.0
$805.9
$817.8
$829.7
$841.6
$853.5
$865.4
$20,342.3
. Change in Producer
Surplus for
Equipment
Manufacturers (103)
-$1.9
-$531.3
-$531.3
-$531.9
-$532.6
-$534.1
-$1,726.8
-$1,727.0
-$1,726.3
-$1,726.7
-$1,726.8
-$1,198.1
-$1,198.2
-$1,198.3
-$1,198.4
-$1,198.5
-$7.7
-$7.8
-$7.9
-$8.0
-$8.1
-$8.2
-$8.3
-$8.4
-$12,244.8
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.
10-81
-------
Draft Regulatory Impact Analysis
Table 10.B-6. Impacts on General Industrial Equipment Market and Manufacturers
(Average Price per Equipment = $132,972)a
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
NPVb
Engineering
Cost/Unit
-$0.04
$51.50
$51.51
$48.95
$485.64
$1,289.38
$1,387.46
$1,496.25
$1,418.49
$1,434.64
$1,426.31
$1,418.48
$1,412.32
$1,409.22
$1,362.03
$1,305.18
$1,297.20
$1,270.65
$1,269.80
$1,268.97
$1,268.17
$1,267.39
$1,266.63
$1,265.89
Absolute
Change in
Price
-$0.51
$46.14
$46.25
$43.66
$425.67
$1,159.24
$1,251.01
$1,329.31
$1,254.49
$1,277.00
$1,275.92
$1,274.87
$1,273.86
$1,272.88
$1,271.92
$1,270.99
$1,270.10
$1,269.22
$1,268.37
$1,267.55
$1,266.75
$1,265.97
$1,265.21
$1,264.47
Change in
Price (%)
0.00%
0.44%
0.44%
0.42%
0.49%
1.52%
2.51%
2.52%
2.06%
2.24%
2.24%
2.24%
2.24%
2.24%
2.24%
2.25%
2.25%
2.25%
2.25%
2.25%
2.25%
2.25%
2.25%
2.26%
1.84%
Change in
Quantity
-0.002%
-0.002%
-0.002%
-0.003%
-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%
£
Total
Engineering
Costs (103) IV
—
$606.1
$611.4
$600.4
$7,449.8
$27,094.6
$29,136.6
$32,268.0
$32,185.8
$32,218.8
$32,152.6
$32,104.6
$32,174.9
$32,485.8
$29,031.1
$24,618.3
$24,349.5
$22,405.9
$22,716.8
$23,027.6
$23,338.4
$23,649.3
$23,960.1
$24,271.0
$333,923.8
"hange in Producer
Surplus for
Equipment
lanufacturers (103)
-$31.7
-$366.3
-$366.9
-$375.5
-$4,335.3
-$9,575.1
-$10,216.5
-$12,715.2
-$12,703.7
-$12,414.3
-$12,038.9
-$11,681.6
-$11,442.5
-$11,444.0
-$7,680.1
-$2,958.0
-$2,380.0
-$127.2
-$128.8
-$130.4
-$132.0
-$133.6
-$135.2
-$136.8
-$85,851.8
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.
10-82
-------
Economic Impact Analysis
Table 10.B-7. Impacts on Lawn and Garden Equipment Market and Manufacturers
(Average Price per Equipment = $12,394)a
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
NPVb
Engineering
Cost/Unit
-$0.01
$138.80
$138.46
$131.75
$168.43
$333.59
$574.91
$581.20
$454.05
$508.37
$506.56
$494.69
$493.20
$491.76
$486.74
$475.01
$465.57
$461.32
$460.55
$459.81
$459.09
$458.39
$457.71
$457.06
Absolute
Change in
Price
-$0.08
$120.56
$120.64
$114.32
$146.92
$299.74
$536.55
$539.53
$413.26
$468.39
$467.36
$466.36
$465.40
$464.48
$463.59
$462.74
$461.91
$461.11
$460.34
$459.60
$458.88
$458.18
$457.51
$456.85
Change in
Price (%)
0.00%
3.07%
3.08%
2.92%
2.95%
3.23%
6.96%
6.96%
4.97%
5.91%
5.90%
5.90%
5.90%
5.90%
5.90%
5.90%
5.89%
5.89%
5.89%
5.89%
5.89%
5.89%
5.89%
5.88%
5.11%
Change in
Quantity
(%)
-0.002%
-0.002%
-0.002%
-0.003%
-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%
£
Total
Engineering
Costs (103) IV
—
$2,233.4
$2,271.9
$2,228.9
$2,520.4
$3,435.9
$5,502.4
$5,774.8
$4,984.6
$5,470.2
$5,538.9
$5,085.3
$5,154.1
$5,222.8
$5,092.9
$4,579.5
$4,172.9
$4,046.5
$4,115.2
$4,183.9
$4,252.6
$4,321.4
$4,390.1
$4,458.8
$63,452.9
"hange in Producer
Surplus for
Equipment
lanufacturers (103)
-$2.9
-$760.0
-$760.1
-$760.9
-$960.6
-$1,545.1
-$1,789.2
-$1,984.8
-$1,983.7
-$1,984.3
-$1,984.5
-$1,462.3
-$1,462.5
-$1,462.7
-$1,264.2
-$682.3
-$207.1
-$12.1
-$12.3
-$12.4
-$12.6
-$12.8
-$12.9
-$13.1
-$15,141.2
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.
10-83
-------
Draft Regulatory Impact Analysis
APPENDIX IOC: Impacts on Application Market Producers and
Consumers
This appendix provides the time series of impacts from 2007 through 2030 for the product and
service application markets included in the model. There are 3 application markets: construction,
agriculture, and manufacturing.
Tables 10C-1 through 10C-3 provide the time series of impacts for the three application
markets. Each table includes the following:
• relative change in market price (%),
• relative change in market quantity (%), and
• change in producer and consumer surplus for each application market.
Price increases range from an average of 0.01 percent in the manufacturing sector to 0.05
percent in the agricultural sector. Even though 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 consumption by producers and consumers within each application market increases
over time.
10-84
-------
Economic Impact Analysis
Table 10C-1. Impacts on Agricultural Application Market and
Agricultural Producers and Consumers"
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
Change in Price (%)
0.009%
0.012%
0.012%
0.021%
0.044%
0.059%
0.064%
0.068%
0.063%
0.065%
0.065%
0.065%
0.065%
0.065%
0.065%
0.065%
0.065%
0.065%
0.065%
0.065%
0.065%
0.065%
0.065%
0.065%
AVG
0.055%
Change in Quantity (%)
-0.002%
-0.002%
-0.002%
-0.004%
-0.009%
-0.012%
-0.013%
-0.014%
-0.013%
-0.013%
-0.013%
-0.013%
-0.013%
-0.013%
-0.013%
-0.013%
-0.013%
-0.013%
-0.013%
-0.013%
-0.013%
-0.013%
-0.013%
-0.013%
AVG
-0.011%
Change in Producer and
Consumer Surplus ($103)
-$43,960.9
-$57,864.1
-$59,188.4
-$108,934.7
-$229,431.7
-$315,186.7
-$348,696.5
-$377,452.5
-$357,954.7
-$374,252.1
-$380,487.4
-$386,722.8
-$392,958.1
-$399,193.5
-$405,428.8
-$411,664.1
-$417,899.4
-$424,134.7
-$430,370.0
-$436,592.0
-$442,814.0
-$449,036.0
-$455,257.9
-$461,479.9
NPVb
-$5,050,376.0
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.
10-85
-------
Draft Regulatory Impact Analysis
Table 10C-2. Impacts on Construction Application Market and
Construction Producers and Consumers'1
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
Change in Price (%)
0.003%
0.004%
0.004%
0.007%
0.015%
0.021%
0.022%
0.024%
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%
AVG
0.020%
Change in Quantity (%)
-0.003%
-0.004%
-0.004%
-0.007%
-0.014%
-0.020%
-0.021%
-0.023%
-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%
AVG
-0.019%
Change in Producer and
Consumer Surplus ($103)
-$58,654.6
-$74,356.5
-$76,058.2
-$143,107.9
-$297,490.6
-$437,573.1
-$468,250.1
-$515,649.1
-$499,896.0
-$516,309.7
-$524,842.8
-$533,375.9
-$541,909.0
-$550,442.1
-$558,975.1
-$567,508.0
-$576,041.0
-$584,573.9
-$593,106.8
-$601,622.0
-$610,137.1
-$618,652.2
-$627,167.4
-$635,682.5
NPVb
-$6,923,515.6
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.
10-86
-------
Economic Impact Analysis
Table 10C-3. Impacts on Manufacturing Application Market and
Manufacturing Producers and Consumers"
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
Change in Price (%)
0.003%
0.004%
0.004%
0.004%
0.005%
0.008%
0.010%
0.010%
0.009%
0.009%
0.009%
0.009%
0.009%
0.009%
0.009%
0.009%
0.009%
0.009%
0.009%
0.009%
0.009%
0.009%
0.009%
0.009%
AVG
0.008%
Change in Quantity (%)
-0.002%
-0.002%
-0.002%
-0.003%
-0.003%
-0.004%
-0.006%
-0.006%
-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%
AVG
-0.005%
Change in Producer and
Consumer Surplus ($103)
-$111,732.2
-$148,130.8
-$150,298.2
-$181,349.6
-$222,767.7
-$316,540.3
-$414,805.7
-$428,962.5
-$388,563.0
-$412,018.5
-$418,249.7
-$424,459.7
-$430,700.7
-$436,821.7
-$443,264.9
-$449,719.9
-$456,187.1
-$462,666.7
-$469,158.8
-$475,656.2
-$482,166.5
-$488,690.2
-$495,227.3
-$501,778.2
NPVb
-$5,770,293.9
a Figures are in 2001 dollars.
b Net present values are calculated using a social discount rate of 3 percent over the 2004 to 2030 time period.
10-87
-------
Draft Regulatory Impact Analysis
APPENDIX 10D: Impacts on the Nonroad Fuel Market
This appendix provides the time series of impacts from 2007 through 2030 for the nonroad
diesel fuel market. Eight nonroad diesel fuel markets were modeled: 2 sulfur content levels (15
ppm and 500 ppm) for each of 4 PADDs (PADDs 1&3, PADD 2, PADD 4, and PADD 5). It
should be noted that PADD 5 includes Alaska and Hawaii.
Tables 10D-1 through 10D-4 provide the time series of impacts for the diesel fuel market for
the four regional fuel markets. Each table includes the following:
• average price per gallon,
• average engineering costs (variable and fixed) per gallon,
• absolute change in the PADDs' nonroad diesel price ($),
- Note that the estimated absolute change in market price is based on average variable and
fixed costs; see Appendix I 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 ($), and
• change in producer surplus for all fuel producers.
About 60 to 65 percent of high-sulfur diesel fuel is consumed by nonroad diesel equipment, the
other 35 to 40 percent is consumed by marine equipment and locomotive engines. 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 $2001, and the real per-gallon prices are assumed to be
constant within each regional fuel market. Initially, nonroad diesel equipment, locomotive, and
marine engines are included in the 500 ppm market. As the proposed rule phases in 2010,
nonroad equipment switches to the 15 ppm market. The engineering compliance costs are
greater to refine 15 ppm (4.6 cents/gal) compared to 500 ppm (2.6 cents/gal), thus the price
change in the 15 ppm market is greater than in the 500 ppm market.
For each regional fuel market, the majority of the cost of the regulation is passed along through
increased fuel prices. Price increases for the 15 ppm market are about an average of 6.24 percent
per gallon in each regional fuel market. Even though the cost per engine and market impacts (in
terms of percentage change in price and quantity) stabilize within the first few years of the
regulation, the engineering costs and producer surplus changes continue to gradually increase
because the projected consumption of diesel fuel increases over time.
10-88
-------
Table 10D-1. Impacts on the Nonroad Fuel Market in PADD 1&3 (Average Price per Gallon = $0.9199)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
NPVb
Engineering
Cost/Unit
$0.0301
$0.0301
$0.0301
$0.0301
$0.0301
$0.0301
$0.0301
$0.0301
$0.0301
$0.0301
$0.0301
$0.0301
$0.0301
$0.0301
$0.0301
$0.0301
$0.0301
$0.0301
$0.0301
$0.0301
$0.0301
15ppm PADD
Absolute Change
Change in in Price
Price (%)
—
—
—
$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
0.00%
0.00%
0.00%
3.25%
3.23%
3.21%
3.20%
3.20%
3.20%
3.20%
3.20%
3.20%
3.20%
3.20%
3.20%
3.20%
3.20%
3.20%
3.20%
3.20%
3.20%
3.20%
3.20%
3.20%
2.802%
1&3
Change in
Quantity
(%)
0.000%
0.000%
0.000%
-0.006%
-0.011%
-0.016%
-0.017%
-0.018%
-0.017%
-0.018%
-0.018%
-0.018%
-0.018%
-0.018%
-0.018%
-0.018%
-0.018%
-0.018%
-0.018%
-0.018%
-0.018%
-0.018%
-0.018%
-0.018%
Total
Engineering
Costs (S106)
—
—
—
$97.9
$100.1
$102.2
$104.4
$106.6
$108.8
$111.0
$113.1
$115.3
$117.4
$119.6
$121.8
$123.9
$126.1
$128.3
$130.4
$132.6
$134.7
$136.9
$139.0
$141.2
$1,559.2
SOOppm PADD
Engineering
Cost/Unit
$0.0209
$0.0209
$0.0209
$0.0159
$0.0159
$0.0159
$0.0159
$0.0159
$0.0159
$0.0159
$0.0159
$0.0159
$0.0159
$0.0159
$0.0159
$0.0159
$0.0159
$0.0159
$0.0159
$0.0159
$0.0159
$0.0159
$0.0159
$0.0159
Absolute
Change
in Price
$0.02
$0.02
$0.02
$0.02
$0.02
$0.02
$0.02
$0.02
$0.02
$0.02
$0.02
$0.02
$0.02
$0.02
$0.02
$0.02
$0.02
$0.02
$0.02
$0.02
$0.02
$0.02
$0.02
$0.02
Change
in Price
(%)
1.64%
1.64%
1.64%
1.72%
1.71%
1.71%
1.70%
1.70%
1.70%
1.70%
1.70%
1.70%
1.70%
1.70%
1.70%
1.70%
1.70%
1.70%
1.70%
1.70%
1.70%
1.70%
1.70%
1.70%
1.697%
1&3
Change in
Quantity
(%)
-0.002%
-0.003%
-0.003%
-0.003%
-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%
Total
Engineering
Costs (S106)
$69.8
$71.1
$72.4
$25.4
$25.8
$26.0
$26.1
$26.3
$26.5
$26.8
$27.0
$27.2
$27.4
$27.6
$27.9
$28.2
$28.4
$28.7
$29.0
$29.3
$29.6
$29.9
$30.2
$30.5
$555.7
Change in
Producer
Surplus for Fuel
Producers (S106)
-$0.4
-$0.5
-$0.6
-$0.9
-$1.6
-$2.4
-$2.7
-$2.9
-$2.8
-$2.9
-$2.9
-$3.0
-$3.0
-$3.1
-$3.1
-$3.2
-$3.2
-$3.3
-$3.3
-$3.4
-$3.4
-$3.5
-$3.5
-$3.6
-$39.0
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.
10-89
-------
Table 10D-2. Impacts on the Nonroad Fuel Market in PADD 2 (Average Price per Gallon = $0.9399)a
ISppm PADD 2
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
NPVb
Engineering
Cost/Unit
$0.0611
$0.0611
$0.0611
$0.0611
$0.0611
$0.0611
$0.0611
$0.0611
$0.0611
$0.0611
$0.0611
$0.0611
$0.0611
$0.0611
$0.0611
$0.0611
$0.0611
$0.0611
$0.0611
$0.0611
$0.0611
Absolute
Change in
Price
—
—
—
$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
(%)
0.00%
0.00%
0.00%
6.48%
6.45%
6.43%
6.43%
6.42%
6.43%
6.43%
6.43%
6.43%
6.43%
6.43%
6.43%
6.43%
6.43%
6.43%
6.43%
6.43%
6.43%
6.43%
6.43%
6.43%
5.626%
Change in
Quantity
(%)
0.000%
0.000%
0.000%
-0.006%
-0.012%
-0.016%
-0.018%
-0.019%
-0.018%
-0.018%
-0.018%
-0.018%
-0.018%
-0.018%
-0.018%
-0.018%
-0.018%
-0.018%
-0.018%
-0.018%
-0.018%
-0.018%
-0.018%
-0.018%
Total
Engineering
Costs (S106)
—
—
—
$132.8
$135.7
$138.7
$141.7
$144.6
$147.6
$150.5
$153.4
$156.4
$159.3
$162.3
$165.2
$168.1
$171.1
$174.0
$176.9
$179.8
$182.8
$185.7
$188.6
$191.5
$2,115.0
SOOppm PADD 2
Engineering
Cost/Unit
$0.0415
$0.0415
$0.0415
$0.0357
$0.0357
$0.0357
$0.0357
$0.0357
$0.0357
$0.0357
$0.0357
$0.0357
$0.0357
$0.0357
$0.0357
$0.0357
$0.0357
$0.0357
$0.0357
$0.0357
$0.0357
$0.0357
$0.0357
$0.0357
Absolute
Change
in Price
$0.03
$0.03
$0.03
$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
Change
in Price
(%)
3.19%
3.19%
3.19%
3.79%
3.78%
3.78%
3.77%
3.77%
3.77%
3.77%
3.77%
3.77%
3.77%
3.77%
3.77%
3.77%
3.77%
3.77%
3.77%
3.77%
3.77%
3.77%
3.77%
3.77%
3.702%
Change in
Quantity
(%)
-0.002%
-0.003%
-0.003%
-0.003%
-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%
Total
Engineering
Costs (S106)
$96.6
$98.3
$100.1
$43.2
$43.9
$44.2
$44.5
$44.7
$45.1
$45.5
$45.9
$46.3
$46.7
$47.0
$47.4
$47.9
$48.4
$48.9
$49.4
$49.9
$50.4
$50.9
$51.4
$51.9
$885.1
Change in
Producer
Surplus for Fuel
Producers (S106)
-$0.3
-$0.4
-$0.4
-$0.6
-$1.2
-$1.7
-$1.9
-$2.1
-$1.9
-$2.0
-$2.1
-$2.1
-$2.1
-$2.2
-$2.2
-$2.2
-$2.3
-$2.3
-$2.4
-$2.4
-$2.4
-$2.5
-$2.5
-$2.5
-$27.7
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.
10-90
-------
Table 10D-3. Impacts on the Nonroad Fuel Market in PADD 4 (Average Price per Gallon = $0.9499)a
ISppm PADD 4
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
NPVb
Engineering
Cost/Unit
$0.0891
$0.0891
$0.0891
$0.0891
$0.0891
$0.0891
$0.0891
$0.0891
$0.0891
$0.0891
$0.0891
$0.0891
$0.0891
$0.0891
$0.0891
$0.0891
$0.0891
$0.0891
$0.0891
$0.0891
$0.0891
Absolute
Change in
Price
—
—
—
$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
Change
in Price
(%)
0.00%
0.00%
0.00%
9.35%
9.33%
9.31%
9.30%
9.30%
9.30%
9.30%
9.30%
9.30%
9.30%
9.30%
9.30%
9.30%
9.30%
9.30%
9.30%
9.30%
9.30%
9.30%
9.30%
9.30%
8.142%
Change in
Quantity
(%)
0.000%
0.000%
0.000%
-0.006%
-0.012%
-0.017%
-0.018%
-0.020%
-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%
-0.019%
Total
Engineering
Costs (S106)
—
—
—
$25.3
$25.9
$26.4
$27.0
$27.6
$28.1
$28.7
$29.2
$29.8
$30.4
$30.9
$31.5
$32.0
$32.6
$33.2
$33.7
$34.3
$34.8
$35.4
$35.9
$36.5
$403.0
SOOppm PADD 4
Engineering
Cost/Unit
$0.0371
$0.0371
$0.0371
$0.0336
$0.0336
$0.0336
$0.0336
$0.0336
$0.0336
$0.0336
$0.0336
$0.0336
$0.0336
$0.0336
$0.0336
$0.0336
$0.0336
$0.0336
$0.0336
$0.0336
$0.0336
$0.0336
$0.0336
$0.0336
Absolute
Change
in Price
$0.04
$0.04
$0.04
$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
Change
in Price
(%)
4.33%
4.33%
4.33%
3.52%
3.52%
3.52%
3.51%
3.51%
3.51%
3.51%
3.51%
3.51%
3.51%
3.51%
3.51%
3.51%
3.51%
3.51%
3.51%
3.51%
3.51%
3.51%
3.51%
3.51%
3.615%
Change in
Quantity
(%)
-0.002%
-0.003%
-0.003%
-0.003%
-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%
Total
Engineering
Costs (S106)
$34.0
$34.4
$34.9
$19.3
$19.5
$19.7
$19.8
$19.9
$20.1
$20.3
$20.5
$20.6
$20.8
$20.9
$21.1
$21.4
$21.6
$21.8
$22.0
$22.2
$22.5
$22.7
$22.9
$23.1
$369.3
Change in
Producer
Surplus for Fuel
Producers (S106)
-$0.1
-$0.1
-$0.1
-$0.1
-$0.2
-$0.3
-$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
-$0.5
-$0.5
-$5.2
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.
10-91
-------
Table 10D-4. Impacts on the Nonroad Fuel Market in PADD 5 (Average Price per Gallon = $0.9599)a
ISppm PADD 5
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
NPVb
Engineering
Cost/Unit
$0.0586
$0.0586
$0.0586
$0.0586
$0.0586
$0.0586
$0.0586
$0.0586
$0.0586
$0.0586
$0.0586
$0.0586
$0.0586
$0.0586
$0.0586
$0.0586
$0.0586
$0.0586
$0.0586
$0.0586
$0.0586
Absolute
Change in
Price
—
—
—
$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
(%)
0.00%
0.00%
0.00%
6.08%
6.06%
6.03%
6.03%
6.02%
6.03%
6.03%
6.03%
6.03%
6.03%
6.03%
6.03%
6.03%
6.03%
6.03%
6.03%
6.03%
6.03%
6.03%
6.03%
6.03%
5.278%
Change in
Quantity
(%)
0.000%
0.000%
0.000%
-0.006%
-0.012%
-0.017%
-0.018%
-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%
-0.018%
Total
Engineering
Costs (S106)
—
—
—
$27.3
$27.9
$28.5
$29.1
$29.7
$30.3
$30.9
$31.5
$32.1
$32.7
$33.3
$33.9
$34.5
$35.1
$35.7
$36.3
$36.9
$37.5
$38.1
$38.7
$39.3
$434.3
SOOppm PADD 5
Engineering
Cost/Unit
$0.0344
$0.0344
$0.0344
$0.0217
$0.0217
$0.0217
$0.0217
$0.0217
$0.0217
$0.0217
$0.0217
$0.0217
$0.0217
$0.0217
$0.0217
$0.0217
$0.0217
$0.0217
$0.0217
$0.0217
$0.0217
$0.0217
$0.0217
$0.0217
Absolute
Change
in Price
$0.03
$0.03
$0.03
$0.02
$0.02
$0.02
$0.02
$0.02
$0.02
$0.02
$0.02
$0.02
$0.02
$0.02
$0.02
$0.02
$0.02
$0.02
$0.02
$0.02
$0.02
$0.02
$0.02
$0.02
Change
in Price
(%)
2.64%
2.63%
2.63%
2.25%
2.25%
2.24%
2.23%
2.23%
2.24%
2.24%
2.24%
2.24%
2.24%
2.24%
2.24%
2.24%
2.24%
2.24%
2.24%
2.24%
2.24%
2.24%
2.24%
2.24%
2.287%
Change in
Quantity
(%)
-0.002%
-0.003%
-0.003%
-0.003%
-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%
Total
Engineering
Costs (S106)
$15.7
$16.0
$16.3
$4.1
$4.1
$4.2
$4.2
$4.2
$4.2
$4.3
$4.3
$4.4
$4.4
$4.4
$4.5
$4.5
$4.6
$4.6
$4.6
$4.7
$4.7
$4.8
$4.8
$4.9
$101.1
Change in
Producer
Surplus for Fuel
Producers (S106)
-$0.1
-$0.1
-$0.1
-$0.1
-$0.3
-$0.4
-$0.4
-$0.4
-$0.4
-$0.4
-$0.4
-$0.5
-$0.5
-$0.5
-$0.5
-$0.5
-$0.5
-$0.5
-$0.5
-$0.5
-$0.5
-$0.5
-$0.5
-$0.5
-$6.0
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.
10-92
-------
Economic Impact Analysis
APPENDIX 10E: Time Series of Social Cost
This appendix provides a time series of the estimated social costs for the proposed program for
the period 2007 through 2030. Costs are presented in 2001 dollars.
10-93
-------
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
Change in Market Surplus
($106/yr)
NR Spillover
Operating and Marker Costs
($106/yr)
Social Costs ($106/yr)
2007
$0.0
$0.5
$0.4
$0.1
$0.1
$214.3
$86.7
$127.7
$58.7
$44.0
$111.7
$1.2
$0.5
$0.4
$0.1
$0.1
$216.1
$44.8
-$221.3
$39.6
2008
$14.0
$8.9
$2.0
$2.4
$4.5
$280.4
$113.0
$167.3
$74.4
$57.9
$148.1
$1.5
$0.7
$0.6
$0.2
$0.1
$304.8
$45.8
-$220.3
$130.4
2009
$14.0
$9.0
$2.0
$2.4
$4.5
$285.5
$115.2
$170.4
$76.1
$59.2
$150.3
$1.6
$0.7
$0.6
$0.2
$0.1
$310.0
$46.9
-$224.7
$132.2
2010
$14.0
$9.5
$2.4
$2.6
$4.6
$433.4
$178.6
$254.8
$143.1
$108.9
$181.3
$2.6
$1.2
$0.9
$0.3
$0.2
$459.5
$48.0
-$245.5
$262.00
2011
$23.5
$73.0
$35.1
$27.7
$10.2
$749.7
$315.7
$434.0
$297.5
$229.4
$222.8
$4.8
$2.2
$1.8
$0.5
$0.3
$851.0
$49.0
-$259.0
$641.1
2012
$29.4
$106.2
$50.1
$37.0
$19.0
$1,069.3
$451. 7
$617.6
$437.6
$315.2
$316.5
$6.9
$3.2
$2.6
$0.7
$0.5
$1,211.8
$50.1
-$251.6
$1,010.3
2013
$30.2
$116.1
$53.0
$39.9
$23.2
$1,231.8
$515. 7
$716.1
$468.3
$348.7
$414.8
$7.8
$3.6
$2.9
$0.8
$0.5
$1,385.8
$51.2
-$234.6
$1,202.4
2014
$35.6
$144.0
$67.8
$49.2
$27.1
$1,322.1
$555.1
$767.0
$515.6
$377.5
$429.0
$8.4
$3.9
$3.1
$0.9
$0.6
$1,510.1
$52.3
-$233.4
$1,329.0
2015
$35.6
$143.8
$67.7
$49.1
$27.0
$1,246.4
$525.1
$721.3
$499.9
$358.0
$388.6
$8.0
$3.7
$2.9
$0.8
$0.6
$1,433.9
$53.3
-$226.6
$1,260.6
2016
$25.5
$124.9
$57.5
$41.0
$26.4
$1,302.6
$548.0
$754.6
$516.3
$374.3
$412.0
$8.4
$3.9
$3.1
$0.8
$0.6
$1,461.4
$54.4
-$217.5
$1,298.3
2017
$20.3
$119.8
$54.9
$39.1
$25.8
$1,323.6
$556.9
$766.7
$524.8
$380.5
$418.2
$8.5
$3.9
$3.1
$0.9
$0.6
$1,472.2
$55.5
-$209.0
$1,318.6
2018
$5.5
$111.3
$53.2
$36.5
$21.7
$1,344.6
$565. 7
$778.8
$533.4
$386.7
$424.5
$8.7
$4.0
$3.2
$0.9
$0.6
$1,470.1
$56.5
-$201.7
$1,324.9
(continued)
10-94
-------
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
Change in Market Surplus
(SlOVyr)
NR Spillover
Operating and Marker Costs
(SlOVyr)
Social Costs ($106/yr)
2019
$0.1
$102.6
$48.2
$33.2
$21.2
$1,365.6
$574.6
$791.0
$541.9
$393.0
$430.7
$8.8
$4.1
$3.2
$0.9
$0.6
$1,477.1
$57.6
-$195.5
$1,339.2
2020
$0.1
$102.6
$48.2
$33.2
$21.2
$1,386.5
$583.4
$803.1
$550.4
$399.2
$436.8
$9.0
$4.1
$3.3
$0.9
$0.6
$1,498.2
$58.6
-$190.1
$1,366.7
2021
$0.1
$60.3
$27.4
$17.0
$16.0
$1,407.7
$592.3
$815.3
$559.0
$405.4
$443.3
$9.1
$4.2
$3.4
$0.9
$0.7
$1,477.2
$59.7
-$185.9
$1,350.9
2022
$0.1
$32.9
$15.3
$9.5
$8.1
$1,428.9
$601.3
$827.6
$567.5
$411.7
$449.7
$9.3
$4.3
$3.4
$0.9
$0.7
$1,471.2
$60.8
-$182.5
$1,349.4
2023
$0.1
$23.6
$12.9
$7.1
$3.6
$1,450.1
$610.2
$839.9
$576.0
$417.9
$456.2
$9.4
$4.3
$3.5
$0.9
$0.7
$1,483.3
$61.8
-$179.7
$1,365.4
2024
$0.1
$4.9
$3.5
$1.2
$0.2
$1,471.4
$619.2
$852.2
$584.6
$424.1
$462.7
$9.6
$4.4
$3.5
$1.0
$0.7
$1,486.0
$62.9
-$177.4
$1,371.5
2025
$0.1
$5.0
$3.6
$1.2
$0.2
$1,492.6
$628.2
$864.5
$593.1
$430.4
$469.2
$9.7
$4.5
$3.6
$1.0
$0.7
$1,507.5
$63.9
-$175.6
$1,395.8
2026
$0.1
$5.1
$3.6
$1.2
$0.2
$1,513.9
$637.1
$876.8
$601.6
$436.6
$475.7
$9.9
$4.5
$3.6
$1.0
$0.7
$1,528.9
$65.0
-$174.3
$1,419.6
2027
$0.1
$5.1
$3.7
$1.3
$0.2
$1,535.1
$646.1
$889.1
$610.1
$442.8
$482.2
$10.0
$4.6
$3.7
$1.0
$0.7
$1,550.4
$66.0
-$173.7
$1,442.8
2028
$0.1
$5.2
$3.7
$1.3
$0.2
$1,556.4
$655.1
$901.4
$618.7
$449.0
$488.7
$10.2
$4.7
$3.7
$1.0
$0.7
$1,571.9
$67.1
-$173.7
$1,465.3
2029
$0.1
$5.3
$3.8
$1.3
$0.2
$1,577.7
$664.0
$913.7
$627.2
$455.3
$495.2
$10.3
$4.7
$3.8
$1.0
$0.7
$1,593.4
$68.1
-$174.0
$1,487.5
2030
$0.1
$5.3
$3.8
$1.3
$0.2
$1,598.9
$672.9
$926.0
$635.7
$461.5
$501.8
$10.5
$4.8
$3.9
$1.0
$0.8
$1,614.9
$69.2
-$174.5
$1,509.6
10-95
-------
Draft Regulatory Impact Analysis
APPENDIX 10F: Model Equations
To enhance understanding of the economic model EPA used in this report, additional details
about the model's structure are provided in this appendix. The equations describing supply, final
demand, and intermediate (i.e., derived) demand relationships are presented below along with a
brief description of the solution algorithm.
10F.1 Model Equations
A constant-elasticity functional form was selected for all supply and final demand functions.
The general form and description of these equations are presented below:
Supply Equation: Qx = a(Px - Ac - Acy)s (10F.1)
Final Demand Equation: Qx = aP^ (1°E2)
where
x = production output,
y = production input,
Qx = quantity of output (x) supplied or demanded,
Px = market price for output (x),
a = constant,
Ac = direct supply shift ($/Qx),
Acy = indirect supply shift resulting from change in the price of input y, and
s,t| = 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:
±1.2-= Eap"^ • £-!- = e. (10F.3)
dp q a
The intermediate input (Qy) demands is specified within the supply chain as a function of
output (Qx). 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:
10-96
-------
Economic Impact Analysis
Constant Calibration: a= Qx° (10F.5)
(PXO)E
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:
AP • Q0
v0__ (10F6)
Q
xO
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 proposal):
• less than 25 hp,
26 to 50 hp,
51 to 75 hp,
76tolOOhp,
101 to 175 hp,
1 76 to 600 hp, and
• 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 manufacturersL and are represented by derived demand functions
(Eqs. [104a] and [10F.4b]).
Captive Domestic Supply Equation: Sengcap =a!(p-c)E (10F.7)
Merchant Domestic Supply Equation: Sengmer = a2(p - c)e (10F.8)
LNote 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.
10-97
-------
Draft Regulatory Impact Analysis
Import Supply Equation: M eng = a(p - c)E (10F.9)
Integrated Demand Equation: Dl = S (Sequip) (10F. 10)
Nonintegrated Demand Equation: DM = S(Sequip) (10F.11)
Market Clearing Condition: Sengcap + Sengmer + Meng = Dl + DM (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,M so the total number of diesel equipment markets is 42,
not 49). The equipment types are:
• construction
• agricultural,
• 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 (construction, agricultural, and manufacturers) is represented by derived demand
functions (Eq. [10F.4a and 10F.4b])
Domestic Integrated Supply Equation: SeqI=a(p-c)E (10F.13)
Domestic Nonintegrated Supply Equation: SeqNI = a(p — c - cy)e (10F.14)
Domestic Demand Equation: Deq =y.Qeq M 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:
M These are: agricultural equipment >600 hp; gensets & welders > 600 hp; refrigeration & A/C >
71 hp (4 hp categories); and lawn & garden >600 hp.
10-98
-------
Economic Impact Analysis
• construction
• agricultural, and
• 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 = a(papp - p0APeq -p!APfuel)e (10F.17)
Foreign (Import) Supply Equation: M app = apappe (10F.18)
Domestic Demand Equation: Dapp=apT1 (10F.19)
Foreign (Export) Demand Equation: Xapp = apn (10F.20)
Market Clearing Condition: Sapp + Mapp = D+Xapp (10F.21)
fuelO
Po and Pj are the baseline input shares of equipment ——— and fuel —
l^QappoJ ^Q
10F.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:
• 500 ppm nonroad diesel fuel, and
• 15 ppm nonroad diesel fuel.
The four regional nonroad diesel fuel markets are
PADD 1 and 3,
PADD 2,
PADD 4, and
• PADD 5 (includes Alaska and Hawaii)
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] are specified below:
Supply Equation: Sj = a(Pj - Ac)s (10F.22)
Derived Demand Equation: Dj = Sqjo l +AQ app (10F.23)
I QappO J
Market Clearing Condition: SJ = DJ (10F.24)
10-99
-------
Draft Regulatory Impact Analysis
10F.6 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.25)
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.
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
Pi+i=Pi' p- (10F.26)
4. Recalculate market supply with new prices, accounting for fuel-switching choices
associated with new energy 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).
10-100
-------
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 implementation of the
proposed rule 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 (construction, agriculture, and
manufacturing),
• supply elasticities for equipment markets,
• supply elasticities for engine markets, and
• supply elasticities for diesel fuel markets.
Note that demand elasticities for the equipment, engine, and diesel fuel 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 contains the demand and supply elasticities used to estimate the
economic impact of the proposed 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
construction and agricultural 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.
10-101
-------
Draft Regulatory Impact Analysis
Table 10G-1
Summary of Market Demand Elasticities Used in the NDEIM
Market
Estimate
Source
Method
Input Data Summary
Applications
Construction
Agriculture
-0.96 EPA econometric
estimate
-0.20 EPA econometric
estimate
Simultaneous equation
(log-log) approach
Productivity shift
approach (Morgenstern,
Pizer, and Shih, 2002)
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)
Manufacturing -0.58
EPA econometric
estimate
Simultaneous equation
(log-log) approach.
Equipment
Construction
Agriculture
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
Annual time series from
1958 - 1995 developed by
Jorgenson et al. (Jorgenson,
1990; Jorgenson, Gollop, and
Fraumeni, 1987)
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
10-102
-------
Economic Impact Analysis
Table 10G-2
Summary of Market Supply Elasticities Used in the NDEIM
Markets
Applications
Construction
Agriculture
Manufacturing
Equipment
Construction
Agriculture
Pumps/
compressors
Generators/
Welder Sets
Refrigeration
Industrial
Lawn and
Garden
Engines
Diesel fuel
Estimate
1.0
0.32
1.0
3.31
2.14
2.83
2.91
2.83
5.37
3.37
3.81
0.24
Source
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
Based on Topel and Rosen,
(1988).a
Production-weighted average
of individual crop estimates
ranging from 0.27 to 0.55.
