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 i