(Lin etal., 2000)
Literature estimates are not
available so assumed same
value as for Construction
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 (2002).b
Input Data Summary
Census data, 1963 -
1983
Agricultural Census
data 1991 - 1995
Not applicable
Census data 1958-
1996; SIC 3531
Census data 1958-
1996; SIC 3523
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 Group
(EIG); 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.
10-103
-------
Draft Regulatory Impact Analysis
10G.1 Application Markets - Demand Elasticities
There are three application markets in the NDEIM: construction, agricultural, 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
10-104
-------
Economic Impact Analysis
PL, = log of cost of labor inputs in year t
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 b l 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 NDEEVI.
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:
Pt = fn(PLt,PKt,PMt,HHt,L)
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.
10-105
-------
Draft Regulatory Impact Analysis
Table 10G-3. Construction Demand Elasticity
Number of Observations = 29
R squared = 0.78
Adjusted R squared = 0.75
Variable
intercept
In price
In number of households
In average income per
household
Estimated Coefficients
18.83
-0.96
-1.73
-1.67
t-statistic
5.19
-3.83
-3.37
5.34
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.
10-106
-------
Economic Impact Analysis
This is a technique that regresses historical data for aggregate output on industry productivity
(Morgenstern, Pizer, and Shih, 2002).
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, + s,
where
output, = output t is the industry output in year t,
prod, = industry productivity in year t, and
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:
10-107
-------
Draft Regulatory Impact Analysis
Aln prod, = £sh d^tllW-i) (lnPsh,t-lnPshiM) - (InPO.-lnPO^) (10.G-2)
d
where
P = input prices,
PO = output prices, and
i) = input shares.
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,
Where
Qo = quantity of output
Q! = quantity of input
Since Q0/ Qj 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.
10-108
-------
Economic Impact Analysis
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 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
10-109
-------
Draft Regulatory Impact Analysis
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,
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
10-110
-------
Economic Impact Analysis
changes in prices. A sensitivity analysis for this assumed elasticity is presented in Appendix I.
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 r1 (10G-3)
where
Qt = output in year t,
Kt = real capital consumed in production in year t,N
Lt = quantify of labor used in year t,
Mt = material inputs in year t, and
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, %, aL, aM, 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 = (a{ + aj I (I- a, - am) (10G-4)
This underlying relationship is derived from the technical production function and the
behavioral profit maximization conditions. The derivation for equation (106) is provided in
Appendix 10H.
NCapital consumed is defined as the value added minus labor expenditures, divided by the price
index for capital.
10-111
-------
Draft Regulatory Impact Analysis
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 aL +
aM = 1, then the supply elasticity will be undefined. Alternatively, if aL + aM > 1, this yields a
negative supply elasticity. Thus, a common assumption is that aK + aL + aM = 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
• 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.
10-112
-------
Economic Impact Analysis
Table 10G-6. Engine Supply Elasticity
Supply Elasticity = 3.81
Number of Observations = 33
R-squared = 0.9978
Goldfeld-QuandtF = 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
13.40
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
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
10-113
-------
Draft Regulatory Impact Analysis
Table 10G-8. Construction Supply Elasticity
Supply Elasticity = 3.31
Number of Observations = 33
R-squared = 0.9926
Goldfeld-QuandtF = 1.76
Note: F(14,14) = 2.46.
Variable
Intercept
InK
InT
InM
InL
Estimated Coefficients
1.172
0.2318
-0.0617
0.1511
0.6172
t-statistic
28.54
5.83
-7.08
4.54
13.97
Table 10G-9. Industrial Supply Elasticity
Supply Elasticity = 5.37
Number of Observations = 33
R-squared = 0.9949
Goldfeld-QuandtF = 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
21.90
10-114
-------
Economic Impact Analysis
Table 10G-10. Garden
Supply Elasticity = 3.37
Number of Observations = 33
R-squared = 0.9963
Goldfeld-QuandtF = 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
Table 10G-11. Gensets
Supply Elasticity = 2.91
Number of Observations = 33
R-squared = 0.9909
Goldfeld-QuandtF = 1.16
Note: F(14,14) = 2.46.
Variable
Intercept
InK
InT
InM
InL
Estimated Coefficients
1.1304
0.2557
0.0325
0.3797
0.3646
t-statistic
11.09
3.60
2.73
4.67
4.51
10-115
-------
Draft Regulatory Impact Analysis
Table 10G-12. Pumps
Supply Elasticity = 2.83
Number of Observations = 33
R-squared = 0.9979
Goldfeld-QuandtF = 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 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.
10-116
-------
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_ = aLoc-lkl-a
9L
L =
Substitute Back into Cobb-Douglas:
y =
PocV
w
Iny = _^inP + _^in( ^ ) + lnk
1 -a 1 -a v w
91ny a
9lnP 1-a
= Supply Elasticity
10-117
-------
Draft 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. Also, the application market results are presented without adjusting by the
operating savings. Instead, these are added into the welfare changes separately.
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, $1,202.4 million, with the exception of Scenario 2
(alternative ways to treat fuel market costs). The base case models fuel market costs based on
average variable + fixed costs. Two alternatives were considered: maximum total costs and
maximum variable costs. In both of these alternatives, the social costs of the rule (less operating
savings and fuel marker and spillover costs) would increase by about 2 percent, to about $1,229
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 82
percent), although there are small differences in the way the costs are borne among the markets.
There are also differences in the way the application market costs are shared among producers
and consumers in that market, especially for Scenario 1. 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 ($1,412.1 million versus $1,229.3 million
for the rule), indicating that the refiners would gain from the proposal (about $146.3 million). In
the maximum variable cost scenario, the share borne by the refiners would increase from $7.8
million to $200.9 million, and the share borne by the application market would decrease from
$1,231.8 million to $1,066 million.
With regard to the market analysis, expected price and quantity changes are about the same as
in the base case. The expected change in engine prices is the same except in Scenario 4 (includes
engine and equipment fixed costs), in which the expected engine price increase goes up about 6
percent (from 22.88 percent to 24.22 percent). The expected change in equipment prices is also
10-118
-------
Economic Impact Analysis
similar across scenarios with the exception of Scenario 4, in which the expected equipment price
increase goes up about 11 percent (from 5.23 percent to 5.83 percent). For the application
market, the expected price increase remains between 0.01 percent and 0.02 percent. Expected
fuel price changes are somewhat more volatile across the scenarios, ranging from a 1.77 percent
increase to a 4.54 increase, compared to 3.09 percent increase in the base case. Finally, expected
decreases in the quantities produced do not change much. The largest expected quantity decrease
is 0.019 for the equipment market in Scenariol; the smallest is 0.006 for the application markets
in Scenario 1.
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
NDEEVI 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
reported in the econometric analysis (see Appendix 10G).
10-119
-------
Draft Regulatory Impact Analysis
Table 101-1. Sensitivity Analysis of the Supply and Demand Elasticities
for the Application Markets
Parameter/Market
Supply elasticity
Construction
Agriculture
Manufacturing
Demand elasticity
Construction
Agriculture
Manufacturing
Elasticity Source
Literature estimate
Literature estimate
Assumed value
EPA estimate
EPA estimate
EPA estimate
Upper Bound
2.3
0.55
2.3
-1.39
-0.35
-1.02
Base Case
1.00
0.32
1.00
-0.96
-0.20
-0.58
Lower Bound
0.50
0.027
0.50
-0.534
-0.054
-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 are reported in Tables 101-
2 and 101-3. As can be seen in those tables, market price and quantity increases vary negligibly
across the upper- and lower-bound sensitivity scenarios.
The change in total social surplus for 2013 also remains essentially unchanged across all
scenarios and is approximately the same as for the proposed program ($1,202.5 million).
However, consumers in the application market bear a smaller share of the social costs when they
are more responsive to price changes relative to producers (supply lower bound and demand
upper bound scenarios). As shown, consumers bear approximately 33 and 46 percent,
respectively, in these scenarios compared to 58 percent in the base case. In contrast, they bear a
higher share (up to 78 percent) when they are less responsive to price changes relative to
producers (supply upper bound and demand lower bound scenarios). While the burden of the
fuel market changes slightly, it always remain below 1 percent of the social costs.
10-120
-------
Economic Impact Analysis
Table 101-2. Application Market Sensitivity Analysis for Supply Elasticities"'13
Scenario
Application Markets
Price ($/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 (units/yr)
Change in Producer Surplus
($106/yr)
Engine Markets
Price ($/q)
Quantity (units/yr)
Change in Producer Surplus
($106/yr)
Fuel Markets
Price ($/q)
Quantity (gal/yr)
Change in Producer Surplus
($106/yr)
Change in Market Surplus
($106/yr)
NR Spillover
Operating and Marker Cost
($106/yr)
Social Costs ($106/yr)
Base Case
Absolute
NA
NA
$716.1
$515.6
$1,231.7
$1,016.49
-118
$116.1
$839.71
-69
$30.2
$0.03
-293,593
$7.8
$1,385.7
$51.2
-$234.6
$1,202.4
Relative
0.02%
-0.010%
NA
NA
NA
5.23%
-0.014%
NA
22.88%
-0.013%
NA
3.09%
-0.014%
NA
NA
NA
NA
NA
Supply
Absolute
NA
NA
$914.1
$313.2
$1,227.3
$1,041.63
-161
$117.6
$839.65
-95
$30.2
$0.03
-401,456
$10.6
$1,385.7
$51.2
-$234.6
$1,202.3
Upper Bound
Relative
0.02%
-0.014%
NA
NA
NA
5.23
-0.019%
NA
22.88%
-0.017%
NA
3.07%
-0.020%
NA
NA
NA
NA
NA
Supply
Absolute
NA
NA
$412.1
$825.0
$1,237.1
$1,018.68
-63
$114.3
$839.78
-40
$30.2
$0.03
-163,005
$4.3
$1,385.8
$5.12
-$234.6
$1,202.4
Lower Bound
Relative
0.01%
-0.007%
NA
NA
NA
5.23%
-0.008%
NA
22.88%
-0.007%
NA
3.12%
-0.008%
NA
NA
NA
NA
NA
a Sensitivity analysis is presented for 2013.
b Figures are in 2001 dollars.
10-121
-------
Table 101-3. Application Market Sensitivity Analysis for Demand Elasticities"'13
Scenario
Application Markets
Price ($/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 (units/yr)
Change in Producer Surplus
($106/yr)
Engine Markets
Price ($/q)
Quantity (units/yr)
Change in Producer Surplus
($106/yr)
Fuel Markets
Price ($/q)
Quantity (gal/yr)
Change in Producer Surplus
($106/yr)
Change in Market Surplus
($106/yr)
NR Spillover
Operating and Marker Cost
($106/yr)
Social Costs ($106/yr)
Base Case
Absolute
NA
NA
$716.1
$515.6
$1,231.7
$1,016.49
-118
$116.1
$839.71
-69
$30.2
$0.03
-293,593
$7.8
$1,385.7
$51.2
-$234.6
$1,202.4
Relative
0.02%
-0.010%
NA
NA
NA
5.23%
-0.014%
NA
22.88%
-0.013%
NA
3.09%
-0.014%
NA
NA
NA
NA
NA
Demand
Absolute
NA
NA
$566.7
$662.4
$1,229.0
$1,015.45
-145
$116.9
$839.67
-86
$30.2
$0.03
-359,059
$9.5
$1,385.7
$51.2
-$234.6
$1,202.3
Upper Bound
Relative
0.01%
-0.013%
NA
NA
NA
5.23%
-0.018%
NA
22.88%
-0.016%
NA
3.08%
-0.018%
NA
NA
NA
NA
NA
Demand
Absolute
NA
NA
$970.9
$265.3
$1,236.1
$1,018.19
-72
$114.7
$839.78
-42
$30.2
$0.03
-184,642
$4.8
$1,385.8
$51.2
-$234.6
$1,202.4
Lower Bound
Relative
0.02%
-0.006%
NA
NA
NA
5.23%
-0.009%
NA
22.88%
-0.008%
NA
3.11%
-0.009%
NA
NA
NA
NA
NA
a Sensitivity analysis is presented for 2013.
b Figures are in 2001 dollars.
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 elasticity values evaluated for each market are provided in
Table 101-4. The engine and equipment market supply elasticities are derived econometrically.
10-122
-------
Economic Impact Analysis
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
Construction
Agriculture
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
6.06
3.72
5.62
12.93
7.96
12.14
5.62
2.00
Base Case
3.81
3.31
2.14
2.83
5.37
3.37
2.91
2.83
0.20
Lower Bound
2.33
2.09
1.31
1.62
2.90
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.
For the engine and equipment markets (Tables 101-5 and 101-6), all quantitative estimates for
both market impacts (price and quantity changes) and social impacts (how the burden is shared
across markets) remain essentially unchanged when compared to the proposed program, 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. This is because the derived demand for
10-123
-------
Draft Regulatory Impact Analysis
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.
For the fuel market (Table 101-7), there is some variation in impacts. As the fuel market
supply elasticity becomes more elastic (supply upper bound; producers become more sensitive to
price changes), the change in fuel prices increases from 3.09 percent in the base case to 3.14
percent in the supply upper bound case, and producer welfare losses fall from $7.8 million to
about $1.0 million. In contrast, as the fuel market supply elasticity becomes less elastic (supply
lower bound; producers become less responsive to price changes), the change in fuel prices
decreases from 3.09 percent in the base case to 2.78 percent in the lower bound case, and
producer welfare losses increase from $7.8 million to $47.7 million.
It should be remembered that the demand elasticities for the equipment and engine diesel fuel
markets are derived as part of the model, and therefore sensitivity analysis was not conducted on
those parameters.0 In other words, the change in the application market quantities determines the
demand responsiveness in the engine, equipment, and diesel fuel markets. As a result, the
demand sensitivity analysis for these markets is indirectly shown in Table 101-2. Nonroad diesel
equipment and fuel expenditures are relatively small shares of total production costs for the
application markets. Therefore changes in these input prices do not significantly alter input
demand (i.e., demand in these markets is highly inelastic).
°For a discussion of the concept of derived demand, see Section 10.2.2.3 Incorporating
Multimarket Interactions.
10-124
-------
Economic Impact Analysis
Table 101-5. Equipment Market Supply Elasticity Sensitivity Analysis
a,b
Scenario
Application Markets
Price ($/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 (units/yr)
Change in Producer Surplus
($106/yr)
Engine Markets
Price ($/q)
Quantity (units/yr)
Change in Producer Surplus
($106/yr)
Fuel Markets
Price ($/q)
Quantity (gal/yr)
Change in Producer
Surplus($106/yr)
Change in Market Surplus
($106/yr)
NR Spillover
Operating and Marker Cost
($106/yr)
Social Costs ($106/yr)
Base Case
Absolute
NA
NA
$716.1
$515.6
$1,231.7
$1,016.49
-118
$116.1
$839.71
-69
$30.2
$0.03
-293,593
$7.8
$1,385.7
$51.2
-$234.6
$1,202.4
Relative
0.02%
-0.010%
NA
NA
NA
5.23%
-0.014%
NA
22.88%
-0.013%
NA
3.09%
-0.014%
NA
NA
NA
NA
NA
Supply
Absolute
NA
NA
$717.1
$516.5
$1,233.5
$1,018.73
-118
$114.2
$839.72
-67
$30.2
$0.03
-294,171
$7.8
$1,385.7
$51.2
-$234.6
$1,202.4
Upper Bound
Relative
0.02%
-0.010%
NA
NA
NA
5.23%
-0.014%
NA
22.88%
-0.012%
NA
3.09%
-0.014%
NA
NA
NA
NA
NA
Supply
Absolute
NA
NA
$714.7
$514.5
$1,229.2
$1,013.48
-118
$118.6
$839.71
-70
$30.2
$0.03
-292,828
$7.7
$1,385.7
$51.2
-$234.6
$1,202.4
Lower Bound
Relative
0.02%
-0.010%
NA
NA
NA
5.22%
-0.014%
NA
22.88%
-0.013%
NA
3.09%
-0.014%
NA
NA
NA
NA
NA
a Sensitivity analysis is presented for 2013.
b Figures are in 2001 dollars.
10-125
-------
Draft Regulatory Impact Analysis
Table 101-6. Engine Market Supply Elasticity Sensitivity Analysis"'13
Scenario
Application Markets
Price ($/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 (units/yr)
Change in Producer Surplus
($106/yr)
Engine Markets
Price ($/q)
Quantity (units/yr)
Change in Producer Surplus
($106/yr)
Fuel Markets
Price ($/q)
Quantity (gal/yr)
Change in Producer Surplus
($106/yr)
Change in Market Surplus
($106/yr)
NR Spillover
Operating and Marker Cost
($106/yr)
Social Costs ($106/yr)
Base Case
Absolute
NA
NA
$716.1
$515.6
$1,231.7
$1,016.49
-118
$116.1
$839.71
-69
$30.2
$0.03
-293,593
$7.8
$1,385.7
$51.2
-$234.6
$1,202.4
Relative
0.02%
-0.010%
NA
NA
NA
5.23%
-0.014%
NA
22.88%
-0.013%
NA
3.09%
-0.014%
NA
NA
NA
NA
NA
Supply
Absolute
NA
NA
$716.1
$515.6
$1,231.7
$1,016.55
-118
$116.1
$839.80
-70
$30.2
$0.03
-293,603
$7.8
$1,385.7
$51.2
-$234.6
$1,202.4
Upper Bound
Relative
0.02%
-0.010%
NA
NA
NA
5.23%
-0.014%
NA
22.88%
-0.013%
NA
3.09%
-0.014%
NA
NA
NA
NA
Supply
Absolute
NA
NA
$716.0
$515.6
$1,231.6
$1,016.43
-118
$116.1
$839.61
-69
$30.3
$0.03
-293,580
$7.8
$1,385.7
$51.2
-$234.6
$1,202.4
Lower Bound
Relative
0.02%
-0.010%
NA
NA
NA
5.23%
-0.014%
NA
22.88%
-0.013%
NA
3.09%
-0.014%
NA
NA
NA
NA
a Sensitivity analysis is presented for 2013.
b Figures are in 2001 dollars.
10-126
-------
Economic Impact Analysis
Table 101-7. Fuel Market Supply Elasticity Sensitivity Analysis'
a,b
Scenario
Application Markets
Price ($/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 (units/yr)
Change in Producer Surplus
($106/yr)
Engine Markets
Price ($/q)
Quantity (units/yr)
Change in Producer Surplus
($106/yr)
Fuel Markets
Price ($/q)
Quantity (gal/yr)
Change in Producer Surplus
($106/yr)
Change in Market Surplus
($106/yr)
NR Spillover
Operating and Marker Cost
($106/yr)
Social Costs ($106/yr)
Base Case
Absolute
NA
NA
$716.1
$515.6
$1,231.7
$1,016.49
-118
$116.1
$839.71
-69
$30.2
$0.03
-293,593
$7.8
$1,385.7
$51.2
-$234.6
$1,202.4
Relative
0.02%
-0.010%
NA
NA
NA
5.23%
-0.014%
NA
22.88%
-0.013%
NA
3.09%
-0.014%
NA
NA
NA
NA
NA
Supply Upper Bound
Absolute
NA
NA
$720.0
$518.5
$1,238.5
$1,016.46
-119
$116.1
$839.71
-70
$30.2
$0.03
-295,287
$0.9
$1,385.7
$51.2
-$234.6
$1,202.4
Relative
0.02%
-0.010%
NA
NA
NA
5.23%
-0.014%
NA
22.88%
-0.013%
NA
3.14%
-0.014%
NA
NA
NA
NA
NA
Supply
Absolute
NA
NA
$692.8
$499.2
$1,191.9
$1,016.66
-114
$115.9
$839.72
-67
$30.2
$0.03
-283,979
$47.7
$1,385.8
$51.2
-$234.6
$1,202.4
Lower Bound
Relative
0.02%
-0.010%
NA
NA
NA
5.23%
-0.014%
NA
22.88%
-0.012%
NA
2.78%
-0.014%
NA
NA
NA
NA
NA
a Sensitivity analysis is presented for 2013.
b Figures are in 2001 dollars.
10-127
-------
Draft 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:
O=r *O +C *O
Vtot '-'max Vmax ^ass Va
(101.2)
10-128
-------
Economic Impact Analysis
where Cave is the average control cost for the total population; Qmax, Cmax, and Qagg, Cagg are the
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 (units/yr)
Change in Producer Surplus
($106/yr)
Engine Markets
Price ($/q)
Quantity (units/yr)
Change in Producer Surplus
($106/yr)
Fuel Markets
Price ($/q)
Quantity (gal/yr)
Change in Producer Surplus
($106/yr)
Change in Market Surplus
($106/yr)
NR Spillover
Operating and Marker Cost
($106/yr)
Social Costs ($106/yr)
Average
Absolute
Change
NA
NA
$716.1
$515.6
$1,231.7
$1,016.49
-118
$116.1
$839.71
-69
$30.2
$0.03
-293,593
$7.8
$1,385.7
$51.2
-$234.6
$1,202.4
Total Scenario
Relative
Change (%)
0.02%
-0.010%
NA
NA
NA
5.23%
-0.014%
NA
22.88%
-0.013%
NA
3.09%
-0.014%
NA
NA
NA
NA
NA
Maximum Total Scenario
Absolute
Change
NA
NA
$814.7
$597.4
$1,412.1
$1,015.75
-136
$116.7
$839.69
-80
$30.2
$0.04
-337,228
$-146.3
$1,412.7
$51.2
-$234.6
$1,229.3
Relative
Change (%)
0.02%
-0.012%
NA
NA
NA
5.23%
-0.016%
NA
22.88%
-0.015%
NA
4.54%
-0.017%
NA
NA
NA
NA
NA
Maximum Variable Scenario
Absolute
Change
NA
NA
$612.3
$453.7
$1,066.0
$1,017.06
-104
$115.6
$839.73
-61
$30.2
$0.02
-259,056
$200.9
$1,412.8
$51.2
-$234.6
$1,229.4
Relative
Change (%)
0.01%
-0.009%
NA
NA
NA
5.23%
-0.013%
NA
22.88%
-0.011%
NA
1.77%
-0.013%
NA
NA
NA
NA
NA
a Sensitivity analysis is presented for 2013.
b Figures are in 2001 dollars.
10-129
-------
Draft Regulatory Impact Analysis
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
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 3.09 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 would be expected to benefit from the rule,
by about $146.3 million. This is because the change in market price would exceed 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 ($1,412.1 million, compared to $1,231.7 million for the welfare costs of the proposed
program less 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. Thus,
in this scenario, the expected refinery welfare loss is greater than for the propose program,
increasing from $7.8 million to $200.9 million. Similarly, the expected welfare loss to the
application markets (without considering the operating savings) decreases from $1,412.1 million
to $1,066 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, although output
does not contract. In this case, consumers and producers in the application market would be
expected to bear more than the cost of the rule. However, if only the highest cost refinery's
variable costs drive the new market price, then prices are expected to increase less, and producers
and consumer will bear less of the burden, with refiners bearing more.
101.3 Operating Cost Scenario
Changes in operating costs resulting from lower sulfur content nonroad diesel fuel are
included in the social cost estimates presented in Section 10.1. However, because of the
10-130
-------
Economic Impact Analysis
uncertainty of how these savings will effect individual equipment purchase decisions, operating
savings were not included in the market and analysis and were added to social costs after changes
in price and quantity were estimated. The results of this analysis are included in Table 101.9.
In this sensitivity analysis, operating saving are modeled as a cost reduction (benefit) for
producers in the application markets. To allow comparison of the results to the base case, the
base case is adjusted by adding all the operating savings to the producer surplus, making it
$273.7 million. This is because application market producers are the users of diesel equipment
and therefore it makes intuitive sense that these benefits accrue to them. In this scenario,
operating cost savings are treated as negative supply shift for the application supply curves.
When the operating costs are included in the total welfare costs, the social costs for this scenario
are about the same as the base case (about $1,202 million). The burden across the markets is also
unchanged. The price increase and quantity decrease in the application markets is expected to be
smaller. This is because by including operating savings in the supply shift, the magnitude of the
shift decreases. This leads to a smaller price and quantity change in the application market.
At the same time, the distribution of costs between producers and consumers in the
application market changes when operating costs are treated differently in the model. In the
NDEEVI, application consumers bear 72 percent of the burden of the loss of welfare surplus in the
application market, while producers bear the other 28 percent. This is because this scenario
assumes that application market consumers do not make market decisions based on operating
costs, and that they expect to run their equipment as before. Producers are not expected to pass
along operating savings to their customers. When the operating savings are included in the
model, the way the cost burden is shared changes 58 percent for the application market
consumers and 42 percent for the application market producers. Prices increase less and output
decreases less. In other words, the impacts of the operating savings are shared among the
producers and consumers.
10-131
-------
Draft Regulatory Impact Analysis
Operating
Table 101-9
Savings Included in the Market Analysis'1
Base Case (2013)
Scenario
Application Markets
Price ($/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 (units/yr)
Change in Producer Surplus
($106/yr)
Engine Markets
Price ($/q)
Quantity (units/yr)
Change in Producer Surplus
($106/yr)
Fuel Markets
Price ($/q)
Quantity (gal/yr)
Change in Producer Surplus
($106/yr)
NR Spillover
Marker Cost
Total Social Cost
Absolute
Change
NA
NA
$716.1
$273.7
$989.8
$1,016.49
-118
$116.1
$839.71
-69
$30.2
$0.03
-293,593
$7.8
$51.2
$7.3
$1,202.4
Relative Change
(%)
0.02%
-0.010%
NA
NA
NA
5.23%
-0.014%
NA
22.88%
-0.013%
NA
3.09%
-0.014%
NA
NA
NA
NA
Adding Operating Savings To App
Absolute
Change
NA
NA
$577.6
$414.7
$992.2
$1,017.50
-95
$115.3
$839.75
-56
$30.2
$0.03
-235,921
$6.2
$51.2
$7.3
$1,202.5
Relative Change
(%)
0.01%
-0.008%
NA
NA
NA
5.23%
-0.011%
NA
22.88%
-0.010%
NA
3.10%
-0.012%
NA
NA
NA
NA
Sensitivity analysis is presented for 201:
10-132
-------
Economic Impact Analysis
101.4 Engine and Equipment Fixed Cost Shift Scenario
As discussed in Section 10.3 only the variable costs are used to shift the supply curve in the
engines and equipment markets. Fixed costs are assumed to be R&D costs that are absorbed by
engine and equipment markets over a 5-year period and hence do not affect market prices or
quantities. As a result, producers are not able to pass any of these costs on and bear all fixed
costs as a decrease in producer surplus.
In this scenario, the supply shift for engine producers includes the fixed and variable
compliance costs. The results are presented in Table 101-10. In this scenario, engine producers
are able to pass along the majority of the fixed compliance costs to the downstream markets
rather than absorb them as a one-to-one reduction in profits. As expected, this scenario leads to a
higher projected price increases for the engine and equipment markets (from 5.2 percent in the
baseline case to 5.8 percent for equipment markets and from 22.9 percent in the baseline case to
24.2 percent for engine markets). These costs are passed on to the application markets, and their
expected share of the compliance burden increases from 90 percent to 99 percent. However, the
total social costs of the regulation are not expected to change measurably as the higher prices
lead to almost no change in the demand for equipment and engines.
10-133
-------
Draft Regulatory Impact Analysis
Table 101-10 Fixed Costs
Added to
Supply Shift in Engine and Equipment
Base Case (2013)
Scenario
Application Markets
Price ($/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 (units/yr)
Change in Producer Surplus
($106/yr)
Engine Markets
Price ($/q)
Quantity (units/yr)
Change in Producer Surplus
($106/yr)
Fuel Markets
Price ($/q)
Quantity (gal/yr)
Change in Producer Surplus
($106/yr)
Change in Market Surplus
($106/yr)
NR Spillover
Operating and Marker Cost
($106/yr)
Social Costs ($106/yr)
Absolute
Change
NA
NA
$716.1
$515.6
$1,231.7
$1,016.49
-118
$116.1
$839.71
-69
$30.2
$0.03
-293,593
$7.8
$1,385.7
$51.2
-$234.6
$1,202.4
Relative Change
(%)
0.02%
-0.010%
NA
NA
NA
5.23%
-0.014%
NA
22.88%
-0.013%
NA
3.09%
-0.014%
NA
NA
NA
NA
NA
Marketsa
Shocking Engine and Equipment
Markets by Total Costs
Absolute
Change
NA
NA
$796.9
$575.4
$1,372.3
$1,187.23
-132
$4.6
$894.93
-78
$0.1
$0.03
-329,511
$8.7
$1,385.7
$51.2
-$234.6
$1,202.3
Relative Change
(%)
0.02%
-0.011%
NA
NA
NA
5.83%
-0.016%
NA
24.22%
-0.014%
NA
3.08%
-0.016%
NA
NA
NA
NA
NA
Sensitivity analysis is presented for 2013.
10-134
-------
Economic Impact Analysis
101.5 Alternative Social Discount Rates
Future benefits and costs are commonly discounted to account for the time value of money.
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 $16.5
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 a 7 percent social discount rate. With the 7 percent social discount rate,
the present value of total social costs decreases from $18.9 billion to $9.2 billion.
Table 101-11. Net Present Values3
Engine Producers Total
Equipment Producers Total
Construction Equipment
Agricultural Equipment
Industrial Equipment
Application Total
Total Consumer
Total Producer
Construction
Agriculture
Manufacturing
Fuel Producers Total
PADD 1&3
PADD2
PADD 4
PADD 5
NR Spillover
Marker Cost
Total
Market
Surplus
(106)
$190.0
$927.4
$433.6
$306.7
$187.1
$17,744.2
$7,450.7
$10,293.5
$6,923.5
$5,050.4
$5,770.3
$113.9
$52.3
$41.9
$11.5
$8.1
$18,975.5
NPV (3%)
Operating Cost
Savings
(106)
-$3,402.4
-$1,094.9
-$629.3
-$1,678.1
$886.48
$63.0
-$2,452.8
Total
$190.0
$927.4
$433.6
$306.7
$187.1
$14,341.8
$5,828.6
$4,421.1
$4,092.2
$113.9
$52.3
$41.9
$11.5
$8.1
$16,522.7
NPV (7%)
Market
Surplus
(106)
$135.4
$595.2
$276.2
$198.0
$120.9
$10,066.8
$4,222.6
$5,844.1
$3,895.1
$2,847.5
$3,324.1
$64.2
$29.5
$23.6
$6.5
$4.6
$10,861.6
Operating
Cost Savings
(106)
-$2,204.9
-$709.6
-$407.8
-$1,087.5
$538.2
$50.93
-$1,615.8
Total
$135.4
$595.2
$276.2
$198.0
$120.9
$7,861.9
$3,185.6
$2,439.7
$2,236.6
$64.2
$29.5
$23.6
$6.5
$4.6
$9,245.9
Figures are in 2001 dollars.
10-135
-------
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 Definition and Description of Small Entities 11-2
11.3.1 Description of Nonroad Diesel Engine and Equipment Manufacturers 11-3
11.3.2 Description of the Nonroad Diesel Fuel Industry 11-3
11.4 Summary of Small Entities to Which the Rulemaking Will Apply 11-4
11.4.1 Nonroad Diesel Engine Manufacturers 11-4
11.4.2 Nonroad Diesel Equipment Manufacturers 11-5
11.4.3 Nonroad Diesel Fuel Refiners 11-5
11.4.4 Nonroad Diesel Fuel Distributors and Marketers 11-6
11.5 Related Federal Rules 11-6
11.6 Projected Reporting, Recordkeeping, and Other Compliance Requirements 11-6
11.7 Projected Economic Effects of the Proposed Rulemaking 11-7
11.8 Regulatory Alternatives 11-8
11.8.1 Small Engine Manufacturers 11-8
11.8.1.1 Flexibility Alternatives for Small Engine Manufacturers 11-9
11.8.1.2 Hardship Provisions for Small Engine Manufacturers 11-10
11.8.1.3 Other Small Engine Manufacturer Issues 11-11
11.8.1.4 SB A Office of Advocacy Observations 11-11
11.8.2 Nonroad Diesel Equipment Manufacturers 11-13
11.8.2.1 Flexibility Alternatives for Small Equipment Manufacturers 11-13
11.8.2.2 Hardship Provisions for Small Equipment Manufacturers 11-16
11.8.3 Nonroad Diesel Fuel Refiners 11-16
11.8.3.1 Flexibility Alternatives for Small Fuel Refiners 11-17
11.8.3.2 Small Refiner Incentives for Early Compliance 11-19
11.8.3.3 Hardship Provisions for Small Fuel Refiners 11-21
11.8.4 Nonroad Diesel Fuel Distributors and Marketers 11-21
-------
Small-Business Flexibility Analysis
CHAPTER 11: Small-Business Flexibility Analysis
This chapter discusses our Initial Regulatory Flexibility Analysis (IRFA) which evaluates the
potential impacts of the proposed standards on small entities. The Regulatory Flexibility Act, as
amended by the Small Business Regulatory Enforcement Fairness Act of 1996 (SBREFA),
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. Pursuant to this requirement, we have prepared an
IRF A for the proposed rule. Throughout the process of developing the IRFA, we conducted
outreach and held meetings with representatives from the various small entities that could be
affected by the rulemaking to gain feedback, including recommendations, on how to reduce the
impact of the rule on these entities. The small business recommendations stated here reflect the
comments of the small entity representatives (SERs) and members of the Small Business
Advocacy Review Panel (SBAR Panel, or 'the Panel').
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 IRFA. A summary of the Panel's recommendations is presented in
the preamble of this proposed rulemaking. Further, a detailed discussion of the Panel's advice
and recommendations is found in the Final Panel Report contained in the docket for this
proposed rulemaking.
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 IRFA under section 603 of the Regulatory Flexibility Act. Key elements of an
IRFA are:
- a description of and, where feasible, an estimate of the number of small entities to which the
proposed rule will apply;
- projected reporting, record keeping, and other compliance requirements of the proposed rule,
including an estimate of the classes of small entities which will be subject to the requirements
and the type of professional skills necessary for preparation of the report or record;
an identification to the extent practicable, of all other relevant Federal rules which may
duplicate, overlap, or conflict with the proposed rule;
any significant alternatives to the proposed rule which accomplish the stated objectives of
applicable statutes and which 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
those entities. Although we are not required by the Clean Air Act to provide special treatment to
11-1
-------
Draft Regulatory Impact Analysis
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 proposed rule are located in the
preamble to the proposed rule. As previously stated, controlling emissions from nonroad engines
and equipment, in conjunction with diesel fuel quality 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 on-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.
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 that the
opportunity for cost-effective emission reductions on a large scale appears to exist.
11.3 Definition and Description of Small Entities
Small entities include small businesses, small organizations, and small governmental
jurisdictions. For the purposes of assessing the impacts of the proposed 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 which 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.
11-2
-------
Small-Business Flexibility Analysis
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
Major SIC Codes"
Major Group 35
Major Group 35
Major Group 35
Major Group 35
2911
a Standard Industrial Classification
b We have included in past fuels rulemakings a provision that, in order to qualify for the small refiner flexibilities, a
refiner must also have a company-wide crude refining capacity of no greater than 155,000 barrels per calendar day. We
have included this criterion in the small refiner definition for a nonroad diesel sulfur program as well.
11.3.1 Description of Nonroad Diesel Engine and Equipment Manufacturers
To assess how many engine and equipment manufacturers would directly be affected by the
proposed rule which may meet these small entity criteria, we first created a database consisting of
firms listed in the Power Systems Research database and compared this with the list of
companies from the analysis performed for the 1998 nonroad rulemaking along 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 which use nonroad diesel engines, there are
numerous SIC codes in which the equipment manufacturers report their sales, though the
maj ority of the firms are listed under the SIC maj or group 3 5xx- Industrial and Commercial
Machinery and Computer Equipment.
11.3.2 Description of the Nonroad Diesel Fuel Industry
The analysis that we developed for the refining industry is built on analysis that was
performed for the gasoline and highway diesel sulfur programs in recent years. Information
about the characteristics of refiners comes from sources including the Energy Information
Administration within the U.S. Department of Energy, and from oil industry literature. Our
current assessment is 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 are the key sources thus far for information about this industry. This industry sector
includes several types of businesses that fall into several different SBA small entity criteria; our
assessment is that the vast majority of these entities are small.
-------
Draft Regulatory Impact Analysis
11.4 Summary of Small Entities to Which the Rulemaking Will Apply
The following sections discuss the small entities - namely nonroad diesel engine
manufacturers, nonroad diesel equipment manufacturers, and nonroad fuel refiners and fuel
marketers/distributors - directly regulated by this proposed rule. Also, Table 11-2 lists our
assessment of the number of small entities that will be directly affected by this rulemaking.
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
Number of Affected Entities
4a
335a
26
see discussion below
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 Nonroad Diesel Engine Manufacturers
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.
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.
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.
A All sales information used for this analysis was 2000 data.
11-4
-------
Small-Business Flexibility Analysis
11.4.2 Nonroad Diesel Equipment Manufacturers
The proposed rule may result in equipment manufacturers incurring increased costs as a
result of the need to make changes to their equipment to accommodate changes to the engine size
and the addition of an aftertreatment package. 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
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.
11.4.3 Nonroad Diesel Fuel Refiners
Our current assessment is that 26 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 currently produce approximately 6
percent of the total high-sulfur diesel fuel. It should be noted that because of the dynamics in the
refining industry (e.g., mergers and acquisitions), the actual number of refiners that ultimately
qualify for small refiner status under a future nonroad diesel sulfur program could be different
from this initial estimate.
11.4.4 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 would meet small entity criteria. Common carrier pipeline
companies are also a part of the distribution system; 10 of them are small businesses.
11-5
-------
Draft Regulatory Impact Analysis
11.5 Related Federal Rules
The proposed certification fees rule, through the Agency's Certification and Compliance
Division, may be in place by the time this rule is being implemented, and EPA took this potential
cost into consideration when assessing the effects that this rule may have on small businesses.
The fuel regulations that we expect to propose would be similar in many respects to the
existing sulfur standard for highway diesel fuel. We are not aware of any area where the
regulations under consideration would directly duplicate or overlap with the existing federal,
state, or local regulations; however, several small refiners will also be subject to the gasoline
sulfur and highway diesel sulfur control requirements, as well as air toxics requirements.
More stringent nonroad diesel sulfur standards may require some refiners to obtain permits
from state and local air pollution control agencies under the Clean Air Act's New Source Review
program prior to constructing the desulfurization equipment needed to meet the standards.
The Internal Revenue Service has an existing rule that levies taxes on highway diesel fuel
only. The rule requires that nonroad diesel (untaxed) fuel be dyed so that regulators and
customers will know which type of fuel is which.
11.6 Projected Reporting, Recordkeeping, and Other Compliance
Requirements
As with any emission control program, the Agency must have the assurance that the regulated
entities will meet the emissions standards and all related provisions. For engine and equipment
manufacturers, EPA proposes to continue the reporting, recordkeeping, and compliance
requirements prescribed for these categories in 40 CFR part 89. Key among these are
certification requirements and provisions related to reporting of production, emissions
information, use of transition provisions, etc.
For any fuel control program, EPA must have the assurance that fuel produced by refiners
meets the applicable standard, and that the fuel continues to meet the standard as it passes
downstream through the distribution system to the ultimate end user. This is particularly
important in the case of diesel fuel, where the aftertreatment technologies expected to be used to
meet the engine standards under consideration are highly sensitive to sulfur. The recordkeeping,
reporting and compliance provisions of the proposed rule 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.
11-6
-------
Small-Business Flexibility Analysis
11.7 Projected Economic Effects of the Proposed Rulemaking
The projected costs of the rulemaking on a per engine basis were independent of the size of
the engine or equipment manufacturer. A full discussion of these costs, and the corresponding
methodology, is located in Chapter 6 of this Draft RIA. Of the 710 entities with publicly
available sales data, our screening analysis found that the average total annual compliance costs
would be $33,000 per small entity and $4.42 million per large entity. Further, a cost-to-sales
ratio test, a ratio of the estimated annualized compliance costs to the value of sales per company,
was performed for these entities.8 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 proposed rule would be greater
than 1 percent, but less than 3 percent, of their sales). 1 percent (4 companies) of small entities
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.
Based on our outreach, fact-finding, and analysis of the potential impacts of our regulations
on small businesses, the Panel concluded that small refiners in general would likely experience a
significant and disproportionate financial hardship in reaching the objectives of the proposed
nonroad diesel fuel sulfur program. One indication of this disproportionate hardship for small
refiners is the relatively high cost per gallon projected for producing nonroad diesel fuel under
the proposed program. Refinery modeling (of all refineries), indicates significantly higher
refining costs for small refiners. Specifically, we project that without special provisions, refining
costs for small refiners on average would be about 5.5 cents per gallon compared to about 4.0
cents per gallon for non-small refiners. Chapter 7 of this Draft RIA further discusses the
estimated costs of production and distribution of low sulfur fuels.
The majority of the fuel-related cost of the proposal is refming-related, with only 15-25
percent of the costs being distribution-related. The proposed allowance that highway and off-
highway diesel engine fuel meeting the same sulfur specification can be shipped fungibly until it
leaves the terminal obviates the need for additional storage tankage in this segment of the
distribution system.0 The proposed rule would also allow 500 ppm off-highway diesel engine
fuel to be mixed with high-sulfur diesel fuel once the fuels are dyed to meet IRS requirements.
This provision would ease the last part of the distribution of high-sulfur nonroad, locomotive,
and marine (NRLM) diesel fuel.
B 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.
c Including the refinery, pipeline, marine tanker, and barge segments of the distribution
system.
11-7
-------
Draft Regulatory Impact Analysis
For a complete discussion of the economic impacts of the proposed rulemaking, see Chapter
10, the economic impact analysis chapter, of this Draft RIA.
11.8 Regulatory Alternatives
The Panel's findings and discussions are based on the information that was available during
the term of the Panel and issues that were raised by the SERs during the outreach meetings and in
their written comments. It was agreed that EPA should consider the issues raised by the SERs
(and issues raised in the course of the Panel) and that EPA should consider the comments on
flexibility alternatives that would help to mitigate any negative impacts on small businesses.
Alternatives discussed throughout the Panel process include those offered in the development of
the upcoming rule. Though some of the recommended flexibilities may be appropriate to apply
to all entities affected by the rulemaking, the Panel's discussions and recommendations are
focused mainly on the impacts, and ways to mitigate adverse impacts, on small businesses. A
summary of the Panel's recommendations, along with those provisions that we are actually
proposing in this action, are detailed below. A full discussion of the regulatory alternatives and
hardship provisions discussed and recommended by the Panel, all written comments received
from SERs, and summaries of the two outreach meetings that were held with the SERs can be
found in the SBREFA Final Panel Report.1 In addition, all of the flexibilities (or 'transition
provisions') that were proposed in the rulemaking for small businesses, as well as those for all
entities that may be affected by the rulemaking, are described in the preamble to the proposed
rule.
11.8.1 Small Engine Manufacturers
The Panel developed a wide range of regulatory alternatives to mitigate the impacts of the
rulemaking on small businesses, and recommended that we propose and seek comment on the
flexibilities. Described below are the flexibility options recommended by the Panel, along with
alternatives that were suggested by some individual Panel members, and our proposed regulatory
alternatives.
11.8.1.1 Flexibility Alternatives for Small Engine Manufacturers
11.8.1.1.1 SBAR Panel Recommendations
Based on the recommendations of the Panel, the transition flexibilities that were under
consideration were dependent upon what approach, or approaches, we proposed for the
rulemaking. Further, each manufacturer would be limited to 2,500 units per year (to allow for
some market growth). The proposed transition provisions are:
1. For an approach with two phases of standards the Panel recommended that:
an engine manufacturer could skip the first phase and comply on time with the
second; or,
11-8
-------
Small-Business Flexibility Analysis
a manufacturer could delay compliance with each phase of standards for up to two
years.
2. For an approach that entails only one phase of standards, the manufacturer could opt to
delay compliance. The Panel recommended that the length of the delay be a three-year
period; 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).
All Panel members believed that the aforementioned options would offer an opportunity to
reduce the burden on small manufacturers while at the same time meeting the regulatory goals of
the Agency. Further, the we believe that these options will not put small manufacturers at a
significant disadvantage as they will be in compliance with the Tier 4 standards in the long run
and the flexibility options will give them more lead time to comply.
11.8.1.1.2 EPA 's Proposed Regulatory Alternatives
We feel that a complete exemption from the upcoming standards (even assuming that such an
exemption could be justified legally) would put these manufacturers at a competitive
disadvantage as the rest of the market will be producing compliant engines and only equipment
able to accommodate compliant engines will be saleable. Due to the structure of the standards
and their timing, as discussed in Section in of the preamble to the proposed rulemaking, we are
proposing regulatory alternatives, or transition provisions, for small engine manufacturers which
encompass both approaches recommended by the Panel (with the inclusion of the 2,500 unit limit
for each manufacturer).
• With regard to PM:
Engines under 25 hp, and those between 75 and 175 hp, have only one standard so the
manufacturer could delay compliance with these standards for up to three years.
Based on available data, we believe that there are no small manufacturers of nonroad
diesel engines above 175 hp.
For engines between 50 and 75 hp, we are proposing a one phase program with the
option to delay compliance for one year if interim standards are met. For this power
category we are treating the PM standard as a two phase standard with the stipulation
that small manufacturers cannot use PM credits to meet the interim standard.
Furthermore, if a small manufacturer elects the optional approach to the standard (i.e.,
opts to skip the interim standard), no further relief will be provided. See Section HI of
the preamble to the proposed rulemaking for further detail on the PM standards for
engine manufacturers.
• With regard to NOx:
There is no change in the NOx standard for engines under 25 hp and those between 50
and 75 hp, therefore, we are not proposing special provisions for these two power
bands.
For engines in the 25-50 hp and the 75-175 hp categories we are proposing a three
year delay in the program consistent with the one-phase approach recommendation
11-9
-------
Draft Regulatory Impact Analysis
above. Again, based on available data, we believe that there are no small
manufacturers of nonroad diesel engines above 175 hp.
11.8.1.2 Hardship Provisions for Small Engine Manufacturers
11.8.1.2.1 SBAR Panel Recommendations
The Panel recommended that two types of hardship provisions be extended to small engine
manufacturers. These provisions are:
1. 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
(i.e., fire, tornado, supplier not fulfilling contract, etc.); and
2. For the case where a manufacturer has taken all reasonable business, technical, and
economic steps to comply but cannot do so.
Either relief provision could provide lead time for up to 2 years—in addition to the
flexibilities listed above in Section 11.8.1.1—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.
For the flexibilities listed above, the Panel recommended that engine manufacturers and
importers must have certified engines in model year 2002 or earlier in order to take advantage of
these provisions. Each manufacturer would be limited to 2,500 units per year (to allow for some
market growth). These provisions were recommended by the Panel in order to prohibit the
misuse of these flexibilities as a tool to enter the nonroad diesel market or to gain unfair market
position relative to other manufacturers.
11.8.1.2.2 EPA 's Proposed Hardship Provisions
We are proposing to adopt the Panel recommendations for hardship provisions for small
engine manufacturers. While perhaps ultimately not necessary given the phase-in schedule
discussed above, we believe that such provisions provide a useful safety valve in the event of
unforeseen extreme hardship.
11.8.1.3 Other Small Engine Manufacturer Issues
11.8.1.3.1 SBAR Panel Recommendations
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.
11-10
-------
Small-Business Flexibility Analysis
11.8.1.3.2 EPA 's Proposal
As discussed in Section Vn of the preamble to this proposal, we are indeed proposing ABT
provisions. ABT is being proposed as it is intended to enhance the flexibility offered to engine
manufacturers that will be of assistance in making the transition to meet the stringent standards
in this proposed rule in the leadtime proposed. As noted in Section VII. A, we are proposing to
retain the basic structure of the current nonroad diesel ABT program, though a number of
changes (which will help to accommodate implementation of the proposed emission standards)
are being proposed with this action.
Though the Panel recommended small engine manufacturer-specific ABT provisions, such
provisions are not being included in this proposal. We do not believe it would be appropriate to
provide a different ABT program for small engine manufacturers, especially given the special
provisions that are proposed above. Discussions during the SBAR process indicated that small
volume manufacturers would need extra time to comply due to cost and personnel constraints,
and there is little reason to believe that small manufacturer specific ABT provisions could create
an incentive to accelerate compliance. Small manufacturers would of course be able to
participate in the general ABT program.
11.8.1.4 SBA Office of Advocacy Observations
11.8.1.4.1 What One Panel Member Observed
The SBA Chief Counsel for the Office of Advocacy offered some observations about the
impacts of the regulatory approaches on affected small engine and equipment manufacturers.
While the other Panel members did not join in these observations, the Panel recommended that
the Administrator carefully consider these points and examine further the factual, legal and
policy questions raised here in developing the proposed rule. First, given the available
information, the Office of Advocacy stated that they had substantial doubts about the technical
feasibility and cost of engineering aftertreatment devices into a wide diversity of nonroad diesel
applications for engines less than 50 kilowatts (70 hp). They stated that considerable concern has
been raised regarding the technical feasibility of aftertreatment devices, even for larger engines,
and particularly in the case of NOx adsorbers. Second, the low retail cost and low annual
production for many of these applications make it extremely difficult for the equipment
manufacturer to absorb these additional costs. Third, Advocacy believes that given the small size
of these engines, and the typically small useful life, and the fact that these engines are already
subject to Tier 2 regulations, the environmental reductions attributable to such engines would be
relatively small. The Office of Advocacy believes that, based on the available information, the
Agency does not have a sufficient basis to move forward with a proposal that would require
nonroad engines under 50 kilowatts to use aftertreatment devices.
Based on the SERs' concerns about the technical feasibility of the Tier 4 standards, and the
technical information discussed in the Panel report, SBA recommended that we include a
11-11
-------
Draft Regulatory Impact Analysis
technological review of the standards in the 2008 timeframe in the rulemaking proposal. The
Panel recommended that we consider this recommendation.
11.8.1.4.2 EPA 's Observations
SB A Office of Advocacy stated that considerable concern has been raised regarding the
technical feasibility of PM and NOx aftertreatment devices, particularly in the case of NOx
adsorbers. As explained in the preamble, we have found no factual basis for this statement with
respect to PM controls based on use of advanced aftertreatment for engines between 25 and 75
hp. We are not proposing standards based on performance of advanced aftertreatment for
engines under 25 hp, and for NOx, for engines 75 hp and under.
With respect to the PM standards for these engines, however, EPA disagrees with the
statement made by the Office of Advocacy that, based on available information, we do not have a
sufficient basis to move forward with this proposed rulemaking requiring nonroad engines under
50 kW to use aftertreatment devices. As we have documented in the preamble and elsewhere in
this Draft RIA, EPA believes that the standards for PM for engines in these power ranges are
feasible at reasonable cost, and will help to improve very important air quality problems,
especially by reducing exposure to diesel PM and by aiding in attainment of the PM 2.5 National
Ambient Air Quality Standards (NAAQS). Indeed, given these facts, EPA is skeptical that an
alternative of no PM standards for these engines would be appropriate under section 213 (a) (4).
Moreover, the statement regarding cost impacts fails to account for transition flexibilities
provided all equipment manufacturers as part of the proposal.
11.8.2 Nonroad Diesel Equipment Manufacturers
11.8.2.1 Flexibility Alternatives for Small Equipment Manufacturers
11.8.2.1.1 SBAR Panel Recommendations
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
potential modifications. The recommended transition provisions for small manufacturers are:
1. Percent-of-Production Allowance: Over a seven model year period, 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 is to be
implemented by power category with the average determined over the period in which the
flexibility is used.
2. Small Volume Allowance: A manufacturer may 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 is limited to one family per power
category. Alternatively, the Panel also recommended, at the manufacturer's choice by hp
11-12
-------
Small-Business Flexibility Analysis
category, a program that eliminates the "single family provision" restriction with revised
total and annual sales limits as shown below:
For categories <175 hp - 525 previous Tier engines (over 7 years) with an annual cap
of 150 units (these engine numbers are separate for each hp category defined in the
regulations).
- For categories of > 175hp - 350 previous Tier engines (over 7 years) with an annual
cap of 100 units (these engine numbers are separate for each hp category defined in
the regulations).
The Panel recommended that we seek comment on the total number of engines and
annual cap values listed above. In contrast to the Tier 2/Tier3 rule promulgated in 1998,
SBA expects the transition to the Tier 4 technology will be more costly and technically
difficult. Therefore, the small equipment manufacturers may need more liberal flexibility
allowances especially for equipment using the lower hp engines. The Panel's
recommended flexibility may 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. Thus, 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 which could be covered
in any hp category.
3. In addition, due to the changing nature of the technology as the manufacturers transition
from Tier 2 to Tier 3 and Tier 4, the Panel recommended that the equipment
manufacturers be permitted to borrow from the Tier3/Tier 4 flexibilities for use in the
Tier 2/Tier 3 time frame.
To maximize the likelihood that the application of these flexibilities will result in the
availability of previous Tier engines for use by the small equipment manufacturers, the Panel
recommended that - similar to the application of flexibility options that are currently in place -
these three flexibilities should be provided to all equipment manufacturers. (See discussion on
transition provisions for all equipment manufacturers in Section Vn.B of the proposed rule
preamble.)
An issue was raised that we establish a provision which would allow manufacturers to
request limited "application specific" alternative standards for equipment configurations which
present unusually challenging technical issues for compliance. The three flexibilities
recommended above would provide latitude, at least in the near term, and a properly structured
emission-credit program for the engine manufacturers would provide long-term latitude. Even if
one were to assume that these flexibilities provide insufficient leeway (which may not be the
case), application specific standards would still be cumbersome for both the small equipment
manufacturers and for the Agency. Nonetheless, the Panel recommended that we seek comment
on the need for and value of special application specific standards for small equipment
manufacturers.
11-13
-------
Draft Regulatory Impact Analysis
11.8.2.1.2 EPA 's Proposed Regulatory Alternatives
We are in fact proposing the Percent-of-Production and Small Volume Allowances for all
equipment manufacturers, and explicitly took the Panel report into account in making that
proposal (see Section VII.B of the preamble). We believe that this proposal should provide the
type of transition leeway recommended by the Panel. We believe that the transition provisions
could allow small equipment manufacturers to postpone any redesign needed on low sales
volume or difficult equipment packages, thus saving both money and strain on limited
engineering staffs. Within limits, small 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.
In regards to the Panel's suggested exemption and annual cap values listed above, we have
requested comment on both of these elements in Section Vn.B of the preamble to the proposed
rule. We have also requested comment on implementing the small volume allowance provision
without the single family limit provision using caps slightly lower than 700 units, with this
provision being applied separately to each engine power category subject to the proposed
standards.
Similar to the discussion in Section VHB of the proposed rule preamble, we are requesting
comment on new proposed 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 that small equipment
manufacturers and importers have reported equipment sales using certified engines in model year
2002 or earlier was sufficient to alleviate this problem. Upon further analysis, EPA 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 US market or more than the small volume allowances identified in the
transition provisions. This also creates a potentially significant disparity between the treatment
of foreign and domestic equipment manufacturers. We did not intend this outcome, and we do
not believe that it is needed to provide reasonable leadtime to foreign equipment manufacturers.
The purpose of these transition provisions is to lessen the burden on small equipment
manufacturers. Therefore, we are requesting comment on the additional requirement that only
the small nonroad diesel equipment manufacturer that is most responsible for the manufacturing
and assembling process, and therefore the burden of complying with the proposed standards,
would qualify for the allowances provided under the small equipment manufacturer transition
provisions. Under this requirement, only a small importer that produces or manufactures
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 complying with the
proposed standard, and therefore would not receive any allowances under these transition
provisions directly, but could import exempt equipment if it is covered by an allowance or
11-14
-------
Small-Business Flexibility Analysis
transition provisions associated with a foreign small equipment manufacturer. We believe that
this requirement transfers the flexibility offered in these transition provisions to the party with
the burden and would allow transition provisions and allowances to be used by foreign small
equipment manufacturers in the same way as domestic small equipment manufacturers, while
avoiding the potential for misuse by importers of unnecessary allowances.
We are also proposing the Panel's recommendation that equipment manufacturers be allowed
to borrow from Tier 4 flexibilities in the Tier2/3 timeframe. See the more extended discussion
on this issue in Section VII.B of the preamble.
With regard to the Panel recommendation for a provision allowing small manufacturers to
request limited "application specific" alternative standards for equipment configurations which
present unusually challenging technical issues for compliance, we have requested comment on
this recommendation (in Section VII.C of the preamble to the proposed rule). We believe that
the need for such a provision has not been established and putting it forth without more
information could provide more lead time than can be justified, and could undermine emission
reductions which are achievable. Moreover, no participant in the SBAR process offered any
empirical support that such a problem even exists. Nor have such issues been demonstrated (or
raised) by equipment manufacturers, small or large, in implementing the current nonroad
standards. Further, we believe that any application-specific difficulties can be accommodated by
the transition provisions the Agency is proposing including ABT. Nonetheless, in keeping with
the SBAR recommendations, we have requested comment on the value of, and need for, special
application specific standards for small equipment manufacturers in the preamble.
11.8.2.2 Hardship Provisions for Small Equipment Manufacturers
11.8.2.2.1 SBAR Panel Recommendations
The Panel also recommended that two types of hardship provisions be extended to small
equipment manufacturers. These provisions are:
1. 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 (i.e., fire, tornado, supplier not fulfilling contract, etc.); and
2. 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 cannot be sought by a manufacturer
which also manufactures the engines for its equipment.
11-15
-------
Draft Regulatory Impact Analysis
11.8.2.2.2 EPA 's Proposed Hardship Provisions
We are proposing that the Panel recommended hardship provisions be extended to small
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 explained earlier (and also in Sections VII.B and VII.C of the
preamble to the proposed rule), this proposal is needed to thwart misuse of these 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.
As explained earlier in Section VII.B of the preamble to the proposed rule, hardship relief
would not be available until other allowances have been exhausted. Either relief provision would
provide additional lead time for small equipment manufacturers for up to two model years based
on the circumstances, but we may require recovery of the lost environmental benefit.
11.8.3 Nonroad Diesel Fuel Refiners
11.8.3.1 Flexibility Alternatives for Small Fuel Refiners
11.8.3.1.1 SBAR Panel Recommendations
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 Panel process, 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.
11-16
-------
Small-Business Flexibility Analysis
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)
Under 1-
Step
Program
Non-Small b
Small
Under 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 non-small refiners for 2008, and for small refiners for
2012 and later.
0 Assumes 500 ppm standard for marine + locomotive fuel for non-small refiners for 2007, and for small refiners for
2010 and later.
11.8.3.1.2 EPA 's Proposed Regulatory Alternatives
We have continued to consider the issues raised during the SBREFA process and have
decided to propose each of the flexibility provisions recommended by the Panel. Because we are
proposing in this rule a two-step approach to fuel implementation, we are thus proposing the
small refiner relief provisions as recommended by the Panel for a two-step program, which are
shown in Table 11-4 below.
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 non-small refiners for 2007 and later and for small refiners
for 2010 and later.
11-17
-------
Draft Regulatory Impact Analysis
Generally, we have structured these proposed provisions to address small refiner hardship
while expeditiously achieving air quality benefits and ensuring that the availability of 15 ppm
nonroad diesel fuel would coincide with the introduction of 2011 model year nonroad diesel
engines and equipment. The following paragraphs review the reasons we believe that the special
provisions for small refiners recommended by the Panel are necessary and appropriate.
First, the proposed compliance schedule for the nonroad diesel program, combined with
flexibility for small refiners, would achieve the air quality benefits of the 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. Therefore, they can benefit
from additional time to accumulate capital internally or to secure capital financing from lenders.
Second, 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. Thus, we believe that allowing
small refiners some additional time for newer technologies to be proven out by other refiners
would have the added benefit of reducing the risks faced by small refiners. The added time
would likely allow for lower costs of these improvements in desulfurization technology (e.g.,
better catalyst technology or lower-pressure hydrotreater technology). This would help to offset
the disproportionate financial burden facing small refiners.
Third, providing small refiners more time to comply would increase the availability of
engineering and construction resources. Most refiners would need to install additional
processing equipment to meet the nonroad diesel sulfur requirements. We anticipate that there
may be significant competition for technology services, engineering resources, and construction
management and labor. In addition, vendors will be more likely to contract their services with
the larger refiners first, as their projects will offer larger profits for the vendors. Temporarily
delaying compliance for small refiners would spread out the demand for these resources and
probably reduce any cost premiums caused by limited supply.
We have also requested comment on a slightly different compliance schedule that would
require 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.
11-18
-------
Small-Business Flexibility Analysis
11.8.3.2 Small Refiner Incentives for Early Compliance
11.8.3.2.1 SBAR Panel Recommendations
The SBAR 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:
1. 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.
2. 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. However, the Panel recommended that the maximum per-gallon sulfur
cap for any small refiner remain at 450 ppm.
11.8.3.2.2 EPA 's Proposal
We agree with the Panel recommendation of encouraging early compliance with the
standards. Some small refiners have indicated that they might find it necessary to produce fuel
meeting the nonroad diesel sulfur standards earlier than they would be required to under the
small refiner program described above, for a variety of reasons: some small refiners could find
that their distribution systems limit the number of grades of diesel fuel that will be carried; others
might find it 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 could 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). Given these situations, we are
proposing that small refiners be able to choose between the two mutually exclusive options, as
recommended by the Panel, to provide incentives for early compliance.
More specifically, with the first option a small refiner could generate NRLM diesel sulfur
credits for production of 500 ppm NRLM diesel fuel prior to June 1, 2010, and for production of
15 ppm nonroad fuel from June 1, 2010 through May 31, 2012. The specifics of the overall
credit program, including how they would be applicable to small refiners, are described in
Section IV of the proposed rulemaking preamble.
A refiner that qualifies for the second option could receive a modest revision in its interim
small refiner gasoline sulfur standards, starting January 1, 2004. Specifically, the applicable
small refiner annual average and per-gallon cap gasoline standards would be revised upward by
11-19
-------
Draft Regulatory Impact Analysis
20 percent for the duration of the small refiner gasoline sulfur interim 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))). However, as recommended by
the Panel, in no case could the per-gallon cap exceed 450 ppm, the highest level allowed under
the gasoline sulfur program.
We believe it is very important to link any such temporary relaxation of a small refiner
gasoline sulfur interim sulfur standards with environmental benefit of early desulfurization of a
significant volume of nonroad diesel fuel. Thus, we are proposing that a small refiner wishing to
use the second option produce a minimum volume of nonroad diesel fuel at 15 ppm by June 1,
2006. Each participating small refiner would need to produce a volume of 15 ppm fuel that was
at least 85 percent of the volume represented by its non-highway distillate baseline percentage. Jf
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 minimum volume requirement, the original small refiner interim gasoline sulfur standard
applicable to that refiner would be reinstated. In addition, the refiner would need to compensate
for the higher gasoline levels that it had enjoyed by purchasing gasoline sulfur credits or
producing an equivalent volume of gasoline below the required sulfur levels. These
compensation provisions are discussed further in Section VIJJ of the preamble. Under this
option, a small refiner could in effect shift some funds from its gasoline sulfur program to
accelerate desulfurization of nonroad diesel fuel. Given the environmental benefit that would
result from the production of 15 ppm fuel earlier than necessary, and the small potential loss of
emission reduction under the gasoline sulfur program from fuel produced by the very few small
refiners that we believe would qualify under this second option, we believe the environmental
impact of this option would be neutral or positive.
11.8.3.3 Hardship Provisions for Small Fuel Refiners
11.8.3.3.1 SBAR Panel Recommendations
The Panel recommended that we propose refiner hardship provisions modeled after those
established under the gasoline sulfur and highway diesel fuel sulfur program (see 40 CFR 80.270
and 80.560). Specifically, the Panel recommended that we propose a process that, like the
hardship provisions of the gasoline and highway diesel rules, allows refiners to seek case-by-case
approval of applications for temporary waivers to the nonroad diesel sulfur standards, based on a
demonstration to the Agency of extreme hardship circumstances. This provision would allow
domestic and foreign refiners, including small refiners, to request additional flexibility based on a
showing of unusual circumstances that result in extreme hardship and significantly affect the
ability of the refiner to comply by the applicable date, despite its best efforts.
11-20
-------
Small-Business Flexibility Analysis
11.8.3.3.2 EPA 's Proposed Hardship Provisions
We believe that providing short-term relief to those refiners that need additional time because
they face hardship circumstances facilitates adoption of an overall program that reduces NRLM
diesel fuel sulfur to 500 ppm beginning in 2007, and nonroad diesel fuel sulfur to 15 ppm in
2010, for the majority of the industry.
11.8.4 Nonroad Diesel Fuel Distributors and Marketers
The diesel fuel approach being considered by the Agency includes the possibility of there
being two grades of nonroad diesel fuel (500/15 ppm) in the market place for at least a transition
period. The distributors support a one-step approach because it has no significant impact on their
operations. The distributors offered some suggestions on how they might deal with this issue,
but indicated that there would be adverse impacts in some circumstances. The Panel
recommended that we study this issue further. Chapter 7 of this Draft Regulatory Impact
Analysis further discusses costs and related issues relevant to fuel distributors under our
proposed program.
We have designed the proposed fuel sulfur program 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
fuel program, it is not possible for us to provide special provisions for particular (i.e., small) fuel
distributors to limit the potential impact of the proposed rule. The benefits of the proposed low
sulfur diesel program can only be achieved if the volume of diesel fuel consumed by NRLM
engines is matched by the production and distribution of at least the same volume of diesel fuel
produced to the appropriate low sulfur levels. The proposed program must also ensure sufficient
availability of 15 ppm diesel fuel for use in nonroad engines in 2010 and not compromise the
availability needs for 15 ppm diesel fuel for use in highway diesel engines under the highway
diesel program, which begins in 2006. Thus, the low sulfur diesel fuel that we are proposing that
refiners produce would need to be carried through the fuel distribution system to the end-user.
In order to allow for a smooth and orderly transition of diesel fuel in the distribution system
to 15 ppm, we are proposing that parties downstream of the refineries be allowed a small amount
of additional time to turnover their tanks to 15 ppm. We are proposing that at the terminal level,
nonroad diesel fuel would be required to meet the 15 ppm sulfur 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 sulfur standard by September 1, 2010. The
proposed transition schedule for compliance with the 15 ppm standard at refineries, terminals,
and secondary distributors are the same as those allowed under the recently promulgated highway
diesel fuel program.
Further, to avoid the costs associated with segregating 500 ppm NRLM diesel fuel from 500
ppm highway fuel, we are proposing that the existing requirement that NRLM diesel fuel be dyed
11-21
-------
Draft Regulatory Impact Analysis
leaving the refinery would need to be made voluntary. This is discussed in Section 11.7 of this
Draft RIA.
11-22
-------
Small-Business Flexibility Analysis
Chapter 11 References
1. Final Panel Report of the Small Business Advocacy Review Panel on EPA's Proposed Rule-
Control of Emission of Air Pollution From Land-Based Nonroad Compression Ignition Engines,
December 23, 2002.
11-23
-------
CHAPTER 12: Regulatory Alternatives
12.1 Range of Options Considered 12-1
12.1.1 One-Step Options 12-1
12.1.2 Two-Step Options 12-6
12.2 Emission Inventory Impacts Comparison 12-18
12.2.1 Assumptions Regarding Fuel Sulfur Content 12-19
12.2.1.1 Certification Fuel 12-19
12.2.1.2 In-Use Fuel 12-20
12.2.2 Emission Inventories for Alternative Program Options 12-23
12.2.2.1 NOx 12-24
12.2.2.2 PM 12-25
12.2.2.3 NMHC 12-28
12.2.2.4 CO 12-30
12.2.2.5 SO2 12-32
12.2.3 Cumulative Emission Reductions for Alternative Program Options 12-35
12.3 Benefits Comparison 12-37
12.4 Cost Analysis for Alternative Options 12-53
12.4.1 One Step Options 12-53
12.4.1.1 Option 1 12-53
12.4.1.2 Option la 12-59
12.4.1.3 Option Ib 12-59
12.4.2 Two Step Options 12-60
12.4.2.1 The Proposal 12-60
12.4.2.2 Option 2a 12-60
12.4.2.3 Option 2b 12-60
12.4.2.4 Option 2c 12-65
12.4.2.5 Option 2d 12-68
12.4.2.6 Option 2e 12-68
12.4.3 Other Options 12-69
12.4.3.1 Options 12-69
12.4.3.2 Option 4 12-70
12.4.3.3 Option 5a 12-71
12.4.3.4 Option 5b 12-73
12.5 Costs per Ton 12-74
12.5.3 Incremental Cost per Ton for Option 2c 12-77
12.5.4 Incremental Cost per Ton for Option 2e 12-77
12.5.5 Incremental Cost per Ton for Option 3 12-78
12.5.6 Incremental Cost per Ton for Option 4 12-78
12.5.7 Incremental Cost per Ton for Option 5a 12-79
12.5.8 Incremental Cost per Ton for Option 5b 12-80
12.6 Summary and Assessment of Alternative Program Options 12-81
12.6.1 Summary of Results of Options Analysis 12-81
12.6.2 Discussion of Rationale, Issues, and Feasibility Assessment of Options 12-84
12.6.2.1 One-Step Options 12-84
12.6.2.2 Two-Step Options 12-89
Appendix 12A: Certification Fuel Sulfur Levels 12-105
Appendix 12B: Incremental Cost, Emission Reductions, Benefits, and Cost Effectiveness 12-113
-------
Regulatory Alternatives
CHAPTER 12: Regulatory Alternatives
Our proposed 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 and discuss the alternative program
options that we evaluated in order to make this determination. These alternatives are cast as
twelve specific Program Options.
For each Option, we first present a full description of the level and timing of fuel and engine
standards. We then present the emissions inventory impacts associated with each Option in
comparison to our proposed program, as well as the monetized health and welfare benefits, costs,
and cost-effectiveness. Finally, we present our assessment of the rationale, feasibility, and issues
associated with each Option in light of the analyses we conducted.
12.1 Range of Options Considered
Our proposed emission control program consists 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 have 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 only a description of what the program options are. Subsequent
sections present the inventory impacts, benefits, costs, and cost-effectiveness. Finally, Section
12.6 summarizes the rationale for each option and our evaluation of the issues and feasibility
associated with the options.
12.1.1 One-Step Options
One-step options are those in which the fuel sulfur standard is applied in a single step; there
are no phase-ins or step changes. In all one-step options, the transient test cycle is required
concurrently with the introduction of the transitional Tier 4 engine standards in any horsepower
group.
Option la differs from Options 1 and Ib in terms of the engine standards and their associated
timing. Because so much time was needed to produce benefits estimates, EPA decided early in
the program development process to use this option as the basis of our benefits analysis (although
EPA ultimately determined not to propose this option). Option Ib differs from Option 1 only in
12-1
-------
Draft Regulatory Impact Analysis
the timing of the fuel sulfur standard, and is 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.
The one-step options are summarized in Table 12.1.1-1. Following this table is a summary of
the existing Tier 1, Tier 2, and Tier 3 standards from 40 CFR §89.112 that form the baseline of
our analyses. The specifics of the three one-step options are shown in the standard charts in
Figures 12.1.1-2, 3, and 4. 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 12.1.1-1
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 2013; no NOx aftertreatment
• >75 hp: PM aftertreatment-based standards phasing in beginning
in 2009; NOx aftertreatment-based standards phasing in beginning
in 2011
See Figure 12.1.1-2
• Fuel sulfur < 15ppm in June 2008
• PM aftertreatment-based standards introduced in 2009-10
• NOx aftertreatment-based standards introduced in 2011-12
See Figure 12.1.1-3
Same as Option la, except fuel sulfur standard required two years
earlier
See Figure 12.1.1-4
12-2
-------
Regulatory Alternatives
Figure 12.1.1-1
Existing Engine and Fuel Standards
hp group
2005 2006 2007 2008 | 2009 2010 | 2011 2012 | 2013 2014 | 2015
Nonroad engine standards (g/bhp-hr)"
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.
12-3
-------
Draft Regulatory Impact Analysis
Figure 12.1.1-2
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)"
hp<25
25 < hp
hp<50
50 750
Tierl
Tier 2
TierS
0.30
0.22
50%: 0
Tier 2
PM
PM
01PM
0.02PM, 3. 3T NOx
0.01 PM
50%: 0.30 NOx
50%: 0.0 1PM, 0.30 NOx
0.30 NOx
Fuel sulfur standard (ppm)p
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.
T 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.
12-4
-------
Regulatory Alternatives
Figure 12.1.1-3
Engine and Fuel Standards Under Option la
hp group
2005
2006
2007
2008
2009
2010
2011
2012 2013 2014 2015
Nonroad engine standards (g/bhp-hr)"
hp<25
25 < hp
hp<50
50 750
Tierl
Tier 2
TierS
Tier 2
0.01
PM
0.30 NOx
Fuel sulfur standard (ppm)15
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.
12-5
-------
Draft Regulatory Impact Analysis
Figure U.I.1-4
Engine and Fuel Standards Under Option Ib
hp group
2005
2006
2007
2008
2009
2010
2011
2012 2013 2014 2015
Nonroad engine standards (g/bhp-hr)"
hp<25
25 < hp
hp<50
50 750
Tierl
Tier 2
TierS
Tier 2
0.01
PM
0.30 NOx
Fuel sulfur standard (ppm)15
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.
12.1.2 Two-Step Options
Two-step options are those in which the fuel sulfur standard is set first at SOOppm for several
years, and then is lowered further to 15ppm. The exact timing of the introduction of the SOOppm
and the 15ppm standards varies 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 is
required concurrently with the introduction of the transitional Tier 4 engine standards. The one
exception is Option 5b, under which the existing steady-state test applies indefinitely for <75 hp
engines.
Our proposed program forms the basis for all of the two-step program options. The two-step
options are summarized in Table 12.1.2-1. Following this table is a summary of the existing Tier
1, Tier 2, and Tier 3 standards from 40 CFR §89.112 that form the baseline of our analyses. The
specifics of the two-step options are shown in the standard charts in Figures 12.1.2-2 through 11.
12-6
-------
Regulatory Alternatives
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 12.1.2-1
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)
See Figure 12.1.2-2
Option 2a
Same as our proposed program, except:
• Transitional sulfur standard of 500 ppm is introduced one year earlier
See Figure 12.1.2-3
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 12.1.2-4
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 12.1.2-5
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 12.1.2-6
Option 2e
Same as our proposed program, except:
• No new Tier 4 NOx standards.
See Figure 12.1.2-7
Option 3
Same as our proposed program, except:
• Above-ground mining equipment >750 hp remains at the Tier 2 standards
See Figure 12.1.2-8
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 12.1.2-9
Option 5 a
Same as our proposed program, except:
• No new Tier 4 standards for <75 hp engines
See Figure 12.1.2-10
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 12.1.2-11
12-7
-------
Draft Regulatory Impact Analysis
Figure 12.1.2-1
Existing Engine and Fuel Standards
hp group
2005 2006 2007 2008 | 2009 2010 | 2011 2012 | 2013 2014 | 2015
Nonroad engine standards (g/bhp-hr)"
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.
12-8
-------
Regulatory Alternatives
Figure 12.1.2-2
Engine and Fuel Standards under the Proposed Program
hp group
2005
2006
2007
2008
2009
2010
2011
2012
2013
2014 2015
Nonroad engine standards (g/bhp-hr)"
hp<25
25 < hp
hp<50
50 750
Tierl
Tier 2
0.30 PM
0.22 PM
TierS
Tier 2
0.02PM, 3. 3e NOx
100%T : 0.01 PM
50%Y: 0.30 NOx
50%6 : 0.01PM, 0.30 NOx
0.01PM
0.30 NOx
Fuel sulfur standard (ppm)15
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.
T 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.
6 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
-------
Draft Regulatory Impact Analysis
Figure 12.1.2-3
Engine and Fuel Standards under Option 2a
hp group
2005
2006
2007
2008
2009
2010
2011
2012
2013
2014 2015
Nonroad engine standards (g/bhp-hr)"
hp<25
25 < hp
hp<50
50 750
Tierl
Tier 2
0.30 PM
0.22 PM
TierS
Tier 2
50%6 : 0
0.02PM, 3. 3e NOx
100%T : 0.01 PM
50%Y: 0.30 NOx
01 PM, 0.
30 NOx
0.01PM
0.30 NOx
Fuel sulfur standard (ppm)15
Loco &
marine
Nonroad
Uncon-
trolled
Uncon-
trolled
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.
T 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.
6 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 U.I.2-4
Engine and Fuel Standards under Option 2b
hp group
2005
2006
2007
2008
2009
2010
2011
2012
2013
2014 2015
Nonroad engine standards (g/bhp-hr)"
hp<25
25 < hp
hp<50
50 750
Tierl
Tier 2
0.30 PM
0.22
TierS
Tier 2
PM
0.01
PM
50%:
0.01
PM
0.01
PM
0.02
PM
0.02PM, 3. 3e NOx
50%Y: 0.30 NOx
50%6 : 0.01 PM,
0.30 NOx
100%:
0.01
PM
0.01PM
0.30 NOx
Fuel sulfur standard (ppm)15
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.
T 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.
6 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
-------
Draft Regulatory Impact Analysis
Figure 12.1.2-5
Engine and Fuel Standards under Option 2c
hp group
2005
2006
2007
2008
2009
2010
2011
2012
2013
2014 2015
Nonroad engine standards (g/bhp-hr)"
hp<25
25 < hp
hp<50
50 750
Tierl
Tier 2
0.30 PM
0.22 PM
TierS
Tier 2
0.01
PM
0.02PM, 3. 3e NOx
100%T : 0.01 PM
50%Y: 0.30 NOx
50%6 : 0.01PM, 0.30 NOx
0.01PM
0.30 NOx
Fuel sulfur standard (ppm)15
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.
T 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.
6 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 12.1.2-6
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)"
hp<25
25 < hp
hp<50
50 750
Tierl
Tier 2
0.30PM
0.22 PM
TierS
Tier 2
100%T :
50%Y :
NOx
50%6 : 0.01 PM,
NOx
0.30
0.02 PM NOx
50%: 0.30 NOx
0.01PM
0.30
0.30
0.01PM
0.30 NOx
Fuel sulfur standard (ppm)15
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.
T 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.
6 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
-------
Draft Regulatory Impact Analysis
Figure 12.1.2-7
Engine and Fuel Standards under Option 2e
hp group
2005
2006
2007
2008 2009 2010 2011 2012 2013 2014 2015
Nonroad engine standards (g/bhp-hr)"
hp<25
25 < hp
hp<50
50 750
Tierl
Tier 2
0.30 PM
0.22 PM 0.02 PM
Tier3 0.01PM
Tier 2 50%6 : 0.01 PM 0.01 PM
Fuel sulfur standard (ppm)15
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.
6 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 12.1.2-8
Engine and Fuel Standards under Option 3
hp group
2005
2006
2007
2008
2009
2010
2011
2012
2013
2014 2015
Nonroad engine standards (g/bhp-hr)"
hp<25
25 < hp
hp<50
50 750
Tierl
Tier 2
0.30 PM
0.22 PM
TierS
Tier 2
0.02PM, 3. 3e NOx
100%T : 0.01 PM
50%Y: 0.30 NOx
50%6 : 0.01 PM, 0.
Mining equipment
at Tier 2
30 NOx
remains
0.01PM
0.30 NOx
0.01PM
0.30 NOx
Mining
equipment
at Tier 2
Fuel sulfur standard (ppm)15
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.
T 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.
6 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
-------
Draft Regulatory Impact Analysis
Figure 12.1.2-9
Engine and Fuel Standards under Option 4
hp group
2005
2006
2007
2008
2009
2010
2011
2012
2013
2014 2015
Nonroad engine standards (g/bhp-hr)"
hp<25
25 < hp
hp<50
50 750
Tierl
Tier 2
0.30 PM
0.22 PM
TierS
Tier 2
0.02PM, 3. 3e NOx
100%T : 0.01 PM
50%Y: 0.30 NOx
50%6 : 0.01PM, 0.30 NOx
0.01PM
0.30 NOx
Fuel sulfur standard (ppm)15
Loco &
marine
Nonroad
Uncontrolled
Uncontrolled
500 ppm
500 ppm
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.
T 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.
6 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 12.1.2-10
Engine and Fuel Standards under Option 5a
hp group
2005
2006
2007
2008
2009 | 2010
2011
2012
2013
2014 2015
Nonroad engine standards (g/bhp-hr)"
hp<25
25 < hp
hp<50
50 750
Tierl
Tier 2
TierS
Tier 2
100%T : 0.01 PM
50%Y: 0.30 NOx
50%6 : 0.01 PM, 0.
30 NOx
0.01PM
0.30 NOx
Fuel sulfur standard (ppm)15
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.
T 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.
6 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
-------
Draft Regulatory Impact Analysis
Figure U.I.2-11
Engine and Fuel Standards under Option 5b
hp group
2005
2006
2007
2008
2009
2010
2011
2012
2013
2014 2015
Nonroad engine standards (g/bhp-hr)"
hp<25
25 < hp
hp<50
50 750
Tierl
Tier 2
0.30 PM
0.22 PM
TierS
Tier 2
100%T : 0.01 PM
50%Y: 0.30 NOx
50%6 : 0.01 PM, 0.
30 NOx
0.01PM
0.30 NOx
Fuel sulfur standard (ppm)15
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.
T 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.
6 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 Emission Inventory Impacts Comparison
This section presents the nonroad inventory impacts of all the program options just set forth
that we considered during development of our proposed program. The methodology and
assumptions used to generate the inventories for all program options are the same as those
described in Chapter 3 for the baseline (no new Tier 4 standards) and our proposed program.
The primary differences between the assumptions made for our proposed program versus those
made for the other program options are related to in-use fuel and certification fuel sulfur levels.
These differences are described in Section 12.2.1 below.
The inventories presented in this section represent all nonroad equipment categories, as well
as locomotive and CI marine which are affected by the fuel standards, although not by the engine
12-18
-------
Regulatory Alternatives
standards. We have not included any potential credits generated under ABT. The PM
inventories include directly emitted sulfate PM (in the form of hydrated sulfuric acid) but do not
include secondary sulfates produced from S02 in the atmosphere.
12.2.1 Assumptions Regarding Fuel Sulfur Content
Among the program options we considered, there are variations in the timing and level of the
fuel sulfur standard. These variations impact both the in-use sulfur level and the certification
sulfur level, which in turn affect the PM and S02 inventories estimated via the NONROAD model.
This section presents our approach to in-use and certification fuel sulfur levels.
12.2.1.1 Certification Fuel
Fuel used to certify new nonroad engines should be representative of the fuel that those
engines will use during their lifetime. Thus the specified maximum sulfur content of nonroad
diesel certification fuel should change in concert with the in-use sulfur standard. For instance,
our proposed program includes a SOOppm in-use sulfur standard that goes into effect in June of
2007, followed by a ISppm sulfur standard that goes into effect in June of 2010. Nonroad engine
manufacturers must therefore show that their engines can meet the standards when tested on fuel
with a sulfur level as high as SOOppm during model years 2008 through 2010, and as high as
ISppm for model years 2011 and beyond.
For most program options, the certification fuel sulfur specification will change in the year
following a change in the in-use fuel sulfur standard. However, we took a different approach for
Options Ib and 2a. Both of these options are intended to show the impact that an earlier change
from uncontrolled to controlled in-use sulfur levels will have on the PM inventories. In order to
generate the full benefits of these options, our modeling does not include a concurrent change to
certification fuel sulfur levels. In other words, we model an in-use reduction in sulfate PM and
SO2 emissions as a result of the in-use fuel having less sulfur than the certification fuel. If the
certification fuel were set at a sulfur level equal to the in-use fuel sulfur level, there would be no
in-use reduction in sulfate PM or SO2 emissions.
A lower maximum sulfur specification for certification fuel makes it easier to comply with
the PM standard, since, as shown in Chapter 4 of this draft RIA, lower fuel sulfur means less
sulfate PM. Manufacturers could take advantage of this benefit of lower sulfur content in
certification fuel by modifying their engines to reduce costs. However, if the change in
certification fuel sulfur level does not exactly coincide with a change in the applicable engine
emission standards, making modifications to an engine family simply to take advantage of the
lower sulfur level of certification fuel may not be cost-effective. Therefore, we have made the
assumption that engines within any horsepower group will only be modified to account for a
lower certification fuel sulfur level when new engine standards become effective. In other
words, for modeling purposes, all engines are assumed to be certified at the sulfur level that
applied when the most recent set of emission standards became effective. This approach results
12-19
-------
Draft Regulatory Impact Analysis
in slightly larger in-use PM benefits, since there will be occasions when manufacturers are in
effect meeting the PM standard using certification fuel with a higher-than-necessary sulfur level.
The assumed cert fuel sulfur levels were used to establish the proper zero-hour emission
factors for new engines. For in-use inventory impacts of these new engines, the emission factors
were further adjusted to account for the assumed in-use sulfur levels. Thus, for instance, engines
certified on 2000ppm sulfur fuel and then operated on SOOppm fuel would realize a PM benefit
relative to the PM certification standard.
The sulfur levels assumed for certification fuel for the purposes of establishing the zero-hour
emission factors are given in Appendix 12A.
12.2.1.2 In-Use Fuel
Section 12.1 presented the sulfur standards that would apply to in-use nonroad fuel under
each of the program options we evaluated. In order to calculate emission inventories using the
NONROAD model, we estimated the likely in-use average sulfur level by calendar year for each
of the options. These average sulfur values were a function of the level and timing of transitional
and final standards, expected refiner compliance margins, and the amount of highway diesel fuel
which is consumed by nonroad engines (so-called "spillover"). The various factors used in the
calculations are listed in Table 12.2.1.2-1, based on the derivations and discussion presented in
Section 7.1.4.2.
Table 12.2.1.2-1
Factors Used to Calculate In-use Sulfur Levels
Average in-use fuel sulfur level for any fuel designed to meet a
standard of 500 ppm
Average in-use fuel sulfur level for fuel designed to meet California's
diesel fuel specifications
Average in-use fuel sulfur level for any fuel designed to meet a
standard of 1 5 ppm
Average in-use sulfur level for fuel intended to be used in nonroad
engines, prior to sulfur control
Nonroad spillover: Fraction of fuel consumed by nonroad engines
which is actually designed to meet on-highway fuel sulfur standards
Locomotive/marine spillover: Fraction of fuel consumed by
locomotives and marine engines which is actually designed to meet on-
highway fuel sulfur standards
340 ppm
120 ppm
11 ppm
3400 ppm
34.9%
32.4%
We first determined the average in-use sulfur level for highway fuel by calender year, using
the factors in Table 12.2.1.2-1 and the phase-in schedule adopted in 2001 (66 FR 5002, January
18, 2001). Table 12.2.1.2-2 presents these sulfur levels.
12-20
-------
Regulatory Alternatives
Table 12.2.1.2-2
Average Sulfur Level for On-highway Fuel
Year
<2005
2006
2007
2008
2009
>2010
Average sulfur (ppm)
300
165
69
69
69
11
Explanation
Nationwide average, including California, prior to introduction of 15ppm
standard. Assumes 10% of nationwide highway diesel meets California's
requirements.
15ppm standard applies beginning in June. Only 80% of the pool meets
the 1 5ppm standard.
Only 80% of the pool meets the 1 5ppm standard.
Only 80% of the pool meets the 1 5ppm standard.
Only 80% of the pool meets the 1 5ppm standard.
100% of the pool meets the 15ppm standard
We then determined the average in-use sulfur level for off-highway fuel. All of the program
options we evaluated include one or more of the following types of transitions, for either nonroad
fuel or locomotive and marine fuel:
• Transition from uncontrolled sulfur levels to a SOOppm standard
• Transition from a SOOppm sulfur standard to a ISppm standard
• Transition from uncontrolled sulfur levels to a ISppm standard
Every one of these transitions is assumed to occur in June, regardless of the calendar year in
which the new standard applies. Using the average sulfur levels presented in Table 12.2.1.2-1,
we generated in-use average sulfur levels for off-highway diesel fuel for the three types of
transitions shown above. Table 12.2.1.2-3 presents the results.
Table 12.2.1.2-3
Average Sulfur Levels for Off-highway Fuel Sulfur Standard Transitions (ppm)
Prior to transition year
Transition year
After transition year
Uncontrolled to
SOOppm standard
3400
1615
340
SOOppm standard to
1 5ppm standard
340
148
11
Uncontrolled to ISppm
standard
3400
1423
11
Finally, to calculate the in-use average sulfur levels under the various program options we
evaluated, we combined the average sulfur levels for on-highway fuel from Table 12.2.1.2-2 with
the average sulfur levels for off-highway fuel from Table 12.2.1.2-3. The spillover fractions
given in Table 12.2.1.2-1 were used to properly weight the on-highway and off-highway average
12-21
-------
Draft Regulatory Impact Analysis
sulfur levels. The results for all program options are given in Tables 12.2.1.2-4 and 12.2.1.2-5,
based on the fuel sulfur standards associated with each option as described in Section 12.1.
Table 12.2.1.2-4
In-use Average Sulfur Levels Used for Modeling Nonroad Engines (ppm)
Baseline
Proposed
program
Option 1
Option la
Option Ib
Option 2a
Option 2b
Option 2c
Option 2d
Option 2e
Option 3
Option 4
Option 5 a
Option 5b
<2005
2318
2318
2318
2318
2318
2318
2318
2318
2318
2318
2318
2318
2318
2318
2006
2271
2271
2271
2271
984
1109
2271
2271
2271
2271
2271
2271
2271
2271
2007
2237
1075
2237
2237
31
245
1075
1075
1075
1075
1075
1075
1075
1075
2008
2237
245
950
950
31
245
245
245
245
245
245
245
245
245
2009
2237
245
31
31
31
245
120
120
245
245
245
245
245
245
2010
2217
100
11
11
11
100
11
11
100
100
100
100
100
100
>2011
2217
11
11
11
11
11
11
11
11
11
11
11
11
11
12-22
-------
Regulatory Alternatives
Table 12.2.1.2-5
In-use Average Sulfur Levels Used for Modeling Locomotive and Marine Engines (ppm)
Baseline
Proposed
program
Option 1
Option la
Option Ib
Option 2a
Option 2b
Option 2c
Option 2d
Option 2e
Option 3
Option 4
Option 5 a
Option 5b
<2005
2396
2396
2396
2396
2396
2396
2396
2396
2396
2396
2396
2396
2396
2396
2006
2352
2352
2352
2352
1016
1145
2352
2352
2352
2352
2352
2352
2352
2352
2007
2321
1114
2321
2321
30
252
1114
1114
1114
1114
1114
1114
1114
1114
2008
2321
252
1114
984
30
252
252
252
252
252
252
252
252
252
2009
2321
252
252
30
30
252
252
252
252
252
252
252
252
252
2010
2302
233
233
11
11
233
233
233
233
233
233
104
233
233
>2011
2302
233
233
11
11
233
233
233
233
233
233
11
233
233
12.2.2 Emission Inventories for Alternative Program Options
This section presents the absolute inventories associated with our proposed program and each
of the program options we evaluated, in short tons per year. All inventories represent only those
off-highway engines affected by our proposed program or each of the alternative program options
- no on-highway, biogenic, or other sources are included. We have presented graphical
illustrations separately for nonroad and locomotive/marine, since we are proposing engine
standards only for the former, and have investigated fuel sulfur standards for locomotives and
marine engines as a way to generate additional PM and SO2 reductions. In addition, there are no
changes to NOx, NMHC, or CO for locomotive and marine under any Option, so we have not
shown separate graphs for these pollutants. Inventory tables include nonroad, locomotive, and
marine sources for PM and SO2, and just nonroad sources for NOx, CO, and NMHC.
Graphic representations of inventories are shown for all years through 2030, and tabulated
values are provided for selected years. All values are presented as 50-state annual tons, and the
particulate matter values are PM10. Note that the emission reductions used for the calculation of
health and welfare benefits were based on 48-state inventories and the relevant particulate matter
12-23
-------
Draft Regulatory Impact Analysis
was PM2.5 due to the fact that the air quality modeling on which these benefits were based
requires the use of these alternative measures of inventory impacts.
12.2.2.1 NOx
This section presents the NOx inventories for nonroad engines affected by our proposed
program and the alternative program options. In general, the options represent little or no change
in the NOx standard levels and timing in comparison with our proposed program. Primary
differences are exhibited for:
• Options la and Ib for which NOx aftertreatment is required for all engines
• Option 2d which adds NOx aftertreatment-based standards for 25-75hp
• Option 2e which assumes no new Tier 4 NOx standards
• Option 3 which exempts large above-ground mining equipment
Figure 12.2.2.1-1
50-State Inventories for nonroad NOx (tons)
1,800,000
1,600,000
1,400,000
1,200,000
1,000,000
800,000
600,000
400,000
200,000
2000
2010
2020
2030
• Baseline
• Proposal
Option 1
Option 1a
• Option 1b
• Option 2a
• Option 2b
• Option 2c
Option 2d
• Option 2e
Option 3
• Option 4
Option 5a
Option 5b
12-24
-------
Regulatory Alternatives
Table 12.2.2.1-1
50-State Inventories for NOx (tons)
Baseline
Proposed
program
Option 1
Option la
Option Ib
Option 2a
Option 2b
Option 2c
Option 2d
Option 2e
Option 3
Option 4
Option 5 a
Option 5b
2010
1,327,000
1,326,000
1,325,000
1,327,000
1,327,000
1,326,000
1,325,000
1,324,000
1,326,000
1,327,000
1,326,000
1,326,000
1,327,000
1,327,000
2015
1,205,000
987,000
986,000
853,000
853,000
987,000
984,000
985,000
974,000
1,205,000
1,020,000
987,000
1,000,000
1,000,000
2020
1,182,000
675,000
675,000
514,000
514,000
675,000
674,000
674,000
605,000
1,182,000
747,000
675,000
703,000
703,000
2025
1,218,000
520,000
520,000
343,000
343,000
520,000
519,000
519,000
411,000
1,218,000
612,000
520,000
555,000
555,000
2030
1,280,000
454,000
454,000
265,000
265,000
454,000
453,000
453,000
320,000
1,280,000
557,000
454,000
495,000
495,000
12.2.2.2 PM
Particulate matter directly affected by our proposed program is included in these inventories.
Although the majority of diesel exhaust PM is fine (<2.5 microns), we have included all PM up
to 10 microns in our inventory estimates to most properly account for the full impacts of our
proposed program. In terms of PM inventory impacts, differences between each of the
alternative program options and our proposed program are exhibited for most of the program
options.
12-25
-------
Draft Regulatory Impact Analysis
Figure 12.2.2.2-1
50-State Inventories for nonroad PM (tons)
200,000
180,000
160,000
140,000
120,000 -
100,000
80,000
60,000
40,000
20,000
2000
2010
2020
2030
-Baseline
• Proposal
Option 1
Option 1a
-Option 1b
• Option 2a
-Option 2b
• Option 2c
Option 2d
-Option 2e
Option 3
-Option 4
Option 5a
Option 5b
12-26
-------
Regulatory Alternatives
Figure 12.2.2.2-2
50-State Inventories for loco/marine PM (tons)
80,000 -
70,000 -
,.
/-* r\ r\f\r\
60,000 -
50,000 -
40,000
30,000 -
20,000
10,000
*m*™ *^^**^ ^^
^M£S^^&2&8&2&^
• Baseline
> — •— Proposal
Option 1
—x- Option 1a
— a— Option 1b
• Option 2a
— i — Option 2b
— —Option 2c
— Option 2d
x Option 2e
—•—Option 3
—A— Option 4
Option 5a
i i
Option 5b
2000 2010 2020 2030 P
12-27
-------
Draft Regulatory Impact Analysis
Table 12.2.2.2-1
50-State Inventories for total PM (tons)
Baseline
Proposed
program
Option 1
Option la
Option Ib
Option 2a
Option 2b
Option 2c
Option 2d
Option 2e
Option 3
Option 4
Option 5 a
Option 5b
2010
198,000
174,000
171,000
165,000
165,000
174,000
171,000
171,000
174,000
174,000
174,000
173,000
177,000
175,000
2015
197,000
140,000
137,000
125,000
125,000
140,000
133,000
137,000
140,000
140,000
141,000
139,000
150,000
145,000
2020
202,000
109,000
108,000
98,000
98,000
109,000
105,000
108,000
109,000
109,000
112,000
108,000
127,000
120,000
2025
212,000
92,000
92,000
84,000
84,000
92,000
90,000
92,000
92,000
92,000
96,000
92,000
116,000
107,000
2030
223,000
84,000
83,000
77,000
77,000
84,000
83,000
83,000
84,000
84,000
88,000
83,000
111,000
101,000
12.2.2.3 NMHC
The new Tier 4 standards realize a significant reduction in NMHC emissions, including toxic
hydrocarbons, due to the use of technologies such as oxidation catalysts and catalyzed diesel
particulate filters. NMHC impacts exhibited by each alternative program option will largely
mimic the PM impacts.
The NONROAD model provides total hydrocarbon emissions for both exhaust and crankcase
emissions, though crankcase HC is typically only 1-2% of total HC. Methane and ethane are also
included in total hydrocarbon output from NONROAD. However, our standards apply to non-
methane hydrocarbons. Thus we have decided to convert total hydrocarbons from the
NONROAD model into NMHC. To do this, total hydrocarbons is multiplied by 0.9841, which
subtracts out methane. Note that our air quality modeling requires volatile organic compounds
(VOC) instead of total hydrocarbons, and many of the inventories with which we have compared
the impacts of our proposed and alternative Tier 4 nonroad programs use VOCs. For these
purposes, we converted total hydrocarbons from the NONROAD model into VOC by multiplying
by 1.053, which subtracts out methane and ethane and simultaneously adds aldehydes.
12-28
-------
Regulatory Alternatives
Figure 12.2.2.3-1
50-State Inventories for nonroad NMHC (tons)
200,000
180,000 -
160,000 -
140,000 -
120,000
100,000
80,000
60,000 -
40,000 -
20,000
-:':-,
/••,
"/;-,
,-••,
' •;;-,
::v,.
"V
—•—Baseline
— •— Proposal
Option 1
—x- Option 1a
— a— Option 1b
• Option 2a
— i — Option 2b
j
— —Option 2c
^^Ei-EiAv^A.A.A A A A — uption za
ILJU-iIiIli K
— 3K— Option 2e
—•—Option 3
—A— Option 4
Option 5a
o Option 5b
2000 2010 2020 2030
12-29
-------
Draft Regulatory Impact Analysis
Table 12.2.2.3-1
50-State Inventories forNMHC (tons)
Baseline
Proposed
program
Option 1
Option la
Option Ib
Option 2a
Option 2b
Option 2c
Option 2d
Option 2e
Option 3
Option 4
Option 5 a
Option 5b
2010
122,000
122,000
121,000
119,000
119,000
122,000
121,000
121,000
122,000
122,000
122,000
122,000
122,000
122,000
2015
100,000
92,000
91,000
87,000
87,000
92,000
90,000
92,000
92,000
92,000
93,000
92,000
94,000
94,000
2020
92,000
75,000
75,000
70,000
70,000
75,000
74,000
75,000
75,000
75,000
76,000
75,000
79,000
79,000
2025
91,000
68,000
67,000
63,000
63,000
68,000
67,000
67,000
68,000
68,000
69,000
68,000
73,000
73,000
2030
93,000
65,000
65,000
61,000
61,000
65,000
65,000
65,000
65,000
65,000
66,000
65,000
72,000
72,000
12.2.2.4 CO
The new Tier 4 standards realize a significant reduction in CO emissions due to the use of
technologies such as oxidation catalysts and catalyzed diesel paniculate filters. The minor
adjustment we are proposing for CO standards is more of a bookeeping correction, as explained
in the preamble. CO emissions are assumed to be reduced 90% for engines having a PM trap.
Thus the CO impacts exhibited by each alternative program option will largely mimic the PM
impacts.
12-30
-------
Regulatory Alternatives
Figure 12.2.2.4-1
50-State Inventories for nonroad CO (tons)
1,000,000
900,000
800,000 -
700,000
600,000
500,000
400,000
300,000
200,000
100,000
2000
2010
2020
2030
• Baseline
• Proposal
Option 1
Option 1a
• Option 1b
• Option 2a
• Option 2b
• Option 2c
Option 2d
• Option 2e
Option 3
• Option 4
Option 5a
Option 5b
12-31
-------
Draft Regulatory Impact Analysis
Table 12.2.2.4-1
50-State Inventories for CO (tons)
Baseline
Proposed
program
Option 1
Option la
Option Ib
Option 2a
Option 2b
Option 2c
Option 2d
Option 2e
Option 3
Option 4
Option 5 a
Option 5b
2010
693,000
693,000
672,000
632,000
632,000
693,000
678,000
680,000
693,000
693,000
693,000
693,000
693,000
693,000
2015
682,000
498,000
472,000
397,000
397,000
498,000
457,000
485,000
498,000
498,000
508,000
498,000
533,000
533,000
2020
709,000
326,000
310,000
254,000
254,000
326,000
301,000
318,000
326,000
326,000
348,000
326,000
413,000
413,000
2025
754,000
230,000
220,000
176,000
176,000
230,000
215,000
226,000
230,000
230,000
258,000
230,000
353,000
353,000
2030
805,000
181,000
177,000
141,000
141,000
181,000
174,000
179,000
181,000
181,000
212,000
181,000
332,000
332,000
12.2.2.5 SO2
Generally SO2 emissions are proportional to fuel sulfur content. Thus differences in SO2
inventories between our proposed program and the alternative program options are primarily a
function of the differences in the assumed fuel programs. However, the assumed engine
programs do play a small role, as the sulfur-to-SO2 conversion rate decreases when
aftertreatment-based standards are introduced, from a current conversion rate of approximately
98% to an ultimate conversion rate closer to 70%. Despite this engine-based impact of our
proposed program on SO2 emissions, we believe it is appropriate to associate all reductions in
SO2 with the costs of fuel sulfur reductions, as described in Chapter 8, since the 99% reduction in
in-use nonroad fuel sulfur levels overwhelms any impact caused by changes in the sulfur-to-SO2
conversion rate.
12-32
-------
Regulatory Alternatives
Figure 12.2.2.5-1
50-State Inventories for nonroad SO2 (tons)
2000
2010
2020
2030
• Baseline
• Proposal
Option 1
Option 1a
• Option 1b
• Option 2a
• Option 2b
• Option 2c
Option 2d
• Option 2e
Option 3
• Option 4
Option 5a
Option 5b
12-33
-------
Draft Regulatory Impact Analysis
Figure 12.2.2.5-2
50-State Inventories for loco/marine SO2 (tons)
120,000
100,000
80,000
60,000
40,000 -
20,000
2000
2010
2020
2030
• Baseline
• Proposal
Option 1
Option 1a
• Option 1b
• Option 2a
-Option 2b
• Option 2c
Option 2d
• Option 2e
Option 3
• Option 4
Option 5a
Option 5b
12-34
-------
Regulatory Alternatives
Table 12.2.2.5-1
50-State Inventories for total SO2 (tons)
Baseline
Proposed
program
Option 1
Option la
Option Ib
Option 2a
Option 2b
Option 2c
Option 2d
Option 2e
Option 3
Option 4
Option 5 a
Option 5b
2010
291,000
18,000
10,000
1,000
1,000
18,000
10,000
10,000
18,000
18,000
18,000
13,000
18,000
18,000
2015
318,000
10,000
10,000
1,000
1,000
10,000
10,000
10,000
10,000
10,000
10,000
1,000
10,000
10,000
2020
345,000
10,000
10,000
1,000
1,000
10,000
10,000
10,000
10,000
10,000
10,000
1,000
10,000
10,000
2025
373,000
11,000
11,000
1,000
1,000
11,000
11,000
11,000
11,000
11,000
11,000
2,000
11,000
11,000
2030
401,000
11,000
11,000
2,000
2,000
11,000
11,000
11,000
11,000
11,000
11,000
2,000
11,000
11,000
12.2.3 Cumulative Emission Reductions for Alternative Program Options
Inventory impacts of our proposed program and the alternative program options can be
compared for individual calendar years or cumulatively over some timeframe. For the
cumulative comparison, we have chosen to calculate the net present value of the annual emission
reductions of each program, in comparison to the baseline, for all years through 2030. For this
calculation we used a 3% discount rate to bring all tons into 2004. These net present value
reductions are shown in Table 12.2.3-1. We also present the net present value of the differences
between the emissions through 2030 for each alternative program option and our proposed
program in Table 12.2.3-2.
12-35
-------
Draft Regulatory Impact Analysis
Table 12.2.3-1
50-State Net Present Value Emission Reductions
In Comparison to Existing Standards Through 2030 (tons)
Proposed
program
Option 1
Option la
Option Ib
Option 2a
Option 2b
Option 2c
Option 2d
Option 2e
Option 3
Option 4
Option 5 a
Option 5b
NOx
5,407,000
5,409,000
7,187,000
7,187,000
5,407,000
5,428,000
5,419,000
6,159,000
0
4,665,000
5,407,000
5,118,000
5,118,000
PM
1,126,000
1,133,000
1,255,000
1,296,000
1,145,000
1,180,000
1,147,000
1,126,000
1,126,000
1,097,000
1,135,000
917,000
1,005,000
NMHC
184,000
194,000
248,000
248,000
184,000
199,000
189,000
184,000
184,000
175,000
184,000
141,000
141,000
CO
4,149,000
4,396,000
5,164,000
5,164,000
4,149,000
4,493,000
4,262,000
4,149,000
4,149,000
3,924,000
4,149,000
3,216,000
3,216,000
SO,
4,952,000
4,761,000
4,890,000
5,395,000
5,180,000
4,969,000
4,969,000
4,952,000
4,952,000
4,952,000
5,067,000
4,952,000
4,952,000
12-36
-------
Regulatory Alternatives
Table 12.2.3-2
50-State Net Present Value Emission Differences With Respect
To The Proposed Program, Through 2030 (tons)*
Option 1
Option la
Option Ib
Option 2a
Option 2b
Option 2c
Option 2d
Option 2e
Option 3
Option 4
Option 5 a
Option 5b
NOx
1,000
1,780,000
1,780,000
0
21,000
11,000
751,000
-5,407,000
-742,000
0
-290,000
-290,000
PM
6,000
129,000
170,000
18,000
54,000
20,000
0
0
-30,000
9,000
-209,000
-121,000
NMHC
10,000
63,000
63,000
0
15,000
5,000
0
0
-9,000
0
-44,000
-43,000
CO
248,000
1,015,000
1,015,000
0
344,000
113,000
0
0
-225,000
0
-933,000
-933,000
SO,
-191,000
-63,000
443,000
228,000
17,000
17,000
0
0
0
114,000
0
0
*Positive values indicate that the Option produces greater environmental benefits, i.e. the Option results in a smaller
cumulative absolute inventory
12.3 Benefits Comparison
We are able to estimate the benefits of various options using the benefit-transfer methodology
developed in Chapter 9 for estimating the monetized benefits of the proposal. 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 each option. 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.
The estimated 48 state emission reductions are given in Table 12.3-1, 12.3-2 and 12.3-3.
Table 12.3-4 and Figure 12.3-1 present the estimated benefits for each of the options.
A key question for each of the options is how the benefits of that option compare with the
benefits of our proposed program. Table 12.3-5 lists the difference in benefits between each of
the options and the proposal. These differences are shown graphically in Figure 12.3-2.
12-37
-------
Draft Regulatory Impact Analysis
12-38
-------
Table 12.3-1A
48 State SO2 Emission Reductions for Program Options 1-2
Year Option 1 Option 1a Option 1b Option 2a Option 2b Option 2c Option 2d Option 2e
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
0
0
0
0
0
0
0
0
156,782
273,998
279,259
285,014
290,208
295,325
300,447
305,653
311,085
316,310
321,511
326,735
331,851
337,241
342,638
348,041
353,452
358,871
364,268
369,673
375,086
380,509
385,941
0
0
0
0
0
0
0
0
161,358
281,907
287,243
293,130
298,392
303,562
308,736
314,001
319,522
324,813
330,079
335,371
340,543
346,020
351,505
356,998
362,500
368,010
373,499
378,998
384,506
390,025
395,555
0
0
0
0
0
0
159,106
271,364
276,554
281,907
287,243
293,130
298,392
303,562
308,736
314,001
319,522
324,813
330,079
335,371
340,543
346,020
351,505
356,998
362,500
368,010
373,499
378,998
384,506
390,025
395,555
0
0
0
0
0
0
140,081
245,048
249,746
254,544
270,977
285,003
290,196
295,312
300,434
305,639
311,073
316,299
321,501
326,725
331,840
337,231
342,628
348,032
353,444
358,863
364,260
369,665
375,078
380,500
385,932
0
0
0
0
0
0
0
142,948
249,746
265,904
279,264
285,025
290,223
295,340
300,461
305,665
311,097
316,319
321,519
326,741
331,854
337,243
342,639
348,042
353,452
358,870
364,266
369,670
375,082
380,504
385,935
0
0
0
0
0
0
0
142,948
249,746
265,860
279,232
285,014
290,208
295,323
300,445
305,650
311,083
316,307
321,508
326,732
331,846
337,236
342,633
348,037
353,447
358,866
364,262
369,667
375,080
380,502
385,934
0
0
0
0
0
0
0
142,948
249,746
254,543
270,977
285,003
290,196
295,312
300,434
305,639
311,073
316,299
321,501
326,725
331,840
337,231
342,628
348,032
353,444
358,863
364,260
369,665
375,078
380,500
385,932
0
0
0
0
0
0
0
142,948
249,746
254,543
270,977
285,003
290,196
295,312
300,434
305,639
311,073
316,299
321,501
326,725
331,840
337,231
342,628
348,032
353,444
358,863
364,260
369,665
375,078
380,500
385,932
-------
Table 12.3-1B
48 State SO2 Emission Reductions for Program Options 3-5
| Year Option 3 Option 4 Option 5a Option 5b
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
0
0
0
0
0
0
0
142,948
249,746
254,543
270,977
285,001
290,193
295,308
300,427
305,630
311,061
316,284
321,484
326,706
331,820
337,209
342,605
348,008
353,418
358,837
364,233
369,637
375,050
380,472
385,903
0
0
0
0
0
0
0
142,948
249,734
254,532
275,593
293,072
298,327
303,496
308,673
313,942
319,467
324,764
330,034
335,330
340,506
345,986
351,474
356,969
362,473
367,985
373,477
378,977
384,487
390,007
395,537
0
0
0
0
0
0
0
142,948
249,746
254,544
270,977
285,003
290,196
295,307
300,424
305,624
311,053
316,274
321,472
326,693
331,804
337,192
342,587
347,988
353,397
358,814
364,209
369,612
375,023
380,444
385,874
0
0
0
0
0
0
0
142,948
249,746
254,544
270,977
285,003
290,196
295,307
300,424
305,624
311,053
316,274
321,472
326,693
331,804
337,192
342,587
347,988
353,397
358,814
364,209
369,612
375,023
380,444
385,874
-------
Table 12.3-2A
48 State NOx Emission Reductions for Program Options 1-2
Year Option 1 Option 1a Option 1b Option 2a Option 2b Option 2c Option 2d Option 2e
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
0
0
0
0
0
0
0
0
0
503
1,766
21,522
52,597
87,976
153,004
217,852
281,454
342,819
399,696
453,617
503,665
548,065
588,591
626,255
660,995
693,689
723,546
750,977
776,413
800,222
821,736
0
0
0
0
0
0
0
0
0
1
5
36,934
115,220
194,212
273,046
350,521
423,557
492,722
554,913
611,895
663,626
711,839
756,359
796,861
834,447
869,952
902,739
932,592
959,480
985,095
1,009,757
0
0
0
0
0
0
0
0
0
1
5
36,934
115,220
194,212
273,046
350,521
423,557
492,722
554,913
611,895
663,626
711,839
756,359
796,861
834,447
869,952
902,739
932,592
959,480
985,095
1,009,757
0
0
0
0
0
0
0
0
301
619
1,007
20,574
52,218
87,616
152,680
217,575
281,270
342,740
399,692
453,643
503,701
548,149
588,685
626,368
661,122
693,857
723,762
751,182
776,574
800,392
821,911
0
0
0
0
0
0
0
0
301
619
2,098
23,185
54,809
89,885
154,892
219,688
283,278
344,625
401,369
455,139
505,133
549,447
589,871
627,461
662,142
694,803
724,582
751,889
777,232
800,997
822,382
0
0
0
0
0
0
0
0
301
619
2,374
21,936
53,563
88,943
153,963
218,816
282,407
343,732
400,568
454,456
504,416
548,807
589,253
626,879
661,590
694,254
724,056
751,441
776,816
800,611
822,114
0
0
0
0
0
0
0
0
301
619
1,007
20,574
52,218
91,884
161,257
230,428
306,499
379,886
448,475
513,588
573,519
626,977
676,038
721,538
762,962
801,885
837,483
870,213
900,551
928,871
954,589
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
-------
Table 12.3-2B
48 State NOx Emission Reductions for Program Options 3-5
Year Option 3 Option 4 Option 5a Option 5b
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
(0)
(0)
(0)
(0)
(0)
(0)
(0)
(0)
301
619
1,007
15,943
42,959
73,685
129,199
184,630
239,010
291,695
341,250
388,195
431,864
471,461
507,554
541,378
572,629
602,207
629,553
654,659
677,917
699,765
719,378
0
0
0
0
0
0
0
0
301
619
1,007
20,574
52,218
87,616
152,680
217,575
281,270
342,740
399,692
453,643
503,701
548,149
588,685
626,368
661,122
693,857
723,762
751,182
776,574
800,392
821,911
0
0
0
0
0
0
0
0
0
0
0
19,175
50,418
81,973
143,208
204,359
264,494
322,902
377,053
428,369
476,010
518,543
557,366
593,437
626,712
658,107
686,773
713,101
737,449
760,270
780,876
0
0
0
0
0
0
0
0
0
0
0
19,175
50,418
81,973
143,208
204,359
264,494
322,902
377,053
428,369
476,010
518,543
557,366
593,437
626,712
658,107
686,773
713,101
737,449
760,270
780,876
-------
Table 12.3-3A
48 State PM2.5 Emission Reductions for Program Options 1-2
Year Option 1 Option 1a Option 1b Option 2a Option 2b Option 2c Option 2d Option 2e
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
0
0
0
0
0
0
0
0
11,630
21,397
24,225
28,235
33,664
40,514
47,663
54,920
62,027
68,710
75,009
80,989
86,591
91,784
96,713
101,364
105,799
109,990
113,855
117,486
120,883
124,049
127,107
0
0
0
0
0
0
0
0
11,969
22,791
29,437
36,451
43,747
51,222
58,464
65,596
72,326
78,595
84,351
89,834
94,962
99,794
104,398
108,827
113,021
116,925
120,414
123,752
126,976
129,945
132,829
0
0
0
0
0
0
11,805
20,131
20,513
22,791
29,437
36,451
43,747
51,222
58,464
65,596
72,326
78,595
84,351
89,834
94,962
99,794
104,398
108,827
113,021
116,925
120,414
123,752
126,976
129,945
132,829
0
0
0
0
0
0
10,394
18,179
19,061
19,998
21,864
25,496
31,233
37,975
45,139
52,476
59,682
66,680
73,288
79,475
85,254
90,651
95,702
100,450
104,977
109,325
113,414
117,166
120,557
123,788
126,910
0
0
0
0
0
0
0
10,605
19,061
20,841
24,363
30,085
36,723
43,772
51,005
58,165
65,096
71,631
77,749
83,475
88,803
93,826
98,536
103,049
107,373
111,463
115,223
118,599
121,819
124,929
127,826
0
0
0
0
0
0
0
10,605
19,061
20,841
24,236
27,341
33,151
39,955
47,128
54,470
61,539
68,290
74,714
80,819
86,448
91,767
96,669
101,334
105,794
110,012
113,904
117,593
120,955
124,147
127,239
0
0
0
0
0
0
0
10,605
19,061
19,998
21,864
25,496
31,233
37,975
45,139
52,476
59,682
66,680
73,288
79,475
85,254
90,651
95,702
100,450
104,977
109,325
113,414
117,166
120,557
123,788
126,910
0
0
0
0
0
0
0
10,605
19,061
19,998
21,864
25,496
31,233
37,975
45,139
52,476
59,682
66,680
73,288
79,475
85,254
90,651
95,702
100,450
104,977
109,325
113,414
117,166
120,557
123,788
126,910
-------
Table 12.3-3B
48 State PM2.5 Emission Reductions for Program Options 3-5
Year Option 3 Option 4 Option 5a Option 5b
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
0
0
0
0
0
0
0
10,605
19,003
19,880
21,685
25,129
30,676
37,288
44,181
51,234
58,148
64,854
71,216
77,163
82,718
87,946
92,842
97,454
101,858
106,094
110,095
113,770
117,089
120,253
123,309
0
0
0
0
0
0
0
10,605
19,060
19,997
22,206
26,093
31,835
38,581
45,748
53,091
60,303
67,307
73,919
80,112
85,896
91,299
96,357
101,111
105,645
110,000
114,096
117,856
121,254
124,492
127,621
0
0
0
0
0
0
0
10,605
18,240
18,304
19,211
21,803
26,572
31,635
37,094
42,743
48,364
53,903
59,112
63,953
68,458
72,715
76,700
80,439
84,033
87,502
90,771
93,774
96,444
99,006
101,490
0
0
0
0
0
0
0
10,605
18,796
19,452
21,003
24,246
29,607
35,247
41,268
47,449
53,523
59,490
65,101
70,323
75,189
79,790
84,096
88,144
92,020
95,753
99,270
102,499
105,389
108,164
110,855
-------
Table 12.3-4A
Monitized Benefit Estimates for Program Options 1-2 (Millions of year 2000 dollars)
Year Option 1 Option 1 a Option 1b Option 2a Option 2b Option 2c Option 2d Option 2e|
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
$0
$0
$0
$0
$0
$0
$0
$0
$5,274
$9,719
$10,814
$13,300
$15,437
$18,260
$21,551
$25,056
$28,683
$32,398
$36,113
$39,932
$43,770
$47,512
$51,384
$55,290
$59,231
$62,916
$66,547
$70,056
$73,641
$77,201
$80,669
$0
$0
$0
$0
$0
$0
$0
$0
$5,425
$10,269
$12,159
$15,665
$18,777
$22,185
$25,736
$29,444
$33,182
$36,994
$40,742
$44,592
$48,453
$52,376
$56,343
$60,370
$64,426
$68,185
$71,759
$75,436
$79,121
$82,678
$86,372
$0
$0
$0
$0
$0
$0
$5,094
$8,935
$9,327
$10,269
$12,159
$15,665
$18,777
$22,185
$25,736
$29,444
$33,182
$36,994
$40,742
$44,592
$48,453
$52,376
$56,343
$60,370
$64,426
$68,185
$71,759
$75,436
$79,121
$82,678
$86,372
$0
$0
$0
$0
$0
$0
$4,497
$8,088
$8,564
$9,072
$10,015
$12,490
$14,794
$17,570
$20,845
$24,253
$27,891
$31,683
$35,478
$39,348
$43,231
$47,131
$51,034
$54,966
$58,933
$62,670
$66,382
$69,935
$73,515
$77,099
$80,591
$0
$0
$0
$0
$0
$0
$0
$4,701
$8,564
$9,456
$10,848
$13,786
$16,257
$19,156
$22,496
$26,000
$29,601
$33,297
$36,980
$40,741
$44,511
$48,315
$52,129
$55,903
$59,819
$63,477
$67,079
$70,498
$74,021
$77,565
$80,971
$0
$0
$0
$0
$0
$0
$0
$4,701
$8,564
$9,455
$10,820
$12,973
$15,308
$18,115
$21,409
$24,934
$28,547
$32,279
$36,029
$39,885
$43,731
$47,514
$51,376
$55,287
$59,237
$62,931
$66,572
$70,103
$73,675
$77,246
$80,728
$0
$0
$0
$0
$0
$0
$0
$4,701
$8,564
$9,072
$10,015
$12,490
$14,794
$17,600
$20,909
$24,454
$28,199
$32,104
$36,020
$40,019
$44,028
$48,055
$52,090
$56,155
$60,251
$64,013
$67,851
$71,630
$75,338
$78,950
$82,670
$0
$0
$0
$0
$0
$0
$0
$4,701
$8,563
$9,068
$10,009
$12,358
$14,441
$16,847
$19,604
$22,545
$25,474
$28,617
$31,752
$34,831
$38,007
$41,202
$44,294
$47,497
$50,725
$53,644
$56,637
$59,469
$62,322
$65,070
$67,929
INPV2004 $550,024 $608,730 $625,176 $557,176
$565,879
$556,177 $559,522 $485,6161
Delta from
Proposal
$186
$58,892
$75,338
$7,338
$16,040
$6,339
$9,683
($64,222)
-------
Table 12.3-4B
Monitized Benefit Estimates for Program Options 3-5
(millions of year 2000 dollars)
Year
Option 3
Option 4 Option 5a Option 5b|
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
NPV 2004
Delta from
Proposal
$0
$0
$0
$0
$0
$0
$0
$4,701
$8,551
$9,045
$9,973
$12,366
$14,585
$17,285
$20,303
$23,639
$27,091
$30,688
$34,303
$37,986
$41,682
$45,416
$49,156
$52,928
$56,638
$60,229
$63,807
$67,329
$70,679
$74,133
$77,493
$531,782
($18,056)
$0
$0
$0
$0
$0
$0
$0
$4,701
$8,564
$9,071
$10,275
$12,800
$15,116
$17,905
$21,194
$24,717
$28,371
$32,179
$35,991
$39,879
$43,780
$47,599
$51,523
$55,476
$59,465
$63,222
$66,953
$70,527
$74,128
$77,733
$81,347
$556,114
$6,276
$0
$0
$0
$0
$0
$0
$0
$4,701
$8,378
$8,680
$9,390
$11,533
$13,555
$15,713
$18,434
$21,362
$24,422
$27,544
$30,766
$34,048
$37,340
$40,671
$44,011
$47,278
$50,689
$53,913
$57,123
$60,190
$63,286
$66,296
$69,418
$479,478
($70,360)
$0
$0
$0
$0
$0
$0
$0
$4,701
$8,503
$8,943
$9,808
$12,161
$14,354
$16,691
$19,596
$22,710
$25,940
$29,335
$32,727
$36,183
$39,650
$43,057
$46,572
$50,116
$53,699
$57,080
$60,343
$63,659
$66,903
$70,063
$73,334
$507,053
($42,785)
-------
$90,000
$80,000
$70,000
$60,000
$50,000
$40,000
$30,000
$20,000
$10,000
$0
Figure 12.3-1A
Monitized Benefit Estimates for Program Options 1-2 (millions of year 2000 dollars)
Option 1
Option 1a
Option 1b
•Option 2a
•Option 2b
•Option 2c
•Option 2d
Option 2e
Proposal
1995
—n H n n H
2000 2005
2010
2015
2020
2025
2030
2035
-------
$90,000
$80,000
$70,000
$60,000
$50,000
$40,000
$30,000
$20,000
$10,000
$0
Figure 12.3-1B
Monitized Benefit Estimates for Program Options 3-5 (millions of year 2000 dollars)
Option 3
Option 4
Option 5a
Option 5b
Proposal
1995
—n n n n n n
2000 2005
2010
2015
2020
2025
2030
2035
-------
Table 12.3-5A
Benefit Increases for Options 1-2 Compared to Proposal (millions of year 2000 dollars)
Year
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
NPV 2004
Option 1
$0
$0
$0
$0
$0
$0
$0
-$4,701.5
-$3,290
$647
$799
$809
$642
$690
$706
$802
$792
$715
$635
$584
$539
$382
$349
$324
$298
$245
$165
$121
$126
$102
$78
$186
Option 1a
$0
$0
$0
$0
$0
$0
$0
-$4,701.5
-$3,139
$1,197
$2,144
$3,175
$3,983
$4,615
$4,891
$5,190
$5,291
$5,311
$5,264
$5,244
$5,221
$5,245
$5,309
$5,404
$5,493
$5,515
$5,377
$5,500
$5,606
$5,579
$5,781
$58,892
Option 1b
$0
$0
$0
$0
$0
$0
$5,094
$4,234
$763
$1,197
$2,144
$3,175
$3,983
$4,615
$4,891
$5,190
$5,291
$5,311
$5,264
$5,244
$5,221
$5,245
$5,309
$5,404
$5,493
$5,515
$5,377
$5,500
$5,606
$5,579
$5,781
$75,338
Option 2a
$0
$0
$0
$0
$0
$0
$4,497
$3,387
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$7,338
Option 2b
$0
$0
$0
$0
$0
$0
$0
$0
$0
$385
$833
$1,295
$1,463
$1,586
$1,651
$1,746
$1,710
$1,615
$1,502
$1,393
$1,280
$1,185
$1,095
$936
$885
$807
$696
$563
$505
$466
$381
$16,040
Option 2c
$0
$0
$0
$0
$0
$0
$0
$0
$0
$384
$805
$483
$514
$545
$564
$681
$656
$596
$551
$537
$499
$383
$341
$320
$303
$261
$190
$168
$160
$147
$137
$6,339
Option 2d
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$30
$64
$201
$308
$421
$542
$670
$796
$924
$1,056
$1,189
$1,318
$1,343
$1,468
$1,695
$1,823
$1,851
$2,079
$9,683
Option 2e|
$0
$0
$0
$0
$0
$0
$0
$0
-$2
-$3
-$6
-$133
-$353
-$723
-$1,241
-$1,708
-$2,417
-$3,065
-$3,726
-$4,517
-$5,224
-$5,929
-$6,741
-$7,469
-$8,208
-$9,026
-$9,745
-$10,467
-$11,193
-$12,029
-$12,662
($64,222)
-------
Table 12.3-5B
Benefit Increases for Options 3-5 Compared to Proposal
(millions of year 2000 dollars)
Year
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
NPV 2004
Option 3
$0
$0
$0
$0
$0
$0
$0
$0.0
-$13
-$27
-$42
-$124
-$209
-$285
-$542
-$615
-$800
-$995
-$1,175
-$1,362
-$1,550
-$1,714
-$1,879
-$2,038
-$2,296
-$2,441
-$2,575
-$2,606
-$2,836
-$2,966
-$3,098
($18,056)
Option 4
$0
$0
$0
$0
$0
$0
$0
$0.0
$0
$0
$260
$309
$322
$335
$349
$464
$480
$497
$513
$531
$549
$469
$489
$510
$532
$551
$571
$592
$613
$634
$756
$6,276
Option 5a
$0
$0
$0
$0
$0
$0
$0
$0
-$186
-$391
-$625
-$957
-$1,239
-$1,856
-$2,411
-$2,892
-$3,468
-$4,139
-$4,712
-$5,301
-$5,891
-$6,459
-$7,023
-$7,688
-$8,244
-$8,757
-$9,259
-$9,745
-$10,230
-$10,803
-$11,173
($70,360)
Option 5b|
$0
$0
$0
$0
$0
$0
$0
$0
-$61
-$129
-$207
-$330
-$440
-$879
-$1,249
-$1,544
-$1,951
-$2,348
-$2,751
-$3,165
-$3,581
-$4,073
-$4,462
-$4,850
-$5,234
-$5,590
-$6,039
-$6,276
-$6,612
-$7,036
-$7,257
($42,785)|
-------
Figure 12.3-2 A
Benefit Increases for Options 1-2 Compared to Proposal (millions of year 2000 dollars)
$6,000
$4,000
Option 1
Option 1a
Option 1b
-*- Option 2a
-•-Option 2b
Option 2c
Option 2d
Option 2e
2035
-$14,000
-------
$6,000
$4,000
Figure 12.3-2B
Benefit Increases for Options 3-5 Compared to Proposal (millions of year 2000 dollars)
$2,000
-$4,000
-$6,000
-$8,000
-$10,000
-$12,000
-•-Option 3
-•-Option 4
-*- Option 5a
-*- Option 5b
2035
-$14,000
-------
Regulatory Alternatives
12.4 Cost Analysis for Alternative Options
This section describes the cost methodology and the estimates used to evaluate the alternative
options. The section describes our estimates for both the fuel impacts and the engine/equipment
impacts of the various options, if applicable.
The presentation of information on fuel costs is summarized in a series of tables showing the
impact on a cost-per-gallon basis for the appropriate fuel alternative, as well as an estimate of the
aggregate fuel cost impact for each alternative option. However, the detailed fuel cost analysis
used to derive the cost-per-gallon estimates is contained in Chapter 7 of this draft RIA. The
presentation of information on engine/equipment costs are detailed in the related sections below.
The engine and equipment cost estimates for the alternative options relies heavily on the
methodology, and in some cases the estimates, used for the proposal. Our discussion of the cost
estimates for the alternative options will focus on those inputs or methods which are different
from the input or method used for the proposal. To the extent the cost estimates are based on the
data used for the proposal, we have not repeated the analysis behind the estimate here, rather, the
reader can refer to Chapter 6 of this draft RIA for the engine/equipment cost estimates for the
proposal.
As noted in Chapter 3.1.5, there are differences in the fuel quantities used for costs and the
fuel quantities used for emissions inventories resulting rom differences in methods. Please see
Chapter 3.1.5 for additional discussion of these differences.
12.4.1 One Step Options
12.4.1.1 Option 1
This option is described in Figure 12.1.1-1 in Section 12.1 of this draft RIA. Option 1
requires 15ppm sulfur fuel in 2008 for nonroad engines only and 500 ppm sulfur fuel in 2008 for
locomotive and marine engines, which allows early introduction of PM filter technology for
some engines.
12.4.1.1.1 Fuel Costs for Option 1
The total fuel costs from Chapter 7 of the draft RIA comprising the refining and distribution
and additive costs for Option 1 are summarized in the following tables.
12-53
-------
Draft Regulatory Impact Analysis
Table 12.4.1.1.1-1
Per-Gallon Fuel Costs for Option 1 (cents per gallon)
Option
One Step
Specification
ISppmNR
500 ppm L & M
500 DDm L & M
Year
2008 +
2008-2011
2012
Refining
Costs
ft/sal)
4.8
2.2
2.2
Distribution &
Additive Costs
ft/sal)
0.4
0.4
0.2
Total
Costs
ft/sal)
5.2
2.6
2.4
Table 12.4.1.1.1-2
Net Operating Costs for Option 1 Incremental To The Proposal (millions)
(Net present values through 2030 at 3% discount rate)
Specification
1 5 ppm fuel
500 ppm fuel
Gallons
11,530
-21,770
Fuel costs*
$1,020
-$550
Net maintenance costs
$250
Change in net
operating costs
$720
* Note that the incremental fuel costs presented here are calculated as: [proposal $/gal] multiplied by [proposal gallons]
minus the [option $/gal] multiplied by [option gallons]. This is not mathematically equivalent to the difference in gallons
multiplied by the difference in $/gal.
These fuel costs and other related operating costs (i.e., maintenance savings) result in an
increase in the net-present value of Option 1 of approximately $720 million as compared to the
proposal through 2030.
12.4.1.1.2 Engine & Equipment Costs for Option 1
Engine Fixed Costs
As discussed in Section 12.6.2.1.1 of this draft RIA, Option 1 presents a number of unique
challenges for engine manufacturers as compared to the proposal. These include up to two years
of overlap with the nonroad Tier 3 development time frame and two fewer years of learning for
the highway to nonroad technology transfer as compared to our proposal. These changes impact
the engine engineering costs are described below. Because of these unique challenges, Option 1
has the potential to result in limited product offerings for certain segments of the nonroad engine
and equipment market. This potential exists primarily because of the overlapping development
time frames between Tier 3 and Tier 4. To the extent that engine and equipment manufacturers
engineering staff and resources (e.g., sufficient laboratory test cells) are unable to cover both
development programs, companies may need to decide to shift resources from one program to the
other, with the result being limited product availability for either Tier 3 or possibly for Tier 4.
Our cost analysis for Option 1 presented below assumes companies do have these resources.
However, to the extent some companies do not have the necessary resources, our cost analysis
does not attempt to estimate the cost impacts of limited product offerings.
12-54
-------
Regulatory Alternatives
Option 1 has significant overlap with Tier 3 engine development. Nonroad engine
manufacturers typically require 3 to 4 years of development in advance of a major new emission
standard or new engine product launch. This period allows for sufficient time for engine
development as well as providing adequate time for equipment manufacturers to redesign
equipment to accommodate the new technology engines. For the 175-750 hp category, a 2009
implementation could require engine development beginning as early as calender year 2005,
which is also the final year of development before the Tier 3 implementation in 2006. There is
also overlap with Option 1's 2010 implementation for the 100-175 hp category, which has a 2007
Tier 3 implementation. Finally, there would be two years of overlap under Option 1 for the 75-
100 hp engines, which have a 2008 Tier 3 start date.
To estimate the cost impacts of these overlapping development programs, we have estimated
that manufacturers would have sufficient staff to address the work load issues associated with
product development of concurrent engine programs (i.e., development of Tier 3 and Tier 4
engines). This of course assumes that manufacturers have the additional staff to perform the
concurrent engine development programs as well as the testing resources (e.g., laboratory
capacity). It is possible that some manufactures do not have the personnel resources and/or the
laboratory resources to cover both Tier 3 and Tier 4 engine development, and this cost analysis
does not attempt to estimate what the impacts of such a short-fall would be. Based on our
experience and discussions with engine manufacturers we have estimated that a typical product
development group consists of 21 workers (9 engineers, 12 technicians). Our annual cost
estimate for each team, including test cell time, is $3 million per year.2 Therefore, for each year
of potential overlap between the Tier 3 program and the Tier 4 program under Option 1 we have
estimated an additional cost of $3 million per engine platform. Consistent with our estimation
of the number of engine platforms in each power category used for the proposal, this would add
approximately $290 million dollars to Option 1 as compared to the proposal.
The second impact on engine engineering costs of Option 1 is the reduced amount of time for
nonroad engine companies to learn from the 2007 highway heavy-duty diesel experience with
aftertreatment systems. There are a number of ways in which nonroad companies can learn from
the extensive research and development effort being expended to achieve the 2007 highway
standards. These include:
• nonroad engine companies can purchase 2007 highway products and reverse engineer
how the products work;
• nonroad engine companies can learn from information available in the public literature
regarding 2007 highway technologies (such as SAE papers and other technical
publications);
• nonroad engine companies can learn by collaboration with technology vendors such as
exhaust aftertreatment companies who are developing PM filters and NOx aftertreatment
systems with on-highway companies;
• nonroad engine manufacturers can work with 3rd party engineering laboratories such as
AVL, FEV, Ricardo, or Southwest Research Institute who through their work with
12-55
-------
Draft Regulatory Impact Analysis
industry and governments will acquire significant expertise with diesel aftertreatment;
and,
• nonroad engine companies can hire engineers and scientists away from highway
companies who have already gone through the engine design experience.
In order to reduce costs for nonroad companies, they must have access to these various
learning channels early enough in time to impact their R&D programs. For our proposal, which
provides at least 4 years after the 2007 program before the first nonroad engines must use
advanced aftertreatment systems, we have estimated this learning can reduce the R&D costs for
nonroad companies by 30 percent compared to what they would incur if there was no 2007
highway program and the companies were required to develop the aftertreatment technologies
without any learning from outside sources, and for nonroad companies who also are developing
engines to comply with the 2007 highway standards we have estimated the learning time
available with our proposal will reduce their R&D costs by 90 percent. We project based on our
engineering judgement that as the time frame for learning is reduced below 4 years, the potential
R&D cost reductions will decrease substantially, as shown in Table 12.4.1.1.2-1 below.
Table 12.4.1.1.2-1
Impact of Tier 4 Implementation of Engine Research and Development Costs
Company Type
Nonroad &
Highway
Companies
Nonroad only
companies
Estimated Reduction in Tier 4 Engine R&D Costs as a Function of the First Year of
Implementation for Nonroad Advanced Aftertreatment
2011
implementation
90%
30%
2010
implementation
63%
21%
2009
implementation
14%
5%
2008 implementation
0%
0%
Option 1 reduces the availability to learn from the highway program by two years for the 175
- 750 hp category. Based on the estimates provided in the table 12.4.1.1.2-1, this would reduce
the learning for highway companies from 90 percent down to 14 percent, and for the nonroad
only engine companies from 30 percent down to 5 percent. For the 75-175 hp category, Option
1 reduces the highway learning by one year. Based on the estimates provided in Table
12.4.1.1.2-1, this will reduce the learning for highway companies to 63 percent and for nonroad
only companies to 21 percent. Consistent with the engine research and development costs
estimated for the proposal and described in detail in Chapter 6 of this RIA, these adjustments
increase the R&D expenditure of Option 1 by approximately $120 million dollars.
Engine Variable Costs
12-56
-------
Regulatory Alternatives
This option relies on the same engine hardware cost estimates as for the proposal, which are
described in Chapter 6 of this draft RIA. Where appropriate, we have shifted the engine variable
hardware costs in time to match the implementation dates of Option 1. Specifically:
- for the <50 hp category, the hardware costs described in Chapter 6 have been delayed by 1
year;
- for the 50-75 hp category, the 2008 transitional standard hardware has been eliminated;
for the 75-175 hp category and the 175 - 750 hp category, the PM filter system hardware has
been pulled forward by two years for 50 percent of the engines; and,
for the >750 hp category, the hardware cost are the same as in the proposal.
The NPV of the engine variable costs through 2030 is approximately $90 million more than
in the proposal. These costs are higher than the proposal because the elimination of the
transitional PM standards for the 50-75 hp engines, combined with a 1 year delay in the standards
for the < 50 hp engines does not off-set the increased hardware costs associated with the one year
pull-ahead of PM filters for the 75 - 750 hp engines. The annual engine variable costs are shown
in Figure Figure 12.4.1.1.2-1, along with the annual engine variable costs for the proposal and the
other alternative options.
12-57
-------
Draft Regulatory Impact Analysis
Figure 12.4.1.1.2-1. Engine Variable Costs for the Proposal and Alternative Options
$1.2
2004-2030
Proposal =
Option 1 =
Option 2b :
Option 2c:
Option 2e :
Option 3 =
Option 5a :
Option 5b :
NPV:
$11.10
$11.19
= $11.51
= $11.26
= $7.75
$10.72
= $8.10
= $9.19
—•—Proposal
—•—Option 1
-*- Option 2b
-X— Option 2c
-*- Option 2e
-•-Option 3
-a— Option 5a
-8- Option 5b
$0.0*
2004
2008
2028
Equipment Fixed Costs
Chapter 6 of this draft RIA presents a detailed discussion of our methodology for estimating
equipment fixed costs, which is dominated by our estimates for equipment redesign costs. In this
sub-section we will discuss the impact of Option 1 on the equipment fixed costs for each of the
engine power categories.
For the <50 hp engine category there is a one year delay in the standards to 2009. We have
not adjusted the costs to redesign the < 50 hp engines, but we have shifted the costs back by one
year in time.
For the 50-75 hp engine category, Option 1 eliminates the 2008 transitional PM standards,
and we have eliminated the equipment redesign costs associated with the proposed 2008
transitional standard.
For the 75-175 hp engine category, Option 1 pulls ahead the proposed 0.01 g/bhp-hr PM
standard ahead by two years to 2010 for 50 percent of the engines. This is followed by 50
percent of the engines meeting the proposed PM and NOx standard in 2012, and finally 50
12-58
-------
Regulatory Alternatives
percent of the engines must meet the final NOx standard. Therefore, we have estimated Option 1
will require a major equipment redesign for 50 percent of the engines 3 times (2010, 2012 and
2014), or a total of 1.5 redesigns for the power category. In effect, this is one-quarter more
redesigns than expected under the proposal which increases redesign costs by approximately
$470 million.
Equipment Variable Costs
We have estimated the impacts on equipment variable costs in the same manner as done for
engine variable costs by eliminated costs where appropriate and shifting them up a year or two or
back a year or two where appropriate. These changes increase the NPV through 2030 by
approximately $20 million relative to the equipment variable costs expected under the proposal.
Total Engine/Equipment Cost
Based on the estimates provided above, we have estimated the Option 1 will result in an
increase in the net-present value of the engine and equipment costs through 2030 of
approximately $990 million dollars.
12.4.1.2 Option la
Option la is described in Figure 12.1.1-2 in Section 12.1 of this draft RIA. Option la
requires 15ppm sulfur fuel in 2008 for nonroad, locomotive and marine engines. The engine
standards, which are also described in Chapter 12.1, consist of a 2 year introduction for a 0.01
g/bhp-hr PM standard for all nonroad engines by power category beginning in 2009, and a two
year introduction of a 0.30 g/bhp-hr NOx standard for all nonroad engines by power category
beginning in 2011.
As discussed in Section 12.6.2.1.2, we do not believe this very aggressive standards program
is technically feasible for either the fuel program or the engine program, and therefore we have
not provided a cost estimate for Option la.
12.4.1.3 Option Ib
Option Ib is described in Figure 12.1.1-3 in Section 12.1 of this draft RIA. Option Ib has the
same engine standards as Option la; however, the fuel program consists of 15ppm for nonroad,
locomotive and marine engines beginning in 2006. Option Ib is identical to Option la with
respect to the engine standards program, and the fuel program is implemented two years earlier in
2006. As discussed in Section 12.6.2.1.3, we do not believe this very aggressive standards
program is technically feasible for either the fuel program or the engine program, and therefore
we have not provided a cost estimate for Option Ib.
12-59
-------
Draft Regulatory Impact Analysis
12.4.2 Two Step Options
12.4.2.1 The Proposal
Our fuel and engine standards proposal is summarized in Figure 12.1.2-1 in Section 12.1 of
this draft RIA. The cost estimation for the proposal is detailed in Chapters 6 (engine &
equipment program) and 7 (fuel program) of this draft RIA, and will not be repeated here.
12.4.2.2 Option 2a
Option 2a is described in Figure 12.1.2-2 in Section 12.1 of this draft RIA. Option 2a
requires the same engine program as our proposal; however, the first-step of the two step fuel
program (500 ppm sulfur fuel for nonroad, locomotive and marine engines) is implemented one
year earlier than in our proposal (2006 rather than 2007).
As discussed in Section 12.6.2.2.2, we do not believe this aggressive fuel program is
technically feasible and therefore we have not provided a cost estimate for Option 2a.
12.4.2.3 Option 2b
This option is described in Figure 12.1.2-3 in Section 12.1 of this draft RIA. Option 2b is
similar to the fuel program for the proposal, except the 15 ppm sulfur nonroad fuel is pulled
ahead one year to 2009. The engine standards program under Option 2b is similar to the
proposal, except that the PM filter based standards for the >25 hp engines is pulled forward by
one year, however the NOx program and the 2008 PM standards for the <75 hp engines are the
same as the proposal.
12.4.2.3.1 Fuel Costs for Option 2b
The total fuel costs from Chapter 7 of the Draft RIA comprising the refining and distribution
and additive costs for Option 2b are summarized in the following tables.
12-60
-------
Regulatory Alternatives
Table 12.4.2.3.1-1
Total Fuel Costs for Option 2b (cents per gallon)
Option
Nonroad goes to 15
ppm in 2009
Specification
500 ppm NR, L&M
500 ppm L&M
15 ppm NR (total incl 2007)
500 ppm L&M
Year
2007-
2008
2009-
2012
2009+
2013+
Refining
Costs
(eVgal)
2.2
2.2
4.6
2.2
Distribution &
Additive Costs
(eYgal)
0.3
0.4
0.4
0.2
Total
Costs
(eVgal)
2.5
2.6
5.0
2.4
Table 12.4.2.3.1-2
Net Operating Costs for Option 2b Incremental To The Proposal (millions)
(Net present values through 2030 at 3% discount rate)
Specification
1 5 ppm fuel
500 ppm fuel
Gallons
4,270
-5,180
Fuel costs*
$430
-$130
Net maintenance costs
$250
Total operating costs
$540
* Note that the incremental fuel costs presented here are calculated as: [proposal $/gal] multiplied by [proposal gallons]
minus the [option $/gal] multiplied by [option gallons]. This is not mathematically equivalent to the difference in gallons
multiplied by the difference in $/gal.
These fuel costs and other related operating costs (e.g., maintenance savings, fuel
consumption impacts) result in an increase in the net-present value of Option 2b of
approximately $540 million as compared to the proposal through 2030.
12.4.2.3.2 Engine and Equipment Costs for Option 2b
Engine Fixed Costs
As discussed in Section 12.6.2.2.3, Option 2b presents a number of unique challenges for
engine manufacturers as compared to the proposal. These include up to one year of overlap with
the nonroad Tier 3 development time frame for one power category, and one less year for
learning for the highway to nonroad technology transfer as compared to our proposal. In
addition, Option 2b presents a significant challenge for engine manufacturers during the
implementation of the standards for NOx and PM in the 2010-2013 time frame which is not
present in our proposal. Specifically, engines >25 hp will have a series of introductions with new
PM standards one year and new NOx standards the next year. We have estimated a cost impact
for each of these engine engineering impacts as compared to our proposal, as described below.
Because of these unique challenges, Option 2b has the potential to result in limited product
offerings for certain segments of the nonroad engine and equipment market. This potential exists
12-61
-------
Draft Regulatory Impact Analysis
primarily from the rapid change in PM and NOx standards for the same engine power categories
in the 2010-2013 time frame, as well as the overlapping development time frames between Tier 3
and Tier 4. To the extent that engine and equipment manufacturers engineering staff and
resources (e.g., sufficient laboratory test cells) are not sufficient to address the workload issues
associated with these engineering requirements, companies may need to decide to focus their
resources on certain products at the expense of others, with the result being limited product
availability for Tier 3 as well as for Tier 4. Our cost analysis for Option 2b presented below
assumes companies do have these resources. However, to the extent some companies do not
have the necessary resources, our cost analysis does not attempt to estimate the cost impacts of
limited product offerings.
Option 2b has up to one year of engine design overlap with Tier 3 engine development,
specifically for engines in the 75-100 hp range. For these engines, Tier 3 is implemented in
2008, and Option 2b's one year pull-ahead of PM standards would begin in 2011. As discussed
in Section 12.4.1.1.2 (Engine & Equipment Costs for Option 1), nonroad engine manufacturers
typically require 3 to 4 years of development in advance of a major new emission standard or
new engine product launch. As discussed in Section 12.4.1.1.2, we have estimated this potential
overlap in Tier 3 and Tier 4 engine development could cost on the order of $3 million per engine
platform. Consistent with our estimation of the number of engine platforms in each power
category used for the proposal, this adds approximately $30 million dollars to Option 2b as
compared to the proposal.
The second impact on engine engineering costs of Option 2b is the reduced amount of time
for nonroad engine companies to learn from the 2007 highway heavy-duty diesel experience with
aftertreatment systems. Compared to our proposal, Option 2b reduces this time frame by one
year because of the pull-ahead of the PM filter based standards. As discussed in Section
12.4.1.1.2 and using the estimates provided in Table 12.4.1.1.2-1, Option 2b will reduce the
engine research and development cost savings due to learning for highway companies from 90 to
63 percent and for nonroad only companies from 70 to 21 percent. Consistent with the engine
research and development costs estimated for the proposal and described in detail in Chapter 6 of
this RIA, these adjustments increase the R&D expenditure of Option 2b by approximately $40
million dollars.
The third impact of Option 2b on the engine engineering costs is the rapid change of PM and
NOx standards in two years for both the 75-175 hp and 175-750 hp categories. Option 2b
implements a 0.01 g/bhp-hr PM standard in 2010 for 100 percent of the engines, and the
following year 50 percent of the engines must meet a 0.30 g/bhp-hr NOx standard, therefore /^ of
the engines will require a redesign in 2009 and 2010. This will present a significant engine
calibration challenge for engine manufacturers. Under Option 2b, we are projecting that in order
to comply with the requirement to produce 50 percent of the engines to a new standard the next
year, companies would need to expend considerable engineering resources (staff and test cell
time) to develop the new calibrations. We have estimated that each engine platform would
require a team of 3 engineers and 4 technicians plus laboratory test cell resources working for
12-62
-------
Regulatory Alternatives
one year to develop the additional calibrations which Option 2b would require (implementation
of Tier 4 NOx standards 1 year after Tier 4 PM standards for 1A of the engines). We estimate the
cost of this team for one year at $1 million.3 Consistent with our estimation of the number of
engine platforms in each power category used for the proposal, this engineering effort ($1 million
per engine platform for /^ of the platforms in the 75-750 hp categories) adds approximately $30
million dollars to Option 2b as compared to the proposal.
Engine Variable Costs
Option 2b relies on the same engine hardware cost estimates as for the proposal, which are
described in Chapter 6 of this draft RIA. Where appropriate, we have shifted the engine variable
hardware costs in time to match the implementation dates of Option 2b. Specifically:
- for the >25 hp engines, 75-175 hp category and the 175 - 750 hp category, the PM filter
system hardware has been pulled forward by one year
- for the >750 hp category, the PM filter system has been pulled forward by one year for 50
percent of the engines.
The NPV of the engine variable costs through 2030 is approximately $410 million more than
the proposal. The annual engine variable costs are shown in Figure 12.4.1.1.2-1.
Equipment Fixed Costs
Chapter 6 of this draft RIA presents a detailed discussion of our methodology for estimating
equipment fixed costs, which is dominated by our estimates for equipment redesign costs. In this
sub-section we will discuss the impact of Option 2b on the equipment fixed costs for each of the
engine power categories.
For the <25 hp engine category, Option 2b is the same as the proposal, so there are no
differences for equipment redesign costs.
For the 25-50 hp engines, Option 2b would require a redesign in 2012 for PM filters ,
followed by a minor equipment update the next year to accommodate the 3.5 g/bhp-hr NOx
standard. We have estimated the 2012 equipment redesign costs as being equivalent to the
redesign costs of the proposal's 2013 program. We have estimated the cost of Option 2b's 2013
NOx standard impact as being 1A of the redesign costs of the proposal's 2013 costs.
For the 50-75 hp engines, Option 2b requires equipment redesign one year earlier than in the
proposal. However, we estimate the equipment redesign effort is identical to the proposal, and
we have estimated the costs to be the same as the proposal.
For the 75-175 hp engine category, Option 2b pulls ahead the proposed 0.01 g/bhp-hr PM
standard ahead by one year to 2011. This is followed by 50 percent of the engines meeting the
12-63
-------
Draft Regulatory Impact Analysis
proposed NOx standard in 2012, and finally 50 percent of the engines must meet the final NOx
standard in 2014. Therefore, we have estimated Option 2b will require a major equipment
redesign for all of the equipment in 2011, followed by a minor redesign effort in 2012 for 50
percent of the equipment and in 2014 for 50 percent of the equipment. We have estimated that
each of these minor redesign efforts will cost /^ of the major redesign costs estimated for the
proposal.
For the 175 - 750 hp engine category, Option 2b pulls ahead the proposed 0.01 g/bhp-hr PM
standard ahead by one year to 2010. This is followed by 50 percent of the engines meeting the
proposed NOx standard in 2011, and finally 50 percent of the engines must meet the final NOx
standard in 2014. Therefore, we have estimated Option 2b will require a major equipment
redesign for all of the equipment in 2010, followed by a minor redesign effort in 2011 for 50
percent of the equipment and in 2014 for 50 percent of the equipment. We have estimated that
each of these minor redesign efforts will cost 1A of the major redesign costs estimated for the
proposal.
For the > 750 hp category, Option 2b pulls ahead the proposed 0.01 g/bhp-hr PM standard
ahead by one year to 2010 for 50 percent of the engines. This is followed by 50 percent of the
engines meeting the proposed NOx standard in 2011, and finally all of the engines must meet the
final PM and NOx standard in 2014. We have estimated that the equipment which goes through
a major redesigned to accommodate the new PM standard engines in 2010 will not redesign
again until 2014, when they would go through a minor equipment redesign related to the NOx
standard. The other half of the equipment fleet would go through a major redesign in 2011 to
accommodate the NOx standard, and this same equipment would also go through a minor
redesign in 2014 to meet the final PM standard. Consistent with the discussion above, we have
estimated the costs of the major redesign to be equivalent to the redesign estimates for the
proposal, and we have estimated that a minor redesign costs /^ of the proposal's major redesign
estimates.
The combined result of the changes listed above for the equipment fixed costs result in an
increase for Option 2b as compared to our proposal of approximately $130 million.
Equipment Variable Costs
We have estimated the impacts on equipment variable costs in the same manner as done for
engine variable costs by eliminated costs where appropriate and shifting them up a year or two or
back a year or two where appropriate. These changes increase the NPV through 2030 by $10
million relative to the equipment variable costs expected under the proposal.
Total Engine/Equipment Cost
12-64
-------
Regulatory Alternatives
Based on the estimations provided above, we have estimated the Option 2b will result in an
increase in the net-present value of the engine and equipment costs through 2030 of
approximately $640 million dollars.
12.4.2.4 Option 2c
This option is described in Figure 12.1.2-4 in Section 12.1 of this draft RIA. Option 2c is
almost identical to Option 2b which is described in section 12.4.2.3 above, with the exception
that the one year pull ahead of the PM standard is only for the 175-750 hp engine category
(Option 2b pulls ahead the PM filter based standard for all engines >25 hp by one year). As with
Option 2b, this will require 15 ppm sulfur nonroad fuel in 2009, one year earlier than in the
proposal.
12.4.2.4.1 Fuel Costs for Option 2c
The total fuel costs from Chapter 7 of the Draft RIA comprising the refining and distribution
and additive costs for Option 2c are summarized in the following tables. These tables are the
same as in Option 2b.
Table 12.4.2.4.1-1
Total Fuel Costs for Option 2c (cents per gallon)
Option
Nonroad goes to 15
ppm in 2009
Specification
500 ppm NR, L&M
500 ppm L&M
15 ppm NR (total incl 2007)
500 ppm L&M
Year
2007-
2008
2009-
2012
2009+
2013+
Refining
Costs
(eVgal)
2.2
2.2
4.6
2.2
Distribution &
Additive Costs
(eYgal)
0.3
0.4
0.4
0.2
Total
Costs
(eVgal)
2.5
2.5
5.0
2.4
Table 12.4.2.4.1-2
Net Operating Costs for Option 2c Incremental To The Proposal (millions)
(Net present values through 2030 at 3% discount rate)
Specification
1 5 ppm fuel
500 ppm fuel
Gallons
4,270
-5,180
Fuel costs*
$430
-$130
Net maintenance costs
$240
Total operating costs
$530
* Note that the incremental fuel costs presented here are calculated as: [proposal $/gal] multiplied by [proposal gallons]
minus the [option $/gal] multiplied by [option gallons]. This is not mathematically equivalent to the difference in gallons
multiplied by the difference in $/gal.
12-65
-------
Draft Regulatory Impact Analysis
These fuel costs and other related operating costs (e.g., maintenance savings, fuel
consumption impacts) result in an increase in the net-present value of Option 2c of
approximately $530 million as compared to the proposal through 2030.
12.4.2.4.2 Engine and Equipment Costs for Option 2c
Engine Fixed Costs
As discussed in Section 12.6.2.2.4, Option 2c represents a number of unique challenges for
engine and manufacturers as compared to the proposal (these challenges are very similar to those
for Option 2b, but only for those engines and equipment in the 175-750 hp category). As
discussed in Section 12.4.2.3.2 (Option 2b), to the extent that engine and equipment
manufacturers engineering staff and resources are not sufficient to address the workload issues
associated with these engineering requirements, companies may need to decide to focus their
resources on certain products at the expense of others, with the result being limited product
availability for Tier 3 as well as for Tier 4. Our cost analysis for Option 2c presented below
assumes companies do have these resources. However, to the extent some companies do not
have the necessary resources, our cost analysis does not attempt to estimate the cost impacts of
limited product offerings.
Option 2c reduces the amount of time for nonroad engine companies to learn from the 2007
highway heavy-duty diesel experience with aftertreatment systems. Compared to our proposal,
Option 2c reduces this time frame by one year because of the pull-ahead of the PM filter based
standards for the 175-750 hp engine category. As discussed in Section 12.4.1.1.2 and using the
estimates provided in Table 12.4.1.1.2-1, Option 2c will reduce the engine research and
development cost savings due to learning for highway companies from 90 to 63 percent and for
nonroad only companies from 70 to 21 percent. Consistent with the engine research and
development costs estimated for the proposal and described in detail in Chapter 6 of this RIA,
these adjustments increase the R&D expenditure of Option 2c by approximately $40 million
dollars.
As discussed under Option 2b, Option 2c also increases the engine engineering costs relative
to the proposal due to the rapid change of PM and NOx standards in two years. For the 175-750
hp category, Option 2c implements a 0.01 g/bhp-hr PM standard in 2010 for 100 percent of the
engines, and the following year 50 percent of the engines must meet a 0.30 g/bhp-hr NOx
standard, therefore 1A of the engines will require a redesign in 2009 and 2010. This will present a
significant engine calibration challenge for engine manufactures. Under Option 2c, we are
projecting that in order to comply with the requirement to produce 50 percent of the engines to a
new standard the next year, companies would need to expend considerable engineering resources
(staff and test cell time) to develop the new calibrations. We have estimated that each engine
platform would require a team of 3 engineers and 4 technicians plus laboratory test cell resources
working for one year to develop the additional calibrations which Option 2c would require
(implementation of Tier 4 NOx standards 1 year after Tier 4 PM standards for 1A of the engines).
12-66
-------
Regulatory Alternatives
We estimate the cost of this team for one year at $1 million.4 Consistent with our estimation of
the number of engine platforms in each power category used for the proposal, this engineering
effort ($1 million per engine platform for /^ of the platforms in the 175-750 hp category) adds
approximately $20 million dollars to Option 2c as compared to the proposal.
Engine Variable Costs
Option 2c relies on the same engine hardware cost estimates as for the proposal, which are
described in Chapter 6 of this draft RIA. Where appropriate, we have shifted the engine variable
hardware costs in time to match the implementation dates of Option 2c. Specifically for 175 -
750 hp category, the PM filter system hardware has been pulled forward by one year. The NPV
of the engine variable costs through 2030 is approximately $160 million more than the proposal.
The annual engine variable costs are shown in Figure 12.4.1.1.2-1.
Equipment Fixed Costs
Chapter 6 of this draft RIA presents a detailed discussion of our methodology for estimating
equipment fixed costs, which is dominated by our estimates for equipment redesign costs. In this
sub-section we will discuss the impact of Option 2c on the equipment fixed costs for the 175-750
hp category equipment.
For the 175 - 750 hp engine category, Option 2b pulls ahead the proposed 0.01 g/bhp-hr PM
standard ahead by one year to 2010. This is followed by 50 percent of the engines meeting the
proposed NOx standard in 2011, and finally 50 percent of the engines must meet the final NOx
standard in 2014. Therefore, we have estimated Option 2b will require a major equipment
redesign for all of the equipment in 2010, followed by a minor redesign effort in 2011 for 50
percent of the equipment and in 2014 for 50 percent of the equipment. We have estimated that
each of these minor redesign efforts will cost !/2 of the major redesign costs estimated for the
proposal. Compared to the proposal, Option 2b increases the equipment redesign costs for the
75-175 hp category by approximately $70 million.
Equipment Variable Costs
We have estimated the impacts on equipment variable costs in the same manner as done for
engine variable costs by eliminated costs where appropriate and shifting them up a year or two or
back a year or two where appropriate. These changes increase the NPV through 2030 by $10
million relative to the equipment variable costs expected under the proposal.
Total Engine/Equipment Cost
Based on the estimations provided above, we have estimated the Option 2b will result in an
increase in the net-present value of the engine and equipment costs through 2030 of
approximately $300 million dollars.
12-67
-------
Draft Regulatory Impact Analysis
12.4.2.5 Option 2d
This option is described in Figure 12.1.2-5 in Section 12.1 of this draft RIA. Option 2d is the
same as the proposal but with the addition of a 0.30 g/bhp-hr NOx standard applied to engines in
the 25-75 hp category. These NOx standards would be phased in over three years from 2013
through 2015. Option 2d has the same fuel program as the proposal.
As discussed in Section 12.6.2.2.5, we do not believe a 0.30 g/bhp-hr NOx standard is
appropriate for engines in this power category, and therefore we have not provided a cost
estimate for Option 2d.
12.4.2.6 Option 2e
This option is described in Figure 12.1.2-6 in Section 12.1 of this draft RIA. Option 2e
requires the same PM standards and implementation schedule as the proposal, but there are no
Tier 4 NOx standards. Option 2e has the same fuel program as the proposal.
12.4.2.6.1 Fuel Costs for Option 2e
Option 2e has no changes in the fuel program compared to our proposal, therefore the
estimated fuel costs (e.g., the cents/gallon estimates) are no different from the proposal.
However, the elimination of the NOx standard does impact our fuel consumption impacts. As
discussed in Chapter 6.2.3 of this draft RIA (Engine Operating Costs), a combined NOx adsorber
- PM filter system can result in a net increase in fuel consumption of as much as one percent,
while a PM filter only system can result in a net increase in fuel consumption of as much as two
percent. Therefore, removal of the NOx control program results in an increase in the operating
costs of Option 2e as compared to our proposal. The net present value of this increase through
2030 is approximately $460 million.
12.4.2.6.2 Engine and Equipment Costs for Option 2e
Engine Fixed Costs
Option 2e requires no NOx related fixed costs as compared to our proposal. Eliminating
these costs reduces the cost of Option 2e relative to our proposal by approximately $130 million.
Engine Variable Costs
Option 2e removes any new NOx related variable costs from the program. The NPV of the
engine variable costs for Option 2e through 2030 is approximately $3.4 billion less than the
proposal. The annual engine variable costs are shown in Figure 12.4.1.1.2-1.
Equipment Fixed Costs
12-68
-------
Regulatory Alternatives
We have estimated that Option 2e has a minimal impact on the equipment redesign costs
compared to the proposal because the equipment manufacturers will be modifying their products
in order to add PM filters under Option 2e, and we believe there are minimal differences for
equipment manufacturers for packaging a NOx adsorber and a PM filter as compared to
packaging only a PM filter. However, the proposal does include a minor redesign cost estimate
for 50 percent of the equipment in the 75-750 hp categories in 2014 due to the implementation of
the 0.30 g/bhp-hr NOx standard for /^ of the engines. We have eliminated this cost from Option
2e. Compared to the proposal, Option 2e reduces the equipment redesign costs by approximately
$120 million.
Equipment Variable Costs
Option 2e removes any new NOx related variable costs from the program. The NPV of the
equipment variable costs for Option 2e through 2030 would be approximately $170 million less
than the proposal due to less sheet metal required to house exhaust emission control devices and
fewer bolts and brackets needed to secure those devices.
Total Engine/Equipment Cost
Based on the estimations provided above, we have estimated the Option 2e will result in a
decrease in the net-present value of the engine and equipment costs through 2030 of
approximately $3.8 billion dollars.
12.4.3 Other Options
12.4.3.1 Option 3
This option is described in Figure 12.1.2-7 in Section 12.1 of this draft RIA. Option 3
imposes no Tier 4 standards for engines used in above-ground mining equipment (AGME).
Option 2e has the same fuel program as the proposal.
12.4.3.1.1 Fuel Costs for Option 3
Option 2e has no changes on the cost of fuel relative to our proposal. However, the operating
costs for AGME are lower than in our proposal due to the elimination of PM filter maintenance
requirements and our estimate of a one percent fuel consumption increase due to PM filters. This
results in a decrease in the net-present value of Option 3 of approximately $80 million as
compared to the proposal through 2030.
12.4.3.1.2 Engine and Equipment Costs for Option 3
12-69
-------
Draft Regulatory Impact Analysis
Mining equipment is defined for this analysis as certain equipment types over 750hp as
described in Section 12.6.2.2.7 of this draft RIA. This includes equipment types such as
excavators, off highway trucks, wheel loaders, crawler tractor/dozers and off-highway tractors.
Engine Fixed Costs
Because these engines are used in equipment other than the AGME, Option 3 has no impact
on the engine fixed costs.
Engine Variable Costs
We have removed the variable costs associated with the Tier 4 proposal from the AGME
engines (i.e., PM filters and NOx adsorbers) to evaluate the impact of Option 3. The NPV of the
engine variable costs for Option 3 through 2030 is approximately $380 million less than the
proposal. The annual engine variable costs are shown in Figure 12.4.1.1.2-1.
Equipment Fixed Costs
Option 3 would remove any equipment redesign requirements for the AGME. This reduces
the costs of Option 3 by approximately $10 million relative to the proposal.
Equipment Variable Costs
We have eliminated the equipment variable costs for the >750 hp AGME for Option 3.
These changes reduce the NPV through 2030 by approximately $20 million relative to the
equipment variable costs expected under the proposal.
Total Engine/Equipment Cost
Based on the estimations provided above, we have estimated that Option 3 would result in a
decrease in the net-present value of the engine and equipment costs through 2030 of
approximately $410 million dollars.
12.4.3.2 Option 4
Option 4 is described in Figure 12.1.2-8 in Section 12.1 of this draft RIA. Option 4 is similar
to the proposal, but it requires locomotive and marine diesel fuel sulfur levels to be controlled to
a level of 15ppm in 2010.
12.4.3.2.1 Fuel Costs for Option 4
The total fuel costs from Chapter 7 of the Draft RIA comprising the refining and distribution
and additive costs for Option 4 are summarized in the following tables.
12-70
-------
Regulatory Alternatives
Table 12.4.3.2.1-1
Total Fuel Costs for Option 4 (cents per gallon)
Option
Nonroad, Locomotive
and Marine go to 15
ppm in 20 10
Specification
500 ppm NR, L&M
15 ppm NR, L&M
(total incl 2007)
Year
2007+
2010+
Refining
Costs
(eVgal)
2.2
4.6
Distribution &
Additive Costs
(eVgal)
0.3
0.4
Total
Costs
(eVgal)
2.5
5.0
Table 12.4.3.2.1-2
Net Operating Costs for Option 4 Incremental To The Proposal (millions)
(Net present values through 2030 at 3% discount rate)
Specification
1 5 ppm fuel
500 ppm fuel
Gallons
57,760
-54,910
Fuel costs*
$3,100
-$1,350
Net maintenance costs
$20
Total operating costs
$1,770
* Note that the incremental fuel costs presented here are calculated as: [proposal $/gal] multiplied by [proposal gallons]
minus the [option $/gal] multiplied by [option gallons]. This is not mathematically equivalent to the difference in gallons
multiplied by the difference in $/gal.
These fuel costs and other related operating costs (i.e., maintenance savings) result in an
increase in the net-present value of Option 4 of approximately $1.8 billion through 2030 as
compared to the proposal.
12.4.3.2.2 Engine and Equipment Costs for Option 4
Option 4 has the same engine standards program and implementation dates as the proposal,
and therefore the same costs.
12.4.3.3 Option 5a
This option is described in Figure 12.1.2-9 in section 12.1 of this draft RIA. Option 5a has
the same fuel program as the proposal but the engine/equipment program differs from the
proposal in that no new standards would be implemented for <75 horsepower engines.
12.4.3.3.1 Fuel Costs for Option 5a
Option 5a has no changes on the cost of fuel relative to our proposal. However, the operating
costs for <75 horsepower engines are lower than in our proposal due to the elimination of some
operating costs for these engines. Specifically, both the PM filter maintenance requirements and
12-71
-------
Draft Regulatory Impact Analysis
our estimate of a two percent fuel consumption increase due to PM filters would be eliminated
for all 25 to 75 horsepower engines. Also, CCV maintenance costs would be eliminated for all
engines <75 horsepower. Note that oil change maintenance savings would still be realized by
these engines as they would be under the proposal. The elimination of these operating costs
would result in a decrease in the net-present value of Option 5a of approximately $530 million as
compared to the proposal through 2030.
12.4.3.3.2 Engine and Equipment Costs for Option 5a
Engine Fixed Costs
Option 5a would eliminate the need for R&D expenditures described in Chapter 6 as CDPF-
only and DOC/engine-out R&D. It would also eliminate the need for tooling expenditures on
those engine platforms having sales strictly in the <75 horsepower category. This option would
also eliminate proposal-related certification costs for all <75 horsepower engines. Together,
these cost reductions would total $140 million relative to the proposal.
Engine Variable Costs
We have removed the variable costs associated with the Tier 4 proposal from the <75
horsepower engines (i.e., DOCs, PM filters, fuel systems, EGR systems, CCV systems) to
evaluate the impact of Option 5a on engine variable costs. The NPV of the engine variable costs
for Option 5a through 2030 is approximately $3 billion less than the proposal. The annual
engine variable costs are shown in Figure 12.4.1.1.2-1.
Equipment Fixed Costs
Option 5a would eliminate any equipment redesign requirements for <75 horsepower
equipment. This would reduce the equipment fixed costs of Option 5a by $80 million relative to
the proposal.
Equipment Variable Costs
We have eliminated the equipment variable costs for <75 hp equipment for Option 5a. These
changes reduce the NPV through 2030 by approximately $70 million relative to the equipment
variable costs expected under the proposal.
Total Engine/Equipment Cost
Based on the estimations provided above, we have estimated that Option 5a would result in a
decrease in the net-present value of the engine and equipment costs through 2030 of
approximately $3.3 billion dollars.
12-72
-------
Regulatory Alternatives
12.4.3.4 Option 5b
This option is described in Figure 12.1.2-10 in section 12.1 of this draft RIA. Option 5b has
the same fuel program as the proposal but the engine/equipment program differs from the
proposal in that the 2008 standards would remain in effect indefinitely and no CDPF forcing
standards would be implemented for 25 to 75 horsepower engines.
12.4.3.4.1 Fuel Costs for Option 5b
Option 5b has no changes on the cost of fuel relative to our proposal. However, the operating
costs for <75 horsepower engines are lower than in our proposal due to the elimination of some
operating costs for these engines. Specifically, both the PM filter maintenance requirements and
our estimate of a two percent fuel consumption increase due to PM filters would be eliminated
for all 25 to 75 horsepower engines. Note that, unlike Option 5a, CCV maintenance costs would
be incurred for all engines <75 horsepower; also, note that oil change maintenance savings would
still be realized by these engines as they would be under the proposal. The elimination of CDPF-
related operating costs would result in a decrease in the net-present value of Option 5b of
approximately $490 million as compared to the proposal through 2030.
12.4.3.4.2 Engine and Equipment Costs for Option 5b
Engine Fixed Costs
Option 5b would eliminate the need for R&D expenditures described in Chapter 6 as CDPF-
only R&D. It would also eliminate the need for CDPF-only tooling expenditures on those engine
platforms having sales strictly in the 25 to 75 horsepower category. This option would also
eliminate proposal-related certification costs for 25 to 75 horsepower engines beyond 2008.
Together, these cost reductions would total $60 million relative to the proposal.
Engine Variable Costs
We have removed the variable costs associated with the 2013 standards of the Tier 4 proposal
for 25 to 75 horsepower engines (i.e., PM filters, fuel systems, EGR systems) to evaluate the
impact of Option 5b on engine variable costs. The NPV of the engine variable costs for Option
5b through 2030 is approximately $1.9 billion less than the proposal. The annual engine variable
costs are shown in Figure 12.4.1.1.2-1.
Equipment Fixed Costs
Option 5b would eliminate any equipment redesign requirements for 25 to 75 horsepower
equipment associated with the 2013 standards. This would reduce the equipment fixed costs of
Option 5b by $40 million relative to the proposal.
12-73
-------
Draft Regulatory Impact Analysis
Equipment Variable Costs
For Option 5b, we have eliminated the equipment variable costs for 25 to 75 hp equipment
associated with the 2013 standards. These changes reduce the NPV through 2030 by
approximately $70 million relative to the equipment variable costs expected under the proposal.
Total Engine/Equipment Cost
Based on the estimations provided above, we have estimated that Option 5b would result in a
decrease in the net-present value of the engine and equipment costs through 2030 of
approximately $2.1 billion dollars.
12.5 Costs per Ton
For those Program Options where both inventory impacts and cost impacts were generated, it
was possible to calculate an incremental cost per ton relative to the proposal. These incremental
costs per ton for the Program Options are shown in Table 12.5-1. Note that the cost in Table
12.5-1 are expressed in billions of dollars and the emission reductions in tons of emissions. A
brief discussion of how the increment costs per ton were determined is presented below. Note
that there is no discussion of cost per ton for Options la, Ib, 2a, and 2d, since these Options were
determined to be impractical due to infeasibility or other significant concerns, and thus, no costs
were calculated.
12-74
-------
Regulatory Alternatives
Table 12.5-1
Incremental Cost per Ton for Alternatives
(Incremental to the Proposal)
Option
1
2b
2c
2e
o
6
4
5a
5b
COSt (Sbillion)
reductions (tons)
cost/ton ($/ton)
cost
reductions
cost/ton
cost
reductions
cost/ton
cost
reductions
cost/ton
cost
reductions
cost/ton
cost
reductions
cost/ton
cost
reductions
cost/ton
cost
reductions
cost/ton
NOx+NMHC
-
11,000
n/a
-
36,000
n/a
-
16,000
n/a
-$3.1
-5,407,00
$600
-$0.2
-751,000
$300
-
-
n/a
-$0.4
-334,000
$1,100
-$0.4
-333,000
$1,100
PM
$1.7
6,000
$265,000
$1.2
54,000
$22,000
$0.8
20,000
$41,000
$12.4b
1,126,000
$ll,000b
-$0.2
-30,000
$8,300
$0.6
9,000
$64,000
-$3.4
-209,000
$16,500
-$2.2
-121,000
$18,300
S02
-
-191,000
n/a
-
17,000
n/a
-
17,000
n/a
-
-
n/a
-
-
n/a
$1.2
114,000
$10,300
-
-
n/a
-
-
n/a
a Qualitative analysis only of options la, Ib, 2a, and 2d due to the options being impractical due to infeasibility or
other significant concerns.
b In the analysis of the proposed program, the cost for 15ppm fuel is split 50/50 between NOx and PM. For
option 2e, with no NOx program, all of the 15 ppm fuel cost is attribute to PM resulting in a new cost
effectiveness estimate for PM. The PM cost here is the proposal total cost less the proposal SOx cost less the
NOx+NMHC savings of Option 2e. For 2e we present the incremental cost effectiveness of the lost NOx tons and
the new cost effectiveness of the Tier 4 PM tons.
12-75
-------
Draft Regulatory Impact Analysis
12.5.1 Incremental Cost per Ton for Option 1
The incremental cost per ton for the lost SO2 tons due to delaying the introduction of 500
ppm fuel by one year should be roughly the same as the long term SO2 cost per ton of the 500
ppm fuel program. The cost per ton of SO2 for that program is $90 (see Table 8.7-1 of this draft
RIA). This value is so low because the costs of the 500 ppm fuel program are estimated to be
essentially zero due to large maintenance savings expected to occur. In other words, the
maintenance savings associated with the 500 ppm sulfur fuel nearly offset the cost of the fuel.
See Section 8.4 of this draft RIA for more detail.
The fundamental goal of Option 1 is to introduce new PM controls earlier than the proposal.
Therefore, the incremental costs associated with this option - for 15 ppm sulfur fuel two years
earlier than the proposal and for PM technology on >75 horsepower engines two years earlier
than the proposal - are all attributed to PM. These costs were presented in section 12.4.1.1 as
$720 million for fuel and other operating costs and $990 million for engines/equipment for a
total of roughly $1.7 billion. The PM tons gained, presented in Table 12.2.3-2, would be 6,000
tons. This results in an incremental cost per ton of PM (i.e., incremental to the proposal) of
$265,000.
For NOx+NMHC, the small change in the emission reduction is due to the implementation of
the transient test two years early. The feasibility and cost for industry to meet the transient test
two years early is not made since this aspect of the option is not a primary consideration in
considering this approach. No cost estimate was made for the additional development cost
necessary to meet a transient test two years early, so no estimate of the cost per ton of
NOx+NMHC is made.
In summary, this alternative gives up virtually free SO2 reductions to gain very expensive PM
tons ($265,000 per ton).
12.5.2 Incremental Cost per Ton for Option 2b
The goal of Option 2b is to introduce new PM controls earlier than the proposal. Therefore,
the incremental costs associated with this option - for 15 ppm sulfur fuel one year earlier than the
proposal and for PM technology on >25 horsepower engines one year earlier than the proposal -
are all attributed to PM. Section 12.4.2.3 discussed the costs of this option as $540 million for
fuel and other operating costs and $640 million for engines/equipment for a total of roughly $1.2
billion more than the proposal. Table 12.2.3-2 shows that Option 2b gets 54,000 more tons of
PM reduction than does the proposal. This results in an incremental cost per ton of PM of
$22,000.
For SO2 and NOx+NMHC, this option has incidental reductions beyond the proposal due to
the sulfur difference between 500 ppm and 15 ppm in 2009 (therefore a larger SO2 reduction) and
12-76
-------
Regulatory Alternatives
the one year early introduction of the transient test procedures (therefore a larger NOx+NMHC
reduction). No cost estimate was made for the additional development cost necessary to meet a
transient test one year early, so no estimate of the cost per ton of NOx+NMHC is made.
In summary, this alternative gets early PM reductions but has to pay more than double the
rate paid under the proposal ($22,000 per ton vs. $9,300 per ton).
12.5.3 Incremental Cost per Ton for Option 2c
The fundamental goal of Option 2c is to introduce new PM controls earlier than the proposal.
Therefore, the incremental costs associated with this option - for 15 ppm sulfur fuel one year
earlier than the proposal and for PM technology on 175 to 750 horsepower engines one year
earlier than the proposal - are all attributed to PM. The costs were presented in section 12.4.2.4
as $530 million for fuel and other operating costs and $300 million for engines/equipment, while
Table 12.2.3-2 shows the foregone PM reductions to be 20,000 tons. This results in an
incremental cost per ton of PM of $41,000.
This option has incidental SO2 and NOx+NMHC reductions beyond the proposal due to the
sulfur difference between 500 ppm and 15 ppm in 2009 (therefore a larger SO2 reduction) and the
one year early introduction of the transient test procedures (therefore a larger NOx+NMHC
reduction). No cost estimate was made for the additional development cost necessary to meet a
transient test one year early, so no estimate of the cost per ton of NOx+NMHC is made.
In summary, this alternative gets early PM reductions but has to pay more than three times
the rate paid under the proposal ($41,000 per ton vs. $9,300 per ton).
12.5.4 Incremental Cost per Ton for Option 2e
Option 2e reduces compliance costs by eliminating new NOx standards. This option presents
legal concerns since we would be giving up achievable NOx emission reductions solely for cost
reasons, and cost considerations are not to be the driving factor in making decisions under CAA
section 213(a)(3); rather, the overriding goal of this CAA section is air quality (see, for example,
Husqvarna AB v. EPA. 254 F. 3d 195, 200 (D.C. Cir. 2001)). Our purpose here, however, is not
to address the legality of such a program, but rather to analyze it's merits. Therefore, for the sake
of illustration, while the resultant compliance costs would be lower than the proposal, all would
be attributed to PM control. The discussion in section 12.4.2.6 noted that the net present value of
Option 2e costs would be roughly $3.3 billion dollars less than the proposal ($3.8 billion less for
engines/equipment but $460 million more for fuel and other operating costs) while giving up
over five million tons of NOx reductions. The cost per ton of these foregone NOx emissions
(i.e., dollars saved divided by tons given up) can be estimated at $600 per ton.
For PM and SO2, there is no change in the reduction realized under this alternative since
neither the fuel program nor the new PM standards are different than the proposal. However, if a
12-77
-------
Draft Regulatory Impact Analysis
new cost per ton estimate for the whole program were made for PM, the cost effectiveness would
change since the total cost of the 15 ppm sulfur reduction (i.e., sulfur reduction to enable
technology) would only be applied to PM. The new cost per ton estimate for PM under this
alternative would be $11,000 (as compared to $9,300 under the proposal). Note that this $11,000
cost per ton represents a cost per ton for such a program, not an incremental cost per ton relative
to the proposal. For SO2, there would be no incremental cost per ton since both costs and SO2
reductions would be equal to the proposal.
In summary, this alternative gives up substantial, feasible NOx reductions at $600 per ton in
the same timeframe as our Tier 2 passenger car program (NOx+NMHC cost per ton >$2,000)
and the HD 2007 program (>$2,000 per ton). As a PM and SO2 program, this option is an
attractive control option, although PM tons are more expensive than they are under the proposal.
12.5.5 Incremental Cost per Ton for Option 3
This option is basically the same as the proposal except that mining equipment >750
horsepower is exempted from all engine standards. As such, this option roughly estimates the
per engine, or equipment, cost per ton for adding or subtracting mining equipment (we do not
address here the legal basis, or lack of one, for this option). The cost savings realized for this
approach include variable costs for engine hardware, and fixed and variable equipment costs for
mining equipment. These savings assume that other nations would also adopt this approach,
otherwise no savings would be realized for equipment fixed costs because one product would
likely be made worldwide (the engine variable cost savings would still be realized). The savings
also include less fuel consumed by these pieces of equipment because without the PM trap they
would not incur the one percent fuel economy impact and no PM trap maintenance for these
pieces of equipment.
Section 12.4.3.1 presented the incremental costs of this option as $80 million saved on fuel
and other operating costs and $410 million saved on engine/equipment costs for a total increment
of $490 million saved. However, these savings come at the expense of lower NOx+NMHC and
lower PM reductions. Table 12.2.3-2 shows the foregone NOx+NMHC and PM reductions to be
751,000 and 30,000 tons, respectively. Assuming a perfect 50/50 split of costs for these
pollutants results in an incremental cost per ton of PM lost of $8,300 an incremental cost per ton
of NOx+NMHC lost of $300.
In summary, this alternative gives up substantial feasible and relatively inexpensive
(compared to other mobile source programs) NOx+NMHC and PM tons.
12.5.6 Incremental Cost per Ton for Option 4
Option 4 leaves the engine program the same as the proposal but includes locomotive and
marine fuel in the requirement for 15 ppm fuel. PM reductions are realized due to the reduced
engine out sulfur to sulfate conversion from existing locomotive and marine engines. SO2
12-78
-------
Regulatory Alternatives
reductions are realized due to the reduced engine out SO2 from the fuel (98% of the fuel sulfur is
exhausted from the engine as SO2).
The incremental costs for this option were presented in section 12.4.3.2 as $1.8 billion for
fuel and other operating costs with no costs for engines/equipment. The PM reductions gained
are shown in Table 12.2.3-2 as 9,000 tons and the SO2 reductions gained are shown as 114,000
tons. To estimate the cost per ton reduction for this alternative, one-third of the incremental 15
ppm fuel cost is attributed to PM with the balance being attributed to SO2. The resulting
incremental cost per ton for PM is $64,000 and the incremental cost per ton of SO2 is $10,300.
In the absence of new engine standards enabled by the 15 ppm sulfur fuel (i.e., for locomotive
and marine engines), the cost per ton of emissions reduction for this option does not look as
favorable as some of the other options listed here. However, we would anticipate that a fuel
program such as this would be done in conjunction with new technology forcing emission
standards enabled by this clean fuel. In fact, as discussed in Section 6.C of the preamble, it is our
intention to develop an Advanced Notice of Proposed Rulemaking (ANPRM) for such a control
option in the near future. Were this option to include new the control technology enabled by 15
ppm sulfur fuel, we believe it is likely that the program would look very favorable. The cost per
ton estimates for Option 3 would likely be a good surrogate for an estimation of such a program
for locomotive and marine engines (i.e., the program would be very cost effective compared to
other PM emission control programs).
12.5.7 Incremental Cost per Ton for Option 5a
This option is similar to the proposal except that no new standards would be implemented for
<75 horsepower engines. In other words, engines <50 horsepower would remain at Tier 2 levels
and 50 to 75 horsepower engines would remain at Tier 3 levels. As such, this option roughly
estimates the per vehicle cost per ton for adding or subtracting the <75 horsepower elements of
the engine program. The cost savings realized for this approach include variable costs for engine
hardware and equipment hardware in the <75 horsepower category, and fixed costs for engine
R&D, tooling, and certification, and equipment redesign in the <75 horsepower category. These
savings assume that other nations would also adopt this approach, otherwise no savings would be
realized for equipment fixed costs because one product would likely be made worldwide (the
engine variable cost savings would still be realized). The savings also include less fuel
consumed by 25 to 75 horsepower pieces of equipment because without the PM trap they would
not incur the two percent fuel economy impact associated with the PM trap. Further, 25 to 75
horsepower pieces of equipment would not incur the PM trap related maintenance costs and all
engines <75 horsepower would not incur the CCV maintenance costs because CCV systems
would not be required.
Section 12.4.3.3 presented the incremental costs of this option as $530 million saved on fuel
and other operating costs (i.e., lower operating costs) and $3.3 billion saved on engine/equipment
costs for a total increment of $3.8 billion saved. However, these savings come at the expense of
12-79
-------
Draft Regulatory Impact Analysis
lower NOx+NMHC and lower PM reductions. Table 12.2.3-2 shows the foregone NOx+NMHC
and PM reductions to be 334,000 and 209,000 tons, respectively. Attributing these costs to
NOx+NMHC and PM according to the cost allocations shown in Table 8.1-2 results in an
incremental cost per ton of PM lost of $16,500 and an incremental cost per ton of NOx+NMHC
lost of $1,100.
In summary, this alternative gives up substantial feasible (compared to other mobile source
programs) NOx+NMHC and PM tons.
12.5.8 Incremental Cost per Ton for Option 5b
This option is similar to the proposal except that the 2008 standards for <75 horsepower
engines would remain in effect indefinitely and no new PM trap forcing standards would be
implemented for 25 to 75 horsepower engines nor new EGR forcing NOx standards for 25 to 50
horsepower engines. As such, this option roughly estimates the per engine, or equipment, cost
per ton for adding or subtracting the 2013 standards for 25 to 75 horsepower engines (we do not
address here the legal basis, or lack of one, for this option). The cost savings realized for this
approach include variable costs for engine hardware and equipment hardware associated with the
2013 standards, and fixed costs for engine R&D, tooling, and certification, and equipment
redesign associated with the 2013 standards. These savings assume that other nations would also
adopt this approach, otherwise no savings would be realized for equipment fixed costs because
one product would likely be made worldwide (the engine variable cost savings would still be
realized). The savings also include less fuel consumed by 25 to 75 horsepower pieces of
equipment because without the PM trap they would not incur the two percent fuel economy
impact associated with the PM trap. Further, 25 to 75 horsepower pieces of equipment would not
incur the PM trap related maintenance costs.
Section 12.4.3.4 presented the incremental costs of this option as $490 million saved on fuel
and other operating costs (i.e., lower operating costs) and $2.1 billion saved on engine/equipment
costs for a total increment of $2.6 billion saved. However, these savings come at the expense of
lower NOx+NMHC and lower PM reductions. Table 12.2.3-2 shows the foregone NOx+NMHC
and PM reductions to be 333,000 and 121,000 tons, respectively. The foregone NOx+NMHC
reduction relative to the proposal is almost identical for Option 5b as it was for Option 5a
although it is slightly lower due to the NMHC reduction realized by the addition of DOCs under
Option 5b that would not be realized under Option 5a. Attributing these costs to NOx+NMHC
and PM according to the cost allocations shown in Table 8.1-2 results in an incremental cost per
ton of PM lost of $18,300 an incremental cost per ton of NOx+NMHC lost of $1,100.
In summary, this alternative gives up substantial feasible (compared to other mobile source
programs) NOx+NMHC and PM tons.
12-80
-------
Regulatory Alternatives
12.6 Summary and Assessment of Alternative Program Options
Having presented each of the alternative Program Options and their associated inventory
impacts, benefits, costs, and cost-effectiveness in the preceding sections, we here provide a
comparative summary of these Options and an assessment of the rationale, issues, and feasibility
of each one.
12.6.1 Summary of Results of Options Analysis
As we developed the program we are proposing in today's Notice of Proposed Rulemaking,
we evaluated a number of alternative Program Options with regard to the scope, level, and timing
of the standards to ensure that we were looking at the full range of possible control options.
Table 12.6.1-1 contains a summary of the alternative Program Options we considered and the
expected emission reductions, costs, and monetized benefits associated with them in comparison
to the proposal. These Program Options cover a broad range of possible approaches and serve to
provide insight into the many other program design alternatives not expressly evaluated further.
While we are interested in comments on all of the alternatives presented, we are especially
interested in comments on two alternative scenarios that we believe merit further consideration in
developing the final rule; a primary one-step program (Option 1), and a requirement that the
second step of sulfur control to 15 ppm in 2010 apply to locomotive and marine diesel fuel in
addition to nonroad diesel fuel (Option 4).
12-81
-------
Table 12.6.1-1
Summary of Alternative Program Options
(Incremental to the Proposal)
Option
Fuel Standards
Engine Standards
Estimated Relative
Inventory Impacts0 (NPV
tons thru 2030; 3%
discount)
Estimated
Cost Impacts -
SBillion
(NPV thru
2030; 3%)
Estimated
Benefits Stream -
$Billione
(NPV thru 2030;
3%)
Proposal (inventory impacts, costs and benefits reported below for the options are compared to the proposal)
. 500 PPM in 2007 for
NR, loco/marine
• 15ppmin2010NRonly
• >25hp: PM AT introduced 20 13
• >75 hp: NOx AT introduced and phased-in
2011-2013
• <25 hp: PM stds in 2008
• 25-75 hp: PM stds in 2008 (optional for 50-75
hp)
Relative to baseline:
1,126,000PM
4,952,000 SO2
5,59 1,000 NOx+NMHC
$16.7
$550b
1-Step Fuel Options
1
la
Ib
• 15ppmin2008forNR
only
• 500 ppm in 2008 for
loco/marine
• 15 ppm in 2008 for NR,
loco/marine
• 1 5 ppm in 2006 for NR,
loco/marine
• < 50 hp: PM stds only in 2009
• 25-75 hp: PM AT stds and EGR or equivalent
NOx technology in 2013; no NOx AT
• >75 hp: PM AT stds phasing in beginning in
2009; NOx AT phasing in beginning in 20 1 1
• PM AT introduced in 2009- 10
• NOx AT introduced in 20 1 1 - 1 2
Same as la
6,000 PM
-191,000 SO2
11, 000 NOx+NMHC
129,000 PM
-63,000 SO2
1,843,000 NOx+NMHC
$1.7d
a
$.2b
$59
a
2-Step Fuel Options
2a
2b
2c
Same as proposal except -
• 500 ppm in 2006 for NR,
loco/marine
Same as proposal except -
• 15 ppm in 2009 for NR
Same as proposal except -
• 15 ppm in 2009 for NR
Same as proposal
Same as proposal except -
• Move PM AT up 1 year for all engines > 25 hp
(phase in starts 20 10)
Same as proposal except -
• Move PM AT up 1 year for all engines 175-750
hp (phase in starts 2010)
18,000 PM
228,000 SO2
0 NOx+NMHC
54,000 PM
17,000 SO2
36,000 NOx+NMHC
20,000 PM
17,000 SO2
16,000 NOx+NMHC
a
$1.2d
$0.8d
$7b
$16b
$6b
-------
Option
2d
Fuel Standards
• Same as proposal
Engine Standards
Same as proposal except -
• Phase-in NOx AT for 25-75hp beginning in
2013
Estimated Relative
Inventory Impacts0 (NPV
tons thru 2030; 3%
discount)
0PM
OSO2
751,OOONOx+NMHC
Estimated
Cost Impacts -
SBillion
(NPV thru
2030; 3%)
a
Estimated
Benefits Stream -
$Billione
(NPV thru 2030;
3%)
$10b
Other Options
3
4
5a
5b
• Same as proposal
Same as proposal except -
• loco/marine fuel to 15
ppmin2010
• Same as proposal
• Same as proposal
Same as proposal except -
• Mining equipment over 750 hp left at Tier 2
Same as proposal
Same as proposal except-
• No Tier 4 standards <75 hp
Same as proposal except-
• No new <75hp standards after 2008
(i.e.,noCDPFsin2013)
-30,000 PM
OSO2
-75 1,000 NOx+NMHC
9,000 PM
1 14,000 SO2
0 NOx+NMHC
-209,000 PM
OSO2
-334,000 NOx+NMHC
-121,000PM
OSO2
-333,000 NOx+NMHC
-$0.5
$1.8
-$3.8
-$2.6
-$18b
$6b
-$70
-$43
aQualitative analysis only. Option is impractical due to infeasibility or other significant concerns. See the draft RIA for a detailed discussion
bBy benefits transfer method
cNet Present (2004) Value impacts through 2030, using a 3% discount rate, relative to the proposed program. Positive values mean that the Option produces
greater emission reductions from baseline than the proposed program.
dCost estimates do not include the costs due to potential for limited product offerings and market disruptions in the engine/equipment and/or fuel markets. See
Section V of this preamble and the draft FIA for a detailed discussion.
eBenefits do not include CO, VOC, air toxics, ozone, and PM welfare benefits. See Section V.F of this preamble and the draft RIA for additional discussion.
-------
Draft Regulatory Impact Analysis
12.6.2 Discussion of Rationale, Issues, and Feasibility Assessment of Options
Each of the Program Options defined and presented in Section 12.1 is discussed here in
terms of the rationale for considering the option, issues surrounding the option, and our
assessment of the feasibility of the option. Inventory impacts for each option are presented in
Section 12.2, health and environmental benefit comparisons are presented in Section 12.3, and
comparative cost and cost-effectiveness for these Program Options is presented in Sections 12.4
and 12.5, respectively.
12.6.2.1 One-Step Options
12.6.2.1.1 Option 1
In defining Option 1 we focused on designing a program with long-term engine standard
levels identical to those being proposed, implemented as early as possible under a one-step
approach to nonroad fuel desulfurization, and structured such that both engine and fuel
requirements and timing would have a high likelihood of being technologically feasible. In doing
so, we recognized the need to account for a number of factors:
• The need for 15 ppm maximum sulfur nonroad diesel fuel to enable highly-sulfur
sensitive emission control technology on nonroad engines,
The need to coordinate refinery investments to desulfurize nonroad diesel fuel with
similar efforts already mandated for this industry for highway diesel fuel and gasoline in
the same general timeframe,
• The need to provide adequate lead time for the migration of relevant emission control
technologies from the highway sector,
• The need to provide adequate stability periods for Tier 3 standards and for Tier 2
standards for engines under 50 hp, and
• The workload of engine and equipment manufacturers in preparing hundreds of engine
models and thousands of machine models for Tier 4 compliance.
The resulting Option 1 program design is reflected in Figure 12.1.1-1. The one-step fuel
change occurs in 2008. This is one year later than the proposal's first step, but it provides 15
ppm maximum sulfur nonroad diesel fuel two years earlier than the proposal's second step does.
In Option 1, locomotive and marine diesel fuel is desulfurized to 500 ppm in 2008 as well, one
year later than under the proposal.
These fuel program differences yield both positive and negative impacts on relative
emissions reductions. Early sulfate PM reductions in the existing fleet would be delayed a year
such that no PM reductions would occur in 2007. The Tier 4 PM standard for <25 hp engines
and the transitional PM standard for 25-50 hp engines would both be delayed a year to 2009, and
the transitional PM standard for 50-75 hp engines would be eliminated. These differences come
about because these PM standards depend on the availability of nonroad diesel fuel with sulfur
12-84
-------
Regulatory Alternatives
levels below 500 ppm, which under the one-step fuel option does not happen until the shift to 15
ppm fuel in 2008. This delays any potential PM standards to 2009 at earliest and, in the case of
50-75 hp engines which have new Tier 3 standards taking effect in 2008, makes it unworkable to
adopt the transitional standard at all because these engines (and the machines using them) would
need to be redesigned for new emission standards in 2008, again in 2009, and yet again in 2013,
as discussed below. Even we were to have the transitional standard take effect in 2010 or 2011
instead of 2009 in order to pace the redesign process more evenly, these rapid redesigns would
likely be unacceptably costly.
The most important impact of this Option 1 fuel regulation schedule is the potential for
high-efficiency exhaust emission control to occur as early as the 2009 model year. Even
accounting for the other factors listed above, such as the need to provide adequate lead time for
the migration of relevant emission control technologies from the highway sector, PM filters can
be introduced earlier on a large segment of nonroad diesel engines under Option 1. Our
consideration of these factors in setting aNOx standard schedule, particularly the need for
technology migration lead time, leads us to conclude, however, that the earlier availability of 15
ppm sulfur fuel would not lead to earlier implementation of NOx adsorber technology. The
completion of the NOx technology phase-in for the highway sector will occur in 2010. We
believe that 2011 would remain as the earliest model year that this technology could begin to
phase into the nonroad diesel sector, as proposed.
Although earlier introduction of PM filter technology is made possible by the earlier
availability of enabling fuel, the need for adequate lead time to transfer PM filter technology
from the highway sector to the wide variety of nonroad diesel applications, and the need for a
coordinated PM/NOx phase-in to avoid large and costly redesign workload burdens, result in a
somewhat complex phase-in schedule for Option 1. (For analysis of an option that does not take
much account of this constraint, see section 12.6.2.1.2 on Option la below.) Specifically, we
would phase in standards as indicated in Figure 12.1.1-1. Engines in the 175-750 hp category
would be subject to the 0.01 g/bhp-hr PM filter-based standard in 2009, when the regulated fuel
becomes available, but only for 50% of a manufacturer's U.S.-directed production. The other
50% would meet this PM standard beginning in 2011, concurrent with initiation of the 0.30
g/bhp-hr NOx adsorber-based standard for 50% of production. This makes it possible to
optimize the PM filter technology transfer process by focusing on the most "highway-like"
engine platforms in this power category first, and also to reduce the engineering workload by
redesigning many engine families, comprising half of production, to meet PM and NOx standards
simultaneously in the 2011 model year. The NOx phase-in would then be completed in 2014, as
under the proposal, allowing five years of stability for the 50% of production redesigned for PM
control in 2009 before the redesign for NOx in 2014. All in all, this approach increases the
opportunity for a manufacturer to coordinate product redesign strategies for new standards with
product redesign cycles driven by marketing and other concerns, while still achieving substantial
PM filter introduction in 2009.
12-85
-------
Draft Regulatory Impact Analysis
The phase-in for engines in the 75-175 hp category would follow the same pattern, but
one year later, to account for the need to spread the workload, and also to provide additional time
to transfer highway technology to engines in this category, as is done under the proposal and in
past tiers of standards. Note that this approach to phasing in standards helps to optimize the
redesign strategies to reduce workload burden, but not as well as under the proposal. It also does
not fully mitigate concerns over shortened Tier 3 stability periods under Option 1, reduced to two
years for some engines (50% of 75-100 hp engines).
For engines over 750 hp, we have retained the proposal's 50% phase-in approach in
2011-2013. We believe that decoupling the PM and NOx phase-in for this category by
implementing the PM standard one or two years earlier could potentially create severe problems.
These engines typically are used in low sales volume machines that have long normal product
cycles. Early PM control would not only result in two Tier 4 redesigns steps for some of these
engines and machines, but would also shorten the Tier 2 stability period.
The implementation issue is somewhat simpler for engines below 75 hp because of the
lack of NOx-adsorber based standards. For the engines below 25 hp it is simplified even further
by the lack of PM filter-based standards. These would be subject to a non-PM filter-based
standard in 2009, when the regulated fuel becomes available. (See Section 4.1.1 for a discussion
of how fuel sulfur degrades the efficiency of diesel oxidation catalysts used for PM control.) We
believe that PM filter technology for 25-75 hp engines is constrained primarily by highway
technology transfer considerations, and thus would be implemented in Option 1 in 2013 as under
the proposal. This is late enough that it would still make sense under Option 1 to adopt
transitional PM standards as in the proposal, even with the one-year delay to 2009 caused by the
delay in fuel regulation. However, transitional standards would not be applied under this option
to the 50-75 hp engines in this category because of the conflict with Tier 3 timing discussed
above.
12.6.2.1.2 Option la
The analysis for Option la shows what added environmental benefits would be possible
under a very aggressive approach to engine standard-setting, compared to the proposal and to the
more technologically feasible Option 1. On the fuel side, Option la would go further than the
proposal and Option 1 by regulating locomotive and marine diesel fuel to the 15 ppm maximum
sulfur level along with other nonroad diesel fuel in 2008. Issues associated with regulating
locomotive and marine fuel to 15 ppm sulfur are discussed in section 12.6.2.2.8. Otherwise the
approach to fuel regulation is identical to that taken in Option 1.
The Option la approach to engine standards applies the 0.01 g/bhp-hr PM standard to
engines of all sizes: in 2009 for engines >175 hp and in 2010 for engines <175 hp. This is 2-5
years earlier than under the proposal for engines above 75 hp. For 25-75 hp engines, it is three
years earlier and at a 50% lower emission level (0.01 compared to 0.02 g/bhp-hr), but without the
proposed 2008 transitional PM standard that is tied to regulating fuel in two steps. For engines
12-86
-------
Regulatory Alternatives
<25 hp, the Option la approach to PM standard-setting is two years later than the proposed Tier
4 standard but at a PM filter-based level 97% lower than the proposed 0.30 g/hp-hr level.
Although Option la's two-year phase-in of the PM standard in 2009-2010 follows the logic that
fuel desulfurization must precede the application of PM filters, and direct!onally addresses the
critical workload and technology transfer issues detailed in section HI of the proposal, we do not
believe that this analytical option is technologically feasible with respect to PM standard-setting,
for reasons discussed in Chapter 4 and in section in of the preamble to the proposal.
For NOx control, Option la applies a similar 2-year phase-in: in 2011 for engines >175
hp and in 2012 for engines <175 hp. These later NOx start dates compared to those of the Option
la PM standards direct!onally reflect the need for additional development time after similar
standards fully phase in for heavy-duty highway diesel engines in 2010, in order to transfer this
technology to nonroad applications. This phase-in of NOx standards results in an Option la Tier
4 program with NOx adsorber-based standards fully phased in three years earlier than under the
proposal for engines >175 hp, and two years earlier than under the proposal for 75-175 hp
engines, although for all of these engines >75 hp, the proposal begins phasing in the NOx
standard (at a 50% of sales level) in the same year that Option la begins its NOx control
requirement (at 100%). For engines <75 hp, Option la's 0.30 g/bhp-hr NOx standard would
yield over 90% better NOx control than the non-NOx adsorber-based standards under the
proposal. As concluded above for PM control, however, we do not believe that this analytical
option is technologically feasible with respect to NOx standards-setting, for reasons discussed in
Chapter 4 and in section in of the preamble to the proposal.
One additional major complication created by Option la's focus on getting PM control as
early as possible is the very large additional workload, especially for equipment manufacturers,
created by having two major Tier 4 redesign steps coming two years apart for every engine, first
for PM in 2009-2010, and then for NOx in 2011-2012. Moreover, these major redesigns follow
quite closely on the major engine and equipment redesign effort in 2006-2008 for Tier 3 (and
Tier 2 for engines >750 hp), with Tier 3 stability periods as short as 2 years for many engines.
Stability periods this short would be unprecedented in EPA mobile source programs for
technology-forcing standards such as those required by Tier 3 and the proposed Tier 4.
Furthermore, the Option la approach would result in an overlap of implementation schedules for
nonroad Tier 4 standards and the highway HDDE emission control program that phases in over
2007-2010. A number of engine manufacturers participate in both markets, and thus would
likely be certifying and marketing new highway engines in 2009 and 2010, concurrent with the
turnover of their entire nonroad engine product line to meet the new nonroad diesel PM standard
in the same years. This could put a serious strain on their engineering resources and add to the
cost of the program, potentially to the extent of making the program infeasible.
Based on the above discussion and the analyses performed for this option described in
this Chapter, we conclude that Option la would not be appropriate for proposal. In particular, we
do not believe that the set of engine standards under Option la would be technologically feasible.
12-87
-------
Draft Regulatory Impact Analysis
12.6.2.1.3 Option lb
This alternative, a variation of Option la, would implement a 15 ppm sulfur cap for all
nonroad, locomotive, and marine diesel fuel starting on June 1, 2006 for refiners and importers.
The rationale behind doing so would be to move up the program for NRLM fuel to coincide with
the initial implementation of the 15 ppm cap for highway diesel fuel. The engine standards
would be unchanged in comparison to Option la. They would still be initiated starting with the
2009 model year for PM and 2011 model year for NOx. Thus, this alternative, relative to Option
la, would be a pure fuel program, moving up the 15 ppm sulfur standard by two years.
We have examined this alternative from a number of angles relative to the proposal and
Option la:
1) The need for further sulfur dioxide and PM emission reductions in this timeframe,
2) Its impact on the desulfurization technology used to meet the 15 and 500 ppm caps,
3) The leadtime available for refiners to meet the 15 and 500 ppm standards in 2006,
4) The impact on the supply of highway diesel fuel, and
5) The potential cost-effectiveness and cost-benefit of the additional sulfur control.
Because this option only affects fuel sulfur content and not engine emission standards, the
only air quality benefits are reduced sulfur dioxide and sulfate PM emissions. The need for these
reductions is just as great in 2006-7 as they are in the 2008-2010 timeframe. As outlined in
Chapter 2, ambient fine PM levels are currently above the NAAQS for fine PM. Ambient fine
PM levels in 2006-2007 are more likely to be near current levels than those in 2008-2010, given
that less time is available for current emission controls, like the 2007 highway diesel program, to
take effect. Thus, moving up the 15 ppm standard should be considered for its air quality
impacts. These emission reductions and their resulting benefits are shown in Sections 12.2.2 and
12.3, respectively.
However, a 2006 implementation date for a 15 ppm sulfur cap on all NRLM fuel does not
appear to allow sufficient leadtime for refiners to design and construct new desulfurization
equipment. Leadtime for the proposed 2007 500 ppm NRLM diesel fuel cap was evaluated in
Chapter 5.3. There it was determined that refiners needed 2.25-3.25 years after the final rule in
order to design and construct new hydrotreaters to produce 500 ppm fuel. This analysis
considered the fact that the 500 ppm cap could be met using well established, conventional
hydrotreating technology. More time would be required to design and construct equipment to
produce 15 ppm nonroad diesel fuel. Even ignoring this additional time, a 2006 implementation
date would only allow refiners facing the minimum required leadtime enough time to meet the
one-step fuel standards. A 2006 implementation date would allow no time for the generation of
early sulfur reduction credits which might allow some refiners additional time to meet the one-
step fuel standards. Also, it is difficult to project that any refiners would be able to meet these
standards early even if the program granted such credits. Thus, we must conclude that the 2006
one-step option would not be technically feasible due to insufficient leadtime for refiners and
12-88
-------
Regulatory Alternatives
importers to meet the 2006 fuel sulfur standards. For this reason, we were unable to develop any
reliable cost estimates for this option.
In addition to leadtime concerns, applying a 15 ppm sulfur cap for NRLM diesel fuel in
2006 to coincide with the implementation of the highway diesel fuel program would raise
workload concerns for the industry, impacting not only the successful implementation of this
rulemaking, but also the highway rule. A 15 ppm cap on NRLM fuel in 2006 could have
seriously adverse consequences on the supply of highway diesel fuel and thus, the successful
implementation of the 2007 highway diesel fuel program. We added the temporary compliance
option to the 2007 highway diesel fuel program to ease implementation in 2006 and assure
sufficient supply of highway diesel fuel. The temporary compliance option allows 20% of
highway diesel fuel to remain at 500 ppm until 2010. Starting a 15 ppm NRLM cap in 2006
would essentially negate the benefit of the temporary compliance option, as the volume of high
sulfur nonroad diesel fuel is roughly 15% of highway diesel fuel volume. We have not evaluated
the degree that highway fuel supply would be negatively impacted, however, the impact would be
directionally negative.
Since this option is not feasible, we were not able to derive costs, and therefore cost per
ton or cost/benefit values that correspond to it. However, under the hypothetical where leadtime
was not a constraint on feasibility, we can still provide some general assessments. Applying a 15
ppm cap in 2006 for all NRLM fuel would reduce refiners' ability to utilize lower cost, advanced
desulfurization technologies to meet the 15 ppm nonroad diesel fuel sulfur cap. This is discussed
in Chapters 5 and 7 above. In 2006, we would project that few if any refiners would utilize
advanced technologies. This would increase the cost of 15 ppm fuel by roughly 10% compared
to Option 1 where 40% of refiners are estimated to be able to take advantage of these
technologies and more than 20% in comparison to today's proposal. This impact on cost would
last for roughly 15 years, or as long as this equipment was in use. Other than this increase in
costs, the incremental cost effectiveness and cost-benefit ratio would be expected to be of a
similar magnitude to that for option 4 as discussed in chapter 12.6.2.2.8 below. Thus, a rough
estimate suggests that if this option were feasible, the benefits would still be substantial and costs
would be reasonable, but not nearly as well as is true for the proposal or a long term 500 ppm
cap.
12.6.2.2 Two-Step Options
12.6.2.2.1 Proposed Program
The proposed program is included in this Chapter for the purpose of comparison with the
alternative regulatory options analyzed. We believe it to be a feasible program that meets the
Agency's requirements under the Clean Air Act. The proposed program is described in detail in
the preamble to the proposal and the feasibility of the proposed engine and fuel requirements is
discussed in detail in Chapters 4 and 5 of the draft RIA.
12-89
-------
Draft Regulatory Impact Analysis
12.6.2.2.2 Option 2a
This alternative would move up the 500 ppm sulfur cap for NRLM diesel fuel one year, to
June 1, 2006 for refiners and importers. The rationale behind doing so would be to move up the
500 ppm cap for NRLM diesel fuel to coincide with the initial implementation of the 15 ppm cap
for highway diesel fuel. The aftertreatment-based engine PM emission standards would not be
moved up. They would still be initiated starting with the 2011 model year. Thus, this
alternative, relative to the proposal, would be a pure fuel program, moving up the 500 ppm fuel
controls of the proposal by one year.
We have examined this alternative from a number of angles relative to the proposal:
1) The need for further sulfur dioxide and PM emission reductions in this timeframe,
2) Its impact on the desulfurization technology used to meet the 15 and 500 ppm caps,
3) The leadtime available for refiners to meet the 15 and 500 ppm standards in 2006,
4) The impact on the supply of highway diesel fuel, and
5) The potential cost-effectiveness and cost-benefit of the additional sulfur control.
Because this option only affects fuel sulfur content and not engine emission standards, the
only air quality benefits are reductions of sulfur dioxide and sulfate PM emissions. The need for
these reductions should be just as great in 2006 as they are in the 2007-2010 timeframe. As
outlined in Chapter 2, ambient fine PM levels are currently above the NAAQS for fine PM in
many areas of the country. Ambient fine PM levels in 2006 are more likely to be near current
levels than those in 2007-2010, given that less time is available for current emission controls,
like the 2007 highway diesel program, to take effect. Thus, moving up the 500 ppm cap should
be considered for its direct air quality impacts. These emission reductions and their resulting
health and welfare benefits are shown in Section 12.2 and 12.3, respectively.
Applying the 500 ppm cap in 2006 as opposed to 2007 should have little impact on the
refining technology used. In Chapter 5, we project that conventional hydrotreating technology
which has been used for over 10 years to produce 500 ppm diesel fuel would be used by refiners
to meet a 500 ppm cap in 2007. This would also be the case for a 2006 standard, if refiners had
sufficient time to build new equipment.
However, a 2006 implementation date for the 500 ppm NRLM sulfur cap does not appear
to allow sufficient leadtime for refiners to design and construct new desulfurization equipment.
Leadtime for the proposed 2007 500 ppm NRLM diesel fuel cap was evaluated in Chapter 5.3.
There it was determined that refiners needed 2.25-3.25 years after the final rule in order to design
and construct new hydrotreaters to produce 500 ppm fuel. A 2006 implementation date would
only allow refiners facing the minimum required leadtime enough time to comply. A 2006
implementation date would allow no time for the generation of early sulfur reduction credits
which might allow some refiners additional time to meet the two-step fuel standards. Also, it is
difficult to project that any refiners would be able to meet these standards early even if the
12-90
-------
Regulatory Alternatives
program granted such credits. Thus, we must conclude that the 2006 two-step option would not
be technically feasible due to insufficient leadtime for refiners and importers to meet the 2006
fuel sulfur standards. For this reason, we were unable to develop any reliable cost estimates for
this option.
In addition to leadtime concerns, as with Option Ib, moving up the 500 ppm standard to
coincide with the implementation of the highway diesel fuel program would also raise workload
concerns for the industry impacting not only the successful implementation of this rulemaking,
but also the highway rule. A 500 ppm standard in 2006 could have an adverse impact on the
supply of highway diesel fuel and thus, the successful implementation of the 2007 highway diesel
fuel program. We added the temporary compliance option to the 2007 highway diesel fuel
program to ease implementation in 2006 and assure sufficient supply of highway diesel fuel. The
temporary compliance option allows 20% of highway diesel fuel to remain at 500 ppm until
2010. Starting the 500 ppm NRLM cap in 2006 would increase the strain on the design and
construction industries, as not only the 2007 highway diesel fuel program, but also the Tier 2
gasoline program are being implemented in the same timeframe. It would also increase the
amount of capital which would need to be raised by the refining industry. We have not evaluated
the degree that highway fuel supply would be negatively impacted. However, the impact would
be directionally negative.
Since this option is not feasible, we were not able to derive costs, and therefore cost per
ton or cost/benefit values that correspond to it. However, were more time given prior to
implementation of the 500 ppm cap, such that leadtime was no longer a constraint on feasibility,
the option essentially turns into the proposed requirement for 500 ppm beginning June 1, 2007
(with its associated costs, emission reductions, and benefits).
12.6.2.2.3 Option 2b
Compared to the proposal, Option 2b pulls the 15 ppm maximum sulfur fuel requirement
forward by one year to 2009. It also pulls all of the PM filter-based PM standards forward by one
year to take advantage of the earlier fuel availability.
Moving up the 15 ppm standard for nonroad diesel fuel by one year would increase
refining costs two ways. One, it would increase the cost of nonroad diesel fuel produced between
June 1, 2009 and June 1, 2010 by 2.4 cents per gallon (from the 2.2 cent per gallon cost of 500
ppm nonroad diesel fuel to the 4.6 cent per gallon cost of 15 ppm nonroad diesel fuel in 2009).
(See Table 12.4.2.3.1-1 above.) Two, it would increase the cost of nonroad diesel fuel produced
after June 1, 2010 by 0.2 cents per gallon, as the cost of producing 15 ppm nonroad diesel fuel
would be 4.4 cents per gallon for the proposed implementation date of June 1, 2010.
Moving up the 15 ppm standard for nonroad diesel fuel by one year would also make the
nonroad diesel fuel sulfur program more stringent than the highway diesel fuel sulfur program,
which does not require 100% of highway diesel fuel to meet a 15 ppm cap until June 1, 2010.
12-91
-------
Draft Regulatory Impact Analysis
Some of the synergies obtained by the proposed program would also be lost. Roughly 20
refineries are projected to start producing both 15 ppm highway diesel fuel and 15 ppm nonroad
diesel fuel in 2010. Requiring the production of 15 ppm nonroad diesel fuel at the same time that
the last 20% of highway diesel fuel must meet this standard would allow these two projects to be
fully coordinated, if not become a single project. Also, the three year interval between the
proposed 500 ppm and 15 ppm caps for nonroad diesel fuel is roughly equal to the life of a
desulfurization catalyst. Thus, many refiners would be bringing their 500 ppm desulfurization
unit down for catalyst replacement right about the time that the additional equipment needed to
meet the 15 ppm cap would need to be tied in. Implementing the 15 ppm cap one year earlier
would require refiners to either replace their existing catalyst earlier than necessary or bring the
unit down the next year again for another catalyst replacement.
In addition, Option 2b would involve a number of engine program considerations beyond
those analyzed for the proposed program. The primary effect of the pull-forward of PM control
is, of course, one-year earlier PM reductions. Over the very long-term the emissions impacts of
phase-in schedule differences diminish to zero, but during the phase-in years and shortly
thereafter, the differences can be substantial, given the over 90% PM reduction achieved by each
new engine entering the fleet meeting the proposed Tier 4 standard. Section 12.2 analyses these
impacts in detail.
The one-year pull-forward of PM standards would decouple PM and NOx control for
many engines. Engines <25 hp would be unaffected because there are no PM filter-based
standards for them. However, 25-50 hp engines would require redesign for PM control in 2012
and redesign for NOx control in 2013. This could create substantial increases in engineering
workload for both engine and equipment manufacturers attempting to carry out the double
redesign for two consecutive model years. This increase might conceivably be mitigated
somewhat by coordinated advance planning, such as by engine manufacturers anticipating NOx-
based changes to their engines and exhaust systems (NOx/PM exhaust emission control device
canning dimensions, for example), and providing these specifications to their equipment
manufacturer customers a year before those changes are actually made to allow for a single
machine redesign effort. Given the large impacts that even modest standards changes have had
on equipment designs in Tier 2, and the difficulty some engine manufacturers have had in
providing their customers with design specifications and prototypes very far ahead of time, it is
not clear that such pre-planning would be very effective.
Like engines <25 hp, engines in the 50-75 hp range would not experience a PM/NOx
standard decoupling under Option 2b because we are not proposing to change the NOx+NMHC
standard from the 2008 Tier 3 standard level for these engines. Engines above 75 hp would
experience this decoupling, however. For 75-175 hp engines, PM filters would be applied in
2011, and NOx adsorbers would begin to phase in in 2012. For 175-750 hp engines, PM filters
would be applied in 2010, and NOx adsorbers would begin to phase in in 2011. For engines
>750 hp, PM filters would be applied to 50% of engines in 2010. In 2011 and 2012, a recoupled
NOx/PM redesign strategy could be pursued with 50% of engines requiring both NOx and PM
12-92
-------
Regulatory Alternatives
Tier 4 controls. However, the standards would then be decoupled again as the remaining 50% of
engines are fitted with PM filters in 2013, and then NOx adsorbers in 2014. As for 25-50 hp
engines, some comprehensive pre-planning could help mitigate the costs of decoupling, but past
experience makes it doubtful that much of this could be assumed. All in all, Option 2b is likely
to result in a large increase in engineering workload for engine and equipment manufacturers.
In addition, earlier long-term PM standards would shorten the stability periods for
previous standards accordingly. The 0.22 g/bhp-hr transitional PM standard for 25-75 hp engines
would be in effect for only four model years, 2008-2011, instead of five. Likewise, previous-tier
standards for >75 hp engines would be in effect for three or four years, depending on engine size.
These shortened stability periods may not directly impact the feasibility of standards, but would
certainly have an adverse impact on manufacturers' ability to accomplish all required redesigns
without increasing engineering staffs and would also reduce the number of years available to
recover fixed costs.
We have not done a detailed analysis of the technological feasibility of PM filter
application one year earlier than under the proposal. The earliest Option 2b application date,
2010 for engines above 175 hp, is three years after similar technologies will be required for
HDDEs. Although we believe that this is likely to provide adequate lead time to accomplish the
transfer of this technology to some nonroad diesel applications, it is not clear that this could be
accomplished for 100% of the 175-750 hp nonroad engines and 50% of the >750 hp engines by
2010, and for all other nonroad diesel engines above 25 hp shortly after this. Even with engine
platforms on which this accelerated schedule could be accomplished, we would anticipate costs
to rise due to the shortened opportunity for learning from highway experience and the resulting
need for basic R&D to develop PM control technology directly in the nonroad sector.
Finally, we expect that under Option 2b, the technology review for engines under 75 hp,
discussed in section ni.G of the proposal, would need to occur in 2006 rather than 2007, to allow
adequate lead time should program adjustments be deemed appropriate. Given the large
experience base expected to accumulate during 2007 as highway engines equipped with
advanced PM and NOx emissions controls take to the road in large numbers, the one-year earlier
review schedule would be unfortunate.
Based on the above discussion and the analyses performed for this option described in
this Chapter, we conclude that Option 2b would not be appropriate for proposal. In particular we
see the large increase in engine and equipment manufacturers' workload for redesign, the
shortened stability periods for previous-tier standards, and the need for additional R&D
expenditures for some degree of parallel nonroad/highway emission control development work,
as large potential barriers to implementing this option.
12-93
-------
Draft Regulatory Impact Analysis
12.6.2.2.4 Option 2c
Option 2c is very similar to Option 2b except that PM filter-based standards would be
pulled forward by one year only for 175-750 hp engines. All points of the above discussion on
Option 2b are relevant here except of course that discussion points specific to <175 hp and >750
hp engines would not apply. Engines in the 175-750 hp category comprise a large segment of the
emissions inventory, of the engine families, and of the total U.S. nonroad engine sales. As a
result the environmental impact of Option 2c, though not as large as that of Option 2b, is
substantial compared to the proposed program, especially in the early years of Tier 4. Likewise,
the adverse impacts of the Option 2c PM pull-forward on the engine and equipment
manufacturing industries would be large, though more focused on those manufacturers with
products in this power range. This is significant because there are many manufacturers who do
not offer products in this range and so would be affected only indirectly. Some of these might
benefit by the added year of experience gained from the use of PM filters on 175-750 hp engines
before PM filters are required on their own products. On the other hand, manufacturers who do
not have ready access to this experience base may find themselves at a disadvantage compared to
their better-connected competitors.
Although these considerations may be significant, we do not see them as critical to the
feasibility and cost impacts of Option 2c. Instead, we believe the primary engine and equipment
issues involved in Option 2c are the above-discussed engineering workload impacts caused by
the decoupling of PM and NOx standards for 175-750 hp engines, the shortened stability periods
for the Tier 3 standards, and the possible feasibility concerns raised by shortened lead time
available for transferring technology from the highway sector.
Based on the above discussion and the analyses performed for this option described in
this Chapter, we conclude that Option 2c would not be appropriate for proposal, for the same key
reasons described above for Option 2b, though to a lesser degree and with a corresponding lesser
emission benefit.
12.6.2.2.5 Option 2d
The proposed program does not apply the NOx adsorber-based 0.30 g/bhp-hr NOx
standard to engines below 75 hp for reasons explained in Chapter 4 and in section in of the
preamble to the proposal. The Option 2d analysis evaluates the environmental and cost impact of
applying this standard to 25-75 hp engines, phased in at 50-50-50-100% over 2013-2016, similar
to the NOx phase-in approach taken for larger engines, though on a later schedule. Although we
do not believe this approach to be appropriate at this time, we have included this matter in the
proposed 2007 technology review as discussed in section in.G of the preamble.
The 25-75 hp category comprises a large and growing segment of the nonroad diesel
engine population. Although on a per-engine basis these engines typically emit far less NOx
over their lifetime than larger engines, they make up a significant NOx source category, as can be
12-94
-------
Regulatory Alternatives
seen in comparing the NOx inventory for Option 2d with that for the proposal (see section 12.2).
In addition to the NOx reductions, the application of NOx adsorbers to 25-75 hp engines would
recover some of the fuel economy impact due to use of actively-regenerated PM filters on these
engines.
The application of NOx adsorbers to 25-75 hp engines would add a sizeable cost to these
engines. However, we would not expect the added cost for advanced NOx control to include the
cost of modifying the engines themselves to accommodate NOx adsorbers (e.g., electronic
common rail fuel systems) because these costs would likely be incurred in meeting our proposed
0.02 g/bhp-hr Tier 4 PM standard, as discussed in Chapter 6. Although under Option 2d the 0.30
g/bhp-hr NOx standard for 25-75 hp engines in 2013 would replace the proposed 3.5 g/bhp-hr
NOx+NMHC Tier 4 standard in the same year, the cost of meeting the 3.5 g standard (via EGR
or equivalent technology) would not be eliminated, because engine-out emissions performance
on this order or better must be achieved to meet the 0.30 g standard employing NOx adsorbers
with control efficiencies on the order of 90%. (In fact the 50-75 hp engines must meet this 3.5 g
standard in 2008 under Tier 3 requirements.)
The Option 2d program would establish a Tier 4 program implementation schedule that
stretches out to 2016, well over a decade from today. Although in principle we support the aim
of the industries we regulate to have long-term regulatory certainty and stability, we must balance
this with the fact that our understanding of how diesel pollution impacts human health and the
environment is the subject of numerous ongoing studies and so is likely to develop and evolve
over the next few years, and also with the likelihood that the rapid pace of emission control
technology development (often with unexpected innovations along the way) will likewise
continue to advance in the years ahead. Standard-setting in this rulemaking with 2016
implementation dates may be inadvisable, and better taken up in the 2007 technology review
planned in the proposal.
Based on the above discussion and the analyses performed for this option described in
this Chapter, as well as the present concerns with technological feasibility voiced in Chapter 4 of
this draft RIA and in Chapter HI of the preamble, we conclude that Option 2d would not be
appropriate for proposal.
12.6.2.2.6 Option 2e
The Option 2e program is identical to the proposal except that no new NOx standards
would be set in Tier 4. Any changes in NOx control from these engines would be incidental,
resulting from adoption of the test procedure changes for the Tier 4 PM control program. This
analytical option obviously presumes that Tier 4 nonroad diesel NOx control would either not be
needed to address air quality concerns or would not be feasible (presumptions we believe are
unfounded). These issues are discussed in detail in Chapters 2 and 4 of this draft RIA, and in
sections n and HI of the preamble to the proposal.
12-95
-------
Draft Regulatory Impact Analysis
We have assumed no changes to the proposed fuel program in analyzing Option 2e
because the proposed fuel desulfurization, though critical to enabling high-efficiency NOx
exhaust emission control, is also needed to enable PM filter technology as explained in section
in.F of the proposal preamble. The first step in the two-step fuel desulfurization proposal is also
primarily a PM-focused action. Finally, the fact that we are proposing PM filter-based standards
before or coincident with Tier 4 NOx standards in all relevant power categories, means that the
timing of fuel changes under a PM-only option would be unchanged from the proposal.
As discussed in Chapter 4, diesel PM filters can be designed to operate effectively with or
without the application of NOx adsorbers in the exhaust stream. In fact under the proposal, some
engines are expected to employ PM filters without NOx adsorbers for phase-in model years or,
for 25-75 hp engines, for all Tier 4 model years. There are economies of integration available to
engine designers working to both the PM and NOx control objectives, such as from combining
PM and NOx control functions into a single can or even into integrated internal structures, but
even so we would expect that PM-only systems would cost significantly less than combined
systems. Some engine designs that do not currently employ sophisticated fuel injection controls
could conceivably continue without these under a PM-only option, but we believe that the need
for active regeneration of PM filters in many nonroad applications, combined with the growing
trend toward application of electronic controls for performance reasons or to meet Tier 2/Tier 3
standards, would tend to mitigate this opportunity. Equipment designers would likely see no or
only modest cost advantages to PM-only systems beyond the NOx control hardware itself
because the Tier 4 program is structured to minimize multiple Tier 4 redesigns as much as
possible, and the likelihood of integrated NOx/PM exhaust emission controls reduces the need
for additional brackets and the like.
A PM-only program would be expected to result in added operating costs compared to the
proposed program due to the increased fuel consumption of PM filter-equipped engines, not
offset by the fuel economy gains of NOx adsorber systems. This matter is discussed in detail in
Chapter 6.
We believe that Option 2e would be highly inappropriate. In particular, we believe that a
lack of new NOx standards in Tier 4 would fail to adequately address the serious air quality
concerns discussed in Chapter 2, and to meet our obligations under section 213(a)(3) of the Clean
Air Act which requires the Agency to develop standards reflecting the greatest emission
reductions feasible, taking cost, noise and safety concerns into consideration. In doing so,
consideration of cost is to be a subordinate consideration, unless costs are somehow exorbitant.
See, e.g. Husqvarna AB v. EPA. 254 F. 3d 195, 200 (D.C. Cir. 2001); Lignite Energy Council v.
EPA. 198 F. 3d 930, 933 (D.C. Cir. 1999). Here, very substantial NOx reductions - on the order
of millions of tons - are technically feasible. The costs of achieving these reductions is not
exorbitant. Moreover, the Tier 4 proposal would set stringent PM standards to be implemented
in the 2011-2014 timeframe, followed by some period of stability before any new standards
beyond Tier 4 would take effect, if found appropriate. Not including new NOx standards in this
same timeframe would leave the nonroad diesel sector as a dominant source of NOx emissions
12-96
-------
Regulatory Alternatives
for many years to come, at a time when other NOx source categories would have finished
implementing stringent measures to deal with NOx-related air quality problems.
12.6.2.2.7 Option 3
As described in section 12.1.2 of this chapter, Option 3 is an exemption from regulation
in this rule for very high power engines ( >750 hp) used in above-ground mining equipment
(AGME). Some have expressed the view that the very large off-highway trucks and earth
movers, over 750 hp, used in above-ground mine and quarry operations may constitute a special
case worthy of special consideration because of a number of factors:
- They operate remote from populated areas;
- They have very low annual sales volumes and therefore high redesign costs;
- They are used in extreme conditions where aftertreatment will not be durable.
However, the above concerns with applying Tier 4 standards to > 750 hp AGME engines
must be balanced with the emissions contribution and the health and welfare concerns from the
engines, as well as EPA's assessment that Tier 4 standards for the >750 hp engines used in
AGME are technologically feasible and otherwise appropriate under the Clean Air Act. It thus
appears that any such exemption would lack a convincing legal rationale, given that mining
engines have already been held to be properly subject to regulation under section 213, see Engine
Manufacturers Ass'n v. EPA. 88 F. 3d 1075, 1098 (D.C. Cir. 1996) and, as explained below,
further reductions in PM and NOx emissions from these engines is technically feasible at
reasonable cost.
Large nonroad mining equipment is used in many areas spread across the United States.
It is often assumed that the very large AGME is concentrated in western states. Information
provided to EPA by a nonroad equipment manufacturer who participates in the >750 hp mining
equipment market indicates that in the past decade the western states (not including the west
coast states) account for nearly 30 percent of the >750 hp AGME sales. However, the eastern US
also has a high share of >750 hp mining equipment. Information provided to EPA by a nonroad
equipment manufacturer who participates in the >750 hp mining equipment market indicates that
in the past decade, more than 40 percent of the >750 hp equipment was sold to the states in the
Ohio River valley. Considering the concentration of coal mining in these states the high use of
these large machines in the Ohio River valley should not be suprising.5
In general, it is reasonable to project that most above-ground mines are not located in
urban areas. However, pollution problems such as ozone and haze are not local but regional
problems due to the long-range transport of emissions. In addition, mines are not always in
remote rural areas but are some times in or near urban areas. In connection with our original
nonroad engine rulemaking in 1994, the American Mining Congress submitted as part of its
public comment a report from the TRC Environmental Corporation which states that 40 mine
sites are located in ozone nonattainment areas.6 See Engine Manufacturers Ass'n v. EPA. 88 F.
12-97
-------
Draft Regulatory Impact Analysis
3d at 1098 (national regulation of nonroad engines used in above-ground mining is justified
under section 213 because some of those engines are used in nonattainment areas).
Even in the western states, air pollution from mining equipment is a concern for state and
tribal air quality agencies. EPA has recently received comments from the Western Regional Air
Partnership supporting further controls on nonroad engines, equipment and fuel, specifically
including mining equipment, in order to comply with EPA's regional haze regulations.7
Another reasons which some have suggested as grounds for exempting >750 hp engines
used in AGME from the proposed Tier 4 standards is the low sales volume and high redesign
costs of the engines and the equipment. It is generally correct that for this category of nonroad
equipment, annual sales volumes are low, typically on the order of 50 or fewer for a given
equipment model, and in many cases fewer than 10. Therefore, the costs of equipment redesign
must be spread over a small number of sales. Our proposal for the >750 hp category provides
significant flexibilites to address these concerns. This includes a phase-in of all standards (not
just NOx and VOC) over three years, as well as the provisions for averaging, banking, and
trading and the transition program for equipment manufacturers which are discussed in section
VII of the proposed preamble. In fact, the >750 hp category is a separate category under the
TPEM which would allow many AGME manufacturers to defer using any Tier 4 technology
engines for a full seven years, until 2019.
In addition, the costs of equipment redesign must be put in the context of the high sales
price of these types of equipment, which is commonly > $1 million. We should also note that
exempting > 750 hp engines used in AGME would not reduce the research & development costs
for engine manufacturers. Many of these large engines would still need to meet the proposed
Tier 4 standards for applications other than AGME, such as cranes, large oil field equipment, and
non-mining applications of off-highway trucks, excavators, etc. Table 12.6.2.2.7-1 below is a list
of the nonroad equipment categories and estimated 1998 U.S. population used in EPA's
NONROAD model which have engines >750 hp, including those we have projected are used in
AGME and those which are not (based on our engineering expertise and discussion with engine
and equipment manufacturers).
12-98
-------
Regulatory Alternatives
Table 12.6.2.2.7-1
Nonroad Equipment Categories Which Use Engines >750hp and Estimated Population
> 750 hp Equipment Category
Crawler Tractor/Dozers
Excavators
Off-Highway Tractors
Off-highway Trucks
Rubber Tire Loaders
Bore/Drill Rigs
Chippers/Stump Grinders
Cranes
Crushing/Processing Equipment
Forest Eqp - Feller/Bunch/Skidder
Other Agricultural Equipment
Other Construction Equipment
Other Oil Field Equipment
Railway Maintenance
Specialty Vehicle Carts
Trenchers
Used in Above-
ground Mining?
Yes
Yes
Yes
Yes
Yes
No
No
No
Yes
No
No
No
No
No
No
No
Est. 1998 U.S. Population3
6,097
408
848
4,574
2,633
911
118
19
4
12
2
29
969
36
50
11
1 Estimated 1998 U.S. populations from EPA's draft NONROAD2002 emission inventory model
Some engine engine manufacturers have argued that the engines used in the largest
mining applications are so large that the aftertreatment systems cannot be scaled up to such sizes
and remain durable (though no manufacturer has provided any specific reasons why this would
be so, nor have any data been presented). As discussed in Section HIE. of the preamble and in
Chapter 4 of this draft RIA, we recognize that many nonroad equipment types experience harsh
and sometimes severe operation conditions. However, as discussed in the preamble and in
Chapter 4 of this draft RIA, existing data already indicate that aftertreatment systems can be
designed to withstand these harsh environments while maintaining their structural integrity. In
fact, many of the actual examples of PM filters which have been used have been for mining
applications. Systems have been used in a number of underground mining applications in Europe
on equipment ranging from 125 to 275 hp for upto 6,900 hours on a single application.8 One
engine manufacturer, Deutz, developed a PM filter system for engines up to 800 hp. The Deutz
system utilized two filters for engines greater than approximately 230 hp, and their largest system
relied on two filters which were 62.5 liters each and have been used on engines with
displacements of 26 liters.9 Finally, one integrated engine/equipment manufacturer offers an
OEM option of a PM filter based system in a number of their equipment types, including mining
equipment.
10
Based on the information available to us and discussed in section HI of the preamble and
chapter 4 of the RIA, we believe that exhaust aftertreatment systems can be designed to be
durable in-use even for the >750hp engines.
12-99
-------
Draft Regulatory Impact Analysis
Emissions from >750 hp AGME are a significant portion of the NOx and PM inventory
from the nonroad diesel engines. Our modeling indicates that these machines, though low in
nationwide sales and population, are not an insignificant part of the NOx and PM inventories.
Table VI-1 in the preamble for this proposal shows AGME >750 hp represents 13 percent of the
net-present value of the NOx reduction and 2 percent of the PM reduction of our proposal. A
graphical representation of the impact on the national inventories of exempting these engines can
also be seen in Figures 12.2.2.1-1 (NOx) and 12.2.2.2-1 (PM).
Table 12.2.2.1-1 shows an increase in NOx emissions in 2030 of approximately 103,000
tons, and Table 12.2.2.2-1 shows an increase in PM emission in 2030 of approximately 4,000
tons if the >750 hp AGME were exempted. Table 12.2.3-2 shows that the cumulative,
undiscounted emission increase which would occur through 2030 if >750 hp AGME engines
were exempted is approximately 742,000 tons of NOx and 30,000 tons of PM.
As discussed in Chapter 12.4, we have estimated the net-present value cost through 2030
of the proposed Tier 4 standards for >750 hp AGME and engines at approximately $490 million.
The estimated aggregate cost per ton for the proposed Tier 4 standards for >750 hp AGME is
$300/ton for NOx+NMHC and $8,300/ton for PM though 2030. The PM cost per ton value is in
line with the estimate for our entire proposal, and the NOx+NMHC estimate is well below the
values for the entire proposal. There is no rational way that such costs could be considered so
hugely exorbitant or disproportionate (the test under the case law cited earlier) as to justify
forgoing the large, achievable emission reductions obtainable from these engines.
Finally, as discussed in Chapter 12.3, we have estimated the net-present value of the
monetized health benefits for our proposed standards for >750 hp engines used in AGME
through 2030 as being approximately $16 billion.
Based on the information available to us, we do not believe this option should be
promulgated. The standards we have proposed for >750 hp AGME engines are feasible and very
cost-effective. AGME contributes to the same health and welfare concerns as other nonroad
diesel engines, massive emission reductions of PM and NOx from these engines are feasible, and
the costs we have estimated for controlling these engines are not excessive, exorbitant, or
otherwise inappropriate.
12.6.2.2.8 Option 4
In order to enable the high efficiency exhaust emission control technology to begin to be
applied to nonroad engines beginning with the 2011 model year, we are proposing that all
nonroad diesel fuel produced or imported after June 1, 2010 would have to meet a 15 ppm sulfur
cap. Although locomotive and marine diesel engines are similar in size to some of the diesel
engines covered in this proposal, there are many differences (e.g., duty cycles, exhaust system
design configurations, size, and rebuild and maintenance practices) that have caused us to treat
12-100
-------
Regulatory Alternatives
them separately in past EPA programs.A For the same reasons, we are not proposing new engine
standards today for these engine categories and as a result, are also not proposing that the second
step of sulfur control to 15 ppm in 2010 be applied to locomotive and marine fuel. We are
proposing to set a sulfur fuel content standard of 500 ppm for fuel used in locomotive and marine
diesel applications. This fuel standard is expected to provide considerable sulfate PM benefits
regardless of whether or not we also set more stringent emission standards for these engines.
As discussed in Section IV of the preamble, we are also seriously considering extending
the 15 ppm standard to locomotive and marine fuel as early as June 1, 2010 as well. There are
several advantages associated with this alternative. First, as reflected in Table 12.2.3-2, it would
provide over 9,000 tons of additional sulfate PM benefits and over 114,000 tons of additional
SO2 benefits from 2007 to 2030 (calculated as net present value in 2004). The cost for these
additional benefits as shown in Section 12.4.3.2 are $1.8 billion. This cost reflects the
incremental cost for reducing the sulfur content of locomotive and marine from 500 ppm to 15
ppm - 2.4 c/gal. The cost also reflects an increase in the long-term per gallon cost of all 15 ppm
NRLM diesel fuel of 0.2 c/gal due to the fact that higher cost refiners are now required to
produce 15 ppm diesel fuel.
Second, reducing sulfur content of locomotive and marine diesel fuel to 15 ppm in 2010
would simplify the fuel distribution system and the design of the fuel program proposed today
since a marker would not be required for locomotive and marine diesel fuel. The marker cost
itself is an estimated 0.2 c/gal. While difficult to quantify, additional cost savings would be
realized by allowing locomotive and marine fuel to be fungible with nonroad and highway diesel
fuel. Furthermore, prices do not necessarily follow costs, and there is reason to believe that the
price for 500 ppm locomotive and marine fuel will not necessarily be appreciably lower than if it
were required to be 15 ppm. Under the proposal, we expect that a certain amount of marine fuel
will be ultra-low sulfur fuel regardless of the standard due to limitations in the production and
distribution of unique fuel grades. Where 500 ppm fuel is available, the possible suppliers of
fuel will likely be more constrained, limiting competition and allowing prices to approach that of
15 ppm fuel. If we were to bring locomotive and marine fuel to 15 ppm, the pool of possible
suppliers could expand beyond those today, since highway diesel fuel will also be at the same
standard. It is difficult to provide any quantitative price comparison, but it is entirely possible
that the price differential between a 15 ppm and 500 ppm standard for locomotive and marine
fuel would be significantly less than the estimated 2.4 c/gal cost differential.
Third, reducing sulfur content of locomotive and marine diesel fuel to 15 ppm in 2010
would help reduce the potential opportunity for misfueling of 2007 and later model year highway
vehicles and 2011 and later model year nonroad equipment with higher sulfur fuel. We do not
A Locomotives, in fact, are treated separately from other nonroad engines and vehicles in the
Clean Air Act, which contains provisions regarding them in section 213(a)(5). Less than 50 hp
marine engines were included in the 1998 final rule for nonroad diesel engines, albeit with some
special provisions to deal with marine-specific engine characteristics and operating cycles.
12-101
-------
Draft Regulatory Impact Analysis
anticipate misfueling to be a significant concern under today's proposal, since by 2010 more than
80% of the total number 2 distillate pool is expected to be 15 ppm (see Table 7.1-16 in Chapter
7). Nevertheless, extending the 15 ppm standard to locomotive and marine would increase this
percentage to more than 85%, further limiting the sources of fuel on which misfueling could
occur either accidentally or intentionally.
Finally, reducing sulfur content of locomotive and marine diesel fuel to 15 ppm in 2010
would allow refiners to coordinate plans to reduce the sulfur content of all of their nonroad diesel
fuel at one time. While in many cases this may not be a significant advantage, it may be a more
important consideration here since it is probably not a question of whether locomotive and
marine fuel must meet a 15 ppm cap, but merely when. As discussed in Section IV of the
preamble, it is the Agency's intention to take action in the near future to set new emission
standards for locomotive and marine engines that could require the use of high efficiency exhaust
emission control technology, and thus, also require the use of 15 ppm sulfur diesel fuel.B We
anticipate that such engine standards would likely take effect in the 2011-13 timeframe, requiring
15 ppm locomotive and marine diesel fuel in the 2010-12 timeframe.
However, discussions with refiners have suggested there are significant advantages to
leaving locomotive and marine diesel fuel at 500 ppm, at least in the near-term and until we set
more stringent standards for those engines. First, the locomotive and marine diesel fuel markets
could provide a market for off-specification product that is important for refiners, particularly
during the transition to 15 ppm for highway and nonroad diesel fuel in 2010. It is possible that
significant volumes of 500 ppm diesel fuel would be created in the distribution system during the
distribution of 15 ppm fuel. For example, the pipeline interface between 15 ppm diesel fuel and
higher sulfur jet fuel would likely contain less than 500 ppm sulfur. Without the ability to sell
this fuel to the locomotive and marine diesel fuel markets, this interface would have to be sold as
heating oil. The available markets for heating oil could be quite limited, particularly outside the
Northeast, causing more fuel to have to be shipped back to refineries for reprocessing at
considerable expense. Maintaining a market for locomotive and marine fuel at 500 ppm would
provide a market across much of the country where off-specification 15 ppm could be marketed.
Waiting just a year or two beyond 2010 for implementing the 15 ppm standard for locomotive
and marine would not address long term desires for outlets for off-specification product, but it
would address the more critical, near term needs during the transition. Second, waiting just
another year or two beyond 2010 is projected to allow virtually all refiners to take advantage of
the new lower cost desulfurization technologies. As discussed in Chapter 6 approximately 80
percent of refineries are projected to be able to take advantage of these new technologies with the
June 1, 2010 implementation date. We project that just a two year delay to 2012 would permit
all refineries to do so, thereby reducing the desulfurization costs for 15 ppm locomotive and
marine fuel. Finally, while the monetized benefits of controlling the sulfur level of locomotive
BThe most recent new standards for locomotives and marine diesel engines (including those
under 50 hp) were set in separate actions (63 FR 18977, April 16, 1998 and 67 FR 68241,
November 8, 2002).
12-102
-------
Regulatory Alternatives
and marine diesel fuel from 500 ppm down to 15 ppm outweigh the costs (even in the absence of
new engine emission standards), the cost per ton for the incremental sulfate PM and SO2
emission reductions as shown in Table 12.5-1 is $64,000 and 10,300 per ton, respectively. These
costs are rather high in comparison to other possible control options.
12.6.2.2.9 Option 5a
The Option 5a program is identical to the proposal except that no new program
requirements would be set in Tier 4 for engines under 75 hp. Instead Tier 2 standards and testing
requirements for engines under 50 hp, and Tier 3 standards and testing requirements for 50-75 hp
engines, would continue indefinitely. This analytical option presumes that Tier 4 nonroad diesel
NOx and PM control from these engines would either not be needed to address air quality
concerns or would not be feasible (presumptions we believe are unfounded). These issues are
discussed in detail in Chapters 2 and 4 of this draft RIA, and in sections II and in of the preamble
to the proposal.
We believe that Option 5a would be inappropriate. As discussed in section in.E of the
proposal preamble, the 0.02 g/bhp-hr PM standard proposed for 25-75 hp engines in 2013 is
feasible, based on the use of high-efficiency PM filters and the availability of nonroad diesel fuel
with sulfur levels capped at 15 ppm. As also discussed in section IHE of the proposal preamble,
the less stringent PM standards proposed for engines under 75 hp in 2008 are feasible, based on
the use of diesel oxidation catalysts and/or engine optimization strategies, and on the availability
of nonroad diesel fuel with sulfur levels capped at 500 ppm. In fact, as discussed in section HIE
of the proposal preamble, some of today's engines already meet the proposed standards. We
believe that the 2008 standards provide a reasonable means of gaining substantial PM reductions
from the nonroad diesel sector in the early years of the Tier 4 program, while managing the
workload, stability, and technology transfer issues involved, but we are also requesting comment
in section HLB.l.d.ii of the proposal preamble on whether it would be better not to set a Tier 4
PM standard in 2008 so that engine designers could instead focus their efforts on meeting a PM-
filter based standard for these engines earlier, say in 2012.
Establishing no Tier 4 PM program at all for engines under 75 hp would, on the other
hand, leave engines under 50 hp at Tier 2 PM standards levels of 0.60 g/bhp-hr (for <25 hp) and
0.45 g/bhp-hr (for 25-50 hp), and would leave 50-75 hp engines at a Tier 3 PM standard level of
0.30 g/bhp-hr. The resulting in-use emissions levels from these engines would be many times
higher than that achieved under the proposed program. As discussed in section 12.2, the overall
loss in Tier 4 PM emissions reductions would be correspondingly large, both in the early and the
long-term timeframes of the program. This option would also fail to address toxic hydrocarbon
concerns, considering the large population of these under 75 hp engines and the fact that they are
often used in populated areas and in equipment without closed cabs.
To take no action on under 75 hp engines in this rulemaking would compromise air
quality goals and would also greatly increase uncertainty for the engine and equipment
12-103
-------
Draft Regulatory Impact Analysis
manufacturing industry. Due to the continuing growth in sales of these smaller engines and the
promising developments that are occurring in diesel emissions control technology, it seems
improbable that putting off action to some point in the future would result in more flexibility,
more leadtime, or less stringent standards than under this proposal. We believe instead that
setting standards now, with plans for a technology review in 2007 for the long-term (2013)
standards, appropriately balances the need for Tier 4 program certainty and leadtime with the
Agency's commitment to reconsider program requirements where necessary in light of
continuing technology progress and demonstration over the next few years.
12.6.2.2.10 Option 5b
The Option 5b program is identical to the proposal except that for engines under 75 hp
only the 2008 engine standards would be set. There would be no additional PM filter-based
standard in 2013 for 25-75 hp engines, and no additional NOx+NMHC standard in 2013 for 25-
50 hp engines. This analytical option presumes that controlling PM from 25-75 hp engines to
levels achievable with PM filters would either not be needed to address air quality concerns or
would not be feasible (presumptions we believe are unfounded). These issues are discussed in
detail in Chapters 2 and 4 of this draft RIA, and in sections n and HI of the preamble to the
proposal.
Although, unlike Option 5a, Option 5b does involve important PM reductions beginning
in 2008, much of the Option 5a discussion in section 12.6.2.2.9 applies here as well. The loss in
long-term Tier 4 PM emissions reductions would be large, as discussed in section 12.2, because
the PM reductions from engines produced after 2008 would be only on the order of 50%
compared to previous-tier engines, instead of the more than 95% reductions available through the
use of PM filters. This option could also leave a large unaddressed toxic hydrocarbon concern,
depending on the degree to which manufacturers choose to meet the 2008 standards through the
use of diesel oxidation catalysts. Overall, we believe that Option 5b would be inappropriate.
12-104
-------
Regulatory Alternatives
Appendix 12A: Certification Fuel Sulfur Levels
The sulfur levels assumed for certification fuel for the purposes of modeling emission
benefits of each program option are presented in this appendix. Note that the Tier 1 standards for
>750hp engines continued through 2005. Manufacturers subject to these Tier 1 standards are
assumed to have certified on fuel having an average sulfur content of 3300ppm, based on
existing records of those tests.
As described in Section 12.2.1.1, the cert fuel sulfur levels in the charts below do not
always coincide with changes in the required maximum sulfur level for certification fuel. Engine
manufacturers are unlikely to make modifications to their engines to take advantage of the lower
sulfur requirement for cert fuel until new engine standards make such modifications necessary.
The assumed cert fuel sulfur levels were used to establish the proper zero-hour emission factors
for new engines. For in-use inventory impacts of these new engines, the emission factors were
further adjusted to account for the assumed in-use sulfur levels. Thus, for instance, engines
certified on 2000ppm sulfur fuel and then operated on SOOppm fuel would realize a PM benefit
relative to the PM certification standard.
Figure 12A-1
Assumed Certification Fuel Sulfur Levels To Establish
Zero Hour Emission Factors Under Option 1 (ppm)
hp group
2005 2006 2007
2008
2009 2010 2011 2012 2013 2014
2015
hp<25
25
-------
Draft Regulatory Impact Analysis
Figure 12A-2
Assumed Certification Fuel Sulfur Levels To Establish
Zero Hour Emission Factors Under Option la (ppm)
hp group
2005 2006 2007
2008
2009
2010
2011
2012
2013
2014
2015
hp<25
25
-------
Regulatory Alternatives
Figure 12A-4
Assumed Certification Fuel Sulfur Levels To Establish
Zero Hour Emission Factors Under Proposed Program (ppm)
hp group
hp<25
25
-------
Draft Regulatory Impact Analysis
Figure 12A-6
Assumed Certification Fuel Sulfur Levels To Establish
Zero Hour Emission Factors Under Option 2b (ppm)
hp group
hp<25
25
-------
Regulatory Alternatives
Figure 12A-8
Assumed Certification Fuel Sulfur Levels To Establish
Zero Hour Emission Factors Under Option 2d (ppm)
hp group
hp<25
25
-------
Draft Regulatory Impact Analysis
Figure 12 A-10
Assumed Certification Fuel Sulfur Levels f To Establish
Zero Hour Emission Factors Under Option 3 (ppm)
hp group
hp<25
25
-------
Regulatory Alternatives
Figure 12 A-12
Assumed Certification Fuel Sulfur Levels To Establish
Zero Hour Emission Factors Under Option 5a (ppm)
hp group
hp<25
25
-------
Draft Regulatory Impact Analysis
12-112
-------
Regulatory Alternatives
Appendix 12B: Incremental Cost, Emission Reductions,
Benefits, and Cost Effectiveness
This Appendix provides incremental costs, incremental emission reductions, marginal
cost per ton of emission reduction, and incremental benefits for each in a series of potential
control steps. The cost, emission reduction, and cost per ton data are presented in Table 12B-1,
and the cost and benefit data are presented in Table 12B-2.
Because the emission reductions represent the change from the preceding baseline level,
the order of the control steps affects the estimate of cost per ton. Some, but not all, of the steps
specified in Table 12B-1 are components of our proposal. The data presented in Table 12B-1
and 12B-2 are provided as additional information for the reader.
For each control step, the baseline emission levels are presented prior to the introduction
of that control step. The first baseline level in the table represents the emissions levels absent
any new controls for nonroad engines or nonroad, locomotive and marine fuels. Subsequent
baseline levels represent the difference between the preceding baseline level and the reductions
from the preceding control steps (i.e., the remaining emissions).
The costs in the table represent approximate costs for each control step, apportioned
among various pollutants. Our method for apportioning costs to a particular pollutant is
described in Chapter 8, Table 8.1-2. In this case, the apportioning of costs is simplified,
somewhat, as each control step has a distinct pollutant focus (i.e., the applications of
DOCs/engine-out reductions and CDPFs for PM, even though some NMHC reductions are
realized). The costs shown here should be considered as rough approximations, because they
have been derived from our program costs by splitting various fixed costs of the program by
pollutant and control step. For example, the R&D costs estimated in Chapter 6, and used here,
for engines larger than 75 hp were roughly split 67 percent to NOx control and 33 percent to PM
control. We have made no estimate of the distinct cost of only doing PM control or only doing
NOx control for engines in this horsepower range. We believe that it is likely that R&D costs for
either step alone would be higher than represented in this analysis. Nevertheless, for comparative
purposes we have presented the costs here.
12-113
-------
Draft Regulatory Impact Analysis
Table 12B-1
Incremental Cost, Emissions Reductions, and Cost Effectiveness for Various Control Steps
(All values are expressed as 2004 NPV using a 3 percent discount rate)
Control Steps
500 ppm Sulfur Nonroad,
Locomotive, and Marine Fuel
in 2007
15 ppm Sulfur Nonroad Fuel
and Tier 4 PM for >75hp
Engines
Transitional PM Standards for
<75hp in 2008
CDPF based PM Standards
for25hp-75hpin2013
Tier 4 NOx Standards
15 ppm Sulfur Locomotive
and Marine Fuel in 2010
Baseline
Cost
Reductions
Cost/Ton
Baseline
Cost
Reductions
Cost/Ton
Baseline
Cost
Reductions
Cost/Ton
Baseline
Cost
Reductions
Cost/Ton
Baseline
Cost
Reductions
Cost/Ton
Baseline
Cost
Reductions
Cost/Ton
PM
(NPV 2007-2030)
3,251
-
374
-
2,877
$9.9
917
$10,800
1,960
$1.2
88
$14,200
1,872
$2.2
121
$18,300
1,751
-
0
-
1,751
$0.6
9
$64,200
Remaining tons NR, Locomotive and Marine| 1,742
NOx+NMHC
(NPV 2007-2030)
21,745
-
0
-
21,745
-
137
-
21,608
-
1
-
21,608
-
0
-
21,608
$3.3
5,407
$600
16,200
-
0
-
16,200
SO2
(NPV 2007-2030)
5,273
$0.5
4,638
$100
635
-
315
-
320
-
0
-
320
-
0
-
320
-
0
-
320
$1.2
114
$10,300
206
Baseline - the NPV of the emission levels prior to the control step (1,000 tons), recalculated after each control step
Cost - the NPV of the annualized costs of the control step ($ billion), apportioned by pollutant
Reductions - the NPV of the emissions reductions from the baseline due to the control step (1,000 tons)
Cost/Ton - the ratio of the Cost and Reductions ($/ton)
The reduction rows in the table represent the emission reductions from the previous
baseline level by pollutant for each of the control steps. The cost per ton row simply reflects the
ratio of the preceding two rows, defining the cost per ton of reduction realized in the control step.
Note that for many of the control steps, reductions in emissions are realized for multiple
pollutants, yet we have attributed cost to only one or two pollutants (depending on the primary
purpose of the control technology, as discussed in Chapter 8.1). This does not mean that the
reductions in the other pollutants can actually be realized for free, only that we have attributed no
costs to those reductions. For example, we have attributed all of the costs of the 15 ppm sulfur
program to PM control, therefore the "Tier 4 NOx Standards" data shows very low $/ton
incremental costs.
12-114
-------
Regulatory Alternatives
Estimates of the cost and dollar benefits of the various control steps are presented in
Table 12B-2, below. The cost estimates are the same as for Table 12B-1 (although summed into
a single value rather than distributed across multiple pollutants). The benefits estimates are an
approximation based upon the benefits estimates for the proposal and the various control options
presented previously in Section 12.3 Benefits Comparison. Each of these control steps can be
approximated by one or more of the options in Table 12.6-1. For example, the PM portion, of
control step, Transitional PM Standards for <75 hp in 2008, can be found as the difference
between options 5a (no control for <75hp engines) and 5b (no CDPF control for 25hp-75hp
engines). As these benefits are based on approximations from other control approaches, the
benefits listed in Table 12B-2 should be considered as approximate estimates to the benefits of
the various control steps.
Table 12B-2
Cost and Benefits of Various Control Steps
(All values are expressed as 2004 NPV using a 3 percent discount rate)
Control Steps
500 ppm Sulfur Nonroad,
Locomotive, and Marine Fuel in
2007
15 ppm Sulfur Nonroad Fuel and
Tier 4 PM for >75 hp Engines
Transitional PM Standards for
<75hp in 2008
CDPF based PM Standards for
25hp-75hpin2013
Tier 4 NOx Standards
15ppm Sulfur Locomotive and
Marine Fuel in 2010
Cost ($ Billion)
NPV(2007-2030)
$0.5
$9.9
$1.2
$2.2
$3.3
$1.8
Benefit ($ Billion)
NPV(2007-2030)
$230
$186
$28
$43
$64
$6
12-115
-------
Draft Regulatory Impact Analysis
Chapter 12 References
1. "Conversion Factors for Hydrocarbon Emission Components," Report No. NR-002, November
24, 1997. EPA Air docket A-2001-28, document number U-A-34.
2. Final Regulatory Impact Analysis: Control of Emissions from Marine Diesel Engines,
November 1999, p. 79 (Docket A-97-50, Document V-B-01).
3. Final Regulatory Impact Analysis: Control of Emissions from Marine Diesel Engines,
November 1999, p. 79 (Docket A-97-50, Document V-B-01).
4. Final Regulatory Impact Analysis: Control of Emissions from Marine Diesel Engines,
November 1999, p. 79 (Docket A-97-50, Document V-B-01).
5. "Information Regarding Mine Locations in the United States", EPA Memorandum. Copy
available in EPA Air Docket A-2001-28
6. "Analysis of Nonroad Engine Emissions in the Mining Industry," TRC Environmental
Corporation, July 1993, p. 1.
7. January 28, 2003 letter from the Western Regional Air Partnership to Administrator Whitman.
Copy available in EPA Air Docket A-2001-28.
8. "Particulate Traps for Retro-Fitting Construction Site Engines VERT" Final Measurements
and Implementation", A. Mayer et. al., SAE paper 1999-01-0116, March 1999. See also
"Particulate Traps for Construction Machines: Properties and Field Experience", J. Czerwinski
et. al., SAE Paper 2000-01-1923.
9. "The Optimized Deutz Service Diesel Particulate Filter System IF', H. Houben et. al., SAE
Technical Paper 942264, 1994. See also "Summary of Conference Call between US EPA and
Deutz Corporation on September 19, 2002 regarding Deutz Diesel Particulate Filter System",
EPA Memorandum to Air Docket A-2001-28".
10. "Particulate Traps for Construction Machines: Properties and Field Experience", J.
Czerwinski et. al., SAE Paper 2000-01-1923.
12-116
------- |